linux-brain/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c

15525 lines
420 KiB
C
Raw Normal View History

/* bnx2x_main.c: QLogic Everest network driver.
*
* Copyright (c) 2007-2013 Broadcom Corporation
* Copyright (c) 2014 QLogic Corporation
* All rights reserved
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation.
*
* Maintained by: Ariel Elior <ariel.elior@qlogic.com>
* Written by: Eliezer Tamir
* Based on code from Michael Chan's bnx2 driver
* UDP CSUM errata workaround by Arik Gendelman
* Slowpath and fastpath rework by Vladislav Zolotarov
* Statistics and Link management by Yitchak Gertner
*
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/kernel.h>
#include <linux/device.h> /* for dev_info() */
#include <linux/timer.h>
#include <linux/errno.h>
#include <linux/ioport.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/pci.h>
#include <linux/aer.h>
#include <linux/init.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/dma-mapping.h>
#include <linux/bitops.h>
#include <linux/irq.h>
#include <linux/delay.h>
#include <asm/byteorder.h>
#include <linux/time.h>
#include <linux/ethtool.h>
#include <linux/mii.h>
#include <linux/if_vlan.h>
#include <linux/crash_dump.h>
#include <net/ip.h>
#include <net/ipv6.h>
#include <net/tcp.h>
#include <net/vxlan.h>
#include <net/checksum.h>
#include <net/ip6_checksum.h>
#include <linux/workqueue.h>
#include <linux/crc32.h>
#include <linux/crc32c.h>
#include <linux/prefetch.h>
#include <linux/zlib.h>
#include <linux/io.h>
#include <linux/semaphore.h>
#include <linux/stringify.h>
#include <linux/vmalloc.h>
#include "bnx2x.h"
#include "bnx2x_init.h"
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
#include "bnx2x_init_ops.h"
#include "bnx2x_cmn.h"
#include "bnx2x_vfpf.h"
#include "bnx2x_dcb.h"
#include "bnx2x_sp.h"
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
#include <linux/firmware.h>
#include "bnx2x_fw_file_hdr.h"
/* FW files */
#define FW_FILE_VERSION \
__stringify(BCM_5710_FW_MAJOR_VERSION) "." \
__stringify(BCM_5710_FW_MINOR_VERSION) "." \
__stringify(BCM_5710_FW_REVISION_VERSION) "." \
__stringify(BCM_5710_FW_ENGINEERING_VERSION)
#define FW_FILE_NAME_E1 "bnx2x/bnx2x-e1-" FW_FILE_VERSION ".fw"
#define FW_FILE_NAME_E1H "bnx2x/bnx2x-e1h-" FW_FILE_VERSION ".fw"
#define FW_FILE_NAME_E2 "bnx2x/bnx2x-e2-" FW_FILE_VERSION ".fw"
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
/* Time in jiffies before concluding the transmitter is hung */
#define TX_TIMEOUT (5*HZ)
static char version[] =
"QLogic 5771x/578xx 10/20-Gigabit Ethernet Driver "
DRV_MODULE_NAME " " DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
MODULE_AUTHOR("Eliezer Tamir");
MODULE_DESCRIPTION("QLogic "
"BCM57710/57711/57711E/"
"57712/57712_MF/57800/57800_MF/57810/57810_MF/"
"57840/57840_MF Driver");
MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_MODULE_VERSION);
MODULE_FIRMWARE(FW_FILE_NAME_E1);
MODULE_FIRMWARE(FW_FILE_NAME_E1H);
MODULE_FIRMWARE(FW_FILE_NAME_E2);
int bnx2x_num_queues;
module_param_named(num_queues, bnx2x_num_queues, int, S_IRUGO);
MODULE_PARM_DESC(num_queues,
" Set number of queues (default is as a number of CPUs)");
static int disable_tpa;
module_param(disable_tpa, int, S_IRUGO);
MODULE_PARM_DESC(disable_tpa, " Disable the TPA (LRO) feature");
static int int_mode;
module_param(int_mode, int, S_IRUGO);
MODULE_PARM_DESC(int_mode, " Force interrupt mode other than MSI-X "
"(1 INT#x; 2 MSI)");
static int dropless_fc;
module_param(dropless_fc, int, S_IRUGO);
MODULE_PARM_DESC(dropless_fc, " Pause on exhausted host ring");
static int mrrs = -1;
module_param(mrrs, int, S_IRUGO);
MODULE_PARM_DESC(mrrs, " Force Max Read Req Size (0..3) (for debug)");
static int debug;
module_param(debug, int, S_IRUGO);
MODULE_PARM_DESC(debug, " Default debug msglevel");
static struct workqueue_struct *bnx2x_wq;
struct workqueue_struct *bnx2x_iov_wq;
struct bnx2x_mac_vals {
u32 xmac_addr;
u32 xmac_val;
u32 emac_addr;
u32 emac_val;
u32 umac_addr[2];
u32 umac_val[2];
u32 bmac_addr;
u32 bmac_val[2];
};
enum bnx2x_board_type {
BCM57710 = 0,
BCM57711,
BCM57711E,
BCM57712,
BCM57712_MF,
BCM57712_VF,
BCM57800,
BCM57800_MF,
BCM57800_VF,
BCM57810,
BCM57810_MF,
BCM57810_VF,
BCM57840_4_10,
BCM57840_2_20,
BCM57840_MF,
BCM57840_VF,
BCM57811,
BCM57811_MF,
BCM57840_O,
BCM57840_MFO,
BCM57811_VF
};
/* indexed by board_type, above */
static struct {
char *name;
} board_info[] = {
[BCM57710] = { "QLogic BCM57710 10 Gigabit PCIe [Everest]" },
[BCM57711] = { "QLogic BCM57711 10 Gigabit PCIe" },
[BCM57711E] = { "QLogic BCM57711E 10 Gigabit PCIe" },
[BCM57712] = { "QLogic BCM57712 10 Gigabit Ethernet" },
[BCM57712_MF] = { "QLogic BCM57712 10 Gigabit Ethernet Multi Function" },
[BCM57712_VF] = { "QLogic BCM57712 10 Gigabit Ethernet Virtual Function" },
[BCM57800] = { "QLogic BCM57800 10 Gigabit Ethernet" },
[BCM57800_MF] = { "QLogic BCM57800 10 Gigabit Ethernet Multi Function" },
[BCM57800_VF] = { "QLogic BCM57800 10 Gigabit Ethernet Virtual Function" },
[BCM57810] = { "QLogic BCM57810 10 Gigabit Ethernet" },
[BCM57810_MF] = { "QLogic BCM57810 10 Gigabit Ethernet Multi Function" },
[BCM57810_VF] = { "QLogic BCM57810 10 Gigabit Ethernet Virtual Function" },
[BCM57840_4_10] = { "QLogic BCM57840 10 Gigabit Ethernet" },
[BCM57840_2_20] = { "QLogic BCM57840 20 Gigabit Ethernet" },
[BCM57840_MF] = { "QLogic BCM57840 10/20 Gigabit Ethernet Multi Function" },
[BCM57840_VF] = { "QLogic BCM57840 10/20 Gigabit Ethernet Virtual Function" },
[BCM57811] = { "QLogic BCM57811 10 Gigabit Ethernet" },
[BCM57811_MF] = { "QLogic BCM57811 10 Gigabit Ethernet Multi Function" },
[BCM57840_O] = { "QLogic BCM57840 10/20 Gigabit Ethernet" },
[BCM57840_MFO] = { "QLogic BCM57840 10/20 Gigabit Ethernet Multi Function" },
[BCM57811_VF] = { "QLogic BCM57840 10/20 Gigabit Ethernet Virtual Function" }
};
#ifndef PCI_DEVICE_ID_NX2_57710
#define PCI_DEVICE_ID_NX2_57710 CHIP_NUM_57710
#endif
#ifndef PCI_DEVICE_ID_NX2_57711
#define PCI_DEVICE_ID_NX2_57711 CHIP_NUM_57711
#endif
#ifndef PCI_DEVICE_ID_NX2_57711E
#define PCI_DEVICE_ID_NX2_57711E CHIP_NUM_57711E
#endif
#ifndef PCI_DEVICE_ID_NX2_57712
#define PCI_DEVICE_ID_NX2_57712 CHIP_NUM_57712
#endif
#ifndef PCI_DEVICE_ID_NX2_57712_MF
#define PCI_DEVICE_ID_NX2_57712_MF CHIP_NUM_57712_MF
#endif
#ifndef PCI_DEVICE_ID_NX2_57712_VF
#define PCI_DEVICE_ID_NX2_57712_VF CHIP_NUM_57712_VF
#endif
#ifndef PCI_DEVICE_ID_NX2_57800
#define PCI_DEVICE_ID_NX2_57800 CHIP_NUM_57800
#endif
#ifndef PCI_DEVICE_ID_NX2_57800_MF
#define PCI_DEVICE_ID_NX2_57800_MF CHIP_NUM_57800_MF
#endif
#ifndef PCI_DEVICE_ID_NX2_57800_VF
#define PCI_DEVICE_ID_NX2_57800_VF CHIP_NUM_57800_VF
#endif
#ifndef PCI_DEVICE_ID_NX2_57810
#define PCI_DEVICE_ID_NX2_57810 CHIP_NUM_57810
#endif
#ifndef PCI_DEVICE_ID_NX2_57810_MF
#define PCI_DEVICE_ID_NX2_57810_MF CHIP_NUM_57810_MF
#endif
#ifndef PCI_DEVICE_ID_NX2_57840_O
#define PCI_DEVICE_ID_NX2_57840_O CHIP_NUM_57840_OBSOLETE
#endif
#ifndef PCI_DEVICE_ID_NX2_57810_VF
#define PCI_DEVICE_ID_NX2_57810_VF CHIP_NUM_57810_VF
#endif
#ifndef PCI_DEVICE_ID_NX2_57840_4_10
#define PCI_DEVICE_ID_NX2_57840_4_10 CHIP_NUM_57840_4_10
#endif
#ifndef PCI_DEVICE_ID_NX2_57840_2_20
#define PCI_DEVICE_ID_NX2_57840_2_20 CHIP_NUM_57840_2_20
#endif
#ifndef PCI_DEVICE_ID_NX2_57840_MFO
#define PCI_DEVICE_ID_NX2_57840_MFO CHIP_NUM_57840_MF_OBSOLETE
#endif
#ifndef PCI_DEVICE_ID_NX2_57840_MF
#define PCI_DEVICE_ID_NX2_57840_MF CHIP_NUM_57840_MF
#endif
#ifndef PCI_DEVICE_ID_NX2_57840_VF
#define PCI_DEVICE_ID_NX2_57840_VF CHIP_NUM_57840_VF
#endif
#ifndef PCI_DEVICE_ID_NX2_57811
#define PCI_DEVICE_ID_NX2_57811 CHIP_NUM_57811
#endif
#ifndef PCI_DEVICE_ID_NX2_57811_MF
#define PCI_DEVICE_ID_NX2_57811_MF CHIP_NUM_57811_MF
#endif
#ifndef PCI_DEVICE_ID_NX2_57811_VF
#define PCI_DEVICE_ID_NX2_57811_VF CHIP_NUM_57811_VF
#endif
static const struct pci_device_id bnx2x_pci_tbl[] = {
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57710), BCM57710 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57711), BCM57711 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57711E), BCM57711E },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57712), BCM57712 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57712_MF), BCM57712_MF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57712_VF), BCM57712_VF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57800), BCM57800 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57800_MF), BCM57800_MF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57800_VF), BCM57800_VF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57810), BCM57810 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57810_MF), BCM57810_MF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57840_O), BCM57840_O },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57840_4_10), BCM57840_4_10 },
{ PCI_VDEVICE(QLOGIC, PCI_DEVICE_ID_NX2_57840_4_10), BCM57840_4_10 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57840_2_20), BCM57840_2_20 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57810_VF), BCM57810_VF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57840_MFO), BCM57840_MFO },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57840_MF), BCM57840_MF },
{ PCI_VDEVICE(QLOGIC, PCI_DEVICE_ID_NX2_57840_MF), BCM57840_MF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57840_VF), BCM57840_VF },
{ PCI_VDEVICE(QLOGIC, PCI_DEVICE_ID_NX2_57840_VF), BCM57840_VF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57811), BCM57811 },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57811_MF), BCM57811_MF },
{ PCI_VDEVICE(BROADCOM, PCI_DEVICE_ID_NX2_57811_VF), BCM57811_VF },
{ 0 }
};
MODULE_DEVICE_TABLE(pci, bnx2x_pci_tbl);
/* Global resources for unloading a previously loaded device */
#define BNX2X_PREV_WAIT_NEEDED 1
static DEFINE_SEMAPHORE(bnx2x_prev_sem);
static LIST_HEAD(bnx2x_prev_list);
/* Forward declaration */
static struct cnic_eth_dev *bnx2x_cnic_probe(struct net_device *dev);
static u32 bnx2x_rx_ustorm_prods_offset(struct bnx2x_fastpath *fp);
static int bnx2x_set_storm_rx_mode(struct bnx2x *bp);
/****************************************************************************
* General service functions
****************************************************************************/
static int bnx2x_hwtstamp_ioctl(struct bnx2x *bp, struct ifreq *ifr);
static void __storm_memset_dma_mapping(struct bnx2x *bp,
u32 addr, dma_addr_t mapping)
{
REG_WR(bp, addr, U64_LO(mapping));
REG_WR(bp, addr + 4, U64_HI(mapping));
}
static void storm_memset_spq_addr(struct bnx2x *bp,
dma_addr_t mapping, u16 abs_fid)
{
u32 addr = XSEM_REG_FAST_MEMORY +
XSTORM_SPQ_PAGE_BASE_OFFSET(abs_fid);
__storm_memset_dma_mapping(bp, addr, mapping);
}
static void storm_memset_vf_to_pf(struct bnx2x *bp, u16 abs_fid,
u16 pf_id)
{
REG_WR8(bp, BAR_XSTRORM_INTMEM + XSTORM_VF_TO_PF_OFFSET(abs_fid),
pf_id);
REG_WR8(bp, BAR_CSTRORM_INTMEM + CSTORM_VF_TO_PF_OFFSET(abs_fid),
pf_id);
REG_WR8(bp, BAR_TSTRORM_INTMEM + TSTORM_VF_TO_PF_OFFSET(abs_fid),
pf_id);
REG_WR8(bp, BAR_USTRORM_INTMEM + USTORM_VF_TO_PF_OFFSET(abs_fid),
pf_id);
}
static void storm_memset_func_en(struct bnx2x *bp, u16 abs_fid,
u8 enable)
{
REG_WR8(bp, BAR_XSTRORM_INTMEM + XSTORM_FUNC_EN_OFFSET(abs_fid),
enable);
REG_WR8(bp, BAR_CSTRORM_INTMEM + CSTORM_FUNC_EN_OFFSET(abs_fid),
enable);
REG_WR8(bp, BAR_TSTRORM_INTMEM + TSTORM_FUNC_EN_OFFSET(abs_fid),
enable);
REG_WR8(bp, BAR_USTRORM_INTMEM + USTORM_FUNC_EN_OFFSET(abs_fid),
enable);
}
static void storm_memset_eq_data(struct bnx2x *bp,
struct event_ring_data *eq_data,
u16 pfid)
{
size_t size = sizeof(struct event_ring_data);
u32 addr = BAR_CSTRORM_INTMEM + CSTORM_EVENT_RING_DATA_OFFSET(pfid);
__storm_memset_struct(bp, addr, size, (u32 *)eq_data);
}
static void storm_memset_eq_prod(struct bnx2x *bp, u16 eq_prod,
u16 pfid)
{
u32 addr = BAR_CSTRORM_INTMEM + CSTORM_EVENT_RING_PROD_OFFSET(pfid);
REG_WR16(bp, addr, eq_prod);
}
/* used only at init
* locking is done by mcp
*/
static void bnx2x_reg_wr_ind(struct bnx2x *bp, u32 addr, u32 val)
{
pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS, addr);
pci_write_config_dword(bp->pdev, PCICFG_GRC_DATA, val);
pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS,
PCICFG_VENDOR_ID_OFFSET);
}
static u32 bnx2x_reg_rd_ind(struct bnx2x *bp, u32 addr)
{
u32 val;
pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS, addr);
pci_read_config_dword(bp->pdev, PCICFG_GRC_DATA, &val);
pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS,
PCICFG_VENDOR_ID_OFFSET);
return val;
}
#define DMAE_DP_SRC_GRC "grc src_addr [%08x]"
#define DMAE_DP_SRC_PCI "pci src_addr [%x:%08x]"
#define DMAE_DP_DST_GRC "grc dst_addr [%08x]"
#define DMAE_DP_DST_PCI "pci dst_addr [%x:%08x]"
#define DMAE_DP_DST_NONE "dst_addr [none]"
static void bnx2x_dp_dmae(struct bnx2x *bp,
struct dmae_command *dmae, int msglvl)
{
u32 src_type = dmae->opcode & DMAE_COMMAND_SRC;
int i;
switch (dmae->opcode & DMAE_COMMAND_DST) {
case DMAE_CMD_DST_PCI:
if (src_type == DMAE_CMD_SRC_PCI)
DP(msglvl, "DMAE: opcode 0x%08x\n"
"src [%x:%08x], len [%d*4], dst [%x:%08x]\n"
"comp_addr [%x:%08x], comp_val 0x%08x\n",
dmae->opcode, dmae->src_addr_hi, dmae->src_addr_lo,
dmae->len, dmae->dst_addr_hi, dmae->dst_addr_lo,
dmae->comp_addr_hi, dmae->comp_addr_lo,
dmae->comp_val);
else
DP(msglvl, "DMAE: opcode 0x%08x\n"
"src [%08x], len [%d*4], dst [%x:%08x]\n"
"comp_addr [%x:%08x], comp_val 0x%08x\n",
dmae->opcode, dmae->src_addr_lo >> 2,
dmae->len, dmae->dst_addr_hi, dmae->dst_addr_lo,
dmae->comp_addr_hi, dmae->comp_addr_lo,
dmae->comp_val);
break;
case DMAE_CMD_DST_GRC:
if (src_type == DMAE_CMD_SRC_PCI)
DP(msglvl, "DMAE: opcode 0x%08x\n"
"src [%x:%08x], len [%d*4], dst_addr [%08x]\n"
"comp_addr [%x:%08x], comp_val 0x%08x\n",
dmae->opcode, dmae->src_addr_hi, dmae->src_addr_lo,
dmae->len, dmae->dst_addr_lo >> 2,
dmae->comp_addr_hi, dmae->comp_addr_lo,
dmae->comp_val);
else
DP(msglvl, "DMAE: opcode 0x%08x\n"
"src [%08x], len [%d*4], dst [%08x]\n"
"comp_addr [%x:%08x], comp_val 0x%08x\n",
dmae->opcode, dmae->src_addr_lo >> 2,
dmae->len, dmae->dst_addr_lo >> 2,
dmae->comp_addr_hi, dmae->comp_addr_lo,
dmae->comp_val);
break;
default:
if (src_type == DMAE_CMD_SRC_PCI)
DP(msglvl, "DMAE: opcode 0x%08x\n"
"src_addr [%x:%08x] len [%d * 4] dst_addr [none]\n"
"comp_addr [%x:%08x] comp_val 0x%08x\n",
dmae->opcode, dmae->src_addr_hi, dmae->src_addr_lo,
dmae->len, dmae->comp_addr_hi, dmae->comp_addr_lo,
dmae->comp_val);
else
DP(msglvl, "DMAE: opcode 0x%08x\n"
"src_addr [%08x] len [%d * 4] dst_addr [none]\n"
"comp_addr [%x:%08x] comp_val 0x%08x\n",
dmae->opcode, dmae->src_addr_lo >> 2,
dmae->len, dmae->comp_addr_hi, dmae->comp_addr_lo,
dmae->comp_val);
break;
}
for (i = 0; i < (sizeof(struct dmae_command)/4); i++)
DP(msglvl, "DMAE RAW [%02d]: 0x%08x\n",
i, *(((u32 *)dmae) + i));
}
/* copy command into DMAE command memory and set DMAE command go */
void bnx2x_post_dmae(struct bnx2x *bp, struct dmae_command *dmae, int idx)
{
u32 cmd_offset;
int i;
cmd_offset = (DMAE_REG_CMD_MEM + sizeof(struct dmae_command) * idx);
for (i = 0; i < (sizeof(struct dmae_command)/4); i++) {
REG_WR(bp, cmd_offset + i*4, *(((u32 *)dmae) + i));
}
REG_WR(bp, dmae_reg_go_c[idx], 1);
}
u32 bnx2x_dmae_opcode_add_comp(u32 opcode, u8 comp_type)
{
return opcode | ((comp_type << DMAE_COMMAND_C_DST_SHIFT) |
DMAE_CMD_C_ENABLE);
}
u32 bnx2x_dmae_opcode_clr_src_reset(u32 opcode)
{
return opcode & ~DMAE_CMD_SRC_RESET;
}
u32 bnx2x_dmae_opcode(struct bnx2x *bp, u8 src_type, u8 dst_type,
bool with_comp, u8 comp_type)
{
u32 opcode = 0;
opcode |= ((src_type << DMAE_COMMAND_SRC_SHIFT) |
(dst_type << DMAE_COMMAND_DST_SHIFT));
opcode |= (DMAE_CMD_SRC_RESET | DMAE_CMD_DST_RESET);
opcode |= (BP_PORT(bp) ? DMAE_CMD_PORT_1 : DMAE_CMD_PORT_0);
opcode |= ((BP_VN(bp) << DMAE_CMD_E1HVN_SHIFT) |
(BP_VN(bp) << DMAE_COMMAND_DST_VN_SHIFT));
opcode |= (DMAE_COM_SET_ERR << DMAE_COMMAND_ERR_POLICY_SHIFT);
#ifdef __BIG_ENDIAN
opcode |= DMAE_CMD_ENDIANITY_B_DW_SWAP;
#else
opcode |= DMAE_CMD_ENDIANITY_DW_SWAP;
#endif
if (with_comp)
opcode = bnx2x_dmae_opcode_add_comp(opcode, comp_type);
return opcode;
}
void bnx2x_prep_dmae_with_comp(struct bnx2x *bp,
struct dmae_command *dmae,
u8 src_type, u8 dst_type)
{
memset(dmae, 0, sizeof(struct dmae_command));
/* set the opcode */
dmae->opcode = bnx2x_dmae_opcode(bp, src_type, dst_type,
true, DMAE_COMP_PCI);
/* fill in the completion parameters */
dmae->comp_addr_lo = U64_LO(bnx2x_sp_mapping(bp, wb_comp));
dmae->comp_addr_hi = U64_HI(bnx2x_sp_mapping(bp, wb_comp));
dmae->comp_val = DMAE_COMP_VAL;
}
/* issue a dmae command over the init-channel and wait for completion */
int bnx2x_issue_dmae_with_comp(struct bnx2x *bp, struct dmae_command *dmae,
u32 *comp)
{
int cnt = CHIP_REV_IS_SLOW(bp) ? (400000) : 4000;
int rc = 0;
bnx2x_dp_dmae(bp, dmae, BNX2X_MSG_DMAE);
/* Lock the dmae channel. Disable BHs to prevent a dead-lock
* as long as this code is called both from syscall context and
* from ndo_set_rx_mode() flow that may be called from BH.
*/
spin_lock_bh(&bp->dmae_lock);
/* reset completion */
*comp = 0;
/* post the command on the channel used for initializations */
bnx2x_post_dmae(bp, dmae, INIT_DMAE_C(bp));
/* wait for completion */
udelay(5);
while ((*comp & ~DMAE_PCI_ERR_FLAG) != DMAE_COMP_VAL) {
if (!cnt ||
(bp->recovery_state != BNX2X_RECOVERY_DONE &&
bp->recovery_state != BNX2X_RECOVERY_NIC_LOADING)) {
BNX2X_ERR("DMAE timeout!\n");
rc = DMAE_TIMEOUT;
goto unlock;
}
cnt--;
udelay(50);
}
if (*comp & DMAE_PCI_ERR_FLAG) {
BNX2X_ERR("DMAE PCI error!\n");
rc = DMAE_PCI_ERROR;
}
unlock:
spin_unlock_bh(&bp->dmae_lock);
return rc;
}
void bnx2x_write_dmae(struct bnx2x *bp, dma_addr_t dma_addr, u32 dst_addr,
u32 len32)
{
int rc;
struct dmae_command dmae;
if (!bp->dmae_ready) {
u32 *data = bnx2x_sp(bp, wb_data[0]);
if (CHIP_IS_E1(bp))
bnx2x_init_ind_wr(bp, dst_addr, data, len32);
else
bnx2x_init_str_wr(bp, dst_addr, data, len32);
return;
}
/* set opcode and fixed command fields */
bnx2x_prep_dmae_with_comp(bp, &dmae, DMAE_SRC_PCI, DMAE_DST_GRC);
/* fill in addresses and len */
dmae.src_addr_lo = U64_LO(dma_addr);
dmae.src_addr_hi = U64_HI(dma_addr);
dmae.dst_addr_lo = dst_addr >> 2;
dmae.dst_addr_hi = 0;
dmae.len = len32;
/* issue the command and wait for completion */
rc = bnx2x_issue_dmae_with_comp(bp, &dmae, bnx2x_sp(bp, wb_comp));
if (rc) {
BNX2X_ERR("DMAE returned failure %d\n", rc);
#ifdef BNX2X_STOP_ON_ERROR
bnx2x_panic();
#endif
}
}
void bnx2x_read_dmae(struct bnx2x *bp, u32 src_addr, u32 len32)
{
int rc;
struct dmae_command dmae;
if (!bp->dmae_ready) {
u32 *data = bnx2x_sp(bp, wb_data[0]);
int i;
if (CHIP_IS_E1(bp))
for (i = 0; i < len32; i++)
data[i] = bnx2x_reg_rd_ind(bp, src_addr + i*4);
else
for (i = 0; i < len32; i++)
data[i] = REG_RD(bp, src_addr + i*4);
return;
}
/* set opcode and fixed command fields */
bnx2x_prep_dmae_with_comp(bp, &dmae, DMAE_SRC_GRC, DMAE_DST_PCI);
/* fill in addresses and len */
dmae.src_addr_lo = src_addr >> 2;
dmae.src_addr_hi = 0;
dmae.dst_addr_lo = U64_LO(bnx2x_sp_mapping(bp, wb_data));
dmae.dst_addr_hi = U64_HI(bnx2x_sp_mapping(bp, wb_data));
dmae.len = len32;
/* issue the command and wait for completion */
rc = bnx2x_issue_dmae_with_comp(bp, &dmae, bnx2x_sp(bp, wb_comp));
if (rc) {
BNX2X_ERR("DMAE returned failure %d\n", rc);
#ifdef BNX2X_STOP_ON_ERROR
bnx2x_panic();
#endif
}
}
static void bnx2x_write_dmae_phys_len(struct bnx2x *bp, dma_addr_t phys_addr,
u32 addr, u32 len)
{
int dmae_wr_max = DMAE_LEN32_WR_MAX(bp);
int offset = 0;
while (len > dmae_wr_max) {
bnx2x_write_dmae(bp, phys_addr + offset,
addr + offset, dmae_wr_max);
offset += dmae_wr_max * 4;
len -= dmae_wr_max;
}
bnx2x_write_dmae(bp, phys_addr + offset, addr + offset, len);
}
enum storms {
XSTORM,
TSTORM,
CSTORM,
USTORM,
MAX_STORMS
};
#define STORMS_NUM 4
#define REGS_IN_ENTRY 4
static inline int bnx2x_get_assert_list_entry(struct bnx2x *bp,
enum storms storm,
int entry)
{
switch (storm) {
case XSTORM:
return XSTORM_ASSERT_LIST_OFFSET(entry);
case TSTORM:
return TSTORM_ASSERT_LIST_OFFSET(entry);
case CSTORM:
return CSTORM_ASSERT_LIST_OFFSET(entry);
case USTORM:
return USTORM_ASSERT_LIST_OFFSET(entry);
case MAX_STORMS:
default:
BNX2X_ERR("unknown storm\n");
}
return -EINVAL;
}
static int bnx2x_mc_assert(struct bnx2x *bp)
{
char last_idx;
int i, j, rc = 0;
enum storms storm;
u32 regs[REGS_IN_ENTRY];
u32 bar_storm_intmem[STORMS_NUM] = {
BAR_XSTRORM_INTMEM,
BAR_TSTRORM_INTMEM,
BAR_CSTRORM_INTMEM,
BAR_USTRORM_INTMEM
};
u32 storm_assert_list_index[STORMS_NUM] = {
XSTORM_ASSERT_LIST_INDEX_OFFSET,
TSTORM_ASSERT_LIST_INDEX_OFFSET,
CSTORM_ASSERT_LIST_INDEX_OFFSET,
USTORM_ASSERT_LIST_INDEX_OFFSET
};
char *storms_string[STORMS_NUM] = {
"XSTORM",
"TSTORM",
"CSTORM",
"USTORM"
};
for (storm = XSTORM; storm < MAX_STORMS; storm++) {
last_idx = REG_RD8(bp, bar_storm_intmem[storm] +
storm_assert_list_index[storm]);
if (last_idx)
BNX2X_ERR("%s_ASSERT_LIST_INDEX 0x%x\n",
storms_string[storm], last_idx);
/* print the asserts */
for (i = 0; i < STROM_ASSERT_ARRAY_SIZE; i++) {
/* read a single assert entry */
for (j = 0; j < REGS_IN_ENTRY; j++)
regs[j] = REG_RD(bp, bar_storm_intmem[storm] +
bnx2x_get_assert_list_entry(bp,
storm,
i) +
sizeof(u32) * j);
/* log entry if it contains a valid assert */
if (regs[0] != COMMON_ASM_INVALID_ASSERT_OPCODE) {
BNX2X_ERR("%s_ASSERT_INDEX 0x%x = 0x%08x 0x%08x 0x%08x 0x%08x\n",
storms_string[storm], i, regs[3],
regs[2], regs[1], regs[0]);
rc++;
} else {
break;
}
}
}
BNX2X_ERR("Chip Revision: %s, FW Version: %d_%d_%d\n",
CHIP_IS_E1(bp) ? "everest1" :
CHIP_IS_E1H(bp) ? "everest1h" :
CHIP_IS_E2(bp) ? "everest2" : "everest3",
BCM_5710_FW_MAJOR_VERSION,
BCM_5710_FW_MINOR_VERSION,
BCM_5710_FW_REVISION_VERSION);
return rc;
}
#define MCPR_TRACE_BUFFER_SIZE (0x800)
#define SCRATCH_BUFFER_SIZE(bp) \
(CHIP_IS_E1(bp) ? 0x10000 : (CHIP_IS_E1H(bp) ? 0x20000 : 0x28000))
void bnx2x_fw_dump_lvl(struct bnx2x *bp, const char *lvl)
{
u32 addr, val;
u32 mark, offset;
__be32 data[9];
int word;
u32 trace_shmem_base;
if (BP_NOMCP(bp)) {
BNX2X_ERR("NO MCP - can not dump\n");
return;
}
netdev_printk(lvl, bp->dev, "bc %d.%d.%d\n",
(bp->common.bc_ver & 0xff0000) >> 16,
(bp->common.bc_ver & 0xff00) >> 8,
(bp->common.bc_ver & 0xff));
bnx2x: don't reset chip on cleanup if PCI function is offline When PCI error is detected, in some architectures (like PowerPC) a slot reset is performed - the driver's error handlers are in charge of "disable" device before the reset, and re-enable it after a successful slot reset. There are two cases though that another path is taken on the code: if the slot reset is not successful or if too many errors already happened in the specific adapter (meaning that possibly the device is experiencing a HW failure that slot reset is not able to solve), the core PCI error mechanism (called EEH in PowerPC) will remove the adapter from the system, since it will consider this as a permanent failure on device. In this case, a path is taken that leads to bnx2x_chip_cleanup() calling bnx2x_reset_hw(), which then tries to perform a HW reset on chip. This reset won't succeed since the HW is in a fault state, which can be seen by multiple messages on kernel log like below: bnx2x: [bnx2x_issue_dmae_with_comp:552(eth1)]DMAE timeout! bnx2x: [bnx2x_write_dmae:600(eth1)]DMAE returned failure -1 After some time, the PCI error mechanism gives up on waiting the driver's correct removal procedure and forcibly remove the adapter from the system. We can see soft lockup while core PCI error mechanism is waiting for driver to accomplish the right removal process. This patch adds a verification to avoid a chip reset whenever the function is in PCI error state - since this case is only reached when we have a device being removed because of a permanent failure, the HW chip reset is not expected to work fine neither is necessary. Also, as a minor improvement in error path, we avoid the MCP information dump in case of non-recoverable PCI error (when adapter is about to be removed), since it will certainly fail. Reported-by: Harsha Thyagaraja <hathyaga@in.ibm.com> Signed-off-by: Guilherme G. Piccoli <gpiccoli@linux.vnet.ibm.com> Acked-By: Yuval Mintz <Yuval.Mintz@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-09-01 00:11:57 +09:00
if (pci_channel_offline(bp->pdev)) {
BNX2X_ERR("Cannot dump MCP info while in PCI error\n");
return;
}
val = REG_RD(bp, MCP_REG_MCPR_CPU_PROGRAM_COUNTER);
if (val == REG_RD(bp, MCP_REG_MCPR_CPU_PROGRAM_COUNTER))
BNX2X_ERR("%s" "MCP PC at 0x%x\n", lvl, val);
if (BP_PATH(bp) == 0)
trace_shmem_base = bp->common.shmem_base;
else
trace_shmem_base = SHMEM2_RD(bp, other_shmem_base_addr);
/* sanity */
if (trace_shmem_base < MCPR_SCRATCH_BASE(bp) + MCPR_TRACE_BUFFER_SIZE ||
trace_shmem_base >= MCPR_SCRATCH_BASE(bp) +
SCRATCH_BUFFER_SIZE(bp)) {
BNX2X_ERR("Unable to dump trace buffer (mark %x)\n",
trace_shmem_base);
return;
}
addr = trace_shmem_base - MCPR_TRACE_BUFFER_SIZE;
/* validate TRCB signature */
mark = REG_RD(bp, addr);
if (mark != MFW_TRACE_SIGNATURE) {
BNX2X_ERR("Trace buffer signature is missing.");
return ;
}
/* read cyclic buffer pointer */
addr += 4;
mark = REG_RD(bp, addr);
mark = MCPR_SCRATCH_BASE(bp) + ((mark + 0x3) & ~0x3) - 0x08000000;
if (mark >= trace_shmem_base || mark < addr + 4) {
BNX2X_ERR("Mark doesn't fall inside Trace Buffer\n");
return;
}
printk("%s" "begin fw dump (mark 0x%x)\n", lvl, mark);
printk("%s", lvl);
/* dump buffer after the mark */
for (offset = mark; offset < trace_shmem_base; offset += 0x8*4) {
for (word = 0; word < 8; word++)
data[word] = htonl(REG_RD(bp, offset + 4*word));
data[8] = 0x0;
pr_cont("%s", (char *)data);
}
/* dump buffer before the mark */
for (offset = addr + 4; offset <= mark; offset += 0x8*4) {
for (word = 0; word < 8; word++)
data[word] = htonl(REG_RD(bp, offset + 4*word));
data[8] = 0x0;
pr_cont("%s", (char *)data);
}
printk("%s" "end of fw dump\n", lvl);
}
static void bnx2x_fw_dump(struct bnx2x *bp)
{
bnx2x_fw_dump_lvl(bp, KERN_ERR);
}
static void bnx2x_hc_int_disable(struct bnx2x *bp)
{
int port = BP_PORT(bp);
u32 addr = port ? HC_REG_CONFIG_1 : HC_REG_CONFIG_0;
u32 val = REG_RD(bp, addr);
/* in E1 we must use only PCI configuration space to disable
* MSI/MSIX capability
* It's forbidden to disable IGU_PF_CONF_MSI_MSIX_EN in HC block
*/
if (CHIP_IS_E1(bp)) {
/* Since IGU_PF_CONF_MSI_MSIX_EN still always on
* Use mask register to prevent from HC sending interrupts
* after we exit the function
*/
REG_WR(bp, HC_REG_INT_MASK + port*4, 0);
val &= ~(HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
HC_CONFIG_0_REG_INT_LINE_EN_0 |
HC_CONFIG_0_REG_ATTN_BIT_EN_0);
} else
val &= ~(HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0 |
HC_CONFIG_0_REG_INT_LINE_EN_0 |
HC_CONFIG_0_REG_ATTN_BIT_EN_0);
DP(NETIF_MSG_IFDOWN,
"write %x to HC %d (addr 0x%x)\n",
val, port, addr);
/* flush all outstanding writes */
mmiowb();
REG_WR(bp, addr, val);
if (REG_RD(bp, addr) != val)
BNX2X_ERR("BUG! Proper val not read from IGU!\n");
}
static void bnx2x_igu_int_disable(struct bnx2x *bp)
{
u32 val = REG_RD(bp, IGU_REG_PF_CONFIGURATION);
val &= ~(IGU_PF_CONF_MSI_MSIX_EN |
IGU_PF_CONF_INT_LINE_EN |
IGU_PF_CONF_ATTN_BIT_EN);
DP(NETIF_MSG_IFDOWN, "write %x to IGU\n", val);
/* flush all outstanding writes */
mmiowb();
REG_WR(bp, IGU_REG_PF_CONFIGURATION, val);
if (REG_RD(bp, IGU_REG_PF_CONFIGURATION) != val)
BNX2X_ERR("BUG! Proper val not read from IGU!\n");
}
static void bnx2x_int_disable(struct bnx2x *bp)
{
if (bp->common.int_block == INT_BLOCK_HC)
bnx2x_hc_int_disable(bp);
else
bnx2x_igu_int_disable(bp);
}
void bnx2x_panic_dump(struct bnx2x *bp, bool disable_int)
{
int i;
u16 j;
struct hc_sp_status_block_data sp_sb_data;
int func = BP_FUNC(bp);
#ifdef BNX2X_STOP_ON_ERROR
u16 start = 0, end = 0;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 cos;
#endif
if (IS_PF(bp) && disable_int)
bnx2x_int_disable(bp);
bp->stats_state = STATS_STATE_DISABLED;
bp->eth_stats.unrecoverable_error++;
DP(BNX2X_MSG_STATS, "stats_state - DISABLED\n");
BNX2X_ERR("begin crash dump -----------------\n");
/* Indices */
/* Common */
if (IS_PF(bp)) {
struct host_sp_status_block *def_sb = bp->def_status_blk;
int data_size, cstorm_offset;
BNX2X_ERR("def_idx(0x%x) def_att_idx(0x%x) attn_state(0x%x) spq_prod_idx(0x%x) next_stats_cnt(0x%x)\n",
bp->def_idx, bp->def_att_idx, bp->attn_state,
bp->spq_prod_idx, bp->stats_counter);
BNX2X_ERR("DSB: attn bits(0x%x) ack(0x%x) id(0x%x) idx(0x%x)\n",
def_sb->atten_status_block.attn_bits,
def_sb->atten_status_block.attn_bits_ack,
def_sb->atten_status_block.status_block_id,
def_sb->atten_status_block.attn_bits_index);
BNX2X_ERR(" def (");
for (i = 0; i < HC_SP_SB_MAX_INDICES; i++)
pr_cont("0x%x%s",
def_sb->sp_sb.index_values[i],
(i == HC_SP_SB_MAX_INDICES - 1) ? ") " : " ");
data_size = sizeof(struct hc_sp_status_block_data) /
sizeof(u32);
cstorm_offset = CSTORM_SP_STATUS_BLOCK_DATA_OFFSET(func);
for (i = 0; i < data_size; i++)
*((u32 *)&sp_sb_data + i) =
REG_RD(bp, BAR_CSTRORM_INTMEM + cstorm_offset +
i * sizeof(u32));
pr_cont("igu_sb_id(0x%x) igu_seg_id(0x%x) pf_id(0x%x) vnic_id(0x%x) vf_id(0x%x) vf_valid (0x%x) state(0x%x)\n",
sp_sb_data.igu_sb_id,
sp_sb_data.igu_seg_id,
sp_sb_data.p_func.pf_id,
sp_sb_data.p_func.vnic_id,
sp_sb_data.p_func.vf_id,
sp_sb_data.p_func.vf_valid,
sp_sb_data.state);
}
for_each_eth_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
int loop;
struct hc_status_block_data_e2 sb_data_e2;
struct hc_status_block_data_e1x sb_data_e1x;
struct hc_status_block_sm *hc_sm_p =
CHIP_IS_E1x(bp) ?
sb_data_e1x.common.state_machine :
sb_data_e2.common.state_machine;
struct hc_index_data *hc_index_p =
CHIP_IS_E1x(bp) ?
sb_data_e1x.index_data :
sb_data_e2.index_data;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 data_size, cos;
u32 *sb_data_p;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
struct bnx2x_fp_txdata txdata;
if (!bp->fp)
break;
if (!fp->rx_cons_sb)
continue;
/* Rx */
BNX2X_ERR("fp%d: rx_bd_prod(0x%x) rx_bd_cons(0x%x) rx_comp_prod(0x%x) rx_comp_cons(0x%x) *rx_cons_sb(0x%x)\n",
i, fp->rx_bd_prod, fp->rx_bd_cons,
fp->rx_comp_prod,
fp->rx_comp_cons, le16_to_cpu(*fp->rx_cons_sb));
BNX2X_ERR(" rx_sge_prod(0x%x) last_max_sge(0x%x) fp_hc_idx(0x%x)\n",
fp->rx_sge_prod, fp->last_max_sge,
le16_to_cpu(fp->fp_hc_idx));
/* Tx */
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
for_each_cos_in_tx_queue(fp, cos)
{
if (!fp->txdata_ptr[cos])
break;
txdata = *fp->txdata_ptr[cos];
if (!txdata.tx_cons_sb)
continue;
BNX2X_ERR("fp%d: tx_pkt_prod(0x%x) tx_pkt_cons(0x%x) tx_bd_prod(0x%x) tx_bd_cons(0x%x) *tx_cons_sb(0x%x)\n",
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
i, txdata.tx_pkt_prod,
txdata.tx_pkt_cons, txdata.tx_bd_prod,
txdata.tx_bd_cons,
le16_to_cpu(*txdata.tx_cons_sb));
}
loop = CHIP_IS_E1x(bp) ?
HC_SB_MAX_INDICES_E1X : HC_SB_MAX_INDICES_E2;
/* host sb data */
if (IS_FCOE_FP(fp))
continue;
BNX2X_ERR(" run indexes (");
for (j = 0; j < HC_SB_MAX_SM; j++)
pr_cont("0x%x%s",
fp->sb_running_index[j],
(j == HC_SB_MAX_SM - 1) ? ")" : " ");
BNX2X_ERR(" indexes (");
for (j = 0; j < loop; j++)
pr_cont("0x%x%s",
fp->sb_index_values[j],
(j == loop - 1) ? ")" : " ");
/* VF cannot access FW refelection for status block */
if (IS_VF(bp))
continue;
/* fw sb data */
data_size = CHIP_IS_E1x(bp) ?
sizeof(struct hc_status_block_data_e1x) :
sizeof(struct hc_status_block_data_e2);
data_size /= sizeof(u32);
sb_data_p = CHIP_IS_E1x(bp) ?
(u32 *)&sb_data_e1x :
(u32 *)&sb_data_e2;
/* copy sb data in here */
for (j = 0; j < data_size; j++)
*(sb_data_p + j) = REG_RD(bp, BAR_CSTRORM_INTMEM +
CSTORM_STATUS_BLOCK_DATA_OFFSET(fp->fw_sb_id) +
j * sizeof(u32));
if (!CHIP_IS_E1x(bp)) {
pr_cont("pf_id(0x%x) vf_id(0x%x) vf_valid(0x%x) vnic_id(0x%x) same_igu_sb_1b(0x%x) state(0x%x)\n",
sb_data_e2.common.p_func.pf_id,
sb_data_e2.common.p_func.vf_id,
sb_data_e2.common.p_func.vf_valid,
sb_data_e2.common.p_func.vnic_id,
sb_data_e2.common.same_igu_sb_1b,
sb_data_e2.common.state);
} else {
pr_cont("pf_id(0x%x) vf_id(0x%x) vf_valid(0x%x) vnic_id(0x%x) same_igu_sb_1b(0x%x) state(0x%x)\n",
sb_data_e1x.common.p_func.pf_id,
sb_data_e1x.common.p_func.vf_id,
sb_data_e1x.common.p_func.vf_valid,
sb_data_e1x.common.p_func.vnic_id,
sb_data_e1x.common.same_igu_sb_1b,
sb_data_e1x.common.state);
}
/* SB_SMs data */
for (j = 0; j < HC_SB_MAX_SM; j++) {
pr_cont("SM[%d] __flags (0x%x) igu_sb_id (0x%x) igu_seg_id(0x%x) time_to_expire (0x%x) timer_value(0x%x)\n",
j, hc_sm_p[j].__flags,
hc_sm_p[j].igu_sb_id,
hc_sm_p[j].igu_seg_id,
hc_sm_p[j].time_to_expire,
hc_sm_p[j].timer_value);
}
/* Indices data */
for (j = 0; j < loop; j++) {
pr_cont("INDEX[%d] flags (0x%x) timeout (0x%x)\n", j,
hc_index_p[j].flags,
hc_index_p[j].timeout);
}
}
#ifdef BNX2X_STOP_ON_ERROR
if (IS_PF(bp)) {
/* event queue */
BNX2X_ERR("eq cons %x prod %x\n", bp->eq_cons, bp->eq_prod);
for (i = 0; i < NUM_EQ_DESC; i++) {
u32 *data = (u32 *)&bp->eq_ring[i].message.data;
BNX2X_ERR("event queue [%d]: header: opcode %d, error %d\n",
i, bp->eq_ring[i].message.opcode,
bp->eq_ring[i].message.error);
BNX2X_ERR("data: %x %x %x\n",
data[0], data[1], data[2]);
}
}
/* Rings */
/* Rx */
for_each_valid_rx_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
if (!bp->fp)
break;
if (!fp->rx_cons_sb)
continue;
start = RX_BD(le16_to_cpu(*fp->rx_cons_sb) - 10);
end = RX_BD(le16_to_cpu(*fp->rx_cons_sb) + 503);
for (j = start; j != end; j = RX_BD(j + 1)) {
u32 *rx_bd = (u32 *)&fp->rx_desc_ring[j];
struct sw_rx_bd *sw_bd = &fp->rx_buf_ring[j];
BNX2X_ERR("fp%d: rx_bd[%x]=[%x:%x] sw_bd=[%p]\n",
i, j, rx_bd[1], rx_bd[0], sw_bd->data);
}
start = RX_SGE(fp->rx_sge_prod);
end = RX_SGE(fp->last_max_sge);
for (j = start; j != end; j = RX_SGE(j + 1)) {
u32 *rx_sge = (u32 *)&fp->rx_sge_ring[j];
struct sw_rx_page *sw_page = &fp->rx_page_ring[j];
BNX2X_ERR("fp%d: rx_sge[%x]=[%x:%x] sw_page=[%p]\n",
i, j, rx_sge[1], rx_sge[0], sw_page->page);
}
start = RCQ_BD(fp->rx_comp_cons - 10);
end = RCQ_BD(fp->rx_comp_cons + 503);
for (j = start; j != end; j = RCQ_BD(j + 1)) {
u32 *cqe = (u32 *)&fp->rx_comp_ring[j];
BNX2X_ERR("fp%d: cqe[%x]=[%x:%x:%x:%x]\n",
i, j, cqe[0], cqe[1], cqe[2], cqe[3]);
}
}
/* Tx */
for_each_valid_tx_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
if (!bp->fp)
break;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
for_each_cos_in_tx_queue(fp, cos) {
struct bnx2x_fp_txdata *txdata = fp->txdata_ptr[cos];
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
if (!fp->txdata_ptr[cos])
break;
if (!txdata->tx_cons_sb)
continue;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
start = TX_BD(le16_to_cpu(*txdata->tx_cons_sb) - 10);
end = TX_BD(le16_to_cpu(*txdata->tx_cons_sb) + 245);
for (j = start; j != end; j = TX_BD(j + 1)) {
struct sw_tx_bd *sw_bd =
&txdata->tx_buf_ring[j];
BNX2X_ERR("fp%d: txdata %d, packet[%x]=[%p,%x]\n",
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
i, cos, j, sw_bd->skb,
sw_bd->first_bd);
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
start = TX_BD(txdata->tx_bd_cons - 10);
end = TX_BD(txdata->tx_bd_cons + 254);
for (j = start; j != end; j = TX_BD(j + 1)) {
u32 *tx_bd = (u32 *)&txdata->tx_desc_ring[j];
BNX2X_ERR("fp%d: txdata %d, tx_bd[%x]=[%x:%x:%x:%x]\n",
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
i, cos, j, tx_bd[0], tx_bd[1],
tx_bd[2], tx_bd[3]);
}
}
}
#endif
if (IS_PF(bp)) {
bnx2x_fw_dump(bp);
bnx2x_mc_assert(bp);
}
BNX2X_ERR("end crash dump -----------------\n");
}
/*
* FLR Support for E2
*
* bnx2x_pf_flr_clnup() is called during nic_load in the per function HW
* initialization.
*/
#define FLR_WAIT_USEC 10000 /* 10 milliseconds */
#define FLR_WAIT_INTERVAL 50 /* usec */
#define FLR_POLL_CNT (FLR_WAIT_USEC/FLR_WAIT_INTERVAL) /* 200 */
struct pbf_pN_buf_regs {
int pN;
u32 init_crd;
u32 crd;
u32 crd_freed;
};
struct pbf_pN_cmd_regs {
int pN;
u32 lines_occup;
u32 lines_freed;
};
static void bnx2x_pbf_pN_buf_flushed(struct bnx2x *bp,
struct pbf_pN_buf_regs *regs,
u32 poll_count)
{
u32 init_crd, crd, crd_start, crd_freed, crd_freed_start;
u32 cur_cnt = poll_count;
crd_freed = crd_freed_start = REG_RD(bp, regs->crd_freed);
crd = crd_start = REG_RD(bp, regs->crd);
init_crd = REG_RD(bp, regs->init_crd);
DP(BNX2X_MSG_SP, "INIT CREDIT[%d] : %x\n", regs->pN, init_crd);
DP(BNX2X_MSG_SP, "CREDIT[%d] : s:%x\n", regs->pN, crd);
DP(BNX2X_MSG_SP, "CREDIT_FREED[%d]: s:%x\n", regs->pN, crd_freed);
while ((crd != init_crd) && ((u32)SUB_S32(crd_freed, crd_freed_start) <
(init_crd - crd_start))) {
if (cur_cnt--) {
udelay(FLR_WAIT_INTERVAL);
crd = REG_RD(bp, regs->crd);
crd_freed = REG_RD(bp, regs->crd_freed);
} else {
DP(BNX2X_MSG_SP, "PBF tx buffer[%d] timed out\n",
regs->pN);
DP(BNX2X_MSG_SP, "CREDIT[%d] : c:%x\n",
regs->pN, crd);
DP(BNX2X_MSG_SP, "CREDIT_FREED[%d]: c:%x\n",
regs->pN, crd_freed);
break;
}
}
DP(BNX2X_MSG_SP, "Waited %d*%d usec for PBF tx buffer[%d]\n",
poll_count-cur_cnt, FLR_WAIT_INTERVAL, regs->pN);
}
static void bnx2x_pbf_pN_cmd_flushed(struct bnx2x *bp,
struct pbf_pN_cmd_regs *regs,
u32 poll_count)
{
u32 occup, to_free, freed, freed_start;
u32 cur_cnt = poll_count;
occup = to_free = REG_RD(bp, regs->lines_occup);
freed = freed_start = REG_RD(bp, regs->lines_freed);
DP(BNX2X_MSG_SP, "OCCUPANCY[%d] : s:%x\n", regs->pN, occup);
DP(BNX2X_MSG_SP, "LINES_FREED[%d] : s:%x\n", regs->pN, freed);
while (occup && ((u32)SUB_S32(freed, freed_start) < to_free)) {
if (cur_cnt--) {
udelay(FLR_WAIT_INTERVAL);
occup = REG_RD(bp, regs->lines_occup);
freed = REG_RD(bp, regs->lines_freed);
} else {
DP(BNX2X_MSG_SP, "PBF cmd queue[%d] timed out\n",
regs->pN);
DP(BNX2X_MSG_SP, "OCCUPANCY[%d] : s:%x\n",
regs->pN, occup);
DP(BNX2X_MSG_SP, "LINES_FREED[%d] : s:%x\n",
regs->pN, freed);
break;
}
}
DP(BNX2X_MSG_SP, "Waited %d*%d usec for PBF cmd queue[%d]\n",
poll_count-cur_cnt, FLR_WAIT_INTERVAL, regs->pN);
}
static u32 bnx2x_flr_clnup_reg_poll(struct bnx2x *bp, u32 reg,
u32 expected, u32 poll_count)
{
u32 cur_cnt = poll_count;
u32 val;
while ((val = REG_RD(bp, reg)) != expected && cur_cnt--)
udelay(FLR_WAIT_INTERVAL);
return val;
}
int bnx2x_flr_clnup_poll_hw_counter(struct bnx2x *bp, u32 reg,
char *msg, u32 poll_cnt)
{
u32 val = bnx2x_flr_clnup_reg_poll(bp, reg, 0, poll_cnt);
if (val != 0) {
BNX2X_ERR("%s usage count=%d\n", msg, val);
return 1;
}
return 0;
}
/* Common routines with VF FLR cleanup */
u32 bnx2x_flr_clnup_poll_count(struct bnx2x *bp)
{
/* adjust polling timeout */
if (CHIP_REV_IS_EMUL(bp))
return FLR_POLL_CNT * 2000;
if (CHIP_REV_IS_FPGA(bp))
return FLR_POLL_CNT * 120;
return FLR_POLL_CNT;
}
void bnx2x_tx_hw_flushed(struct bnx2x *bp, u32 poll_count)
{
struct pbf_pN_cmd_regs cmd_regs[] = {
{0, (CHIP_IS_E3B0(bp)) ?
PBF_REG_TQ_OCCUPANCY_Q0 :
PBF_REG_P0_TQ_OCCUPANCY,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_TQ_LINES_FREED_CNT_Q0 :
PBF_REG_P0_TQ_LINES_FREED_CNT},
{1, (CHIP_IS_E3B0(bp)) ?
PBF_REG_TQ_OCCUPANCY_Q1 :
PBF_REG_P1_TQ_OCCUPANCY,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_TQ_LINES_FREED_CNT_Q1 :
PBF_REG_P1_TQ_LINES_FREED_CNT},
{4, (CHIP_IS_E3B0(bp)) ?
PBF_REG_TQ_OCCUPANCY_LB_Q :
PBF_REG_P4_TQ_OCCUPANCY,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_TQ_LINES_FREED_CNT_LB_Q :
PBF_REG_P4_TQ_LINES_FREED_CNT}
};
struct pbf_pN_buf_regs buf_regs[] = {
{0, (CHIP_IS_E3B0(bp)) ?
PBF_REG_INIT_CRD_Q0 :
PBF_REG_P0_INIT_CRD ,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_CREDIT_Q0 :
PBF_REG_P0_CREDIT,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_INTERNAL_CRD_FREED_CNT_Q0 :
PBF_REG_P0_INTERNAL_CRD_FREED_CNT},
{1, (CHIP_IS_E3B0(bp)) ?
PBF_REG_INIT_CRD_Q1 :
PBF_REG_P1_INIT_CRD,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_CREDIT_Q1 :
PBF_REG_P1_CREDIT,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_INTERNAL_CRD_FREED_CNT_Q1 :
PBF_REG_P1_INTERNAL_CRD_FREED_CNT},
{4, (CHIP_IS_E3B0(bp)) ?
PBF_REG_INIT_CRD_LB_Q :
PBF_REG_P4_INIT_CRD,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_CREDIT_LB_Q :
PBF_REG_P4_CREDIT,
(CHIP_IS_E3B0(bp)) ?
PBF_REG_INTERNAL_CRD_FREED_CNT_LB_Q :
PBF_REG_P4_INTERNAL_CRD_FREED_CNT},
};
int i;
/* Verify the command queues are flushed P0, P1, P4 */
for (i = 0; i < ARRAY_SIZE(cmd_regs); i++)
bnx2x_pbf_pN_cmd_flushed(bp, &cmd_regs[i], poll_count);
/* Verify the transmission buffers are flushed P0, P1, P4 */
for (i = 0; i < ARRAY_SIZE(buf_regs); i++)
bnx2x_pbf_pN_buf_flushed(bp, &buf_regs[i], poll_count);
}
#define OP_GEN_PARAM(param) \
(((param) << SDM_OP_GEN_COMP_PARAM_SHIFT) & SDM_OP_GEN_COMP_PARAM)
#define OP_GEN_TYPE(type) \
(((type) << SDM_OP_GEN_COMP_TYPE_SHIFT) & SDM_OP_GEN_COMP_TYPE)
#define OP_GEN_AGG_VECT(index) \
(((index) << SDM_OP_GEN_AGG_VECT_IDX_SHIFT) & SDM_OP_GEN_AGG_VECT_IDX)
int bnx2x_send_final_clnup(struct bnx2x *bp, u8 clnup_func, u32 poll_cnt)
{
u32 op_gen_command = 0;
u32 comp_addr = BAR_CSTRORM_INTMEM +
CSTORM_FINAL_CLEANUP_COMPLETE_OFFSET(clnup_func);
int ret = 0;
if (REG_RD(bp, comp_addr)) {
BNX2X_ERR("Cleanup complete was not 0 before sending\n");
return 1;
}
op_gen_command |= OP_GEN_PARAM(XSTORM_AGG_INT_FINAL_CLEANUP_INDEX);
op_gen_command |= OP_GEN_TYPE(XSTORM_AGG_INT_FINAL_CLEANUP_COMP_TYPE);
op_gen_command |= OP_GEN_AGG_VECT(clnup_func);
op_gen_command |= 1 << SDM_OP_GEN_AGG_VECT_IDX_VALID_SHIFT;
DP(BNX2X_MSG_SP, "sending FW Final cleanup\n");
REG_WR(bp, XSDM_REG_OPERATION_GEN, op_gen_command);
if (bnx2x_flr_clnup_reg_poll(bp, comp_addr, 1, poll_cnt) != 1) {
BNX2X_ERR("FW final cleanup did not succeed\n");
DP(BNX2X_MSG_SP, "At timeout completion address contained %x\n",
(REG_RD(bp, comp_addr)));
bnx2x_panic();
return 1;
}
/* Zero completion for next FLR */
REG_WR(bp, comp_addr, 0);
return ret;
}
u8 bnx2x_is_pcie_pending(struct pci_dev *dev)
{
u16 status;
pcie_capability_read_word(dev, PCI_EXP_DEVSTA, &status);
return status & PCI_EXP_DEVSTA_TRPND;
}
/* PF FLR specific routines
*/
static int bnx2x_poll_hw_usage_counters(struct bnx2x *bp, u32 poll_cnt)
{
/* wait for CFC PF usage-counter to zero (includes all the VFs) */
if (bnx2x_flr_clnup_poll_hw_counter(bp,
CFC_REG_NUM_LCIDS_INSIDE_PF,
"CFC PF usage counter timed out",
poll_cnt))
return 1;
/* Wait for DQ PF usage-counter to zero (until DQ cleanup) */
if (bnx2x_flr_clnup_poll_hw_counter(bp,
DORQ_REG_PF_USAGE_CNT,
"DQ PF usage counter timed out",
poll_cnt))
return 1;
/* Wait for QM PF usage-counter to zero (until DQ cleanup) */
if (bnx2x_flr_clnup_poll_hw_counter(bp,
QM_REG_PF_USG_CNT_0 + 4*BP_FUNC(bp),
"QM PF usage counter timed out",
poll_cnt))
return 1;
/* Wait for Timer PF usage-counters to zero (until DQ cleanup) */
if (bnx2x_flr_clnup_poll_hw_counter(bp,
TM_REG_LIN0_VNIC_UC + 4*BP_PORT(bp),
"Timers VNIC usage counter timed out",
poll_cnt))
return 1;
if (bnx2x_flr_clnup_poll_hw_counter(bp,
TM_REG_LIN0_NUM_SCANS + 4*BP_PORT(bp),
"Timers NUM_SCANS usage counter timed out",
poll_cnt))
return 1;
/* Wait DMAE PF usage counter to zero */
if (bnx2x_flr_clnup_poll_hw_counter(bp,
dmae_reg_go_c[INIT_DMAE_C(bp)],
"DMAE command register timed out",
poll_cnt))
return 1;
return 0;
}
static void bnx2x_hw_enable_status(struct bnx2x *bp)
{
u32 val;
val = REG_RD(bp, CFC_REG_WEAK_ENABLE_PF);
DP(BNX2X_MSG_SP, "CFC_REG_WEAK_ENABLE_PF is 0x%x\n", val);
val = REG_RD(bp, PBF_REG_DISABLE_PF);
DP(BNX2X_MSG_SP, "PBF_REG_DISABLE_PF is 0x%x\n", val);
val = REG_RD(bp, IGU_REG_PCI_PF_MSI_EN);
DP(BNX2X_MSG_SP, "IGU_REG_PCI_PF_MSI_EN is 0x%x\n", val);
val = REG_RD(bp, IGU_REG_PCI_PF_MSIX_EN);
DP(BNX2X_MSG_SP, "IGU_REG_PCI_PF_MSIX_EN is 0x%x\n", val);
val = REG_RD(bp, IGU_REG_PCI_PF_MSIX_FUNC_MASK);
DP(BNX2X_MSG_SP, "IGU_REG_PCI_PF_MSIX_FUNC_MASK is 0x%x\n", val);
val = REG_RD(bp, PGLUE_B_REG_SHADOW_BME_PF_7_0_CLR);
DP(BNX2X_MSG_SP, "PGLUE_B_REG_SHADOW_BME_PF_7_0_CLR is 0x%x\n", val);
val = REG_RD(bp, PGLUE_B_REG_FLR_REQUEST_PF_7_0_CLR);
DP(BNX2X_MSG_SP, "PGLUE_B_REG_FLR_REQUEST_PF_7_0_CLR is 0x%x\n", val);
val = REG_RD(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER);
DP(BNX2X_MSG_SP, "PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER is 0x%x\n",
val);
}
static int bnx2x_pf_flr_clnup(struct bnx2x *bp)
{
u32 poll_cnt = bnx2x_flr_clnup_poll_count(bp);
DP(BNX2X_MSG_SP, "Cleanup after FLR PF[%d]\n", BP_ABS_FUNC(bp));
/* Re-enable PF target read access */
REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_TARGET_READ, 1);
/* Poll HW usage counters */
DP(BNX2X_MSG_SP, "Polling usage counters\n");
if (bnx2x_poll_hw_usage_counters(bp, poll_cnt))
return -EBUSY;
/* Zero the igu 'trailing edge' and 'leading edge' */
/* Send the FW cleanup command */
if (bnx2x_send_final_clnup(bp, (u8)BP_FUNC(bp), poll_cnt))
return -EBUSY;
/* ATC cleanup */
/* Verify TX hw is flushed */
bnx2x_tx_hw_flushed(bp, poll_cnt);
/* Wait 100ms (not adjusted according to platform) */
msleep(100);
/* Verify no pending pci transactions */
if (bnx2x_is_pcie_pending(bp->pdev))
BNX2X_ERR("PCIE Transactions still pending\n");
/* Debug */
bnx2x_hw_enable_status(bp);
/*
* Master enable - Due to WB DMAE writes performed before this
* register is re-initialized as part of the regular function init
*/
REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, 1);
return 0;
}
static void bnx2x_hc_int_enable(struct bnx2x *bp)
{
int port = BP_PORT(bp);
u32 addr = port ? HC_REG_CONFIG_1 : HC_REG_CONFIG_0;
u32 val = REG_RD(bp, addr);
bool msix = (bp->flags & USING_MSIX_FLAG) ? true : false;
bool single_msix = (bp->flags & USING_SINGLE_MSIX_FLAG) ? true : false;
bool msi = (bp->flags & USING_MSI_FLAG) ? true : false;
if (msix) {
val &= ~(HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
HC_CONFIG_0_REG_INT_LINE_EN_0);
val |= (HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0 |
HC_CONFIG_0_REG_ATTN_BIT_EN_0);
if (single_msix)
val |= HC_CONFIG_0_REG_SINGLE_ISR_EN_0;
} else if (msi) {
val &= ~HC_CONFIG_0_REG_INT_LINE_EN_0;
val |= (HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0 |
HC_CONFIG_0_REG_ATTN_BIT_EN_0);
} else {
val |= (HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0 |
HC_CONFIG_0_REG_INT_LINE_EN_0 |
HC_CONFIG_0_REG_ATTN_BIT_EN_0);
if (!CHIP_IS_E1(bp)) {
DP(NETIF_MSG_IFUP,
"write %x to HC %d (addr 0x%x)\n", val, port, addr);
REG_WR(bp, addr, val);
val &= ~HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0;
}
}
if (CHIP_IS_E1(bp))
REG_WR(bp, HC_REG_INT_MASK + port*4, 0x1FFFF);
DP(NETIF_MSG_IFUP,
"write %x to HC %d (addr 0x%x) mode %s\n", val, port, addr,
(msix ? "MSI-X" : (msi ? "MSI" : "INTx")));
REG_WR(bp, addr, val);
/*
* Ensure that HC_CONFIG is written before leading/trailing edge config
*/
mmiowb();
barrier();
if (!CHIP_IS_E1(bp)) {
/* init leading/trailing edge */
if (IS_MF(bp)) {
val = (0xee0f | (1 << (BP_VN(bp) + 4)));
if (bp->port.pmf)
/* enable nig and gpio3 attention */
val |= 0x1100;
} else
val = 0xffff;
REG_WR(bp, HC_REG_TRAILING_EDGE_0 + port*8, val);
REG_WR(bp, HC_REG_LEADING_EDGE_0 + port*8, val);
}
/* Make sure that interrupts are indeed enabled from here on */
mmiowb();
}
static void bnx2x_igu_int_enable(struct bnx2x *bp)
{
u32 val;
bool msix = (bp->flags & USING_MSIX_FLAG) ? true : false;
bool single_msix = (bp->flags & USING_SINGLE_MSIX_FLAG) ? true : false;
bool msi = (bp->flags & USING_MSI_FLAG) ? true : false;
val = REG_RD(bp, IGU_REG_PF_CONFIGURATION);
if (msix) {
val &= ~(IGU_PF_CONF_INT_LINE_EN |
IGU_PF_CONF_SINGLE_ISR_EN);
val |= (IGU_PF_CONF_MSI_MSIX_EN |
IGU_PF_CONF_ATTN_BIT_EN);
if (single_msix)
val |= IGU_PF_CONF_SINGLE_ISR_EN;
} else if (msi) {
val &= ~IGU_PF_CONF_INT_LINE_EN;
val |= (IGU_PF_CONF_MSI_MSIX_EN |
IGU_PF_CONF_ATTN_BIT_EN |
IGU_PF_CONF_SINGLE_ISR_EN);
} else {
val &= ~IGU_PF_CONF_MSI_MSIX_EN;
val |= (IGU_PF_CONF_INT_LINE_EN |
IGU_PF_CONF_ATTN_BIT_EN |
IGU_PF_CONF_SINGLE_ISR_EN);
}
/* Clean previous status - need to configure igu prior to ack*/
if ((!msix) || single_msix) {
REG_WR(bp, IGU_REG_PF_CONFIGURATION, val);
bnx2x_ack_int(bp);
}
val |= IGU_PF_CONF_FUNC_EN;
DP(NETIF_MSG_IFUP, "write 0x%x to IGU mode %s\n",
val, (msix ? "MSI-X" : (msi ? "MSI" : "INTx")));
REG_WR(bp, IGU_REG_PF_CONFIGURATION, val);
if (val & IGU_PF_CONF_INT_LINE_EN)
pci_intx(bp->pdev, true);
barrier();
/* init leading/trailing edge */
if (IS_MF(bp)) {
val = (0xee0f | (1 << (BP_VN(bp) + 4)));
if (bp->port.pmf)
/* enable nig and gpio3 attention */
val |= 0x1100;
} else
val = 0xffff;
REG_WR(bp, IGU_REG_TRAILING_EDGE_LATCH, val);
REG_WR(bp, IGU_REG_LEADING_EDGE_LATCH, val);
/* Make sure that interrupts are indeed enabled from here on */
mmiowb();
}
void bnx2x_int_enable(struct bnx2x *bp)
{
if (bp->common.int_block == INT_BLOCK_HC)
bnx2x_hc_int_enable(bp);
else
bnx2x_igu_int_enable(bp);
}
void bnx2x_int_disable_sync(struct bnx2x *bp, int disable_hw)
{
int msix = (bp->flags & USING_MSIX_FLAG) ? 1 : 0;
int i, offset;
if (disable_hw)
/* prevent the HW from sending interrupts */
bnx2x_int_disable(bp);
/* make sure all ISRs are done */
if (msix) {
synchronize_irq(bp->msix_table[0].vector);
offset = 1;
if (CNIC_SUPPORT(bp))
offset++;
for_each_eth_queue(bp, i)
synchronize_irq(bp->msix_table[offset++].vector);
} else
synchronize_irq(bp->pdev->irq);
/* make sure sp_task is not running */
cancel_delayed_work(&bp->sp_task);
cancel_delayed_work(&bp->period_task);
flush_workqueue(bnx2x_wq);
}
/* fast path */
/*
* General service functions
*/
/* Return true if succeeded to acquire the lock */
static bool bnx2x_trylock_hw_lock(struct bnx2x *bp, u32 resource)
{
u32 lock_status;
u32 resource_bit = (1 << resource);
int func = BP_FUNC(bp);
u32 hw_lock_control_reg;
DP(NETIF_MSG_HW | NETIF_MSG_IFUP,
"Trying to take a lock on resource %d\n", resource);
/* Validating that the resource is within range */
if (resource > HW_LOCK_MAX_RESOURCE_VALUE) {
DP(NETIF_MSG_HW | NETIF_MSG_IFUP,
"resource(0x%x) > HW_LOCK_MAX_RESOURCE_VALUE(0x%x)\n",
resource, HW_LOCK_MAX_RESOURCE_VALUE);
return false;
}
if (func <= 5)
hw_lock_control_reg = (MISC_REG_DRIVER_CONTROL_1 + func*8);
else
hw_lock_control_reg =
(MISC_REG_DRIVER_CONTROL_7 + (func - 6)*8);
/* Try to acquire the lock */
REG_WR(bp, hw_lock_control_reg + 4, resource_bit);
lock_status = REG_RD(bp, hw_lock_control_reg);
if (lock_status & resource_bit)
return true;
DP(NETIF_MSG_HW | NETIF_MSG_IFUP,
"Failed to get a lock on resource %d\n", resource);
return false;
}
/**
* bnx2x_get_leader_lock_resource - get the recovery leader resource id
*
* @bp: driver handle
*
* Returns the recovery leader resource id according to the engine this function
* belongs to. Currently only only 2 engines is supported.
*/
static int bnx2x_get_leader_lock_resource(struct bnx2x *bp)
{
if (BP_PATH(bp))
return HW_LOCK_RESOURCE_RECOVERY_LEADER_1;
else
return HW_LOCK_RESOURCE_RECOVERY_LEADER_0;
}
/**
* bnx2x_trylock_leader_lock- try to acquire a leader lock.
*
* @bp: driver handle
*
* Tries to acquire a leader lock for current engine.
*/
static bool bnx2x_trylock_leader_lock(struct bnx2x *bp)
{
return bnx2x_trylock_hw_lock(bp, bnx2x_get_leader_lock_resource(bp));
}
static void bnx2x_cnic_cfc_comp(struct bnx2x *bp, int cid, u8 err);
/* schedule the sp task and mark that interrupt occurred (runs from ISR) */
static int bnx2x_schedule_sp_task(struct bnx2x *bp)
{
/* Set the interrupt occurred bit for the sp-task to recognize it
* must ack the interrupt and transition according to the IGU
* state machine.
*/
atomic_set(&bp->interrupt_occurred, 1);
/* The sp_task must execute only after this bit
* is set, otherwise we will get out of sync and miss all
* further interrupts. Hence, the barrier.
*/
smp_wmb();
/* schedule sp_task to workqueue */
return queue_delayed_work(bnx2x_wq, &bp->sp_task, 0);
}
void bnx2x_sp_event(struct bnx2x_fastpath *fp, union eth_rx_cqe *rr_cqe)
{
struct bnx2x *bp = fp->bp;
int cid = SW_CID(rr_cqe->ramrod_cqe.conn_and_cmd_data);
int command = CQE_CMD(rr_cqe->ramrod_cqe.conn_and_cmd_data);
enum bnx2x_queue_cmd drv_cmd = BNX2X_Q_CMD_MAX;
struct bnx2x_queue_sp_obj *q_obj = &bnx2x_sp_obj(bp, fp).q_obj;
DP(BNX2X_MSG_SP,
"fp %d cid %d got ramrod #%d state is %x type is %d\n",
fp->index, cid, command, bp->state,
rr_cqe->ramrod_cqe.ramrod_type);
/* If cid is within VF range, replace the slowpath object with the
* one corresponding to this VF
*/
if (cid >= BNX2X_FIRST_VF_CID &&
cid < BNX2X_FIRST_VF_CID + BNX2X_VF_CIDS)
bnx2x_iov_set_queue_sp_obj(bp, cid, &q_obj);
switch (command) {
case (RAMROD_CMD_ID_ETH_CLIENT_UPDATE):
DP(BNX2X_MSG_SP, "got UPDATE ramrod. CID %d\n", cid);
drv_cmd = BNX2X_Q_CMD_UPDATE;
break;
case (RAMROD_CMD_ID_ETH_CLIENT_SETUP):
DP(BNX2X_MSG_SP, "got MULTI[%d] setup ramrod\n", cid);
drv_cmd = BNX2X_Q_CMD_SETUP;
break;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
case (RAMROD_CMD_ID_ETH_TX_QUEUE_SETUP):
DP(BNX2X_MSG_SP, "got MULTI[%d] tx-only setup ramrod\n", cid);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
drv_cmd = BNX2X_Q_CMD_SETUP_TX_ONLY;
break;
case (RAMROD_CMD_ID_ETH_HALT):
DP(BNX2X_MSG_SP, "got MULTI[%d] halt ramrod\n", cid);
drv_cmd = BNX2X_Q_CMD_HALT;
break;
case (RAMROD_CMD_ID_ETH_TERMINATE):
DP(BNX2X_MSG_SP, "got MULTI[%d] terminate ramrod\n", cid);
drv_cmd = BNX2X_Q_CMD_TERMINATE;
break;
case (RAMROD_CMD_ID_ETH_EMPTY):
DP(BNX2X_MSG_SP, "got MULTI[%d] empty ramrod\n", cid);
drv_cmd = BNX2X_Q_CMD_EMPTY;
break;
case (RAMROD_CMD_ID_ETH_TPA_UPDATE):
DP(BNX2X_MSG_SP, "got tpa update ramrod CID=%d\n", cid);
drv_cmd = BNX2X_Q_CMD_UPDATE_TPA;
break;
default:
BNX2X_ERR("unexpected MC reply (%d) on fp[%d]\n",
command, fp->index);
return;
}
if ((drv_cmd != BNX2X_Q_CMD_MAX) &&
q_obj->complete_cmd(bp, q_obj, drv_cmd))
/* q_obj->complete_cmd() failure means that this was
* an unexpected completion.
*
* In this case we don't want to increase the bp->spq_left
* because apparently we haven't sent this command the first
* place.
*/
#ifdef BNX2X_STOP_ON_ERROR
bnx2x_panic();
#else
return;
#endif
smp_mb__before_atomic();
atomic_inc(&bp->cq_spq_left);
/* push the change in bp->spq_left and towards the memory */
smp_mb__after_atomic();
DP(BNX2X_MSG_SP, "bp->cq_spq_left %x\n", atomic_read(&bp->cq_spq_left));
if ((drv_cmd == BNX2X_Q_CMD_UPDATE) && (IS_FCOE_FP(fp)) &&
(!!test_bit(BNX2X_AFEX_FCOE_Q_UPDATE_PENDING, &bp->sp_state))) {
/* if Q update ramrod is completed for last Q in AFEX vif set
* flow, then ACK MCP at the end
*
* mark pending ACK to MCP bit.
* prevent case that both bits are cleared.
* At the end of load/unload driver checks that
* sp_state is cleared, and this order prevents
* races
*/
smp_mb__before_atomic();
set_bit(BNX2X_AFEX_PENDING_VIFSET_MCP_ACK, &bp->sp_state);
wmb();
clear_bit(BNX2X_AFEX_FCOE_Q_UPDATE_PENDING, &bp->sp_state);
smp_mb__after_atomic();
/* schedule the sp task as mcp ack is required */
bnx2x_schedule_sp_task(bp);
}
return;
}
irqreturn_t bnx2x_interrupt(int irq, void *dev_instance)
{
struct bnx2x *bp = netdev_priv(dev_instance);
u16 status = bnx2x_ack_int(bp);
u16 mask;
int i;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 cos;
/* Return here if interrupt is shared and it's not for us */
if (unlikely(status == 0)) {
DP(NETIF_MSG_INTR, "not our interrupt!\n");
return IRQ_NONE;
}
DP(NETIF_MSG_INTR, "got an interrupt status 0x%x\n", status);
#ifdef BNX2X_STOP_ON_ERROR
if (unlikely(bp->panic))
return IRQ_HANDLED;
#endif
for_each_eth_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
mask = 0x2 << (fp->index + CNIC_SUPPORT(bp));
if (status & mask) {
/* Handle Rx or Tx according to SB id */
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
for_each_cos_in_tx_queue(fp, cos)
prefetch(fp->txdata_ptr[cos]->tx_cons_sb);
prefetch(&fp->sb_running_index[SM_RX_ID]);
napi_schedule_irqoff(&bnx2x_fp(bp, fp->index, napi));
status &= ~mask;
}
}
if (CNIC_SUPPORT(bp)) {
mask = 0x2;
if (status & (mask | 0x1)) {
struct cnic_ops *c_ops = NULL;
rcu_read_lock();
c_ops = rcu_dereference(bp->cnic_ops);
if (c_ops && (bp->cnic_eth_dev.drv_state &
CNIC_DRV_STATE_HANDLES_IRQ))
c_ops->cnic_handler(bp->cnic_data, NULL);
rcu_read_unlock();
status &= ~mask;
}
}
if (unlikely(status & 0x1)) {
/* schedule sp task to perform default status block work, ack
* attentions and enable interrupts.
*/
bnx2x_schedule_sp_task(bp);
status &= ~0x1;
if (!status)
return IRQ_HANDLED;
}
if (unlikely(status))
DP(NETIF_MSG_INTR, "got an unknown interrupt! (status 0x%x)\n",
status);
return IRQ_HANDLED;
}
/* Link */
/*
* General service functions
*/
int bnx2x_acquire_hw_lock(struct bnx2x *bp, u32 resource)
{
u32 lock_status;
u32 resource_bit = (1 << resource);
int func = BP_FUNC(bp);
u32 hw_lock_control_reg;
int cnt;
/* Validating that the resource is within range */
if (resource > HW_LOCK_MAX_RESOURCE_VALUE) {
BNX2X_ERR("resource(0x%x) > HW_LOCK_MAX_RESOURCE_VALUE(0x%x)\n",
resource, HW_LOCK_MAX_RESOURCE_VALUE);
return -EINVAL;
}
if (func <= 5) {
hw_lock_control_reg = (MISC_REG_DRIVER_CONTROL_1 + func*8);
} else {
hw_lock_control_reg =
(MISC_REG_DRIVER_CONTROL_7 + (func - 6)*8);
}
/* Validating that the resource is not already taken */
lock_status = REG_RD(bp, hw_lock_control_reg);
if (lock_status & resource_bit) {
BNX2X_ERR("lock_status 0x%x resource_bit 0x%x\n",
lock_status, resource_bit);
return -EEXIST;
}
/* Try for 5 second every 5ms */
for (cnt = 0; cnt < 1000; cnt++) {
/* Try to acquire the lock */
REG_WR(bp, hw_lock_control_reg + 4, resource_bit);
lock_status = REG_RD(bp, hw_lock_control_reg);
if (lock_status & resource_bit)
return 0;
usleep_range(5000, 10000);
}
BNX2X_ERR("Timeout\n");
return -EAGAIN;
}
int bnx2x_release_leader_lock(struct bnx2x *bp)
{
return bnx2x_release_hw_lock(bp, bnx2x_get_leader_lock_resource(bp));
}
int bnx2x_release_hw_lock(struct bnx2x *bp, u32 resource)
{
u32 lock_status;
u32 resource_bit = (1 << resource);
int func = BP_FUNC(bp);
u32 hw_lock_control_reg;
/* Validating that the resource is within range */
if (resource > HW_LOCK_MAX_RESOURCE_VALUE) {
BNX2X_ERR("resource(0x%x) > HW_LOCK_MAX_RESOURCE_VALUE(0x%x)\n",
resource, HW_LOCK_MAX_RESOURCE_VALUE);
return -EINVAL;
}
if (func <= 5) {
hw_lock_control_reg = (MISC_REG_DRIVER_CONTROL_1 + func*8);
} else {
hw_lock_control_reg =
(MISC_REG_DRIVER_CONTROL_7 + (func - 6)*8);
}
/* Validating that the resource is currently taken */
lock_status = REG_RD(bp, hw_lock_control_reg);
if (!(lock_status & resource_bit)) {
BNX2X_ERR("lock_status 0x%x resource_bit 0x%x. Unlock was called but lock wasn't taken!\n",
lock_status, resource_bit);
return -EFAULT;
}
REG_WR(bp, hw_lock_control_reg, resource_bit);
return 0;
}
int bnx2x_get_gpio(struct bnx2x *bp, int gpio_num, u8 port)
{
/* The GPIO should be swapped if swap register is set and active */
int gpio_port = (REG_RD(bp, NIG_REG_PORT_SWAP) &&
REG_RD(bp, NIG_REG_STRAP_OVERRIDE)) ^ port;
int gpio_shift = gpio_num +
(gpio_port ? MISC_REGISTERS_GPIO_PORT_SHIFT : 0);
u32 gpio_mask = (1 << gpio_shift);
u32 gpio_reg;
int value;
if (gpio_num > MISC_REGISTERS_GPIO_3) {
BNX2X_ERR("Invalid GPIO %d\n", gpio_num);
return -EINVAL;
}
/* read GPIO value */
gpio_reg = REG_RD(bp, MISC_REG_GPIO);
/* get the requested pin value */
if ((gpio_reg & gpio_mask) == gpio_mask)
value = 1;
else
value = 0;
return value;
}
int bnx2x_set_gpio(struct bnx2x *bp, int gpio_num, u32 mode, u8 port)
{
/* The GPIO should be swapped if swap register is set and active */
int gpio_port = (REG_RD(bp, NIG_REG_PORT_SWAP) &&
REG_RD(bp, NIG_REG_STRAP_OVERRIDE)) ^ port;
int gpio_shift = gpio_num +
(gpio_port ? MISC_REGISTERS_GPIO_PORT_SHIFT : 0);
u32 gpio_mask = (1 << gpio_shift);
u32 gpio_reg;
if (gpio_num > MISC_REGISTERS_GPIO_3) {
BNX2X_ERR("Invalid GPIO %d\n", gpio_num);
return -EINVAL;
}
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_GPIO);
/* read GPIO and mask except the float bits */
gpio_reg = (REG_RD(bp, MISC_REG_GPIO) & MISC_REGISTERS_GPIO_FLOAT);
switch (mode) {
case MISC_REGISTERS_GPIO_OUTPUT_LOW:
DP(NETIF_MSG_LINK,
"Set GPIO %d (shift %d) -> output low\n",
gpio_num, gpio_shift);
/* clear FLOAT and set CLR */
gpio_reg &= ~(gpio_mask << MISC_REGISTERS_GPIO_FLOAT_POS);
gpio_reg |= (gpio_mask << MISC_REGISTERS_GPIO_CLR_POS);
break;
case MISC_REGISTERS_GPIO_OUTPUT_HIGH:
DP(NETIF_MSG_LINK,
"Set GPIO %d (shift %d) -> output high\n",
gpio_num, gpio_shift);
/* clear FLOAT and set SET */
gpio_reg &= ~(gpio_mask << MISC_REGISTERS_GPIO_FLOAT_POS);
gpio_reg |= (gpio_mask << MISC_REGISTERS_GPIO_SET_POS);
break;
case MISC_REGISTERS_GPIO_INPUT_HI_Z:
DP(NETIF_MSG_LINK,
"Set GPIO %d (shift %d) -> input\n",
gpio_num, gpio_shift);
/* set FLOAT */
gpio_reg |= (gpio_mask << MISC_REGISTERS_GPIO_FLOAT_POS);
break;
default:
break;
}
REG_WR(bp, MISC_REG_GPIO, gpio_reg);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_GPIO);
return 0;
}
int bnx2x_set_mult_gpio(struct bnx2x *bp, u8 pins, u32 mode)
{
u32 gpio_reg = 0;
int rc = 0;
/* Any port swapping should be handled by caller. */
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_GPIO);
/* read GPIO and mask except the float bits */
gpio_reg = REG_RD(bp, MISC_REG_GPIO);
gpio_reg &= ~(pins << MISC_REGISTERS_GPIO_FLOAT_POS);
gpio_reg &= ~(pins << MISC_REGISTERS_GPIO_CLR_POS);
gpio_reg &= ~(pins << MISC_REGISTERS_GPIO_SET_POS);
switch (mode) {
case MISC_REGISTERS_GPIO_OUTPUT_LOW:
DP(NETIF_MSG_LINK, "Set GPIO 0x%x -> output low\n", pins);
/* set CLR */
gpio_reg |= (pins << MISC_REGISTERS_GPIO_CLR_POS);
break;
case MISC_REGISTERS_GPIO_OUTPUT_HIGH:
DP(NETIF_MSG_LINK, "Set GPIO 0x%x -> output high\n", pins);
/* set SET */
gpio_reg |= (pins << MISC_REGISTERS_GPIO_SET_POS);
break;
case MISC_REGISTERS_GPIO_INPUT_HI_Z:
DP(NETIF_MSG_LINK, "Set GPIO 0x%x -> input\n", pins);
/* set FLOAT */
gpio_reg |= (pins << MISC_REGISTERS_GPIO_FLOAT_POS);
break;
default:
BNX2X_ERR("Invalid GPIO mode assignment %d\n", mode);
rc = -EINVAL;
break;
}
if (rc == 0)
REG_WR(bp, MISC_REG_GPIO, gpio_reg);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_GPIO);
return rc;
}
int bnx2x_set_gpio_int(struct bnx2x *bp, int gpio_num, u32 mode, u8 port)
{
/* The GPIO should be swapped if swap register is set and active */
int gpio_port = (REG_RD(bp, NIG_REG_PORT_SWAP) &&
REG_RD(bp, NIG_REG_STRAP_OVERRIDE)) ^ port;
int gpio_shift = gpio_num +
(gpio_port ? MISC_REGISTERS_GPIO_PORT_SHIFT : 0);
u32 gpio_mask = (1 << gpio_shift);
u32 gpio_reg;
if (gpio_num > MISC_REGISTERS_GPIO_3) {
BNX2X_ERR("Invalid GPIO %d\n", gpio_num);
return -EINVAL;
}
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_GPIO);
/* read GPIO int */
gpio_reg = REG_RD(bp, MISC_REG_GPIO_INT);
switch (mode) {
case MISC_REGISTERS_GPIO_INT_OUTPUT_CLR:
DP(NETIF_MSG_LINK,
"Clear GPIO INT %d (shift %d) -> output low\n",
gpio_num, gpio_shift);
/* clear SET and set CLR */
gpio_reg &= ~(gpio_mask << MISC_REGISTERS_GPIO_INT_SET_POS);
gpio_reg |= (gpio_mask << MISC_REGISTERS_GPIO_INT_CLR_POS);
break;
case MISC_REGISTERS_GPIO_INT_OUTPUT_SET:
DP(NETIF_MSG_LINK,
"Set GPIO INT %d (shift %d) -> output high\n",
gpio_num, gpio_shift);
/* clear CLR and set SET */
gpio_reg &= ~(gpio_mask << MISC_REGISTERS_GPIO_INT_CLR_POS);
gpio_reg |= (gpio_mask << MISC_REGISTERS_GPIO_INT_SET_POS);
break;
default:
break;
}
REG_WR(bp, MISC_REG_GPIO_INT, gpio_reg);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_GPIO);
return 0;
}
static int bnx2x_set_spio(struct bnx2x *bp, int spio, u32 mode)
{
u32 spio_reg;
/* Only 2 SPIOs are configurable */
if ((spio != MISC_SPIO_SPIO4) && (spio != MISC_SPIO_SPIO5)) {
BNX2X_ERR("Invalid SPIO 0x%x\n", spio);
return -EINVAL;
}
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_SPIO);
/* read SPIO and mask except the float bits */
spio_reg = (REG_RD(bp, MISC_REG_SPIO) & MISC_SPIO_FLOAT);
switch (mode) {
case MISC_SPIO_OUTPUT_LOW:
DP(NETIF_MSG_HW, "Set SPIO 0x%x -> output low\n", spio);
/* clear FLOAT and set CLR */
spio_reg &= ~(spio << MISC_SPIO_FLOAT_POS);
spio_reg |= (spio << MISC_SPIO_CLR_POS);
break;
case MISC_SPIO_OUTPUT_HIGH:
DP(NETIF_MSG_HW, "Set SPIO 0x%x -> output high\n", spio);
/* clear FLOAT and set SET */
spio_reg &= ~(spio << MISC_SPIO_FLOAT_POS);
spio_reg |= (spio << MISC_SPIO_SET_POS);
break;
case MISC_SPIO_INPUT_HI_Z:
DP(NETIF_MSG_HW, "Set SPIO 0x%x -> input\n", spio);
/* set FLOAT */
spio_reg |= (spio << MISC_SPIO_FLOAT_POS);
break;
default:
break;
}
REG_WR(bp, MISC_REG_SPIO, spio_reg);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_SPIO);
return 0;
}
void bnx2x_calc_fc_adv(struct bnx2x *bp)
{
u8 cfg_idx = bnx2x_get_link_cfg_idx(bp);
bp->port.advertising[cfg_idx] &= ~(ADVERTISED_Asym_Pause |
ADVERTISED_Pause);
switch (bp->link_vars.ieee_fc &
MDIO_COMBO_IEEE0_AUTO_NEG_ADV_PAUSE_MASK) {
case MDIO_COMBO_IEEE0_AUTO_NEG_ADV_PAUSE_BOTH:
bp->port.advertising[cfg_idx] |= (ADVERTISED_Asym_Pause |
ADVERTISED_Pause);
break;
case MDIO_COMBO_IEEE0_AUTO_NEG_ADV_PAUSE_ASYMMETRIC:
bp->port.advertising[cfg_idx] |= ADVERTISED_Asym_Pause;
break;
default:
break;
}
}
static void bnx2x_set_requested_fc(struct bnx2x *bp)
{
/* Initialize link parameters structure variables
* It is recommended to turn off RX FC for jumbo frames
* for better performance
*/
if (CHIP_IS_E1x(bp) && (bp->dev->mtu > 5000))
bp->link_params.req_fc_auto_adv = BNX2X_FLOW_CTRL_TX;
else
bp->link_params.req_fc_auto_adv = BNX2X_FLOW_CTRL_BOTH;
}
static void bnx2x_init_dropless_fc(struct bnx2x *bp)
{
u32 pause_enabled = 0;
if (!CHIP_IS_E1(bp) && bp->dropless_fc && bp->link_vars.link_up) {
if (bp->link_vars.flow_ctrl & BNX2X_FLOW_CTRL_TX)
pause_enabled = 1;
REG_WR(bp, BAR_USTRORM_INTMEM +
USTORM_ETH_PAUSE_ENABLED_OFFSET(BP_PORT(bp)),
pause_enabled);
}
DP(NETIF_MSG_IFUP | NETIF_MSG_LINK, "dropless_fc is %s\n",
pause_enabled ? "enabled" : "disabled");
}
int bnx2x_initial_phy_init(struct bnx2x *bp, int load_mode)
{
int rc, cfx_idx = bnx2x_get_link_cfg_idx(bp);
u16 req_line_speed = bp->link_params.req_line_speed[cfx_idx];
if (!BP_NOMCP(bp)) {
bnx2x_set_requested_fc(bp);
bnx2x_acquire_phy_lock(bp);
if (load_mode == LOAD_DIAG) {
struct link_params *lp = &bp->link_params;
lp->loopback_mode = LOOPBACK_XGXS;
/* Prefer doing PHY loopback at highest speed */
if (lp->req_line_speed[cfx_idx] < SPEED_20000) {
if (lp->speed_cap_mask[cfx_idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_20G)
lp->req_line_speed[cfx_idx] =
SPEED_20000;
else if (lp->speed_cap_mask[cfx_idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_10G)
lp->req_line_speed[cfx_idx] =
SPEED_10000;
else
lp->req_line_speed[cfx_idx] =
SPEED_1000;
}
}
if (load_mode == LOAD_LOOPBACK_EXT) {
struct link_params *lp = &bp->link_params;
lp->loopback_mode = LOOPBACK_EXT;
}
rc = bnx2x_phy_init(&bp->link_params, &bp->link_vars);
bnx2x_release_phy_lock(bp);
bnx2x_init_dropless_fc(bp);
bnx2x_calc_fc_adv(bp);
if (bp->link_vars.link_up) {
bnx2x_stats_handle(bp, STATS_EVENT_LINK_UP);
bnx2x_link_report(bp);
}
queue_delayed_work(bnx2x_wq, &bp->period_task, 0);
bp->link_params.req_line_speed[cfx_idx] = req_line_speed;
return rc;
}
BNX2X_ERR("Bootcode is missing - can not initialize link\n");
return -EINVAL;
}
void bnx2x_link_set(struct bnx2x *bp)
{
if (!BP_NOMCP(bp)) {
bnx2x_acquire_phy_lock(bp);
bnx2x_phy_init(&bp->link_params, &bp->link_vars);
bnx2x_release_phy_lock(bp);
bnx2x_init_dropless_fc(bp);
bnx2x_calc_fc_adv(bp);
} else
BNX2X_ERR("Bootcode is missing - can not set link\n");
}
static void bnx2x__link_reset(struct bnx2x *bp)
{
if (!BP_NOMCP(bp)) {
bnx2x_acquire_phy_lock(bp);
bnx2x_lfa_reset(&bp->link_params, &bp->link_vars);
bnx2x_release_phy_lock(bp);
} else
BNX2X_ERR("Bootcode is missing - can not reset link\n");
}
void bnx2x_force_link_reset(struct bnx2x *bp)
{
bnx2x_acquire_phy_lock(bp);
bnx2x_link_reset(&bp->link_params, &bp->link_vars, 1);
bnx2x_release_phy_lock(bp);
}
u8 bnx2x_link_test(struct bnx2x *bp, u8 is_serdes)
{
u8 rc = 0;
if (!BP_NOMCP(bp)) {
bnx2x_acquire_phy_lock(bp);
rc = bnx2x_test_link(&bp->link_params, &bp->link_vars,
is_serdes);
bnx2x_release_phy_lock(bp);
} else
BNX2X_ERR("Bootcode is missing - can not test link\n");
return rc;
}
/* Calculates the sum of vn_min_rates.
It's needed for further normalizing of the min_rates.
Returns:
sum of vn_min_rates.
or
0 - if all the min_rates are 0.
In the later case fairness algorithm should be deactivated.
If not all min_rates are zero then those that are zeroes will be set to 1.
*/
static void bnx2x_calc_vn_min(struct bnx2x *bp,
struct cmng_init_input *input)
{
int all_zero = 1;
int vn;
for (vn = VN_0; vn < BP_MAX_VN_NUM(bp); vn++) {
u32 vn_cfg = bp->mf_config[vn];
u32 vn_min_rate = ((vn_cfg & FUNC_MF_CFG_MIN_BW_MASK) >>
FUNC_MF_CFG_MIN_BW_SHIFT) * 100;
/* Skip hidden vns */
if (vn_cfg & FUNC_MF_CFG_FUNC_HIDE)
vn_min_rate = 0;
/* If min rate is zero - set it to 1 */
else if (!vn_min_rate)
vn_min_rate = DEF_MIN_RATE;
else
all_zero = 0;
input->vnic_min_rate[vn] = vn_min_rate;
}
/* if ETS or all min rates are zeros - disable fairness */
if (BNX2X_IS_ETS_ENABLED(bp)) {
input->flags.cmng_enables &=
~CMNG_FLAGS_PER_PORT_FAIRNESS_VN;
DP(NETIF_MSG_IFUP, "Fairness will be disabled due to ETS\n");
} else if (all_zero) {
input->flags.cmng_enables &=
~CMNG_FLAGS_PER_PORT_FAIRNESS_VN;
DP(NETIF_MSG_IFUP,
"All MIN values are zeroes fairness will be disabled\n");
} else
input->flags.cmng_enables |=
CMNG_FLAGS_PER_PORT_FAIRNESS_VN;
}
static void bnx2x_calc_vn_max(struct bnx2x *bp, int vn,
struct cmng_init_input *input)
{
u16 vn_max_rate;
u32 vn_cfg = bp->mf_config[vn];
if (vn_cfg & FUNC_MF_CFG_FUNC_HIDE)
vn_max_rate = 0;
else {
u32 maxCfg = bnx2x_extract_max_cfg(bp, vn_cfg);
if (IS_MF_PERCENT_BW(bp)) {
/* maxCfg in percents of linkspeed */
vn_max_rate = (bp->link_vars.line_speed * maxCfg) / 100;
} else /* SD modes */
/* maxCfg is absolute in 100Mb units */
vn_max_rate = maxCfg * 100;
}
DP(NETIF_MSG_IFUP, "vn %d: vn_max_rate %d\n", vn, vn_max_rate);
input->vnic_max_rate[vn] = vn_max_rate;
}
static int bnx2x_get_cmng_fns_mode(struct bnx2x *bp)
{
if (CHIP_REV_IS_SLOW(bp))
return CMNG_FNS_NONE;
if (IS_MF(bp))
return CMNG_FNS_MINMAX;
return CMNG_FNS_NONE;
}
void bnx2x_read_mf_cfg(struct bnx2x *bp)
{
int vn, n = (CHIP_MODE_IS_4_PORT(bp) ? 2 : 1);
if (BP_NOMCP(bp))
return; /* what should be the default value in this case */
/* For 2 port configuration the absolute function number formula
* is:
* abs_func = 2 * vn + BP_PORT + BP_PATH
*
* and there are 4 functions per port
*
* For 4 port configuration it is
* abs_func = 4 * vn + 2 * BP_PORT + BP_PATH
*
* and there are 2 functions per port
*/
for (vn = VN_0; vn < BP_MAX_VN_NUM(bp); vn++) {
int /*abs*/func = n * (2 * vn + BP_PORT(bp)) + BP_PATH(bp);
if (func >= E1H_FUNC_MAX)
break;
bp->mf_config[vn] =
MF_CFG_RD(bp, func_mf_config[func].config);
}
if (bp->mf_config[BP_VN(bp)] & FUNC_MF_CFG_FUNC_DISABLED) {
DP(NETIF_MSG_IFUP, "mf_cfg function disabled\n");
bp->flags |= MF_FUNC_DIS;
} else {
DP(NETIF_MSG_IFUP, "mf_cfg function enabled\n");
bp->flags &= ~MF_FUNC_DIS;
}
}
static void bnx2x_cmng_fns_init(struct bnx2x *bp, u8 read_cfg, u8 cmng_type)
{
struct cmng_init_input input;
memset(&input, 0, sizeof(struct cmng_init_input));
input.port_rate = bp->link_vars.line_speed;
if (cmng_type == CMNG_FNS_MINMAX && input.port_rate) {
int vn;
/* read mf conf from shmem */
if (read_cfg)
bnx2x_read_mf_cfg(bp);
/* vn_weight_sum and enable fairness if not 0 */
bnx2x_calc_vn_min(bp, &input);
/* calculate and set min-max rate for each vn */
if (bp->port.pmf)
for (vn = VN_0; vn < BP_MAX_VN_NUM(bp); vn++)
bnx2x_calc_vn_max(bp, vn, &input);
/* always enable rate shaping and fairness */
input.flags.cmng_enables |=
CMNG_FLAGS_PER_PORT_RATE_SHAPING_VN;
bnx2x_init_cmng(&input, &bp->cmng);
return;
}
/* rate shaping and fairness are disabled */
DP(NETIF_MSG_IFUP,
"rate shaping and fairness are disabled\n");
}
static void storm_memset_cmng(struct bnx2x *bp,
struct cmng_init *cmng,
u8 port)
{
int vn;
size_t size = sizeof(struct cmng_struct_per_port);
u32 addr = BAR_XSTRORM_INTMEM +
XSTORM_CMNG_PER_PORT_VARS_OFFSET(port);
__storm_memset_struct(bp, addr, size, (u32 *)&cmng->port);
for (vn = VN_0; vn < BP_MAX_VN_NUM(bp); vn++) {
int func = func_by_vn(bp, vn);
addr = BAR_XSTRORM_INTMEM +
XSTORM_RATE_SHAPING_PER_VN_VARS_OFFSET(func);
size = sizeof(struct rate_shaping_vars_per_vn);
__storm_memset_struct(bp, addr, size,
(u32 *)&cmng->vnic.vnic_max_rate[vn]);
addr = BAR_XSTRORM_INTMEM +
XSTORM_FAIRNESS_PER_VN_VARS_OFFSET(func);
size = sizeof(struct fairness_vars_per_vn);
__storm_memset_struct(bp, addr, size,
(u32 *)&cmng->vnic.vnic_min_rate[vn]);
}
}
/* init cmng mode in HW according to local configuration */
void bnx2x_set_local_cmng(struct bnx2x *bp)
{
int cmng_fns = bnx2x_get_cmng_fns_mode(bp);
if (cmng_fns != CMNG_FNS_NONE) {
bnx2x_cmng_fns_init(bp, false, cmng_fns);
storm_memset_cmng(bp, &bp->cmng, BP_PORT(bp));
} else {
/* rate shaping and fairness are disabled */
DP(NETIF_MSG_IFUP,
"single function mode without fairness\n");
}
}
/* This function is called upon link interrupt */
static void bnx2x_link_attn(struct bnx2x *bp)
{
/* Make sure that we are synced with the current statistics */
bnx2x_stats_handle(bp, STATS_EVENT_STOP);
bnx2x_link_update(&bp->link_params, &bp->link_vars);
bnx2x_init_dropless_fc(bp);
if (bp->link_vars.link_up) {
if (bp->link_vars.mac_type != MAC_TYPE_EMAC) {
struct host_port_stats *pstats;
pstats = bnx2x_sp(bp, port_stats);
/* reset old mac stats */
memset(&(pstats->mac_stx[0]), 0,
sizeof(struct mac_stx));
}
if (bp->state == BNX2X_STATE_OPEN)
bnx2x_stats_handle(bp, STATS_EVENT_LINK_UP);
}
if (bp->link_vars.link_up && bp->link_vars.line_speed)
bnx2x_set_local_cmng(bp);
__bnx2x_link_report(bp);
if (IS_MF(bp))
bnx2x_link_sync_notify(bp);
}
void bnx2x__link_status_update(struct bnx2x *bp)
{
if (bp->state != BNX2X_STATE_OPEN)
return;
/* read updated dcb configuration */
if (IS_PF(bp)) {
bnx2x_dcbx_pmf_update(bp);
bnx2x_link_status_update(&bp->link_params, &bp->link_vars);
if (bp->link_vars.link_up)
bnx2x_stats_handle(bp, STATS_EVENT_LINK_UP);
else
bnx2x_stats_handle(bp, STATS_EVENT_STOP);
/* indicate link status */
bnx2x_link_report(bp);
} else { /* VF */
bp->port.supported[0] |= (SUPPORTED_10baseT_Half |
SUPPORTED_10baseT_Full |
SUPPORTED_100baseT_Half |
SUPPORTED_100baseT_Full |
SUPPORTED_1000baseT_Full |
SUPPORTED_2500baseX_Full |
SUPPORTED_10000baseT_Full |
SUPPORTED_TP |
SUPPORTED_FIBRE |
SUPPORTED_Autoneg |
SUPPORTED_Pause |
SUPPORTED_Asym_Pause);
bp->port.advertising[0] = bp->port.supported[0];
bp->link_params.bp = bp;
bp->link_params.port = BP_PORT(bp);
bp->link_params.req_duplex[0] = DUPLEX_FULL;
bp->link_params.req_flow_ctrl[0] = BNX2X_FLOW_CTRL_NONE;
bp->link_params.req_line_speed[0] = SPEED_10000;
bp->link_params.speed_cap_mask[0] = 0x7f0000;
bp->link_params.switch_cfg = SWITCH_CFG_10G;
bp->link_vars.mac_type = MAC_TYPE_BMAC;
bp->link_vars.line_speed = SPEED_10000;
bp->link_vars.link_status =
(LINK_STATUS_LINK_UP |
LINK_STATUS_SPEED_AND_DUPLEX_10GTFD);
bp->link_vars.link_up = 1;
bp->link_vars.duplex = DUPLEX_FULL;
bp->link_vars.flow_ctrl = BNX2X_FLOW_CTRL_NONE;
__bnx2x_link_report(bp);
bnx2x_sample_bulletin(bp);
/* if bulletin board did not have an update for link status
* __bnx2x_link_report will report current status
* but it will NOT duplicate report in case of already reported
* during sampling bulletin board.
*/
bnx2x_stats_handle(bp, STATS_EVENT_LINK_UP);
}
}
static int bnx2x_afex_func_update(struct bnx2x *bp, u16 vifid,
u16 vlan_val, u8 allowed_prio)
{
struct bnx2x_func_state_params func_params = {NULL};
struct bnx2x_func_afex_update_params *f_update_params =
&func_params.params.afex_update;
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_AFEX_UPDATE;
/* no need to wait for RAMROD completion, so don't
* set RAMROD_COMP_WAIT flag
*/
f_update_params->vif_id = vifid;
f_update_params->afex_default_vlan = vlan_val;
f_update_params->allowed_priorities = allowed_prio;
/* if ramrod can not be sent, response to MCP immediately */
if (bnx2x_func_state_change(bp, &func_params) < 0)
bnx2x_fw_command(bp, DRV_MSG_CODE_AFEX_VIFSET_ACK, 0);
return 0;
}
static int bnx2x_afex_handle_vif_list_cmd(struct bnx2x *bp, u8 cmd_type,
u16 vif_index, u8 func_bit_map)
{
struct bnx2x_func_state_params func_params = {NULL};
struct bnx2x_func_afex_viflists_params *update_params =
&func_params.params.afex_viflists;
int rc;
u32 drv_msg_code;
/* validate only LIST_SET and LIST_GET are received from switch */
if ((cmd_type != VIF_LIST_RULE_GET) && (cmd_type != VIF_LIST_RULE_SET))
BNX2X_ERR("BUG! afex_handle_vif_list_cmd invalid type 0x%x\n",
cmd_type);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_AFEX_VIFLISTS;
/* set parameters according to cmd_type */
update_params->afex_vif_list_command = cmd_type;
update_params->vif_list_index = vif_index;
update_params->func_bit_map =
(cmd_type == VIF_LIST_RULE_GET) ? 0 : func_bit_map;
update_params->func_to_clear = 0;
drv_msg_code =
(cmd_type == VIF_LIST_RULE_GET) ?
DRV_MSG_CODE_AFEX_LISTGET_ACK :
DRV_MSG_CODE_AFEX_LISTSET_ACK;
/* if ramrod can not be sent, respond to MCP immediately for
* SET and GET requests (other are not triggered from MCP)
*/
rc = bnx2x_func_state_change(bp, &func_params);
if (rc < 0)
bnx2x_fw_command(bp, drv_msg_code, 0);
return 0;
}
static void bnx2x_handle_afex_cmd(struct bnx2x *bp, u32 cmd)
{
struct afex_stats afex_stats;
u32 func = BP_ABS_FUNC(bp);
u32 mf_config;
u16 vlan_val;
u32 vlan_prio;
u16 vif_id;
u8 allowed_prio;
u8 vlan_mode;
u32 addr_to_write, vifid, addrs, stats_type, i;
if (cmd & DRV_STATUS_AFEX_LISTGET_REQ) {
vifid = SHMEM2_RD(bp, afex_param1_to_driver[BP_FW_MB_IDX(bp)]);
DP(BNX2X_MSG_MCP,
"afex: got MCP req LISTGET_REQ for vifid 0x%x\n", vifid);
bnx2x_afex_handle_vif_list_cmd(bp, VIF_LIST_RULE_GET, vifid, 0);
}
if (cmd & DRV_STATUS_AFEX_LISTSET_REQ) {
vifid = SHMEM2_RD(bp, afex_param1_to_driver[BP_FW_MB_IDX(bp)]);
addrs = SHMEM2_RD(bp, afex_param2_to_driver[BP_FW_MB_IDX(bp)]);
DP(BNX2X_MSG_MCP,
"afex: got MCP req LISTSET_REQ for vifid 0x%x addrs 0x%x\n",
vifid, addrs);
bnx2x_afex_handle_vif_list_cmd(bp, VIF_LIST_RULE_SET, vifid,
addrs);
}
if (cmd & DRV_STATUS_AFEX_STATSGET_REQ) {
addr_to_write = SHMEM2_RD(bp,
afex_scratchpad_addr_to_write[BP_FW_MB_IDX(bp)]);
stats_type = SHMEM2_RD(bp,
afex_param1_to_driver[BP_FW_MB_IDX(bp)]);
DP(BNX2X_MSG_MCP,
"afex: got MCP req STATSGET_REQ, write to addr 0x%x\n",
addr_to_write);
bnx2x_afex_collect_stats(bp, (void *)&afex_stats, stats_type);
/* write response to scratchpad, for MCP */
for (i = 0; i < (sizeof(struct afex_stats)/sizeof(u32)); i++)
REG_WR(bp, addr_to_write + i*sizeof(u32),
*(((u32 *)(&afex_stats))+i));
/* send ack message to MCP */
bnx2x_fw_command(bp, DRV_MSG_CODE_AFEX_STATSGET_ACK, 0);
}
if (cmd & DRV_STATUS_AFEX_VIFSET_REQ) {
mf_config = MF_CFG_RD(bp, func_mf_config[func].config);
bp->mf_config[BP_VN(bp)] = mf_config;
DP(BNX2X_MSG_MCP,
"afex: got MCP req VIFSET_REQ, mf_config 0x%x\n",
mf_config);
/* if VIF_SET is "enabled" */
if (!(mf_config & FUNC_MF_CFG_FUNC_DISABLED)) {
/* set rate limit directly to internal RAM */
struct cmng_init_input cmng_input;
struct rate_shaping_vars_per_vn m_rs_vn;
size_t size = sizeof(struct rate_shaping_vars_per_vn);
u32 addr = BAR_XSTRORM_INTMEM +
XSTORM_RATE_SHAPING_PER_VN_VARS_OFFSET(BP_FUNC(bp));
bp->mf_config[BP_VN(bp)] = mf_config;
bnx2x_calc_vn_max(bp, BP_VN(bp), &cmng_input);
m_rs_vn.vn_counter.rate =
cmng_input.vnic_max_rate[BP_VN(bp)];
m_rs_vn.vn_counter.quota =
(m_rs_vn.vn_counter.rate *
RS_PERIODIC_TIMEOUT_USEC) / 8;
__storm_memset_struct(bp, addr, size, (u32 *)&m_rs_vn);
/* read relevant values from mf_cfg struct in shmem */
vif_id =
(MF_CFG_RD(bp, func_mf_config[func].e1hov_tag) &
FUNC_MF_CFG_E1HOV_TAG_MASK) >>
FUNC_MF_CFG_E1HOV_TAG_SHIFT;
vlan_val =
(MF_CFG_RD(bp, func_mf_config[func].e1hov_tag) &
FUNC_MF_CFG_AFEX_VLAN_MASK) >>
FUNC_MF_CFG_AFEX_VLAN_SHIFT;
vlan_prio = (mf_config &
FUNC_MF_CFG_TRANSMIT_PRIORITY_MASK) >>
FUNC_MF_CFG_TRANSMIT_PRIORITY_SHIFT;
vlan_val |= (vlan_prio << VLAN_PRIO_SHIFT);
vlan_mode =
(MF_CFG_RD(bp,
func_mf_config[func].afex_config) &
FUNC_MF_CFG_AFEX_VLAN_MODE_MASK) >>
FUNC_MF_CFG_AFEX_VLAN_MODE_SHIFT;
allowed_prio =
(MF_CFG_RD(bp,
func_mf_config[func].afex_config) &
FUNC_MF_CFG_AFEX_COS_FILTER_MASK) >>
FUNC_MF_CFG_AFEX_COS_FILTER_SHIFT;
/* send ramrod to FW, return in case of failure */
if (bnx2x_afex_func_update(bp, vif_id, vlan_val,
allowed_prio))
return;
bp->afex_def_vlan_tag = vlan_val;
bp->afex_vlan_mode = vlan_mode;
} else {
/* notify link down because BP->flags is disabled */
bnx2x_link_report(bp);
/* send INVALID VIF ramrod to FW */
bnx2x_afex_func_update(bp, 0xFFFF, 0, 0);
/* Reset the default afex VLAN */
bp->afex_def_vlan_tag = -1;
}
}
}
static void bnx2x_handle_update_svid_cmd(struct bnx2x *bp)
{
struct bnx2x_func_switch_update_params *switch_update_params;
struct bnx2x_func_state_params func_params;
memset(&func_params, 0, sizeof(struct bnx2x_func_state_params));
switch_update_params = &func_params.params.switch_update;
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_SWITCH_UPDATE;
if (IS_MF_UFP(bp) || IS_MF_BD(bp)) {
int func = BP_ABS_FUNC(bp);
u32 val;
/* Re-learn the S-tag from shmem */
val = MF_CFG_RD(bp, func_mf_config[func].e1hov_tag) &
FUNC_MF_CFG_E1HOV_TAG_MASK;
if (val != FUNC_MF_CFG_E1HOV_TAG_DEFAULT) {
bp->mf_ov = val;
} else {
BNX2X_ERR("Got an SVID event, but no tag is configured in shmem\n");
goto fail;
}
/* Configure new S-tag in LLH */
REG_WR(bp, NIG_REG_LLH0_FUNC_VLAN_ID + BP_PORT(bp) * 8,
bp->mf_ov);
/* Send Ramrod to update FW of change */
__set_bit(BNX2X_F_UPDATE_SD_VLAN_TAG_CHNG,
&switch_update_params->changes);
switch_update_params->vlan = bp->mf_ov;
if (bnx2x_func_state_change(bp, &func_params) < 0) {
BNX2X_ERR("Failed to configure FW of S-tag Change to %02x\n",
bp->mf_ov);
goto fail;
} else {
DP(BNX2X_MSG_MCP, "Configured S-tag %02x\n",
bp->mf_ov);
}
} else {
goto fail;
}
bnx2x_fw_command(bp, DRV_MSG_CODE_OEM_UPDATE_SVID_OK, 0);
return;
fail:
bnx2x_fw_command(bp, DRV_MSG_CODE_OEM_UPDATE_SVID_FAILURE, 0);
}
static void bnx2x_pmf_update(struct bnx2x *bp)
{
int port = BP_PORT(bp);
u32 val;
bp->port.pmf = 1;
DP(BNX2X_MSG_MCP, "pmf %d\n", bp->port.pmf);
/*
* We need the mb() to ensure the ordering between the writing to
* bp->port.pmf here and reading it from the bnx2x_periodic_task().
*/
smp_mb();
/* queue a periodic task */
queue_delayed_work(bnx2x_wq, &bp->period_task, 0);
bnx2x_dcbx_pmf_update(bp);
/* enable nig attention */
val = (0xff0f | (1 << (BP_VN(bp) + 4)));
if (bp->common.int_block == INT_BLOCK_HC) {
REG_WR(bp, HC_REG_TRAILING_EDGE_0 + port*8, val);
REG_WR(bp, HC_REG_LEADING_EDGE_0 + port*8, val);
} else if (!CHIP_IS_E1x(bp)) {
REG_WR(bp, IGU_REG_TRAILING_EDGE_LATCH, val);
REG_WR(bp, IGU_REG_LEADING_EDGE_LATCH, val);
}
bnx2x_stats_handle(bp, STATS_EVENT_PMF);
}
/* end of Link */
/* slow path */
/*
* General service functions
*/
/* send the MCP a request, block until there is a reply */
u32 bnx2x_fw_command(struct bnx2x *bp, u32 command, u32 param)
{
int mb_idx = BP_FW_MB_IDX(bp);
u32 seq;
u32 rc = 0;
u32 cnt = 1;
u8 delay = CHIP_REV_IS_SLOW(bp) ? 100 : 10;
mutex_lock(&bp->fw_mb_mutex);
seq = ++bp->fw_seq;
SHMEM_WR(bp, func_mb[mb_idx].drv_mb_param, param);
SHMEM_WR(bp, func_mb[mb_idx].drv_mb_header, (command | seq));
DP(BNX2X_MSG_MCP, "wrote command (%x) to FW MB param 0x%08x\n",
(command | seq), param);
do {
/* let the FW do it's magic ... */
msleep(delay);
rc = SHMEM_RD(bp, func_mb[mb_idx].fw_mb_header);
/* Give the FW up to 5 second (500*10ms) */
} while ((seq != (rc & FW_MSG_SEQ_NUMBER_MASK)) && (cnt++ < 500));
DP(BNX2X_MSG_MCP, "[after %d ms] read (%x) seq is (%x) from FW MB\n",
cnt*delay, rc, seq);
/* is this a reply to our command? */
if (seq == (rc & FW_MSG_SEQ_NUMBER_MASK))
rc &= FW_MSG_CODE_MASK;
else {
/* FW BUG! */
BNX2X_ERR("FW failed to respond!\n");
bnx2x_fw_dump(bp);
rc = 0;
}
mutex_unlock(&bp->fw_mb_mutex);
return rc;
}
static void storm_memset_func_cfg(struct bnx2x *bp,
struct tstorm_eth_function_common_config *tcfg,
u16 abs_fid)
{
size_t size = sizeof(struct tstorm_eth_function_common_config);
u32 addr = BAR_TSTRORM_INTMEM +
TSTORM_FUNCTION_COMMON_CONFIG_OFFSET(abs_fid);
__storm_memset_struct(bp, addr, size, (u32 *)tcfg);
}
void bnx2x_func_init(struct bnx2x *bp, struct bnx2x_func_init_params *p)
{
if (CHIP_IS_E1x(bp)) {
struct tstorm_eth_function_common_config tcfg = {0};
storm_memset_func_cfg(bp, &tcfg, p->func_id);
}
/* Enable the function in the FW */
storm_memset_vf_to_pf(bp, p->func_id, p->pf_id);
storm_memset_func_en(bp, p->func_id, 1);
/* spq */
if (p->spq_active) {
storm_memset_spq_addr(bp, p->spq_map, p->func_id);
REG_WR(bp, XSEM_REG_FAST_MEMORY +
XSTORM_SPQ_PROD_OFFSET(p->func_id), p->spq_prod);
}
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/**
* bnx2x_get_common_flags - Return common flags
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
*
* @bp device handle
* @fp queue handle
* @zero_stats TRUE if statistics zeroing is needed
*
* Return the flags that are common for the Tx-only and not normal connections.
*/
static unsigned long bnx2x_get_common_flags(struct bnx2x *bp,
struct bnx2x_fastpath *fp,
bool zero_stats)
{
unsigned long flags = 0;
/* PF driver will always initialize the Queue to an ACTIVE state */
__set_bit(BNX2X_Q_FLG_ACTIVE, &flags);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* tx only connections collect statistics (on the same index as the
* parent connection). The statistics are zeroed when the parent
* connection is initialized.
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
*/
__set_bit(BNX2X_Q_FLG_STATS, &flags);
if (zero_stats)
__set_bit(BNX2X_Q_FLG_ZERO_STATS, &flags);
if (bp->flags & TX_SWITCHING)
__set_bit(BNX2X_Q_FLG_TX_SWITCH, &flags);
__set_bit(BNX2X_Q_FLG_PCSUM_ON_PKT, &flags);
__set_bit(BNX2X_Q_FLG_TUN_INC_INNER_IP_ID, &flags);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
#ifdef BNX2X_STOP_ON_ERROR
__set_bit(BNX2X_Q_FLG_TX_SEC, &flags);
#endif
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
return flags;
}
static unsigned long bnx2x_get_q_flags(struct bnx2x *bp,
struct bnx2x_fastpath *fp,
bool leading)
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
{
unsigned long flags = 0;
/* calculate other queue flags */
if (IS_MF_SD(bp))
__set_bit(BNX2X_Q_FLG_OV, &flags);
if (IS_FCOE_FP(fp)) {
__set_bit(BNX2X_Q_FLG_FCOE, &flags);
/* For FCoE - force usage of default priority (for afex) */
__set_bit(BNX2X_Q_FLG_FORCE_DEFAULT_PRI, &flags);
}
if (fp->mode != TPA_MODE_DISABLED) {
__set_bit(BNX2X_Q_FLG_TPA, &flags);
__set_bit(BNX2X_Q_FLG_TPA_IPV6, &flags);
use FW 7.2.16 The patch integrates FW 7.2.16 HSI and implements driver part of GRO flow. FW 7.2.16 adds the ability to aggregate packets for GRO (and not just LRO) and also fixes some bugs. 1. Added new aggregation mode: GRO. In this mode packets are aggregated such that the original packets can be reconstructed by the OS. 2. 57712 HW bug workaround - initialized all CAM TM registers to 0x32. 3. Adding the FCoE statistics structures to the BNX2X HSI. 4. Wrong configuration of TX HW input buffer size may cause theoretical performance effect. Performed configuration fix. 5. FCOE - Arrival of packets beyond task IO size can lead to crash. Fix firmware data-in flow. 6. iSCSI - In rare cases of on-chip termination the graceful termination timer hangs, and the termination doesn't complete. Firmware fix to MSL timer tolerance. 7. iSCSI - Chip hangs when target sends FIN out-of-order or with isles open at the initiator side. Firmware implementation corrected to drop FIN received out-of-order or with isles still open. 8. iSCSI - Chip hangs when in case of retransmission not aligned to 4-bytes from the beginning of iSCSI PDU. Firmware implementation corrected to support arbitrary aligned retransmissions. 9. iSCSI - Arrival of target-initiated NOP-IN during intense ISCSI traffic might lead to crash. Firmware fix to relevant flow. Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-02-20 18:59:08 +09:00
if (fp->mode == TPA_MODE_GRO)
__set_bit(BNX2X_Q_FLG_TPA_GRO, &flags);
}
if (leading) {
__set_bit(BNX2X_Q_FLG_LEADING_RSS, &flags);
__set_bit(BNX2X_Q_FLG_MCAST, &flags);
}
/* Always set HW VLAN stripping */
__set_bit(BNX2X_Q_FLG_VLAN, &flags);
/* configure silent vlan removal */
if (IS_MF_AFEX(bp))
__set_bit(BNX2X_Q_FLG_SILENT_VLAN_REM, &flags);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
return flags | bnx2x_get_common_flags(bp, fp, true);
}
static void bnx2x_pf_q_prep_general(struct bnx2x *bp,
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
struct bnx2x_fastpath *fp, struct bnx2x_general_setup_params *gen_init,
u8 cos)
{
gen_init->stat_id = bnx2x_stats_id(fp);
gen_init->spcl_id = fp->cl_id;
/* Always use mini-jumbo MTU for FCoE L2 ring */
if (IS_FCOE_FP(fp))
gen_init->mtu = BNX2X_FCOE_MINI_JUMBO_MTU;
else
gen_init->mtu = bp->dev->mtu;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
gen_init->cos = cos;
gen_init->fp_hsi = ETH_FP_HSI_VERSION;
}
static void bnx2x_pf_rx_q_prep(struct bnx2x *bp,
struct bnx2x_fastpath *fp, struct rxq_pause_params *pause,
struct bnx2x_rxq_setup_params *rxq_init)
{
u8 max_sge = 0;
u16 sge_sz = 0;
u16 tpa_agg_size = 0;
if (fp->mode != TPA_MODE_DISABLED) {
pause->sge_th_lo = SGE_TH_LO(bp);
pause->sge_th_hi = SGE_TH_HI(bp);
/* validate SGE ring has enough to cross high threshold */
WARN_ON(bp->dropless_fc &&
pause->sge_th_hi + FW_PREFETCH_CNT >
MAX_RX_SGE_CNT * NUM_RX_SGE_PAGES);
tpa_agg_size = TPA_AGG_SIZE;
max_sge = SGE_PAGE_ALIGN(bp->dev->mtu) >>
SGE_PAGE_SHIFT;
max_sge = ((max_sge + PAGES_PER_SGE - 1) &
(~(PAGES_PER_SGE-1))) >> PAGES_PER_SGE_SHIFT;
sge_sz = (u16)min_t(u32, SGE_PAGES, 0xffff);
}
/* pause - not for e1 */
if (!CHIP_IS_E1(bp)) {
pause->bd_th_lo = BD_TH_LO(bp);
pause->bd_th_hi = BD_TH_HI(bp);
pause->rcq_th_lo = RCQ_TH_LO(bp);
pause->rcq_th_hi = RCQ_TH_HI(bp);
/*
* validate that rings have enough entries to cross
* high thresholds
*/
WARN_ON(bp->dropless_fc &&
pause->bd_th_hi + FW_PREFETCH_CNT >
bp->rx_ring_size);
WARN_ON(bp->dropless_fc &&
pause->rcq_th_hi + FW_PREFETCH_CNT >
NUM_RCQ_RINGS * MAX_RCQ_DESC_CNT);
pause->pri_map = 1;
}
/* rxq setup */
rxq_init->dscr_map = fp->rx_desc_mapping;
rxq_init->sge_map = fp->rx_sge_mapping;
rxq_init->rcq_map = fp->rx_comp_mapping;
rxq_init->rcq_np_map = fp->rx_comp_mapping + BCM_PAGE_SIZE;
/* This should be a maximum number of data bytes that may be
* placed on the BD (not including paddings).
*/
bnx2x: uses build_skb() in receive path bnx2x uses following formula to compute its rx_buf_sz : dev->mtu + 2*L1_CACHE_BYTES + 14 + 8 + 8 + 2 Then core network adds NET_SKB_PAD and SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) Final allocated size for skb head on x86_64 (L1_CACHE_BYTES = 64, MTU=1500) : 2112 bytes : SLUB/SLAB round this to 4096 bytes. Since skb truesize is then bigger than SK_MEM_QUANTUM, we have lot of false sharing because of mem_reclaim in UDP stack. One possible way to half truesize is to reduce the need by 64 bytes (2112 -> 2048 bytes) Instead of allocating a full cache line at the end of packet for alignment, we can use the fact that skb_shared_info sits at the end of skb->head, and we can use this room, if we convert bnx2x to new build_skb() infrastructure. skb_shared_info will be initialized after hardware finished its transfert, so we can eventually overwrite the final padding. Using build_skb() also reduces cache line misses in the driver, since we use cache hot skb instead of cold ones. Number of in-flight sk_buff structures is lower, they are recycled while still hot. Performance results : (820.000 pps on a rx UDP monothread benchmark, instead of 720.000 pps) Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> CC: Eilon Greenstein <eilong@broadcom.com> CC: Ben Hutchings <bhutchings@solarflare.com> CC: Tom Herbert <therbert@google.com> CC: Jamal Hadi Salim <hadi@mojatatu.com> CC: Stephen Hemminger <shemminger@vyatta.com> CC: Thomas Graf <tgraf@infradead.org> CC: Herbert Xu <herbert@gondor.apana.org.au> CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Acked-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-14 15:05:34 +09:00
rxq_init->buf_sz = fp->rx_buf_size - BNX2X_FW_RX_ALIGN_START -
BNX2X_FW_RX_ALIGN_END - IP_HEADER_ALIGNMENT_PADDING;
rxq_init->cl_qzone_id = fp->cl_qzone_id;
rxq_init->tpa_agg_sz = tpa_agg_size;
rxq_init->sge_buf_sz = sge_sz;
rxq_init->max_sges_pkt = max_sge;
rxq_init->rss_engine_id = BP_FUNC(bp);
rxq_init->mcast_engine_id = BP_FUNC(bp);
/* Maximum number or simultaneous TPA aggregation for this Queue.
*
* For PF Clients it should be the maximum available number.
* VF driver(s) may want to define it to a smaller value.
*/
rxq_init->max_tpa_queues = MAX_AGG_QS(bp);
rxq_init->cache_line_log = BNX2X_RX_ALIGN_SHIFT;
rxq_init->fw_sb_id = fp->fw_sb_id;
if (IS_FCOE_FP(fp))
rxq_init->sb_cq_index = HC_SP_INDEX_ETH_FCOE_RX_CQ_CONS;
else
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
rxq_init->sb_cq_index = HC_INDEX_ETH_RX_CQ_CONS;
/* configure silent vlan removal
* if multi function mode is afex, then mask default vlan
*/
if (IS_MF_AFEX(bp)) {
rxq_init->silent_removal_value = bp->afex_def_vlan_tag;
rxq_init->silent_removal_mask = VLAN_VID_MASK;
}
}
static void bnx2x_pf_tx_q_prep(struct bnx2x *bp,
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
struct bnx2x_fastpath *fp, struct bnx2x_txq_setup_params *txq_init,
u8 cos)
{
txq_init->dscr_map = fp->txdata_ptr[cos]->tx_desc_mapping;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
txq_init->sb_cq_index = HC_INDEX_ETH_FIRST_TX_CQ_CONS + cos;
txq_init->traffic_type = LLFC_TRAFFIC_TYPE_NW;
txq_init->fw_sb_id = fp->fw_sb_id;
/*
* set the tss leading client id for TX classification ==
* leading RSS client id
*/
txq_init->tss_leading_cl_id = bnx2x_fp(bp, 0, cl_id);
if (IS_FCOE_FP(fp)) {
txq_init->sb_cq_index = HC_SP_INDEX_ETH_FCOE_TX_CQ_CONS;
txq_init->traffic_type = LLFC_TRAFFIC_TYPE_FCOE;
}
}
static void bnx2x_pf_init(struct bnx2x *bp)
{
struct bnx2x_func_init_params func_init = {0};
struct event_ring_data eq_data = { {0} };
if (!CHIP_IS_E1x(bp)) {
/* reset IGU PF statistics: MSIX + ATTN */
/* PF */
REG_WR(bp, IGU_REG_STATISTIC_NUM_MESSAGE_SENT +
BNX2X_IGU_STAS_MSG_VF_CNT*4 +
(CHIP_MODE_IS_4_PORT(bp) ?
BP_FUNC(bp) : BP_VN(bp))*4, 0);
/* ATTN */
REG_WR(bp, IGU_REG_STATISTIC_NUM_MESSAGE_SENT +
BNX2X_IGU_STAS_MSG_VF_CNT*4 +
BNX2X_IGU_STAS_MSG_PF_CNT*4 +
(CHIP_MODE_IS_4_PORT(bp) ?
BP_FUNC(bp) : BP_VN(bp))*4, 0);
}
func_init.spq_active = true;
func_init.pf_id = BP_FUNC(bp);
func_init.func_id = BP_FUNC(bp);
func_init.spq_map = bp->spq_mapping;
func_init.spq_prod = bp->spq_prod_idx;
bnx2x_func_init(bp, &func_init);
memset(&(bp->cmng), 0, sizeof(struct cmng_struct_per_port));
/*
* Congestion management values depend on the link rate
* There is no active link so initial link rate is set to 10 Gbps.
* When the link comes up The congestion management values are
* re-calculated according to the actual link rate.
*/
bp->link_vars.line_speed = SPEED_10000;
bnx2x_cmng_fns_init(bp, true, bnx2x_get_cmng_fns_mode(bp));
/* Only the PMF sets the HW */
if (bp->port.pmf)
storm_memset_cmng(bp, &bp->cmng, BP_PORT(bp));
/* init Event Queue - PCI bus guarantees correct endianity*/
eq_data.base_addr.hi = U64_HI(bp->eq_mapping);
eq_data.base_addr.lo = U64_LO(bp->eq_mapping);
eq_data.producer = bp->eq_prod;
eq_data.index_id = HC_SP_INDEX_EQ_CONS;
eq_data.sb_id = DEF_SB_ID;
storm_memset_eq_data(bp, &eq_data, BP_FUNC(bp));
}
static void bnx2x_e1h_disable(struct bnx2x *bp)
{
int port = BP_PORT(bp);
bnx2x_tx_disable(bp);
REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port*8, 0);
}
static void bnx2x_e1h_enable(struct bnx2x *bp)
{
int port = BP_PORT(bp);
if (!(IS_MF_UFP(bp) && BNX2X_IS_MF_SD_PROTOCOL_FCOE(bp)))
REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port * 8, 1);
/* Tx queue should be only re-enabled */
netif_tx_wake_all_queues(bp->dev);
/*
* Should not call netif_carrier_on since it will be called if the link
* is up when checking for link state
*/
}
#define DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED 3
static void bnx2x_drv_info_ether_stat(struct bnx2x *bp)
{
struct eth_stats_info *ether_stat =
&bp->slowpath->drv_info_to_mcp.ether_stat;
struct bnx2x_vlan_mac_obj *mac_obj =
&bp->sp_objs->mac_obj;
int i;
strlcpy(ether_stat->version, DRV_MODULE_VERSION,
ETH_STAT_INFO_VERSION_LEN);
/* get DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED macs, placing them in the
* mac_local field in ether_stat struct. The base address is offset by 2
* bytes to account for the field being 8 bytes but a mac address is
* only 6 bytes. Likewise, the stride for the get_n_elements function is
* 2 bytes to compensate from the 6 bytes of a mac to the 8 bytes
* allocated by the ether_stat struct, so the macs will land in their
* proper positions.
*/
for (i = 0; i < DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED; i++)
memset(ether_stat->mac_local + i, 0,
sizeof(ether_stat->mac_local[0]));
mac_obj->get_n_elements(bp, &bp->sp_objs[0].mac_obj,
DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED,
ether_stat->mac_local + MAC_PAD, MAC_PAD,
ETH_ALEN);
ether_stat->mtu_size = bp->dev->mtu;
if (bp->dev->features & NETIF_F_RXCSUM)
ether_stat->feature_flags |= FEATURE_ETH_CHKSUM_OFFLOAD_MASK;
if (bp->dev->features & NETIF_F_TSO)
ether_stat->feature_flags |= FEATURE_ETH_LSO_MASK;
ether_stat->feature_flags |= bp->common.boot_mode;
ether_stat->promiscuous_mode = (bp->dev->flags & IFF_PROMISC) ? 1 : 0;
ether_stat->txq_size = bp->tx_ring_size;
ether_stat->rxq_size = bp->rx_ring_size;
#ifdef CONFIG_BNX2X_SRIOV
ether_stat->vf_cnt = IS_SRIOV(bp) ? bp->vfdb->sriov.nr_virtfn : 0;
#endif
}
static void bnx2x_drv_info_fcoe_stat(struct bnx2x *bp)
{
struct bnx2x_dcbx_app_params *app = &bp->dcbx_port_params.app;
struct fcoe_stats_info *fcoe_stat =
&bp->slowpath->drv_info_to_mcp.fcoe_stat;
if (!CNIC_LOADED(bp))
return;
memcpy(fcoe_stat->mac_local + MAC_PAD, bp->fip_mac, ETH_ALEN);
fcoe_stat->qos_priority =
app->traffic_type_priority[LLFC_TRAFFIC_TYPE_FCOE];
/* insert FCoE stats from ramrod response */
if (!NO_FCOE(bp)) {
struct tstorm_per_queue_stats *fcoe_q_tstorm_stats =
&bp->fw_stats_data->queue_stats[FCOE_IDX(bp)].
tstorm_queue_statistics;
struct xstorm_per_queue_stats *fcoe_q_xstorm_stats =
&bp->fw_stats_data->queue_stats[FCOE_IDX(bp)].
xstorm_queue_statistics;
struct fcoe_statistics_params *fw_fcoe_stat =
&bp->fw_stats_data->fcoe;
ADD_64_LE(fcoe_stat->rx_bytes_hi, LE32_0,
fcoe_stat->rx_bytes_lo,
fw_fcoe_stat->rx_stat0.fcoe_rx_byte_cnt);
ADD_64_LE(fcoe_stat->rx_bytes_hi,
fcoe_q_tstorm_stats->rcv_ucast_bytes.hi,
fcoe_stat->rx_bytes_lo,
fcoe_q_tstorm_stats->rcv_ucast_bytes.lo);
ADD_64_LE(fcoe_stat->rx_bytes_hi,
fcoe_q_tstorm_stats->rcv_bcast_bytes.hi,
fcoe_stat->rx_bytes_lo,
fcoe_q_tstorm_stats->rcv_bcast_bytes.lo);
ADD_64_LE(fcoe_stat->rx_bytes_hi,
fcoe_q_tstorm_stats->rcv_mcast_bytes.hi,
fcoe_stat->rx_bytes_lo,
fcoe_q_tstorm_stats->rcv_mcast_bytes.lo);
ADD_64_LE(fcoe_stat->rx_frames_hi, LE32_0,
fcoe_stat->rx_frames_lo,
fw_fcoe_stat->rx_stat0.fcoe_rx_pkt_cnt);
ADD_64_LE(fcoe_stat->rx_frames_hi, LE32_0,
fcoe_stat->rx_frames_lo,
fcoe_q_tstorm_stats->rcv_ucast_pkts);
ADD_64_LE(fcoe_stat->rx_frames_hi, LE32_0,
fcoe_stat->rx_frames_lo,
fcoe_q_tstorm_stats->rcv_bcast_pkts);
ADD_64_LE(fcoe_stat->rx_frames_hi, LE32_0,
fcoe_stat->rx_frames_lo,
fcoe_q_tstorm_stats->rcv_mcast_pkts);
ADD_64_LE(fcoe_stat->tx_bytes_hi, LE32_0,
fcoe_stat->tx_bytes_lo,
fw_fcoe_stat->tx_stat.fcoe_tx_byte_cnt);
ADD_64_LE(fcoe_stat->tx_bytes_hi,
fcoe_q_xstorm_stats->ucast_bytes_sent.hi,
fcoe_stat->tx_bytes_lo,
fcoe_q_xstorm_stats->ucast_bytes_sent.lo);
ADD_64_LE(fcoe_stat->tx_bytes_hi,
fcoe_q_xstorm_stats->bcast_bytes_sent.hi,
fcoe_stat->tx_bytes_lo,
fcoe_q_xstorm_stats->bcast_bytes_sent.lo);
ADD_64_LE(fcoe_stat->tx_bytes_hi,
fcoe_q_xstorm_stats->mcast_bytes_sent.hi,
fcoe_stat->tx_bytes_lo,
fcoe_q_xstorm_stats->mcast_bytes_sent.lo);
ADD_64_LE(fcoe_stat->tx_frames_hi, LE32_0,
fcoe_stat->tx_frames_lo,
fw_fcoe_stat->tx_stat.fcoe_tx_pkt_cnt);
ADD_64_LE(fcoe_stat->tx_frames_hi, LE32_0,
fcoe_stat->tx_frames_lo,
fcoe_q_xstorm_stats->ucast_pkts_sent);
ADD_64_LE(fcoe_stat->tx_frames_hi, LE32_0,
fcoe_stat->tx_frames_lo,
fcoe_q_xstorm_stats->bcast_pkts_sent);
ADD_64_LE(fcoe_stat->tx_frames_hi, LE32_0,
fcoe_stat->tx_frames_lo,
fcoe_q_xstorm_stats->mcast_pkts_sent);
}
/* ask L5 driver to add data to the struct */
bnx2x_cnic_notify(bp, CNIC_CTL_FCOE_STATS_GET_CMD);
}
static void bnx2x_drv_info_iscsi_stat(struct bnx2x *bp)
{
struct bnx2x_dcbx_app_params *app = &bp->dcbx_port_params.app;
struct iscsi_stats_info *iscsi_stat =
&bp->slowpath->drv_info_to_mcp.iscsi_stat;
if (!CNIC_LOADED(bp))
return;
memcpy(iscsi_stat->mac_local + MAC_PAD, bp->cnic_eth_dev.iscsi_mac,
ETH_ALEN);
iscsi_stat->qos_priority =
app->traffic_type_priority[LLFC_TRAFFIC_TYPE_ISCSI];
/* ask L5 driver to add data to the struct */
bnx2x_cnic_notify(bp, CNIC_CTL_ISCSI_STATS_GET_CMD);
}
/* called due to MCP event (on pmf):
* reread new bandwidth configuration
* configure FW
* notify others function about the change
*/
static void bnx2x_config_mf_bw(struct bnx2x *bp)
{
if (bp->link_vars.link_up) {
bnx2x_cmng_fns_init(bp, true, CMNG_FNS_MINMAX);
bnx2x_link_sync_notify(bp);
}
storm_memset_cmng(bp, &bp->cmng, BP_PORT(bp));
}
static void bnx2x_set_mf_bw(struct bnx2x *bp)
{
bnx2x_config_mf_bw(bp);
bnx2x_fw_command(bp, DRV_MSG_CODE_SET_MF_BW_ACK, 0);
}
static void bnx2x_handle_eee_event(struct bnx2x *bp)
{
DP(BNX2X_MSG_MCP, "EEE - LLDP event\n");
bnx2x_fw_command(bp, DRV_MSG_CODE_EEE_RESULTS_ACK, 0);
}
#define BNX2X_UPDATE_DRV_INFO_IND_LENGTH (20)
#define BNX2X_UPDATE_DRV_INFO_IND_COUNT (25)
static void bnx2x_handle_drv_info_req(struct bnx2x *bp)
{
enum drv_info_opcode op_code;
u32 drv_info_ctl = SHMEM2_RD(bp, drv_info_control);
bool release = false;
int wait;
/* if drv_info version supported by MFW doesn't match - send NACK */
if ((drv_info_ctl & DRV_INFO_CONTROL_VER_MASK) != DRV_INFO_CUR_VER) {
bnx2x_fw_command(bp, DRV_MSG_CODE_DRV_INFO_NACK, 0);
return;
}
op_code = (drv_info_ctl & DRV_INFO_CONTROL_OP_CODE_MASK) >>
DRV_INFO_CONTROL_OP_CODE_SHIFT;
/* Must prevent other flows from accessing drv_info_to_mcp */
mutex_lock(&bp->drv_info_mutex);
memset(&bp->slowpath->drv_info_to_mcp, 0,
sizeof(union drv_info_to_mcp));
switch (op_code) {
case ETH_STATS_OPCODE:
bnx2x_drv_info_ether_stat(bp);
break;
case FCOE_STATS_OPCODE:
bnx2x_drv_info_fcoe_stat(bp);
break;
case ISCSI_STATS_OPCODE:
bnx2x_drv_info_iscsi_stat(bp);
break;
default:
/* if op code isn't supported - send NACK */
bnx2x_fw_command(bp, DRV_MSG_CODE_DRV_INFO_NACK, 0);
goto out;
}
/* if we got drv_info attn from MFW then these fields are defined in
* shmem2 for sure
*/
SHMEM2_WR(bp, drv_info_host_addr_lo,
U64_LO(bnx2x_sp_mapping(bp, drv_info_to_mcp)));
SHMEM2_WR(bp, drv_info_host_addr_hi,
U64_HI(bnx2x_sp_mapping(bp, drv_info_to_mcp)));
bnx2x_fw_command(bp, DRV_MSG_CODE_DRV_INFO_ACK, 0);
/* Since possible management wants both this and get_driver_version
* need to wait until management notifies us it finished utilizing
* the buffer.
*/
if (!SHMEM2_HAS(bp, mfw_drv_indication)) {
DP(BNX2X_MSG_MCP, "Management does not support indication\n");
} else if (!bp->drv_info_mng_owner) {
u32 bit = MFW_DRV_IND_READ_DONE_OFFSET((BP_ABS_FUNC(bp) >> 1));
for (wait = 0; wait < BNX2X_UPDATE_DRV_INFO_IND_COUNT; wait++) {
u32 indication = SHMEM2_RD(bp, mfw_drv_indication);
/* Management is done; need to clear indication */
if (indication & bit) {
SHMEM2_WR(bp, mfw_drv_indication,
indication & ~bit);
release = true;
break;
}
msleep(BNX2X_UPDATE_DRV_INFO_IND_LENGTH);
}
}
if (!release) {
DP(BNX2X_MSG_MCP, "Management did not release indication\n");
bp->drv_info_mng_owner = true;
}
out:
mutex_unlock(&bp->drv_info_mutex);
}
static u32 bnx2x_update_mng_version_utility(u8 *version, bool bnx2x_format)
{
u8 vals[4];
int i = 0;
if (bnx2x_format) {
i = sscanf(version, "1.%c%hhd.%hhd.%hhd",
&vals[0], &vals[1], &vals[2], &vals[3]);
if (i > 0)
vals[0] -= '0';
} else {
i = sscanf(version, "%hhd.%hhd.%hhd.%hhd",
&vals[0], &vals[1], &vals[2], &vals[3]);
}
while (i < 4)
vals[i++] = 0;
return (vals[0] << 24) | (vals[1] << 16) | (vals[2] << 8) | vals[3];
}
void bnx2x_update_mng_version(struct bnx2x *bp)
{
u32 iscsiver = DRV_VER_NOT_LOADED;
u32 fcoever = DRV_VER_NOT_LOADED;
u32 ethver = DRV_VER_NOT_LOADED;
int idx = BP_FW_MB_IDX(bp);
u8 *version;
if (!SHMEM2_HAS(bp, func_os_drv_ver))
return;
mutex_lock(&bp->drv_info_mutex);
/* Must not proceed when `bnx2x_handle_drv_info_req' is feasible */
if (bp->drv_info_mng_owner)
goto out;
if (bp->state != BNX2X_STATE_OPEN)
goto out;
/* Parse ethernet driver version */
ethver = bnx2x_update_mng_version_utility(DRV_MODULE_VERSION, true);
if (!CNIC_LOADED(bp))
goto out;
/* Try getting storage driver version via cnic */
memset(&bp->slowpath->drv_info_to_mcp, 0,
sizeof(union drv_info_to_mcp));
bnx2x_drv_info_iscsi_stat(bp);
version = bp->slowpath->drv_info_to_mcp.iscsi_stat.version;
iscsiver = bnx2x_update_mng_version_utility(version, false);
memset(&bp->slowpath->drv_info_to_mcp, 0,
sizeof(union drv_info_to_mcp));
bnx2x_drv_info_fcoe_stat(bp);
version = bp->slowpath->drv_info_to_mcp.fcoe_stat.version;
fcoever = bnx2x_update_mng_version_utility(version, false);
out:
SHMEM2_WR(bp, func_os_drv_ver[idx].versions[DRV_PERS_ETHERNET], ethver);
SHMEM2_WR(bp, func_os_drv_ver[idx].versions[DRV_PERS_ISCSI], iscsiver);
SHMEM2_WR(bp, func_os_drv_ver[idx].versions[DRV_PERS_FCOE], fcoever);
mutex_unlock(&bp->drv_info_mutex);
DP(BNX2X_MSG_MCP, "Setting driver version: ETH [%08x] iSCSI [%08x] FCoE [%08x]\n",
ethver, iscsiver, fcoever);
}
void bnx2x_update_mfw_dump(struct bnx2x *bp)
{
u32 drv_ver;
u32 valid_dump;
if (!SHMEM2_HAS(bp, drv_info))
return;
/* Update Driver load time, possibly broken in y2038 */
SHMEM2_WR(bp, drv_info.epoc, (u32)ktime_get_real_seconds());
drv_ver = bnx2x_update_mng_version_utility(DRV_MODULE_VERSION, true);
SHMEM2_WR(bp, drv_info.drv_ver, drv_ver);
SHMEM2_WR(bp, drv_info.fw_ver, REG_RD(bp, XSEM_REG_PRAM));
/* Check & notify On-Chip dump. */
valid_dump = SHMEM2_RD(bp, drv_info.valid_dump);
if (valid_dump & FIRST_DUMP_VALID)
DP(NETIF_MSG_IFUP, "A valid On-Chip MFW dump found on 1st partition\n");
if (valid_dump & SECOND_DUMP_VALID)
DP(NETIF_MSG_IFUP, "A valid On-Chip MFW dump found on 2nd partition\n");
}
static void bnx2x_oem_event(struct bnx2x *bp, u32 event)
{
u32 cmd_ok, cmd_fail;
/* sanity */
if (event & DRV_STATUS_DCC_EVENT_MASK &&
event & DRV_STATUS_OEM_EVENT_MASK) {
BNX2X_ERR("Received simultaneous events %08x\n", event);
return;
}
if (event & DRV_STATUS_DCC_EVENT_MASK) {
cmd_fail = DRV_MSG_CODE_DCC_FAILURE;
cmd_ok = DRV_MSG_CODE_DCC_OK;
} else /* if (event & DRV_STATUS_OEM_EVENT_MASK) */ {
cmd_fail = DRV_MSG_CODE_OEM_FAILURE;
cmd_ok = DRV_MSG_CODE_OEM_OK;
}
DP(BNX2X_MSG_MCP, "oem_event 0x%x\n", event);
if (event & (DRV_STATUS_DCC_DISABLE_ENABLE_PF |
DRV_STATUS_OEM_DISABLE_ENABLE_PF)) {
/* This is the only place besides the function initialization
* where the bp->flags can change so it is done without any
* locks
*/
if (bp->mf_config[BP_VN(bp)] & FUNC_MF_CFG_FUNC_DISABLED) {
DP(BNX2X_MSG_MCP, "mf_cfg function disabled\n");
bp->flags |= MF_FUNC_DIS;
bnx2x_e1h_disable(bp);
} else {
DP(BNX2X_MSG_MCP, "mf_cfg function enabled\n");
bp->flags &= ~MF_FUNC_DIS;
bnx2x_e1h_enable(bp);
}
event &= ~(DRV_STATUS_DCC_DISABLE_ENABLE_PF |
DRV_STATUS_OEM_DISABLE_ENABLE_PF);
}
if (event & (DRV_STATUS_DCC_BANDWIDTH_ALLOCATION |
DRV_STATUS_OEM_BANDWIDTH_ALLOCATION)) {
bnx2x_config_mf_bw(bp);
event &= ~(DRV_STATUS_DCC_BANDWIDTH_ALLOCATION |
DRV_STATUS_OEM_BANDWIDTH_ALLOCATION);
}
/* Report results to MCP */
if (event)
bnx2x_fw_command(bp, cmd_fail, 0);
else
bnx2x_fw_command(bp, cmd_ok, 0);
}
/* must be called under the spq lock */
static struct eth_spe *bnx2x_sp_get_next(struct bnx2x *bp)
{
struct eth_spe *next_spe = bp->spq_prod_bd;
if (bp->spq_prod_bd == bp->spq_last_bd) {
bp->spq_prod_bd = bp->spq;
bp->spq_prod_idx = 0;
DP(BNX2X_MSG_SP, "end of spq\n");
} else {
bp->spq_prod_bd++;
bp->spq_prod_idx++;
}
return next_spe;
}
/* must be called under the spq lock */
static void bnx2x_sp_prod_update(struct bnx2x *bp)
{
int func = BP_FUNC(bp);
/*
* Make sure that BD data is updated before writing the producer:
* BD data is written to the memory, the producer is read from the
* memory, thus we need a full memory barrier to ensure the ordering.
*/
mb();
REG_WR16(bp, BAR_XSTRORM_INTMEM + XSTORM_SPQ_PROD_OFFSET(func),
bp->spq_prod_idx);
mmiowb();
}
/**
* bnx2x_is_contextless_ramrod - check if the current command ends on EQ
*
* @cmd: command to check
* @cmd_type: command type
*/
static bool bnx2x_is_contextless_ramrod(int cmd, int cmd_type)
{
if ((cmd_type == NONE_CONNECTION_TYPE) ||
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
(cmd == RAMROD_CMD_ID_ETH_FORWARD_SETUP) ||
(cmd == RAMROD_CMD_ID_ETH_CLASSIFICATION_RULES) ||
(cmd == RAMROD_CMD_ID_ETH_FILTER_RULES) ||
(cmd == RAMROD_CMD_ID_ETH_MULTICAST_RULES) ||
(cmd == RAMROD_CMD_ID_ETH_SET_MAC) ||
(cmd == RAMROD_CMD_ID_ETH_RSS_UPDATE))
return true;
else
return false;
}
/**
* bnx2x_sp_post - place a single command on an SP ring
*
* @bp: driver handle
* @command: command to place (e.g. SETUP, FILTER_RULES, etc.)
* @cid: SW CID the command is related to
* @data_hi: command private data address (high 32 bits)
* @data_lo: command private data address (low 32 bits)
* @cmd_type: command type (e.g. NONE, ETH)
*
* SP data is handled as if it's always an address pair, thus data fields are
* not swapped to little endian in upper functions. Instead this function swaps
* data as if it's two u32 fields.
*/
int bnx2x_sp_post(struct bnx2x *bp, int command, int cid,
u32 data_hi, u32 data_lo, int cmd_type)
{
struct eth_spe *spe;
u16 type;
bool common = bnx2x_is_contextless_ramrod(command, cmd_type);
#ifdef BNX2X_STOP_ON_ERROR
if (unlikely(bp->panic)) {
BNX2X_ERR("Can't post SP when there is panic\n");
return -EIO;
}
#endif
spin_lock_bh(&bp->spq_lock);
if (common) {
if (!atomic_read(&bp->eq_spq_left)) {
BNX2X_ERR("BUG! EQ ring full!\n");
spin_unlock_bh(&bp->spq_lock);
bnx2x_panic();
return -EBUSY;
}
} else if (!atomic_read(&bp->cq_spq_left)) {
BNX2X_ERR("BUG! SPQ ring full!\n");
spin_unlock_bh(&bp->spq_lock);
bnx2x_panic();
return -EBUSY;
}
spe = bnx2x_sp_get_next(bp);
/* CID needs port number to be encoded int it */
spe->hdr.conn_and_cmd_data =
cpu_to_le32((command << SPE_HDR_CMD_ID_SHIFT) |
HW_CID(bp, cid));
/* In some cases, type may already contain the func-id
* mainly in SRIOV related use cases, so we add it here only
* if it's not already set.
*/
if (!(cmd_type & SPE_HDR_FUNCTION_ID)) {
type = (cmd_type << SPE_HDR_CONN_TYPE_SHIFT) &
SPE_HDR_CONN_TYPE;
type |= ((BP_FUNC(bp) << SPE_HDR_FUNCTION_ID_SHIFT) &
SPE_HDR_FUNCTION_ID);
} else {
type = cmd_type;
}
spe->hdr.type = cpu_to_le16(type);
spe->data.update_data_addr.hi = cpu_to_le32(data_hi);
spe->data.update_data_addr.lo = cpu_to_le32(data_lo);
/*
* It's ok if the actual decrement is issued towards the memory
* somewhere between the spin_lock and spin_unlock. Thus no
* more explicit memory barrier is needed.
*/
if (common)
atomic_dec(&bp->eq_spq_left);
else
atomic_dec(&bp->cq_spq_left);
DP(BNX2X_MSG_SP,
"SPQE[%x] (%x:%x) (cmd, common?) (%d,%d) hw_cid %x data (%x:%x) type(0x%x) left (CQ, EQ) (%x,%x)\n",
bp->spq_prod_idx, (u32)U64_HI(bp->spq_mapping),
(u32)(U64_LO(bp->spq_mapping) +
(void *)bp->spq_prod_bd - (void *)bp->spq), command, common,
HW_CID(bp, cid), data_hi, data_lo, type,
atomic_read(&bp->cq_spq_left), atomic_read(&bp->eq_spq_left));
bnx2x_sp_prod_update(bp);
spin_unlock_bh(&bp->spq_lock);
return 0;
}
/* acquire split MCP access lock register */
static int bnx2x_acquire_alr(struct bnx2x *bp)
{
u32 j, val;
int rc = 0;
might_sleep();
for (j = 0; j < 1000; j++) {
REG_WR(bp, MCP_REG_MCPR_ACCESS_LOCK, MCPR_ACCESS_LOCK_LOCK);
val = REG_RD(bp, MCP_REG_MCPR_ACCESS_LOCK);
if (val & MCPR_ACCESS_LOCK_LOCK)
break;
usleep_range(5000, 10000);
}
if (!(val & MCPR_ACCESS_LOCK_LOCK)) {
BNX2X_ERR("Cannot acquire MCP access lock register\n");
rc = -EBUSY;
}
return rc;
}
/* release split MCP access lock register */
static void bnx2x_release_alr(struct bnx2x *bp)
{
REG_WR(bp, MCP_REG_MCPR_ACCESS_LOCK, 0);
}
#define BNX2X_DEF_SB_ATT_IDX 0x0001
#define BNX2X_DEF_SB_IDX 0x0002
static u16 bnx2x_update_dsb_idx(struct bnx2x *bp)
{
struct host_sp_status_block *def_sb = bp->def_status_blk;
u16 rc = 0;
barrier(); /* status block is written to by the chip */
if (bp->def_att_idx != def_sb->atten_status_block.attn_bits_index) {
bp->def_att_idx = def_sb->atten_status_block.attn_bits_index;
rc |= BNX2X_DEF_SB_ATT_IDX;
}
if (bp->def_idx != def_sb->sp_sb.running_index) {
bp->def_idx = def_sb->sp_sb.running_index;
rc |= BNX2X_DEF_SB_IDX;
}
/* Do not reorder: indices reading should complete before handling */
barrier();
return rc;
}
/*
* slow path service functions
*/
static void bnx2x_attn_int_asserted(struct bnx2x *bp, u32 asserted)
{
int port = BP_PORT(bp);
u32 aeu_addr = port ? MISC_REG_AEU_MASK_ATTN_FUNC_1 :
MISC_REG_AEU_MASK_ATTN_FUNC_0;
u32 nig_int_mask_addr = port ? NIG_REG_MASK_INTERRUPT_PORT1 :
NIG_REG_MASK_INTERRUPT_PORT0;
u32 aeu_mask;
u32 nig_mask = 0;
u32 reg_addr;
if (bp->attn_state & asserted)
BNX2X_ERR("IGU ERROR\n");
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_PORT0_ATT_MASK + port);
aeu_mask = REG_RD(bp, aeu_addr);
DP(NETIF_MSG_HW, "aeu_mask %x newly asserted %x\n",
aeu_mask, asserted);
aeu_mask &= ~(asserted & 0x3ff);
DP(NETIF_MSG_HW, "new mask %x\n", aeu_mask);
REG_WR(bp, aeu_addr, aeu_mask);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_PORT0_ATT_MASK + port);
DP(NETIF_MSG_HW, "attn_state %x\n", bp->attn_state);
bp->attn_state |= asserted;
DP(NETIF_MSG_HW, "new state %x\n", bp->attn_state);
if (asserted & ATTN_HARD_WIRED_MASK) {
if (asserted & ATTN_NIG_FOR_FUNC) {
bnx2x_acquire_phy_lock(bp);
/* save nig interrupt mask */
nig_mask = REG_RD(bp, nig_int_mask_addr);
/* If nig_mask is not set, no need to call the update
* function.
*/
if (nig_mask) {
REG_WR(bp, nig_int_mask_addr, 0);
bnx2x_link_attn(bp);
}
/* handle unicore attn? */
}
if (asserted & ATTN_SW_TIMER_4_FUNC)
DP(NETIF_MSG_HW, "ATTN_SW_TIMER_4_FUNC!\n");
if (asserted & GPIO_2_FUNC)
DP(NETIF_MSG_HW, "GPIO_2_FUNC!\n");
if (asserted & GPIO_3_FUNC)
DP(NETIF_MSG_HW, "GPIO_3_FUNC!\n");
if (asserted & GPIO_4_FUNC)
DP(NETIF_MSG_HW, "GPIO_4_FUNC!\n");
if (port == 0) {
if (asserted & ATTN_GENERAL_ATTN_1) {
DP(NETIF_MSG_HW, "ATTN_GENERAL_ATTN_1!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_1, 0x0);
}
if (asserted & ATTN_GENERAL_ATTN_2) {
DP(NETIF_MSG_HW, "ATTN_GENERAL_ATTN_2!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_2, 0x0);
}
if (asserted & ATTN_GENERAL_ATTN_3) {
DP(NETIF_MSG_HW, "ATTN_GENERAL_ATTN_3!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_3, 0x0);
}
} else {
if (asserted & ATTN_GENERAL_ATTN_4) {
DP(NETIF_MSG_HW, "ATTN_GENERAL_ATTN_4!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_4, 0x0);
}
if (asserted & ATTN_GENERAL_ATTN_5) {
DP(NETIF_MSG_HW, "ATTN_GENERAL_ATTN_5!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_5, 0x0);
}
if (asserted & ATTN_GENERAL_ATTN_6) {
DP(NETIF_MSG_HW, "ATTN_GENERAL_ATTN_6!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_6, 0x0);
}
}
} /* if hardwired */
if (bp->common.int_block == INT_BLOCK_HC)
reg_addr = (HC_REG_COMMAND_REG + port*32 +
COMMAND_REG_ATTN_BITS_SET);
else
reg_addr = (BAR_IGU_INTMEM + IGU_CMD_ATTN_BIT_SET_UPPER*8);
DP(NETIF_MSG_HW, "about to mask 0x%08x at %s addr 0x%x\n", asserted,
(bp->common.int_block == INT_BLOCK_HC) ? "HC" : "IGU", reg_addr);
REG_WR(bp, reg_addr, asserted);
/* now set back the mask */
if (asserted & ATTN_NIG_FOR_FUNC) {
/* Verify that IGU ack through BAR was written before restoring
* NIG mask. This loop should exit after 2-3 iterations max.
*/
if (bp->common.int_block != INT_BLOCK_HC) {
u32 cnt = 0, igu_acked;
do {
igu_acked = REG_RD(bp,
IGU_REG_ATTENTION_ACK_BITS);
} while (((igu_acked & ATTN_NIG_FOR_FUNC) == 0) &&
(++cnt < MAX_IGU_ATTN_ACK_TO));
if (!igu_acked)
DP(NETIF_MSG_HW,
"Failed to verify IGU ack on time\n");
barrier();
}
REG_WR(bp, nig_int_mask_addr, nig_mask);
bnx2x_release_phy_lock(bp);
}
}
static void bnx2x_fan_failure(struct bnx2x *bp)
{
int port = BP_PORT(bp);
u32 ext_phy_config;
/* mark the failure */
ext_phy_config =
SHMEM_RD(bp,
dev_info.port_hw_config[port].external_phy_config);
ext_phy_config &= ~PORT_HW_CFG_XGXS_EXT_PHY_TYPE_MASK;
ext_phy_config |= PORT_HW_CFG_XGXS_EXT_PHY_TYPE_FAILURE;
SHMEM_WR(bp, dev_info.port_hw_config[port].external_phy_config,
ext_phy_config);
/* log the failure */
netdev_err(bp->dev, "Fan Failure on Network Controller has caused the driver to shutdown the card to prevent permanent damage.\n"
"Please contact OEM Support for assistance\n");
/* Schedule device reset (unload)
* This is due to some boards consuming sufficient power when driver is
* up to overheat if fan fails.
*/
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_FAN_FAILURE, 0);
}
static void bnx2x_attn_int_deasserted0(struct bnx2x *bp, u32 attn)
{
int port = BP_PORT(bp);
int reg_offset;
u32 val;
reg_offset = (port ? MISC_REG_AEU_ENABLE1_FUNC_1_OUT_0 :
MISC_REG_AEU_ENABLE1_FUNC_0_OUT_0);
if (attn & AEU_INPUTS_ATTN_BITS_SPIO5) {
val = REG_RD(bp, reg_offset);
val &= ~AEU_INPUTS_ATTN_BITS_SPIO5;
REG_WR(bp, reg_offset, val);
BNX2X_ERR("SPIO5 hw attention\n");
/* Fan failure attention */
bnx2x_hw_reset_phy(&bp->link_params);
bnx2x_fan_failure(bp);
}
if ((attn & bp->link_vars.aeu_int_mask) && bp->port.pmf) {
bnx2x_acquire_phy_lock(bp);
bnx2x_handle_module_detect_int(&bp->link_params);
bnx2x_release_phy_lock(bp);
}
if (attn & HW_INTERRUT_ASSERT_SET_0) {
val = REG_RD(bp, reg_offset);
val &= ~(attn & HW_INTERRUT_ASSERT_SET_0);
REG_WR(bp, reg_offset, val);
BNX2X_ERR("FATAL HW block attention set0 0x%x\n",
(u32)(attn & HW_INTERRUT_ASSERT_SET_0));
bnx2x_panic();
}
}
static void bnx2x_attn_int_deasserted1(struct bnx2x *bp, u32 attn)
{
u32 val;
if (attn & AEU_INPUTS_ATTN_BITS_DOORBELLQ_HW_INTERRUPT) {
val = REG_RD(bp, DORQ_REG_DORQ_INT_STS_CLR);
BNX2X_ERR("DB hw attention 0x%x\n", val);
/* DORQ discard attention */
if (val & 0x2)
BNX2X_ERR("FATAL error from DORQ\n");
}
if (attn & HW_INTERRUT_ASSERT_SET_1) {
int port = BP_PORT(bp);
int reg_offset;
reg_offset = (port ? MISC_REG_AEU_ENABLE1_FUNC_1_OUT_1 :
MISC_REG_AEU_ENABLE1_FUNC_0_OUT_1);
val = REG_RD(bp, reg_offset);
val &= ~(attn & HW_INTERRUT_ASSERT_SET_1);
REG_WR(bp, reg_offset, val);
BNX2X_ERR("FATAL HW block attention set1 0x%x\n",
(u32)(attn & HW_INTERRUT_ASSERT_SET_1));
bnx2x_panic();
}
}
static void bnx2x_attn_int_deasserted2(struct bnx2x *bp, u32 attn)
{
u32 val;
if (attn & AEU_INPUTS_ATTN_BITS_CFC_HW_INTERRUPT) {
val = REG_RD(bp, CFC_REG_CFC_INT_STS_CLR);
BNX2X_ERR("CFC hw attention 0x%x\n", val);
/* CFC error attention */
if (val & 0x2)
BNX2X_ERR("FATAL error from CFC\n");
}
if (attn & AEU_INPUTS_ATTN_BITS_PXP_HW_INTERRUPT) {
val = REG_RD(bp, PXP_REG_PXP_INT_STS_CLR_0);
BNX2X_ERR("PXP hw attention-0 0x%x\n", val);
/* RQ_USDMDP_FIFO_OVERFLOW */
if (val & 0x18000)
BNX2X_ERR("FATAL error from PXP\n");
if (!CHIP_IS_E1x(bp)) {
val = REG_RD(bp, PXP_REG_PXP_INT_STS_CLR_1);
BNX2X_ERR("PXP hw attention-1 0x%x\n", val);
}
}
if (attn & HW_INTERRUT_ASSERT_SET_2) {
int port = BP_PORT(bp);
int reg_offset;
reg_offset = (port ? MISC_REG_AEU_ENABLE1_FUNC_1_OUT_2 :
MISC_REG_AEU_ENABLE1_FUNC_0_OUT_2);
val = REG_RD(bp, reg_offset);
val &= ~(attn & HW_INTERRUT_ASSERT_SET_2);
REG_WR(bp, reg_offset, val);
BNX2X_ERR("FATAL HW block attention set2 0x%x\n",
(u32)(attn & HW_INTERRUT_ASSERT_SET_2));
bnx2x_panic();
}
}
static void bnx2x_attn_int_deasserted3(struct bnx2x *bp, u32 attn)
{
u32 val;
if (attn & EVEREST_GEN_ATTN_IN_USE_MASK) {
if (attn & BNX2X_PMF_LINK_ASSERT) {
int func = BP_FUNC(bp);
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_12 + func*4, 0);
bnx2x_read_mf_cfg(bp);
bp->mf_config[BP_VN(bp)] = MF_CFG_RD(bp,
func_mf_config[BP_ABS_FUNC(bp)].config);
val = SHMEM_RD(bp,
func_mb[BP_FW_MB_IDX(bp)].drv_status);
if (val & (DRV_STATUS_DCC_EVENT_MASK |
DRV_STATUS_OEM_EVENT_MASK))
bnx2x_oem_event(bp,
(val & (DRV_STATUS_DCC_EVENT_MASK |
DRV_STATUS_OEM_EVENT_MASK)));
if (val & DRV_STATUS_SET_MF_BW)
bnx2x_set_mf_bw(bp);
if (val & DRV_STATUS_DRV_INFO_REQ)
bnx2x_handle_drv_info_req(bp);
if (val & DRV_STATUS_VF_DISABLED)
bnx2x_schedule_iov_task(bp,
BNX2X_IOV_HANDLE_FLR);
if ((bp->port.pmf == 0) && (val & DRV_STATUS_PMF))
bnx2x_pmf_update(bp);
if (bp->port.pmf &&
(val & DRV_STATUS_DCBX_NEGOTIATION_RESULTS) &&
bp->dcbx_enabled > 0)
/* start dcbx state machine */
bnx2x_dcbx_set_params(bp,
BNX2X_DCBX_STATE_NEG_RECEIVED);
if (val & DRV_STATUS_AFEX_EVENT_MASK)
bnx2x_handle_afex_cmd(bp,
val & DRV_STATUS_AFEX_EVENT_MASK);
if (val & DRV_STATUS_EEE_NEGOTIATION_RESULTS)
bnx2x_handle_eee_event(bp);
if (val & DRV_STATUS_OEM_UPDATE_SVID)
bnx2x_handle_update_svid_cmd(bp);
if (bp->link_vars.periodic_flags &
PERIODIC_FLAGS_LINK_EVENT) {
/* sync with link */
bnx2x_acquire_phy_lock(bp);
bp->link_vars.periodic_flags &=
~PERIODIC_FLAGS_LINK_EVENT;
bnx2x_release_phy_lock(bp);
if (IS_MF(bp))
bnx2x_link_sync_notify(bp);
bnx2x_link_report(bp);
}
/* Always call it here: bnx2x_link_report() will
* prevent the link indication duplication.
*/
bnx2x__link_status_update(bp);
} else if (attn & BNX2X_MC_ASSERT_BITS) {
BNX2X_ERR("MC assert!\n");
bnx2x_mc_assert(bp);
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_10, 0);
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_9, 0);
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_8, 0);
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_7, 0);
bnx2x_panic();
} else if (attn & BNX2X_MCP_ASSERT) {
BNX2X_ERR("MCP assert!\n");
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_11, 0);
bnx2x_fw_dump(bp);
} else
BNX2X_ERR("Unknown HW assert! (attn 0x%x)\n", attn);
}
if (attn & EVEREST_LATCHED_ATTN_IN_USE_MASK) {
BNX2X_ERR("LATCHED attention 0x%08x (masked)\n", attn);
if (attn & BNX2X_GRC_TIMEOUT) {
val = CHIP_IS_E1(bp) ? 0 :
REG_RD(bp, MISC_REG_GRC_TIMEOUT_ATTN);
BNX2X_ERR("GRC time-out 0x%08x\n", val);
}
if (attn & BNX2X_GRC_RSV) {
val = CHIP_IS_E1(bp) ? 0 :
REG_RD(bp, MISC_REG_GRC_RSV_ATTN);
BNX2X_ERR("GRC reserved 0x%08x\n", val);
}
REG_WR(bp, MISC_REG_AEU_CLR_LATCH_SIGNAL, 0x7ff);
}
}
/*
* Bits map:
* 0-7 - Engine0 load counter.
* 8-15 - Engine1 load counter.
* 16 - Engine0 RESET_IN_PROGRESS bit.
* 17 - Engine1 RESET_IN_PROGRESS bit.
* 18 - Engine0 ONE_IS_LOADED. Set when there is at least one active function
* on the engine
* 19 - Engine1 ONE_IS_LOADED.
* 20 - Chip reset flow bit. When set none-leader must wait for both engines
* leader to complete (check for both RESET_IN_PROGRESS bits and not for
* just the one belonging to its engine).
*
*/
#define BNX2X_RECOVERY_GLOB_REG MISC_REG_GENERIC_POR_1
#define BNX2X_PATH0_LOAD_CNT_MASK 0x000000ff
#define BNX2X_PATH0_LOAD_CNT_SHIFT 0
#define BNX2X_PATH1_LOAD_CNT_MASK 0x0000ff00
#define BNX2X_PATH1_LOAD_CNT_SHIFT 8
#define BNX2X_PATH0_RST_IN_PROG_BIT 0x00010000
#define BNX2X_PATH1_RST_IN_PROG_BIT 0x00020000
#define BNX2X_GLOBAL_RESET_BIT 0x00040000
/*
* Set the GLOBAL_RESET bit.
*
* Should be run under rtnl lock
*/
void bnx2x_set_reset_global(struct bnx2x *bp)
{
u32 val;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
REG_WR(bp, BNX2X_RECOVERY_GLOB_REG, val | BNX2X_GLOBAL_RESET_BIT);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
}
/*
* Clear the GLOBAL_RESET bit.
*
* Should be run under rtnl lock
*/
static void bnx2x_clear_reset_global(struct bnx2x *bp)
{
u32 val;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
REG_WR(bp, BNX2X_RECOVERY_GLOB_REG, val & (~BNX2X_GLOBAL_RESET_BIT));
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
}
/*
* Checks the GLOBAL_RESET bit.
*
* should be run under rtnl lock
*/
static bool bnx2x_reset_is_global(struct bnx2x *bp)
{
u32 val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
DP(NETIF_MSG_HW, "GEN_REG_VAL=0x%08x\n", val);
return (val & BNX2X_GLOBAL_RESET_BIT) ? true : false;
}
/*
* Clear RESET_IN_PROGRESS bit for the current engine.
*
* Should be run under rtnl lock
*/
static void bnx2x_set_reset_done(struct bnx2x *bp)
{
u32 val;
u32 bit = BP_PATH(bp) ?
BNX2X_PATH1_RST_IN_PROG_BIT : BNX2X_PATH0_RST_IN_PROG_BIT;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
/* Clear the bit */
val &= ~bit;
REG_WR(bp, BNX2X_RECOVERY_GLOB_REG, val);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
}
/*
* Set RESET_IN_PROGRESS for the current engine.
*
* should be run under rtnl lock
*/
void bnx2x_set_reset_in_progress(struct bnx2x *bp)
{
u32 val;
u32 bit = BP_PATH(bp) ?
BNX2X_PATH1_RST_IN_PROG_BIT : BNX2X_PATH0_RST_IN_PROG_BIT;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
/* Set the bit */
val |= bit;
REG_WR(bp, BNX2X_RECOVERY_GLOB_REG, val);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
}
/*
* Checks the RESET_IN_PROGRESS bit for the given engine.
* should be run under rtnl lock
*/
bool bnx2x_reset_is_done(struct bnx2x *bp, int engine)
{
u32 val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
u32 bit = engine ?
BNX2X_PATH1_RST_IN_PROG_BIT : BNX2X_PATH0_RST_IN_PROG_BIT;
/* return false if bit is set */
return (val & bit) ? false : true;
}
/*
* set pf load for the current pf.
*
* should be run under rtnl lock
*/
void bnx2x_set_pf_load(struct bnx2x *bp)
{
u32 val1, val;
u32 mask = BP_PATH(bp) ? BNX2X_PATH1_LOAD_CNT_MASK :
BNX2X_PATH0_LOAD_CNT_MASK;
u32 shift = BP_PATH(bp) ? BNX2X_PATH1_LOAD_CNT_SHIFT :
BNX2X_PATH0_LOAD_CNT_SHIFT;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
DP(NETIF_MSG_IFUP, "Old GEN_REG_VAL=0x%08x\n", val);
/* get the current counter value */
val1 = (val & mask) >> shift;
/* set bit of that PF */
val1 |= (1 << bp->pf_num);
/* clear the old value */
val &= ~mask;
/* set the new one */
val |= ((val1 << shift) & mask);
REG_WR(bp, BNX2X_RECOVERY_GLOB_REG, val);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
}
/**
* bnx2x_clear_pf_load - clear pf load mark
*
* @bp: driver handle
*
* Should be run under rtnl lock.
* Decrements the load counter for the current engine. Returns
* whether other functions are still loaded
*/
bool bnx2x_clear_pf_load(struct bnx2x *bp)
{
u32 val1, val;
u32 mask = BP_PATH(bp) ? BNX2X_PATH1_LOAD_CNT_MASK :
BNX2X_PATH0_LOAD_CNT_MASK;
u32 shift = BP_PATH(bp) ? BNX2X_PATH1_LOAD_CNT_SHIFT :
BNX2X_PATH0_LOAD_CNT_SHIFT;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
DP(NETIF_MSG_IFDOWN, "Old GEN_REG_VAL=0x%08x\n", val);
/* get the current counter value */
val1 = (val & mask) >> shift;
/* clear bit of that PF */
val1 &= ~(1 << bp->pf_num);
/* clear the old value */
val &= ~mask;
/* set the new one */
val |= ((val1 << shift) & mask);
REG_WR(bp, BNX2X_RECOVERY_GLOB_REG, val);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RECOVERY_REG);
return val1 != 0;
}
/*
* Read the load status for the current engine.
*
* should be run under rtnl lock
*/
static bool bnx2x_get_load_status(struct bnx2x *bp, int engine)
{
u32 mask = (engine ? BNX2X_PATH1_LOAD_CNT_MASK :
BNX2X_PATH0_LOAD_CNT_MASK);
u32 shift = (engine ? BNX2X_PATH1_LOAD_CNT_SHIFT :
BNX2X_PATH0_LOAD_CNT_SHIFT);
u32 val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
DP(NETIF_MSG_HW | NETIF_MSG_IFUP, "GLOB_REG=0x%08x\n", val);
val = (val & mask) >> shift;
DP(NETIF_MSG_HW | NETIF_MSG_IFUP, "load mask for engine %d = 0x%x\n",
engine, val);
return val != 0;
}
static void _print_parity(struct bnx2x *bp, u32 reg)
{
pr_cont(" [0x%08x] ", REG_RD(bp, reg));
}
static void _print_next_block(int idx, const char *blk)
{
pr_cont("%s%s", idx ? ", " : "", blk);
}
static bool bnx2x_check_blocks_with_parity0(struct bnx2x *bp, u32 sig,
int *par_num, bool print)
{
u32 cur_bit;
bool res;
int i;
res = false;
for (i = 0; sig; i++) {
cur_bit = (0x1UL << i);
if (sig & cur_bit) {
res |= true; /* Each bit is real error! */
if (print) {
switch (cur_bit) {
case AEU_INPUTS_ATTN_BITS_BRB_PARITY_ERROR:
_print_next_block((*par_num)++, "BRB");
_print_parity(bp,
BRB1_REG_BRB1_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_PARSER_PARITY_ERROR:
_print_next_block((*par_num)++,
"PARSER");
_print_parity(bp, PRS_REG_PRS_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_TSDM_PARITY_ERROR:
_print_next_block((*par_num)++, "TSDM");
_print_parity(bp,
TSDM_REG_TSDM_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_SEARCHER_PARITY_ERROR:
_print_next_block((*par_num)++,
"SEARCHER");
_print_parity(bp, SRC_REG_SRC_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_TCM_PARITY_ERROR:
_print_next_block((*par_num)++, "TCM");
_print_parity(bp, TCM_REG_TCM_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_TSEMI_PARITY_ERROR:
_print_next_block((*par_num)++,
"TSEMI");
_print_parity(bp,
TSEM_REG_TSEM_PRTY_STS_0);
_print_parity(bp,
TSEM_REG_TSEM_PRTY_STS_1);
break;
case AEU_INPUTS_ATTN_BITS_PBCLIENT_PARITY_ERROR:
_print_next_block((*par_num)++, "XPB");
_print_parity(bp, GRCBASE_XPB +
PB_REG_PB_PRTY_STS);
break;
}
}
/* Clear the bit */
sig &= ~cur_bit;
}
}
return res;
}
static bool bnx2x_check_blocks_with_parity1(struct bnx2x *bp, u32 sig,
int *par_num, bool *global,
bool print)
{
u32 cur_bit;
bool res;
int i;
res = false;
for (i = 0; sig; i++) {
cur_bit = (0x1UL << i);
if (sig & cur_bit) {
res |= true; /* Each bit is real error! */
switch (cur_bit) {
case AEU_INPUTS_ATTN_BITS_PBF_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "PBF");
_print_parity(bp, PBF_REG_PBF_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_QM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "QM");
_print_parity(bp, QM_REG_QM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_TIMERS_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "TM");
_print_parity(bp, TM_REG_TM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_XSDM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "XSDM");
_print_parity(bp,
XSDM_REG_XSDM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_XCM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "XCM");
_print_parity(bp, XCM_REG_XCM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_XSEMI_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++,
"XSEMI");
_print_parity(bp,
XSEM_REG_XSEM_PRTY_STS_0);
_print_parity(bp,
XSEM_REG_XSEM_PRTY_STS_1);
}
break;
case AEU_INPUTS_ATTN_BITS_DOORBELLQ_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++,
"DOORBELLQ");
_print_parity(bp,
DORQ_REG_DORQ_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_NIG_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "NIG");
if (CHIP_IS_E1x(bp)) {
_print_parity(bp,
NIG_REG_NIG_PRTY_STS);
} else {
_print_parity(bp,
NIG_REG_NIG_PRTY_STS_0);
_print_parity(bp,
NIG_REG_NIG_PRTY_STS_1);
}
}
break;
case AEU_INPUTS_ATTN_BITS_VAUX_PCI_CORE_PARITY_ERROR:
if (print)
_print_next_block((*par_num)++,
"VAUX PCI CORE");
*global = true;
break;
case AEU_INPUTS_ATTN_BITS_DEBUG_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++,
"DEBUG");
_print_parity(bp, DBG_REG_DBG_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_USDM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "USDM");
_print_parity(bp,
USDM_REG_USDM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_UCM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "UCM");
_print_parity(bp, UCM_REG_UCM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_USEMI_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++,
"USEMI");
_print_parity(bp,
USEM_REG_USEM_PRTY_STS_0);
_print_parity(bp,
USEM_REG_USEM_PRTY_STS_1);
}
break;
case AEU_INPUTS_ATTN_BITS_UPB_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "UPB");
_print_parity(bp, GRCBASE_UPB +
PB_REG_PB_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_CSDM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "CSDM");
_print_parity(bp,
CSDM_REG_CSDM_PRTY_STS);
}
break;
case AEU_INPUTS_ATTN_BITS_CCM_PARITY_ERROR:
if (print) {
_print_next_block((*par_num)++, "CCM");
_print_parity(bp, CCM_REG_CCM_PRTY_STS);
}
break;
}
/* Clear the bit */
sig &= ~cur_bit;
}
}
return res;
}
static bool bnx2x_check_blocks_with_parity2(struct bnx2x *bp, u32 sig,
int *par_num, bool print)
{
u32 cur_bit;
bool res;
int i;
res = false;
for (i = 0; sig; i++) {
cur_bit = (0x1UL << i);
if (sig & cur_bit) {
res = true; /* Each bit is real error! */
if (print) {
switch (cur_bit) {
case AEU_INPUTS_ATTN_BITS_CSEMI_PARITY_ERROR:
_print_next_block((*par_num)++,
"CSEMI");
_print_parity(bp,
CSEM_REG_CSEM_PRTY_STS_0);
_print_parity(bp,
CSEM_REG_CSEM_PRTY_STS_1);
break;
case AEU_INPUTS_ATTN_BITS_PXP_PARITY_ERROR:
_print_next_block((*par_num)++, "PXP");
_print_parity(bp, PXP_REG_PXP_PRTY_STS);
_print_parity(bp,
PXP2_REG_PXP2_PRTY_STS_0);
_print_parity(bp,
PXP2_REG_PXP2_PRTY_STS_1);
break;
case AEU_IN_ATTN_BITS_PXPPCICLOCKCLIENT_PARITY_ERROR:
_print_next_block((*par_num)++,
"PXPPCICLOCKCLIENT");
break;
case AEU_INPUTS_ATTN_BITS_CFC_PARITY_ERROR:
_print_next_block((*par_num)++, "CFC");
_print_parity(bp,
CFC_REG_CFC_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_CDU_PARITY_ERROR:
_print_next_block((*par_num)++, "CDU");
_print_parity(bp, CDU_REG_CDU_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_DMAE_PARITY_ERROR:
_print_next_block((*par_num)++, "DMAE");
_print_parity(bp,
DMAE_REG_DMAE_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_IGU_PARITY_ERROR:
_print_next_block((*par_num)++, "IGU");
if (CHIP_IS_E1x(bp))
_print_parity(bp,
HC_REG_HC_PRTY_STS);
else
_print_parity(bp,
IGU_REG_IGU_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_MISC_PARITY_ERROR:
_print_next_block((*par_num)++, "MISC");
_print_parity(bp,
MISC_REG_MISC_PRTY_STS);
break;
}
}
/* Clear the bit */
sig &= ~cur_bit;
}
}
return res;
}
static bool bnx2x_check_blocks_with_parity3(struct bnx2x *bp, u32 sig,
int *par_num, bool *global,
bool print)
{
bool res = false;
u32 cur_bit;
int i;
for (i = 0; sig; i++) {
cur_bit = (0x1UL << i);
if (sig & cur_bit) {
switch (cur_bit) {
case AEU_INPUTS_ATTN_BITS_MCP_LATCHED_ROM_PARITY:
if (print)
_print_next_block((*par_num)++,
"MCP ROM");
*global = true;
res = true;
break;
case AEU_INPUTS_ATTN_BITS_MCP_LATCHED_UMP_RX_PARITY:
if (print)
_print_next_block((*par_num)++,
"MCP UMP RX");
*global = true;
res = true;
break;
case AEU_INPUTS_ATTN_BITS_MCP_LATCHED_UMP_TX_PARITY:
if (print)
_print_next_block((*par_num)++,
"MCP UMP TX");
*global = true;
res = true;
break;
case AEU_INPUTS_ATTN_BITS_MCP_LATCHED_SCPAD_PARITY:
(*par_num)++;
/* clear latched SCPAD PATIRY from MCP */
REG_WR(bp, MISC_REG_AEU_CLR_LATCH_SIGNAL,
1UL << 10);
break;
}
/* Clear the bit */
sig &= ~cur_bit;
}
}
return res;
}
static bool bnx2x_check_blocks_with_parity4(struct bnx2x *bp, u32 sig,
int *par_num, bool print)
{
u32 cur_bit;
bool res;
int i;
res = false;
for (i = 0; sig; i++) {
cur_bit = (0x1UL << i);
if (sig & cur_bit) {
res = true; /* Each bit is real error! */
if (print) {
switch (cur_bit) {
case AEU_INPUTS_ATTN_BITS_PGLUE_PARITY_ERROR:
_print_next_block((*par_num)++,
"PGLUE_B");
_print_parity(bp,
PGLUE_B_REG_PGLUE_B_PRTY_STS);
break;
case AEU_INPUTS_ATTN_BITS_ATC_PARITY_ERROR:
_print_next_block((*par_num)++, "ATC");
_print_parity(bp,
ATC_REG_ATC_PRTY_STS);
break;
}
}
/* Clear the bit */
sig &= ~cur_bit;
}
}
return res;
}
static bool bnx2x_parity_attn(struct bnx2x *bp, bool *global, bool print,
u32 *sig)
{
bool res = false;
if ((sig[0] & HW_PRTY_ASSERT_SET_0) ||
(sig[1] & HW_PRTY_ASSERT_SET_1) ||
(sig[2] & HW_PRTY_ASSERT_SET_2) ||
(sig[3] & HW_PRTY_ASSERT_SET_3) ||
(sig[4] & HW_PRTY_ASSERT_SET_4)) {
int par_num = 0;
DP(NETIF_MSG_HW, "Was parity error: HW block parity attention:\n"
"[0]:0x%08x [1]:0x%08x [2]:0x%08x [3]:0x%08x [4]:0x%08x\n",
sig[0] & HW_PRTY_ASSERT_SET_0,
sig[1] & HW_PRTY_ASSERT_SET_1,
sig[2] & HW_PRTY_ASSERT_SET_2,
sig[3] & HW_PRTY_ASSERT_SET_3,
sig[4] & HW_PRTY_ASSERT_SET_4);
if (print) {
if (((sig[0] & HW_PRTY_ASSERT_SET_0) ||
(sig[1] & HW_PRTY_ASSERT_SET_1) ||
(sig[2] & HW_PRTY_ASSERT_SET_2) ||
(sig[4] & HW_PRTY_ASSERT_SET_4)) ||
(sig[3] & HW_PRTY_ASSERT_SET_3_WITHOUT_SCPAD)) {
netdev_err(bp->dev,
"Parity errors detected in blocks: ");
} else {
print = false;
}
}
res |= bnx2x_check_blocks_with_parity0(bp,
sig[0] & HW_PRTY_ASSERT_SET_0, &par_num, print);
res |= bnx2x_check_blocks_with_parity1(bp,
sig[1] & HW_PRTY_ASSERT_SET_1, &par_num, global, print);
res |= bnx2x_check_blocks_with_parity2(bp,
sig[2] & HW_PRTY_ASSERT_SET_2, &par_num, print);
res |= bnx2x_check_blocks_with_parity3(bp,
sig[3] & HW_PRTY_ASSERT_SET_3, &par_num, global, print);
res |= bnx2x_check_blocks_with_parity4(bp,
sig[4] & HW_PRTY_ASSERT_SET_4, &par_num, print);
if (print)
pr_cont("\n");
}
return res;
}
/**
* bnx2x_chk_parity_attn - checks for parity attentions.
*
* @bp: driver handle
* @global: true if there was a global attention
* @print: show parity attention in syslog
*/
bool bnx2x_chk_parity_attn(struct bnx2x *bp, bool *global, bool print)
{
struct attn_route attn = { {0} };
int port = BP_PORT(bp);
attn.sig[0] = REG_RD(bp,
MISC_REG_AEU_AFTER_INVERT_1_FUNC_0 +
port*4);
attn.sig[1] = REG_RD(bp,
MISC_REG_AEU_AFTER_INVERT_2_FUNC_0 +
port*4);
attn.sig[2] = REG_RD(bp,
MISC_REG_AEU_AFTER_INVERT_3_FUNC_0 +
port*4);
attn.sig[3] = REG_RD(bp,
MISC_REG_AEU_AFTER_INVERT_4_FUNC_0 +
port*4);
/* Since MCP attentions can't be disabled inside the block, we need to
* read AEU registers to see whether they're currently disabled
*/
attn.sig[3] &= ((REG_RD(bp,
!port ? MISC_REG_AEU_ENABLE4_FUNC_0_OUT_0
: MISC_REG_AEU_ENABLE4_FUNC_1_OUT_0) &
MISC_AEU_ENABLE_MCP_PRTY_BITS) |
~MISC_AEU_ENABLE_MCP_PRTY_BITS);
if (!CHIP_IS_E1x(bp))
attn.sig[4] = REG_RD(bp,
MISC_REG_AEU_AFTER_INVERT_5_FUNC_0 +
port*4);
return bnx2x_parity_attn(bp, global, print, attn.sig);
}
static void bnx2x_attn_int_deasserted4(struct bnx2x *bp, u32 attn)
{
u32 val;
if (attn & AEU_INPUTS_ATTN_BITS_PGLUE_HW_INTERRUPT) {
val = REG_RD(bp, PGLUE_B_REG_PGLUE_B_INT_STS_CLR);
BNX2X_ERR("PGLUE hw attention 0x%x\n", val);
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_ADDRESS_ERROR)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_ADDRESS_ERROR\n");
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_INCORRECT_RCV_BEHAVIOR)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_INCORRECT_RCV_BEHAVIOR\n");
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_WAS_ERROR_ATTN)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_WAS_ERROR_ATTN\n");
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_VF_LENGTH_VIOLATION_ATTN)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_VF_LENGTH_VIOLATION_ATTN\n");
if (val &
PGLUE_B_PGLUE_B_INT_STS_REG_VF_GRC_SPACE_VIOLATION_ATTN)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_VF_GRC_SPACE_VIOLATION_ATTN\n");
if (val &
PGLUE_B_PGLUE_B_INT_STS_REG_VF_MSIX_BAR_VIOLATION_ATTN)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_VF_MSIX_BAR_VIOLATION_ATTN\n");
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_TCPL_ERROR_ATTN)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_TCPL_ERROR_ATTN\n");
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_TCPL_IN_TWO_RCBS_ATTN)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_TCPL_IN_TWO_RCBS_ATTN\n");
if (val & PGLUE_B_PGLUE_B_INT_STS_REG_CSSNOOP_FIFO_OVERFLOW)
BNX2X_ERR("PGLUE_B_PGLUE_B_INT_STS_REG_CSSNOOP_FIFO_OVERFLOW\n");
}
if (attn & AEU_INPUTS_ATTN_BITS_ATC_HW_INTERRUPT) {
val = REG_RD(bp, ATC_REG_ATC_INT_STS_CLR);
BNX2X_ERR("ATC hw attention 0x%x\n", val);
if (val & ATC_ATC_INT_STS_REG_ADDRESS_ERROR)
BNX2X_ERR("ATC_ATC_INT_STS_REG_ADDRESS_ERROR\n");
if (val & ATC_ATC_INT_STS_REG_ATC_TCPL_TO_NOT_PEND)
BNX2X_ERR("ATC_ATC_INT_STS_REG_ATC_TCPL_TO_NOT_PEND\n");
if (val & ATC_ATC_INT_STS_REG_ATC_GPA_MULTIPLE_HITS)
BNX2X_ERR("ATC_ATC_INT_STS_REG_ATC_GPA_MULTIPLE_HITS\n");
if (val & ATC_ATC_INT_STS_REG_ATC_RCPL_TO_EMPTY_CNT)
BNX2X_ERR("ATC_ATC_INT_STS_REG_ATC_RCPL_TO_EMPTY_CNT\n");
if (val & ATC_ATC_INT_STS_REG_ATC_TCPL_ERROR)
BNX2X_ERR("ATC_ATC_INT_STS_REG_ATC_TCPL_ERROR\n");
if (val & ATC_ATC_INT_STS_REG_ATC_IREQ_LESS_THAN_STU)
BNX2X_ERR("ATC_ATC_INT_STS_REG_ATC_IREQ_LESS_THAN_STU\n");
}
if (attn & (AEU_INPUTS_ATTN_BITS_PGLUE_PARITY_ERROR |
AEU_INPUTS_ATTN_BITS_ATC_PARITY_ERROR)) {
BNX2X_ERR("FATAL parity attention set4 0x%x\n",
(u32)(attn & (AEU_INPUTS_ATTN_BITS_PGLUE_PARITY_ERROR |
AEU_INPUTS_ATTN_BITS_ATC_PARITY_ERROR)));
}
}
static void bnx2x_attn_int_deasserted(struct bnx2x *bp, u32 deasserted)
{
struct attn_route attn, *group_mask;
int port = BP_PORT(bp);
int index;
u32 reg_addr;
u32 val;
u32 aeu_mask;
bool global = false;
/* need to take HW lock because MCP or other port might also
try to handle this event */
bnx2x_acquire_alr(bp);
if (bnx2x_chk_parity_attn(bp, &global, true)) {
#ifndef BNX2X_STOP_ON_ERROR
bp->recovery_state = BNX2X_RECOVERY_INIT;
schedule_delayed_work(&bp->sp_rtnl_task, 0);
/* Disable HW interrupts */
bnx2x_int_disable(bp);
/* In case of parity errors don't handle attentions so that
* other function would "see" parity errors.
*/
#else
bnx2x_panic();
#endif
bnx2x_release_alr(bp);
return;
}
attn.sig[0] = REG_RD(bp, MISC_REG_AEU_AFTER_INVERT_1_FUNC_0 + port*4);
attn.sig[1] = REG_RD(bp, MISC_REG_AEU_AFTER_INVERT_2_FUNC_0 + port*4);
attn.sig[2] = REG_RD(bp, MISC_REG_AEU_AFTER_INVERT_3_FUNC_0 + port*4);
attn.sig[3] = REG_RD(bp, MISC_REG_AEU_AFTER_INVERT_4_FUNC_0 + port*4);
if (!CHIP_IS_E1x(bp))
attn.sig[4] =
REG_RD(bp, MISC_REG_AEU_AFTER_INVERT_5_FUNC_0 + port*4);
else
attn.sig[4] = 0;
DP(NETIF_MSG_HW, "attn: %08x %08x %08x %08x %08x\n",
attn.sig[0], attn.sig[1], attn.sig[2], attn.sig[3], attn.sig[4]);
for (index = 0; index < MAX_DYNAMIC_ATTN_GRPS; index++) {
if (deasserted & (1 << index)) {
group_mask = &bp->attn_group[index];
DP(NETIF_MSG_HW, "group[%d]: %08x %08x %08x %08x %08x\n",
index,
group_mask->sig[0], group_mask->sig[1],
group_mask->sig[2], group_mask->sig[3],
group_mask->sig[4]);
bnx2x_attn_int_deasserted4(bp,
attn.sig[4] & group_mask->sig[4]);
bnx2x_attn_int_deasserted3(bp,
attn.sig[3] & group_mask->sig[3]);
bnx2x_attn_int_deasserted1(bp,
attn.sig[1] & group_mask->sig[1]);
bnx2x_attn_int_deasserted2(bp,
attn.sig[2] & group_mask->sig[2]);
bnx2x_attn_int_deasserted0(bp,
attn.sig[0] & group_mask->sig[0]);
}
}
bnx2x_release_alr(bp);
if (bp->common.int_block == INT_BLOCK_HC)
reg_addr = (HC_REG_COMMAND_REG + port*32 +
COMMAND_REG_ATTN_BITS_CLR);
else
reg_addr = (BAR_IGU_INTMEM + IGU_CMD_ATTN_BIT_CLR_UPPER*8);
val = ~deasserted;
DP(NETIF_MSG_HW, "about to mask 0x%08x at %s addr 0x%x\n", val,
(bp->common.int_block == INT_BLOCK_HC) ? "HC" : "IGU", reg_addr);
REG_WR(bp, reg_addr, val);
if (~bp->attn_state & deasserted)
BNX2X_ERR("IGU ERROR\n");
reg_addr = port ? MISC_REG_AEU_MASK_ATTN_FUNC_1 :
MISC_REG_AEU_MASK_ATTN_FUNC_0;
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_PORT0_ATT_MASK + port);
aeu_mask = REG_RD(bp, reg_addr);
DP(NETIF_MSG_HW, "aeu_mask %x newly deasserted %x\n",
aeu_mask, deasserted);
aeu_mask |= (deasserted & 0x3ff);
DP(NETIF_MSG_HW, "new mask %x\n", aeu_mask);
REG_WR(bp, reg_addr, aeu_mask);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_PORT0_ATT_MASK + port);
DP(NETIF_MSG_HW, "attn_state %x\n", bp->attn_state);
bp->attn_state &= ~deasserted;
DP(NETIF_MSG_HW, "new state %x\n", bp->attn_state);
}
static void bnx2x_attn_int(struct bnx2x *bp)
{
/* read local copy of bits */
u32 attn_bits = le32_to_cpu(bp->def_status_blk->atten_status_block.
attn_bits);
u32 attn_ack = le32_to_cpu(bp->def_status_blk->atten_status_block.
attn_bits_ack);
u32 attn_state = bp->attn_state;
/* look for changed bits */
u32 asserted = attn_bits & ~attn_ack & ~attn_state;
u32 deasserted = ~attn_bits & attn_ack & attn_state;
DP(NETIF_MSG_HW,
"attn_bits %x attn_ack %x asserted %x deasserted %x\n",
attn_bits, attn_ack, asserted, deasserted);
if (~(attn_bits ^ attn_ack) & (attn_bits ^ attn_state))
BNX2X_ERR("BAD attention state\n");
/* handle bits that were raised */
if (asserted)
bnx2x_attn_int_asserted(bp, asserted);
if (deasserted)
bnx2x_attn_int_deasserted(bp, deasserted);
}
void bnx2x_igu_ack_sb(struct bnx2x *bp, u8 igu_sb_id, u8 segment,
u16 index, u8 op, u8 update)
{
u32 igu_addr = bp->igu_base_addr;
igu_addr += (IGU_CMD_INT_ACK_BASE + igu_sb_id)*8;
bnx2x_igu_ack_sb_gen(bp, igu_sb_id, segment, index, op, update,
igu_addr);
}
static void bnx2x_update_eq_prod(struct bnx2x *bp, u16 prod)
{
/* No memory barriers */
storm_memset_eq_prod(bp, prod, BP_FUNC(bp));
mmiowb(); /* keep prod updates ordered */
}
static int bnx2x_cnic_handle_cfc_del(struct bnx2x *bp, u32 cid,
union event_ring_elem *elem)
{
u8 err = elem->message.error;
if (!bp->cnic_eth_dev.starting_cid ||
(cid < bp->cnic_eth_dev.starting_cid &&
cid != bp->cnic_eth_dev.iscsi_l2_cid))
return 1;
DP(BNX2X_MSG_SP, "got delete ramrod for CNIC CID %d\n", cid);
if (unlikely(err)) {
BNX2X_ERR("got delete ramrod for CNIC CID %d with error!\n",
cid);
bnx2x_panic_dump(bp, false);
}
bnx2x_cnic_cfc_comp(bp, cid, err);
return 0;
}
static void bnx2x_handle_mcast_eqe(struct bnx2x *bp)
{
struct bnx2x_mcast_ramrod_params rparam;
int rc;
memset(&rparam, 0, sizeof(rparam));
rparam.mcast_obj = &bp->mcast_obj;
netif_addr_lock_bh(bp->dev);
/* Clear pending state for the last command */
bp->mcast_obj.raw.clear_pending(&bp->mcast_obj.raw);
/* If there are pending mcast commands - send them */
if (bp->mcast_obj.check_pending(&bp->mcast_obj)) {
rc = bnx2x_config_mcast(bp, &rparam, BNX2X_MCAST_CMD_CONT);
if (rc < 0)
BNX2X_ERR("Failed to send pending mcast commands: %d\n",
rc);
}
netif_addr_unlock_bh(bp->dev);
}
static void bnx2x_handle_classification_eqe(struct bnx2x *bp,
union event_ring_elem *elem)
{
unsigned long ramrod_flags = 0;
int rc = 0;
u32 echo = le32_to_cpu(elem->message.data.eth_event.echo);
u32 cid = echo & BNX2X_SWCID_MASK;
struct bnx2x_vlan_mac_obj *vlan_mac_obj;
/* Always push next commands out, don't wait here */
__set_bit(RAMROD_CONT, &ramrod_flags);
switch (echo >> BNX2X_SWCID_SHIFT) {
case BNX2X_FILTER_MAC_PENDING:
DP(BNX2X_MSG_SP, "Got SETUP_MAC completions\n");
if (CNIC_LOADED(bp) && (cid == BNX2X_ISCSI_ETH_CID(bp)))
vlan_mac_obj = &bp->iscsi_l2_mac_obj;
else
vlan_mac_obj = &bp->sp_objs[cid].mac_obj;
break;
case BNX2X_FILTER_VLAN_PENDING:
DP(BNX2X_MSG_SP, "Got SETUP_VLAN completions\n");
vlan_mac_obj = &bp->sp_objs[cid].vlan_obj;
break;
case BNX2X_FILTER_MCAST_PENDING:
DP(BNX2X_MSG_SP, "Got SETUP_MCAST completions\n");
/* This is only relevant for 57710 where multicast MACs are
* configured as unicast MACs using the same ramrod.
*/
bnx2x_handle_mcast_eqe(bp);
return;
default:
BNX2X_ERR("Unsupported classification command: 0x%x\n", echo);
return;
}
rc = vlan_mac_obj->complete(bp, vlan_mac_obj, elem, &ramrod_flags);
if (rc < 0)
BNX2X_ERR("Failed to schedule new commands: %d\n", rc);
else if (rc > 0)
DP(BNX2X_MSG_SP, "Scheduled next pending commands...\n");
}
static void bnx2x_set_iscsi_eth_rx_mode(struct bnx2x *bp, bool start);
static void bnx2x_handle_rx_mode_eqe(struct bnx2x *bp)
{
netif_addr_lock_bh(bp->dev);
clear_bit(BNX2X_FILTER_RX_MODE_PENDING, &bp->sp_state);
/* Send rx_mode command again if was requested */
if (test_and_clear_bit(BNX2X_FILTER_RX_MODE_SCHED, &bp->sp_state))
bnx2x_set_storm_rx_mode(bp);
else if (test_and_clear_bit(BNX2X_FILTER_ISCSI_ETH_START_SCHED,
&bp->sp_state))
bnx2x_set_iscsi_eth_rx_mode(bp, true);
else if (test_and_clear_bit(BNX2X_FILTER_ISCSI_ETH_STOP_SCHED,
&bp->sp_state))
bnx2x_set_iscsi_eth_rx_mode(bp, false);
netif_addr_unlock_bh(bp->dev);
}
static void bnx2x_after_afex_vif_lists(struct bnx2x *bp,
union event_ring_elem *elem)
{
if (elem->message.data.vif_list_event.echo == VIF_LIST_RULE_GET) {
DP(BNX2X_MSG_SP,
"afex: ramrod completed VIF LIST_GET, addrs 0x%x\n",
elem->message.data.vif_list_event.func_bit_map);
bnx2x_fw_command(bp, DRV_MSG_CODE_AFEX_LISTGET_ACK,
elem->message.data.vif_list_event.func_bit_map);
} else if (elem->message.data.vif_list_event.echo ==
VIF_LIST_RULE_SET) {
DP(BNX2X_MSG_SP, "afex: ramrod completed VIF LIST_SET\n");
bnx2x_fw_command(bp, DRV_MSG_CODE_AFEX_LISTSET_ACK, 0);
}
}
/* called with rtnl_lock */
static void bnx2x_after_function_update(struct bnx2x *bp)
{
int q, rc;
struct bnx2x_fastpath *fp;
struct bnx2x_queue_state_params queue_params = {NULL};
struct bnx2x_queue_update_params *q_update_params =
&queue_params.params.update;
/* Send Q update command with afex vlan removal values for all Qs */
queue_params.cmd = BNX2X_Q_CMD_UPDATE;
/* set silent vlan removal values according to vlan mode */
__set_bit(BNX2X_Q_UPDATE_SILENT_VLAN_REM_CHNG,
&q_update_params->update_flags);
__set_bit(BNX2X_Q_UPDATE_SILENT_VLAN_REM,
&q_update_params->update_flags);
__set_bit(RAMROD_COMP_WAIT, &queue_params.ramrod_flags);
/* in access mode mark mask and value are 0 to strip all vlans */
if (bp->afex_vlan_mode == FUNC_MF_CFG_AFEX_VLAN_ACCESS_MODE) {
q_update_params->silent_removal_value = 0;
q_update_params->silent_removal_mask = 0;
} else {
q_update_params->silent_removal_value =
(bp->afex_def_vlan_tag & VLAN_VID_MASK);
q_update_params->silent_removal_mask = VLAN_VID_MASK;
}
for_each_eth_queue(bp, q) {
/* Set the appropriate Queue object */
fp = &bp->fp[q];
queue_params.q_obj = &bnx2x_sp_obj(bp, fp).q_obj;
/* send the ramrod */
rc = bnx2x_queue_state_change(bp, &queue_params);
if (rc < 0)
BNX2X_ERR("Failed to config silent vlan rem for Q %d\n",
q);
}
if (!NO_FCOE(bp) && CNIC_ENABLED(bp)) {
fp = &bp->fp[FCOE_IDX(bp)];
queue_params.q_obj = &bnx2x_sp_obj(bp, fp).q_obj;
/* clear pending completion bit */
__clear_bit(RAMROD_COMP_WAIT, &queue_params.ramrod_flags);
/* mark latest Q bit */
smp_mb__before_atomic();
set_bit(BNX2X_AFEX_FCOE_Q_UPDATE_PENDING, &bp->sp_state);
smp_mb__after_atomic();
/* send Q update ramrod for FCoE Q */
rc = bnx2x_queue_state_change(bp, &queue_params);
if (rc < 0)
BNX2X_ERR("Failed to config silent vlan rem for Q %d\n",
q);
} else {
/* If no FCoE ring - ACK MCP now */
bnx2x_link_report(bp);
bnx2x_fw_command(bp, DRV_MSG_CODE_AFEX_VIFSET_ACK, 0);
}
}
static struct bnx2x_queue_sp_obj *bnx2x_cid_to_q_obj(
struct bnx2x *bp, u32 cid)
{
DP(BNX2X_MSG_SP, "retrieving fp from cid %d\n", cid);
if (CNIC_LOADED(bp) && (cid == BNX2X_FCOE_ETH_CID(bp)))
return &bnx2x_fcoe_sp_obj(bp, q_obj);
else
return &bp->sp_objs[CID_TO_FP(cid, bp)].q_obj;
}
static void bnx2x_eq_int(struct bnx2x *bp)
{
u16 hw_cons, sw_cons, sw_prod;
union event_ring_elem *elem;
u8 echo;
u32 cid;
u8 opcode;
int rc, spqe_cnt = 0;
struct bnx2x_queue_sp_obj *q_obj;
struct bnx2x_func_sp_obj *f_obj = &bp->func_obj;
struct bnx2x_raw_obj *rss_raw = &bp->rss_conf_obj.raw;
hw_cons = le16_to_cpu(*bp->eq_cons_sb);
/* The hw_cos range is 1-255, 257 - the sw_cons range is 0-254, 256.
* when we get the next-page we need to adjust so the loop
* condition below will be met. The next element is the size of a
* regular element and hence incrementing by 1
*/
if ((hw_cons & EQ_DESC_MAX_PAGE) == EQ_DESC_MAX_PAGE)
hw_cons++;
/* This function may never run in parallel with itself for a
* specific bp, thus there is no need in "paired" read memory
* barrier here.
*/
sw_cons = bp->eq_cons;
sw_prod = bp->eq_prod;
DP(BNX2X_MSG_SP, "EQ: hw_cons %u sw_cons %u bp->eq_spq_left %x\n",
hw_cons, sw_cons, atomic_read(&bp->eq_spq_left));
for (; sw_cons != hw_cons;
sw_prod = NEXT_EQ_IDX(sw_prod), sw_cons = NEXT_EQ_IDX(sw_cons)) {
elem = &bp->eq_ring[EQ_DESC(sw_cons)];
rc = bnx2x_iov_eq_sp_event(bp, elem);
if (!rc) {
DP(BNX2X_MSG_IOV, "bnx2x_iov_eq_sp_event returned %d\n",
rc);
goto next_spqe;
}
opcode = elem->message.opcode;
/* handle eq element */
switch (opcode) {
case EVENT_RING_OPCODE_VF_PF_CHANNEL:
bnx2x_vf_mbx_schedule(bp,
&elem->message.data.vf_pf_event);
continue;
case EVENT_RING_OPCODE_STAT_QUERY:
DP_AND((BNX2X_MSG_SP | BNX2X_MSG_STATS),
"got statistics comp event %d\n",
bp->stats_comp++);
/* nothing to do with stats comp */
goto next_spqe;
case EVENT_RING_OPCODE_CFC_DEL:
/* handle according to cid range */
/*
* we may want to verify here that the bp state is
* HALTING
*/
/* elem CID originates from FW; actually LE */
cid = SW_CID(elem->message.data.cfc_del_event.cid);
DP(BNX2X_MSG_SP,
"got delete ramrod for MULTI[%d]\n", cid);
if (CNIC_LOADED(bp) &&
!bnx2x_cnic_handle_cfc_del(bp, cid, elem))
goto next_spqe;
q_obj = bnx2x_cid_to_q_obj(bp, cid);
if (q_obj->complete_cmd(bp, q_obj, BNX2X_Q_CMD_CFC_DEL))
break;
goto next_spqe;
case EVENT_RING_OPCODE_STOP_TRAFFIC:
DP(BNX2X_MSG_SP | BNX2X_MSG_DCB, "got STOP TRAFFIC\n");
bnx2x_dcbx_set_params(bp, BNX2X_DCBX_STATE_TX_PAUSED);
if (f_obj->complete_cmd(bp, f_obj,
BNX2X_F_CMD_TX_STOP))
break;
goto next_spqe;
case EVENT_RING_OPCODE_START_TRAFFIC:
DP(BNX2X_MSG_SP | BNX2X_MSG_DCB, "got START TRAFFIC\n");
bnx2x_dcbx_set_params(bp, BNX2X_DCBX_STATE_TX_RELEASED);
if (f_obj->complete_cmd(bp, f_obj,
BNX2X_F_CMD_TX_START))
break;
goto next_spqe;
case EVENT_RING_OPCODE_FUNCTION_UPDATE:
echo = elem->message.data.function_update_event.echo;
if (echo == SWITCH_UPDATE) {
DP(BNX2X_MSG_SP | NETIF_MSG_IFUP,
"got FUNC_SWITCH_UPDATE ramrod\n");
if (f_obj->complete_cmd(
bp, f_obj, BNX2X_F_CMD_SWITCH_UPDATE))
break;
} else {
int cmd = BNX2X_SP_RTNL_AFEX_F_UPDATE;
DP(BNX2X_MSG_SP | BNX2X_MSG_MCP,
"AFEX: ramrod completed FUNCTION_UPDATE\n");
f_obj->complete_cmd(bp, f_obj,
BNX2X_F_CMD_AFEX_UPDATE);
/* We will perform the Queues update from
* sp_rtnl task as all Queue SP operations
* should run under rtnl_lock.
*/
bnx2x_schedule_sp_rtnl(bp, cmd, 0);
}
goto next_spqe;
case EVENT_RING_OPCODE_AFEX_VIF_LISTS:
f_obj->complete_cmd(bp, f_obj,
BNX2X_F_CMD_AFEX_VIFLISTS);
bnx2x_after_afex_vif_lists(bp, elem);
goto next_spqe;
case EVENT_RING_OPCODE_FUNCTION_START:
DP(BNX2X_MSG_SP | NETIF_MSG_IFUP,
"got FUNC_START ramrod\n");
if (f_obj->complete_cmd(bp, f_obj, BNX2X_F_CMD_START))
break;
goto next_spqe;
case EVENT_RING_OPCODE_FUNCTION_STOP:
DP(BNX2X_MSG_SP | NETIF_MSG_IFUP,
"got FUNC_STOP ramrod\n");
if (f_obj->complete_cmd(bp, f_obj, BNX2X_F_CMD_STOP))
break;
goto next_spqe;
case EVENT_RING_OPCODE_SET_TIMESYNC:
DP(BNX2X_MSG_SP | BNX2X_MSG_PTP,
"got set_timesync ramrod completion\n");
if (f_obj->complete_cmd(bp, f_obj,
BNX2X_F_CMD_SET_TIMESYNC))
break;
goto next_spqe;
}
switch (opcode | bp->state) {
case (EVENT_RING_OPCODE_RSS_UPDATE_RULES |
BNX2X_STATE_OPEN):
case (EVENT_RING_OPCODE_RSS_UPDATE_RULES |
BNX2X_STATE_OPENING_WAIT4_PORT):
case (EVENT_RING_OPCODE_RSS_UPDATE_RULES |
BNX2X_STATE_CLOSING_WAIT4_HALT):
DP(BNX2X_MSG_SP, "got RSS_UPDATE ramrod. CID %d\n",
SW_CID(elem->message.data.eth_event.echo));
rss_raw->clear_pending(rss_raw);
break;
case (EVENT_RING_OPCODE_SET_MAC | BNX2X_STATE_OPEN):
case (EVENT_RING_OPCODE_SET_MAC | BNX2X_STATE_DIAG):
case (EVENT_RING_OPCODE_SET_MAC |
BNX2X_STATE_CLOSING_WAIT4_HALT):
case (EVENT_RING_OPCODE_CLASSIFICATION_RULES |
BNX2X_STATE_OPEN):
case (EVENT_RING_OPCODE_CLASSIFICATION_RULES |
BNX2X_STATE_DIAG):
case (EVENT_RING_OPCODE_CLASSIFICATION_RULES |
BNX2X_STATE_CLOSING_WAIT4_HALT):
DP(BNX2X_MSG_SP, "got (un)set vlan/mac ramrod\n");
bnx2x_handle_classification_eqe(bp, elem);
break;
case (EVENT_RING_OPCODE_MULTICAST_RULES |
BNX2X_STATE_OPEN):
case (EVENT_RING_OPCODE_MULTICAST_RULES |
BNX2X_STATE_DIAG):
case (EVENT_RING_OPCODE_MULTICAST_RULES |
BNX2X_STATE_CLOSING_WAIT4_HALT):
DP(BNX2X_MSG_SP, "got mcast ramrod\n");
bnx2x_handle_mcast_eqe(bp);
break;
case (EVENT_RING_OPCODE_FILTERS_RULES |
BNX2X_STATE_OPEN):
case (EVENT_RING_OPCODE_FILTERS_RULES |
BNX2X_STATE_DIAG):
case (EVENT_RING_OPCODE_FILTERS_RULES |
BNX2X_STATE_CLOSING_WAIT4_HALT):
DP(BNX2X_MSG_SP, "got rx_mode ramrod\n");
bnx2x_handle_rx_mode_eqe(bp);
break;
default:
/* unknown event log error and continue */
BNX2X_ERR("Unknown EQ event %d, bp->state 0x%x\n",
elem->message.opcode, bp->state);
}
next_spqe:
spqe_cnt++;
} /* for */
smp_mb__before_atomic();
atomic_add(spqe_cnt, &bp->eq_spq_left);
bp->eq_cons = sw_cons;
bp->eq_prod = sw_prod;
/* Make sure that above mem writes were issued towards the memory */
smp_wmb();
/* update producer */
bnx2x_update_eq_prod(bp, bp->eq_prod);
}
static void bnx2x_sp_task(struct work_struct *work)
{
struct bnx2x *bp = container_of(work, struct bnx2x, sp_task.work);
DP(BNX2X_MSG_SP, "sp task invoked\n");
/* make sure the atomic interrupt_occurred has been written */
smp_rmb();
if (atomic_read(&bp->interrupt_occurred)) {
/* what work needs to be performed? */
u16 status = bnx2x_update_dsb_idx(bp);
DP(BNX2X_MSG_SP, "status %x\n", status);
DP(BNX2X_MSG_SP, "setting interrupt_occurred to 0\n");
atomic_set(&bp->interrupt_occurred, 0);
/* HW attentions */
if (status & BNX2X_DEF_SB_ATT_IDX) {
bnx2x_attn_int(bp);
status &= ~BNX2X_DEF_SB_ATT_IDX;
}
/* SP events: STAT_QUERY and others */
if (status & BNX2X_DEF_SB_IDX) {
struct bnx2x_fastpath *fp = bnx2x_fcoe_fp(bp);
if (FCOE_INIT(bp) &&
(bnx2x_has_rx_work(fp) || bnx2x_has_tx_work(fp))) {
/* Prevent local bottom-halves from running as
* we are going to change the local NAPI list.
*/
local_bh_disable();
napi_schedule(&bnx2x_fcoe(bp, napi));
local_bh_enable();
}
/* Handle EQ completions */
bnx2x_eq_int(bp);
bnx2x_ack_sb(bp, bp->igu_dsb_id, USTORM_ID,
le16_to_cpu(bp->def_idx), IGU_INT_NOP, 1);
status &= ~BNX2X_DEF_SB_IDX;
}
/* if status is non zero then perhaps something went wrong */
if (unlikely(status))
DP(BNX2X_MSG_SP,
"got an unknown interrupt! (status 0x%x)\n", status);
/* ack status block only if something was actually handled */
bnx2x_ack_sb(bp, bp->igu_dsb_id, ATTENTION_ID,
le16_to_cpu(bp->def_att_idx), IGU_INT_ENABLE, 1);
}
/* afex - poll to check if VIFSET_ACK should be sent to MFW */
if (test_and_clear_bit(BNX2X_AFEX_PENDING_VIFSET_MCP_ACK,
&bp->sp_state)) {
bnx2x_link_report(bp);
bnx2x_fw_command(bp, DRV_MSG_CODE_AFEX_VIFSET_ACK, 0);
}
}
irqreturn_t bnx2x_msix_sp_int(int irq, void *dev_instance)
{
struct net_device *dev = dev_instance;
struct bnx2x *bp = netdev_priv(dev);
bnx2x_ack_sb(bp, bp->igu_dsb_id, USTORM_ID, 0,
IGU_INT_DISABLE, 0);
#ifdef BNX2X_STOP_ON_ERROR
if (unlikely(bp->panic))
return IRQ_HANDLED;
#endif
if (CNIC_LOADED(bp)) {
struct cnic_ops *c_ops;
rcu_read_lock();
c_ops = rcu_dereference(bp->cnic_ops);
if (c_ops)
c_ops->cnic_handler(bp->cnic_data, NULL);
rcu_read_unlock();
}
/* schedule sp task to perform default status block work, ack
* attentions and enable interrupts.
*/
bnx2x_schedule_sp_task(bp);
return IRQ_HANDLED;
}
/* end of slow path */
void bnx2x_drv_pulse(struct bnx2x *bp)
{
SHMEM_WR(bp, func_mb[BP_FW_MB_IDX(bp)].drv_pulse_mb,
bp->fw_drv_pulse_wr_seq);
}
static void bnx2x_timer(unsigned long data)
{
struct bnx2x *bp = (struct bnx2x *) data;
if (!netif_running(bp->dev))
return;
if (IS_PF(bp) &&
!BP_NOMCP(bp)) {
int mb_idx = BP_FW_MB_IDX(bp);
u16 drv_pulse;
u16 mcp_pulse;
++bp->fw_drv_pulse_wr_seq;
bp->fw_drv_pulse_wr_seq &= DRV_PULSE_SEQ_MASK;
drv_pulse = bp->fw_drv_pulse_wr_seq;
bnx2x_drv_pulse(bp);
mcp_pulse = (SHMEM_RD(bp, func_mb[mb_idx].mcp_pulse_mb) &
MCP_PULSE_SEQ_MASK);
/* The delta between driver pulse and mcp response
* should not get too big. If the MFW is more than 5 pulses
* behind, we should worry about it enough to generate an error
* log.
*/
if (((drv_pulse - mcp_pulse) & MCP_PULSE_SEQ_MASK) > 5)
BNX2X_ERR("MFW seems hanged: drv_pulse (0x%x) != mcp_pulse (0x%x)\n",
drv_pulse, mcp_pulse);
}
if (bp->state == BNX2X_STATE_OPEN)
bnx2x_stats_handle(bp, STATS_EVENT_UPDATE);
/* sample pf vf bulletin board for new posts from pf */
if (IS_VF(bp))
bnx2x_timer_sriov(bp);
mod_timer(&bp->timer, jiffies + bp->current_interval);
}
/* end of Statistics */
/* nic init */
/*
* nic init service functions
*/
static void bnx2x_fill(struct bnx2x *bp, u32 addr, int fill, u32 len)
{
u32 i;
if (!(len%4) && !(addr%4))
for (i = 0; i < len; i += 4)
REG_WR(bp, addr + i, fill);
else
for (i = 0; i < len; i++)
REG_WR8(bp, addr + i, fill);
}
/* helper: writes FP SP data to FW - data_size in dwords */
static void bnx2x_wr_fp_sb_data(struct bnx2x *bp,
int fw_sb_id,
u32 *sb_data_p,
u32 data_size)
{
int index;
for (index = 0; index < data_size; index++)
REG_WR(bp, BAR_CSTRORM_INTMEM +
CSTORM_STATUS_BLOCK_DATA_OFFSET(fw_sb_id) +
sizeof(u32)*index,
*(sb_data_p + index));
}
static void bnx2x_zero_fp_sb(struct bnx2x *bp, int fw_sb_id)
{
u32 *sb_data_p;
u32 data_size = 0;
struct hc_status_block_data_e2 sb_data_e2;
struct hc_status_block_data_e1x sb_data_e1x;
/* disable the function first */
if (!CHIP_IS_E1x(bp)) {
memset(&sb_data_e2, 0, sizeof(struct hc_status_block_data_e2));
sb_data_e2.common.state = SB_DISABLED;
sb_data_e2.common.p_func.vf_valid = false;
sb_data_p = (u32 *)&sb_data_e2;
data_size = sizeof(struct hc_status_block_data_e2)/sizeof(u32);
} else {
memset(&sb_data_e1x, 0,
sizeof(struct hc_status_block_data_e1x));
sb_data_e1x.common.state = SB_DISABLED;
sb_data_e1x.common.p_func.vf_valid = false;
sb_data_p = (u32 *)&sb_data_e1x;
data_size = sizeof(struct hc_status_block_data_e1x)/sizeof(u32);
}
bnx2x_wr_fp_sb_data(bp, fw_sb_id, sb_data_p, data_size);
bnx2x_fill(bp, BAR_CSTRORM_INTMEM +
CSTORM_STATUS_BLOCK_OFFSET(fw_sb_id), 0,
CSTORM_STATUS_BLOCK_SIZE);
bnx2x_fill(bp, BAR_CSTRORM_INTMEM +
CSTORM_SYNC_BLOCK_OFFSET(fw_sb_id), 0,
CSTORM_SYNC_BLOCK_SIZE);
}
/* helper: writes SP SB data to FW */
static void bnx2x_wr_sp_sb_data(struct bnx2x *bp,
struct hc_sp_status_block_data *sp_sb_data)
{
int func = BP_FUNC(bp);
int i;
for (i = 0; i < sizeof(struct hc_sp_status_block_data)/sizeof(u32); i++)
REG_WR(bp, BAR_CSTRORM_INTMEM +
CSTORM_SP_STATUS_BLOCK_DATA_OFFSET(func) +
i*sizeof(u32),
*((u32 *)sp_sb_data + i));
}
static void bnx2x_zero_sp_sb(struct bnx2x *bp)
{
int func = BP_FUNC(bp);
struct hc_sp_status_block_data sp_sb_data;
memset(&sp_sb_data, 0, sizeof(struct hc_sp_status_block_data));
sp_sb_data.state = SB_DISABLED;
sp_sb_data.p_func.vf_valid = false;
bnx2x_wr_sp_sb_data(bp, &sp_sb_data);
bnx2x_fill(bp, BAR_CSTRORM_INTMEM +
CSTORM_SP_STATUS_BLOCK_OFFSET(func), 0,
CSTORM_SP_STATUS_BLOCK_SIZE);
bnx2x_fill(bp, BAR_CSTRORM_INTMEM +
CSTORM_SP_SYNC_BLOCK_OFFSET(func), 0,
CSTORM_SP_SYNC_BLOCK_SIZE);
}
static void bnx2x_setup_ndsb_state_machine(struct hc_status_block_sm *hc_sm,
int igu_sb_id, int igu_seg_id)
{
hc_sm->igu_sb_id = igu_sb_id;
hc_sm->igu_seg_id = igu_seg_id;
hc_sm->timer_value = 0xFF;
hc_sm->time_to_expire = 0xFFFFFFFF;
}
/* allocates state machine ids. */
static void bnx2x_map_sb_state_machines(struct hc_index_data *index_data)
{
/* zero out state machine indices */
/* rx indices */
index_data[HC_INDEX_ETH_RX_CQ_CONS].flags &= ~HC_INDEX_DATA_SM_ID;
/* tx indices */
index_data[HC_INDEX_OOO_TX_CQ_CONS].flags &= ~HC_INDEX_DATA_SM_ID;
index_data[HC_INDEX_ETH_TX_CQ_CONS_COS0].flags &= ~HC_INDEX_DATA_SM_ID;
index_data[HC_INDEX_ETH_TX_CQ_CONS_COS1].flags &= ~HC_INDEX_DATA_SM_ID;
index_data[HC_INDEX_ETH_TX_CQ_CONS_COS2].flags &= ~HC_INDEX_DATA_SM_ID;
/* map indices */
/* rx indices */
index_data[HC_INDEX_ETH_RX_CQ_CONS].flags |=
SM_RX_ID << HC_INDEX_DATA_SM_ID_SHIFT;
/* tx indices */
index_data[HC_INDEX_OOO_TX_CQ_CONS].flags |=
SM_TX_ID << HC_INDEX_DATA_SM_ID_SHIFT;
index_data[HC_INDEX_ETH_TX_CQ_CONS_COS0].flags |=
SM_TX_ID << HC_INDEX_DATA_SM_ID_SHIFT;
index_data[HC_INDEX_ETH_TX_CQ_CONS_COS1].flags |=
SM_TX_ID << HC_INDEX_DATA_SM_ID_SHIFT;
index_data[HC_INDEX_ETH_TX_CQ_CONS_COS2].flags |=
SM_TX_ID << HC_INDEX_DATA_SM_ID_SHIFT;
}
void bnx2x_init_sb(struct bnx2x *bp, dma_addr_t mapping, int vfid,
u8 vf_valid, int fw_sb_id, int igu_sb_id)
{
int igu_seg_id;
struct hc_status_block_data_e2 sb_data_e2;
struct hc_status_block_data_e1x sb_data_e1x;
struct hc_status_block_sm *hc_sm_p;
int data_size;
u32 *sb_data_p;
if (CHIP_INT_MODE_IS_BC(bp))
igu_seg_id = HC_SEG_ACCESS_NORM;
else
igu_seg_id = IGU_SEG_ACCESS_NORM;
bnx2x_zero_fp_sb(bp, fw_sb_id);
if (!CHIP_IS_E1x(bp)) {
memset(&sb_data_e2, 0, sizeof(struct hc_status_block_data_e2));
sb_data_e2.common.state = SB_ENABLED;
sb_data_e2.common.p_func.pf_id = BP_FUNC(bp);
sb_data_e2.common.p_func.vf_id = vfid;
sb_data_e2.common.p_func.vf_valid = vf_valid;
sb_data_e2.common.p_func.vnic_id = BP_VN(bp);
sb_data_e2.common.same_igu_sb_1b = true;
sb_data_e2.common.host_sb_addr.hi = U64_HI(mapping);
sb_data_e2.common.host_sb_addr.lo = U64_LO(mapping);
hc_sm_p = sb_data_e2.common.state_machine;
sb_data_p = (u32 *)&sb_data_e2;
data_size = sizeof(struct hc_status_block_data_e2)/sizeof(u32);
bnx2x_map_sb_state_machines(sb_data_e2.index_data);
} else {
memset(&sb_data_e1x, 0,
sizeof(struct hc_status_block_data_e1x));
sb_data_e1x.common.state = SB_ENABLED;
sb_data_e1x.common.p_func.pf_id = BP_FUNC(bp);
sb_data_e1x.common.p_func.vf_id = 0xff;
sb_data_e1x.common.p_func.vf_valid = false;
sb_data_e1x.common.p_func.vnic_id = BP_VN(bp);
sb_data_e1x.common.same_igu_sb_1b = true;
sb_data_e1x.common.host_sb_addr.hi = U64_HI(mapping);
sb_data_e1x.common.host_sb_addr.lo = U64_LO(mapping);
hc_sm_p = sb_data_e1x.common.state_machine;
sb_data_p = (u32 *)&sb_data_e1x;
data_size = sizeof(struct hc_status_block_data_e1x)/sizeof(u32);
bnx2x_map_sb_state_machines(sb_data_e1x.index_data);
}
bnx2x_setup_ndsb_state_machine(&hc_sm_p[SM_RX_ID],
igu_sb_id, igu_seg_id);
bnx2x_setup_ndsb_state_machine(&hc_sm_p[SM_TX_ID],
igu_sb_id, igu_seg_id);
DP(NETIF_MSG_IFUP, "Init FW SB %d\n", fw_sb_id);
/* write indices to HW - PCI guarantees endianity of regpairs */
bnx2x_wr_fp_sb_data(bp, fw_sb_id, sb_data_p, data_size);
}
static void bnx2x_update_coalesce_sb(struct bnx2x *bp, u8 fw_sb_id,
u16 tx_usec, u16 rx_usec)
{
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bnx2x_update_coalesce_sb_index(bp, fw_sb_id, HC_INDEX_ETH_RX_CQ_CONS,
false, rx_usec);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bnx2x_update_coalesce_sb_index(bp, fw_sb_id,
HC_INDEX_ETH_TX_CQ_CONS_COS0, false,
tx_usec);
bnx2x_update_coalesce_sb_index(bp, fw_sb_id,
HC_INDEX_ETH_TX_CQ_CONS_COS1, false,
tx_usec);
bnx2x_update_coalesce_sb_index(bp, fw_sb_id,
HC_INDEX_ETH_TX_CQ_CONS_COS2, false,
tx_usec);
}
static void bnx2x_init_def_sb(struct bnx2x *bp)
{
struct host_sp_status_block *def_sb = bp->def_status_blk;
dma_addr_t mapping = bp->def_status_blk_mapping;
int igu_sp_sb_index;
int igu_seg_id;
int port = BP_PORT(bp);
int func = BP_FUNC(bp);
int reg_offset, reg_offset_en5;
u64 section;
int index;
struct hc_sp_status_block_data sp_sb_data;
memset(&sp_sb_data, 0, sizeof(struct hc_sp_status_block_data));
if (CHIP_INT_MODE_IS_BC(bp)) {
igu_sp_sb_index = DEF_SB_IGU_ID;
igu_seg_id = HC_SEG_ACCESS_DEF;
} else {
igu_sp_sb_index = bp->igu_dsb_id;
igu_seg_id = IGU_SEG_ACCESS_DEF;
}
/* ATTN */
section = ((u64)mapping) + offsetof(struct host_sp_status_block,
atten_status_block);
def_sb->atten_status_block.status_block_id = igu_sp_sb_index;
bp->attn_state = 0;
reg_offset = (port ? MISC_REG_AEU_ENABLE1_FUNC_1_OUT_0 :
MISC_REG_AEU_ENABLE1_FUNC_0_OUT_0);
reg_offset_en5 = (port ? MISC_REG_AEU_ENABLE5_FUNC_1_OUT_0 :
MISC_REG_AEU_ENABLE5_FUNC_0_OUT_0);
for (index = 0; index < MAX_DYNAMIC_ATTN_GRPS; index++) {
int sindex;
/* take care of sig[0]..sig[4] */
for (sindex = 0; sindex < 4; sindex++)
bp->attn_group[index].sig[sindex] =
REG_RD(bp, reg_offset + sindex*0x4 + 0x10*index);
if (!CHIP_IS_E1x(bp))
/*
* enable5 is separate from the rest of the registers,
* and therefore the address skip is 4
* and not 16 between the different groups
*/
bp->attn_group[index].sig[4] = REG_RD(bp,
reg_offset_en5 + 0x4*index);
else
bp->attn_group[index].sig[4] = 0;
}
if (bp->common.int_block == INT_BLOCK_HC) {
reg_offset = (port ? HC_REG_ATTN_MSG1_ADDR_L :
HC_REG_ATTN_MSG0_ADDR_L);
REG_WR(bp, reg_offset, U64_LO(section));
REG_WR(bp, reg_offset + 4, U64_HI(section));
} else if (!CHIP_IS_E1x(bp)) {
REG_WR(bp, IGU_REG_ATTN_MSG_ADDR_L, U64_LO(section));
REG_WR(bp, IGU_REG_ATTN_MSG_ADDR_H, U64_HI(section));
}
section = ((u64)mapping) + offsetof(struct host_sp_status_block,
sp_sb);
bnx2x_zero_sp_sb(bp);
/* PCI guarantees endianity of regpairs */
sp_sb_data.state = SB_ENABLED;
sp_sb_data.host_sb_addr.lo = U64_LO(section);
sp_sb_data.host_sb_addr.hi = U64_HI(section);
sp_sb_data.igu_sb_id = igu_sp_sb_index;
sp_sb_data.igu_seg_id = igu_seg_id;
sp_sb_data.p_func.pf_id = func;
sp_sb_data.p_func.vnic_id = BP_VN(bp);
sp_sb_data.p_func.vf_id = 0xff;
bnx2x_wr_sp_sb_data(bp, &sp_sb_data);
bnx2x_ack_sb(bp, bp->igu_dsb_id, USTORM_ID, 0, IGU_INT_ENABLE, 0);
}
void bnx2x_update_coalesce(struct bnx2x *bp)
{
int i;
for_each_eth_queue(bp, i)
bnx2x_update_coalesce_sb(bp, bp->fp[i].fw_sb_id,
bp->tx_ticks, bp->rx_ticks);
}
static void bnx2x_init_sp_ring(struct bnx2x *bp)
{
spin_lock_init(&bp->spq_lock);
atomic_set(&bp->cq_spq_left, MAX_SPQ_PENDING);
bp->spq_prod_idx = 0;
bp->dsb_sp_prod = BNX2X_SP_DSB_INDEX;
bp->spq_prod_bd = bp->spq;
bp->spq_last_bd = bp->spq_prod_bd + MAX_SP_DESC_CNT;
}
static void bnx2x_init_eq_ring(struct bnx2x *bp)
{
int i;
for (i = 1; i <= NUM_EQ_PAGES; i++) {
union event_ring_elem *elem =
&bp->eq_ring[EQ_DESC_CNT_PAGE * i - 1];
elem->next_page.addr.hi =
cpu_to_le32(U64_HI(bp->eq_mapping +
BCM_PAGE_SIZE * (i % NUM_EQ_PAGES)));
elem->next_page.addr.lo =
cpu_to_le32(U64_LO(bp->eq_mapping +
BCM_PAGE_SIZE*(i % NUM_EQ_PAGES)));
}
bp->eq_cons = 0;
bp->eq_prod = NUM_EQ_DESC;
bp->eq_cons_sb = BNX2X_EQ_INDEX;
/* we want a warning message before it gets wrought... */
atomic_set(&bp->eq_spq_left,
min_t(int, MAX_SP_DESC_CNT - MAX_SPQ_PENDING, NUM_EQ_DESC) - 1);
}
/* called with netif_addr_lock_bh() */
static int bnx2x_set_q_rx_mode(struct bnx2x *bp, u8 cl_id,
unsigned long rx_mode_flags,
unsigned long rx_accept_flags,
unsigned long tx_accept_flags,
unsigned long ramrod_flags)
{
struct bnx2x_rx_mode_ramrod_params ramrod_param;
int rc;
memset(&ramrod_param, 0, sizeof(ramrod_param));
/* Prepare ramrod parameters */
ramrod_param.cid = 0;
ramrod_param.cl_id = cl_id;
ramrod_param.rx_mode_obj = &bp->rx_mode_obj;
ramrod_param.func_id = BP_FUNC(bp);
ramrod_param.pstate = &bp->sp_state;
ramrod_param.state = BNX2X_FILTER_RX_MODE_PENDING;
ramrod_param.rdata = bnx2x_sp(bp, rx_mode_rdata);
ramrod_param.rdata_mapping = bnx2x_sp_mapping(bp, rx_mode_rdata);
set_bit(BNX2X_FILTER_RX_MODE_PENDING, &bp->sp_state);
ramrod_param.ramrod_flags = ramrod_flags;
ramrod_param.rx_mode_flags = rx_mode_flags;
ramrod_param.rx_accept_flags = rx_accept_flags;
ramrod_param.tx_accept_flags = tx_accept_flags;
rc = bnx2x_config_rx_mode(bp, &ramrod_param);
if (rc < 0) {
BNX2X_ERR("Set rx_mode %d failed\n", bp->rx_mode);
return rc;
}
return 0;
}
static int bnx2x_fill_accept_flags(struct bnx2x *bp, u32 rx_mode,
unsigned long *rx_accept_flags,
unsigned long *tx_accept_flags)
{
/* Clear the flags first */
*rx_accept_flags = 0;
*tx_accept_flags = 0;
switch (rx_mode) {
case BNX2X_RX_MODE_NONE:
/*
* 'drop all' supersedes any accept flags that may have been
* passed to the function.
*/
break;
case BNX2X_RX_MODE_NORMAL:
__set_bit(BNX2X_ACCEPT_UNICAST, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_MULTICAST, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, rx_accept_flags);
/* internal switching mode */
__set_bit(BNX2X_ACCEPT_UNICAST, tx_accept_flags);
__set_bit(BNX2X_ACCEPT_MULTICAST, tx_accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, tx_accept_flags);
if (bp->accept_any_vlan) {
__set_bit(BNX2X_ACCEPT_ANY_VLAN, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_ANY_VLAN, tx_accept_flags);
}
break;
case BNX2X_RX_MODE_ALLMULTI:
__set_bit(BNX2X_ACCEPT_UNICAST, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_ALL_MULTICAST, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, rx_accept_flags);
/* internal switching mode */
__set_bit(BNX2X_ACCEPT_UNICAST, tx_accept_flags);
__set_bit(BNX2X_ACCEPT_ALL_MULTICAST, tx_accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, tx_accept_flags);
if (bp->accept_any_vlan) {
__set_bit(BNX2X_ACCEPT_ANY_VLAN, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_ANY_VLAN, tx_accept_flags);
}
break;
case BNX2X_RX_MODE_PROMISC:
/* According to definition of SI mode, iface in promisc mode
* should receive matched and unmatched (in resolution of port)
* unicast packets.
*/
__set_bit(BNX2X_ACCEPT_UNMATCHED, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_UNICAST, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_ALL_MULTICAST, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, rx_accept_flags);
/* internal switching mode */
__set_bit(BNX2X_ACCEPT_ALL_MULTICAST, tx_accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, tx_accept_flags);
if (IS_MF_SI(bp))
__set_bit(BNX2X_ACCEPT_ALL_UNICAST, tx_accept_flags);
else
__set_bit(BNX2X_ACCEPT_UNICAST, tx_accept_flags);
__set_bit(BNX2X_ACCEPT_ANY_VLAN, rx_accept_flags);
__set_bit(BNX2X_ACCEPT_ANY_VLAN, tx_accept_flags);
break;
default:
BNX2X_ERR("Unknown rx_mode: %d\n", rx_mode);
return -EINVAL;
}
return 0;
}
/* called with netif_addr_lock_bh() */
static int bnx2x_set_storm_rx_mode(struct bnx2x *bp)
{
unsigned long rx_mode_flags = 0, ramrod_flags = 0;
unsigned long rx_accept_flags = 0, tx_accept_flags = 0;
int rc;
if (!NO_FCOE(bp))
/* Configure rx_mode of FCoE Queue */
__set_bit(BNX2X_RX_MODE_FCOE_ETH, &rx_mode_flags);
rc = bnx2x_fill_accept_flags(bp, bp->rx_mode, &rx_accept_flags,
&tx_accept_flags);
if (rc)
return rc;
__set_bit(RAMROD_RX, &ramrod_flags);
__set_bit(RAMROD_TX, &ramrod_flags);
return bnx2x_set_q_rx_mode(bp, bp->fp->cl_id, rx_mode_flags,
rx_accept_flags, tx_accept_flags,
ramrod_flags);
}
static void bnx2x_init_internal_common(struct bnx2x *bp)
{
int i;
/* Zero this manually as its initialization is
currently missing in the initTool */
for (i = 0; i < (USTORM_AGG_DATA_SIZE >> 2); i++)
REG_WR(bp, BAR_USTRORM_INTMEM +
USTORM_AGG_DATA_OFFSET + i * 4, 0);
if (!CHIP_IS_E1x(bp)) {
REG_WR8(bp, BAR_CSTRORM_INTMEM + CSTORM_IGU_MODE_OFFSET,
CHIP_INT_MODE_IS_BC(bp) ?
HC_IGU_BC_MODE : HC_IGU_NBC_MODE);
}
}
static void bnx2x_init_internal(struct bnx2x *bp, u32 load_code)
{
switch (load_code) {
case FW_MSG_CODE_DRV_LOAD_COMMON:
case FW_MSG_CODE_DRV_LOAD_COMMON_CHIP:
bnx2x_init_internal_common(bp);
/* no break */
case FW_MSG_CODE_DRV_LOAD_PORT:
/* nothing to do */
/* no break */
case FW_MSG_CODE_DRV_LOAD_FUNCTION:
/* internal memory per function is
initialized inside bnx2x_pf_init */
break;
default:
BNX2X_ERR("Unknown load_code (0x%x) from MCP\n", load_code);
break;
}
}
static inline u8 bnx2x_fp_igu_sb_id(struct bnx2x_fastpath *fp)
{
return fp->bp->igu_base_sb + fp->index + CNIC_SUPPORT(fp->bp);
}
static inline u8 bnx2x_fp_fw_sb_id(struct bnx2x_fastpath *fp)
{
return fp->bp->base_fw_ndsb + fp->index + CNIC_SUPPORT(fp->bp);
}
static u8 bnx2x_fp_cl_id(struct bnx2x_fastpath *fp)
{
if (CHIP_IS_E1x(fp->bp))
return BP_L_ID(fp->bp) + fp->index;
else /* We want Client ID to be the same as IGU SB ID for 57712 */
return bnx2x_fp_igu_sb_id(fp);
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
static void bnx2x_init_eth_fp(struct bnx2x *bp, int fp_idx)
{
struct bnx2x_fastpath *fp = &bp->fp[fp_idx];
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 cos;
unsigned long q_type = 0;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u32 cids[BNX2X_MULTI_TX_COS] = { 0 };
fp->rx_queue = fp_idx;
fp->cid = fp_idx;
fp->cl_id = bnx2x_fp_cl_id(fp);
fp->fw_sb_id = bnx2x_fp_fw_sb_id(fp);
fp->igu_sb_id = bnx2x_fp_igu_sb_id(fp);
/* qZone id equals to FW (per path) client id */
fp->cl_qzone_id = bnx2x_fp_qzone_id(fp);
/* init shortcut */
fp->ustorm_rx_prods_offset = bnx2x_rx_ustorm_prods_offset(fp);
/* Setup SB indices */
fp->rx_cons_sb = BNX2X_RX_SB_INDEX;
/* Configure Queue State object */
__set_bit(BNX2X_Q_TYPE_HAS_RX, &q_type);
__set_bit(BNX2X_Q_TYPE_HAS_TX, &q_type);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
BUG_ON(fp->max_cos > BNX2X_MULTI_TX_COS);
/* init tx data */
for_each_cos_in_tx_queue(fp, cos) {
bnx2x_init_txdata(bp, fp->txdata_ptr[cos],
CID_COS_TO_TX_ONLY_CID(fp->cid, cos, bp),
FP_COS_TO_TXQ(fp, cos, bp),
BNX2X_TX_SB_INDEX_BASE + cos, fp);
cids[cos] = fp->txdata_ptr[cos]->cid;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
}
/* nothing more for vf to do here */
if (IS_VF(bp))
return;
bnx2x_init_sb(bp, fp->status_blk_mapping, BNX2X_VF_ID_INVALID, false,
fp->fw_sb_id, fp->igu_sb_id);
bnx2x_update_fpsb_idx(fp);
bnx2x_init_queue_obj(bp, &bnx2x_sp_obj(bp, fp).q_obj, fp->cl_id, cids,
fp->max_cos, BP_FUNC(bp), bnx2x_sp(bp, q_rdata),
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bnx2x_sp_mapping(bp, q_rdata), q_type);
/**
* Configure classification DBs: Always enable Tx switching
*/
bnx2x_init_vlan_mac_fp_objs(fp, BNX2X_OBJ_TYPE_RX_TX);
DP(NETIF_MSG_IFUP,
"queue[%d]: bnx2x_init_sb(%p,%p) cl_id %d fw_sb %d igu_sb %d\n",
fp_idx, bp, fp->status_blk.e2_sb, fp->cl_id, fp->fw_sb_id,
fp->igu_sb_id);
}
static void bnx2x_init_tx_ring_one(struct bnx2x_fp_txdata *txdata)
{
int i;
for (i = 1; i <= NUM_TX_RINGS; i++) {
struct eth_tx_next_bd *tx_next_bd =
&txdata->tx_desc_ring[TX_DESC_CNT * i - 1].next_bd;
tx_next_bd->addr_hi =
cpu_to_le32(U64_HI(txdata->tx_desc_mapping +
BCM_PAGE_SIZE*(i % NUM_TX_RINGS)));
tx_next_bd->addr_lo =
cpu_to_le32(U64_LO(txdata->tx_desc_mapping +
BCM_PAGE_SIZE*(i % NUM_TX_RINGS)));
}
*txdata->tx_cons_sb = cpu_to_le16(0);
SET_FLAG(txdata->tx_db.data.header.header, DOORBELL_HDR_DB_TYPE, 1);
txdata->tx_db.data.zero_fill1 = 0;
txdata->tx_db.data.prod = 0;
txdata->tx_pkt_prod = 0;
txdata->tx_pkt_cons = 0;
txdata->tx_bd_prod = 0;
txdata->tx_bd_cons = 0;
txdata->tx_pkt = 0;
}
static void bnx2x_init_tx_rings_cnic(struct bnx2x *bp)
{
int i;
for_each_tx_queue_cnic(bp, i)
bnx2x_init_tx_ring_one(bp->fp[i].txdata_ptr[0]);
}
static void bnx2x_init_tx_rings(struct bnx2x *bp)
{
int i;
u8 cos;
for_each_eth_queue(bp, i)
for_each_cos_in_tx_queue(&bp->fp[i], cos)
bnx2x_init_tx_ring_one(bp->fp[i].txdata_ptr[cos]);
}
static void bnx2x_init_fcoe_fp(struct bnx2x *bp)
{
struct bnx2x_fastpath *fp = bnx2x_fcoe_fp(bp);
unsigned long q_type = 0;
bnx2x_fcoe(bp, rx_queue) = BNX2X_NUM_ETH_QUEUES(bp);
bnx2x_fcoe(bp, cl_id) = bnx2x_cnic_eth_cl_id(bp,
BNX2X_FCOE_ETH_CL_ID_IDX);
bnx2x_fcoe(bp, cid) = BNX2X_FCOE_ETH_CID(bp);
bnx2x_fcoe(bp, fw_sb_id) = DEF_SB_ID;
bnx2x_fcoe(bp, igu_sb_id) = bp->igu_dsb_id;
bnx2x_fcoe(bp, rx_cons_sb) = BNX2X_FCOE_L2_RX_INDEX;
bnx2x_init_txdata(bp, bnx2x_fcoe(bp, txdata_ptr[0]),
fp->cid, FCOE_TXQ_IDX(bp), BNX2X_FCOE_L2_TX_INDEX,
fp);
DP(NETIF_MSG_IFUP, "created fcoe tx data (fp index %d)\n", fp->index);
/* qZone id equals to FW (per path) client id */
bnx2x_fcoe(bp, cl_qzone_id) = bnx2x_fp_qzone_id(fp);
/* init shortcut */
bnx2x_fcoe(bp, ustorm_rx_prods_offset) =
bnx2x_rx_ustorm_prods_offset(fp);
/* Configure Queue State object */
__set_bit(BNX2X_Q_TYPE_HAS_RX, &q_type);
__set_bit(BNX2X_Q_TYPE_HAS_TX, &q_type);
/* No multi-CoS for FCoE L2 client */
BUG_ON(fp->max_cos != 1);
bnx2x_init_queue_obj(bp, &bnx2x_sp_obj(bp, fp).q_obj, fp->cl_id,
&fp->cid, 1, BP_FUNC(bp), bnx2x_sp(bp, q_rdata),
bnx2x_sp_mapping(bp, q_rdata), q_type);
DP(NETIF_MSG_IFUP,
"queue[%d]: bnx2x_init_sb(%p,%p) cl_id %d fw_sb %d igu_sb %d\n",
fp->index, bp, fp->status_blk.e2_sb, fp->cl_id, fp->fw_sb_id,
fp->igu_sb_id);
}
void bnx2x_nic_init_cnic(struct bnx2x *bp)
{
if (!NO_FCOE(bp))
bnx2x_init_fcoe_fp(bp);
bnx2x_init_sb(bp, bp->cnic_sb_mapping,
BNX2X_VF_ID_INVALID, false,
bnx2x_cnic_fw_sb_id(bp), bnx2x_cnic_igu_sb_id(bp));
/* ensure status block indices were read */
rmb();
bnx2x_init_rx_rings_cnic(bp);
bnx2x_init_tx_rings_cnic(bp);
/* flush all */
mb();
mmiowb();
}
void bnx2x_pre_irq_nic_init(struct bnx2x *bp)
{
int i;
/* Setup NIC internals and enable interrupts */
for_each_eth_queue(bp, i)
bnx2x_init_eth_fp(bp, i);
/* ensure status block indices were read */
rmb();
bnx2x_init_rx_rings(bp);
bnx2x_init_tx_rings(bp);
if (IS_PF(bp)) {
/* Initialize MOD_ABS interrupts */
bnx2x_init_mod_abs_int(bp, &bp->link_vars, bp->common.chip_id,
bp->common.shmem_base,
bp->common.shmem2_base, BP_PORT(bp));
/* initialize the default status block and sp ring */
bnx2x_init_def_sb(bp);
bnx2x_update_dsb_idx(bp);
bnx2x_init_sp_ring(bp);
} else {
bnx2x_memset_stats(bp);
}
}
void bnx2x_post_irq_nic_init(struct bnx2x *bp, u32 load_code)
{
bnx2x_init_eq_ring(bp);
bnx2x_init_internal(bp, load_code);
bnx2x_pf_init(bp);
bnx2x_stats_init(bp);
/* flush all before enabling interrupts */
mb();
mmiowb();
bnx2x_int_enable(bp);
/* Check for SPIO5 */
bnx2x_attn_int_deasserted0(bp,
REG_RD(bp, MISC_REG_AEU_AFTER_INVERT_1_FUNC_0 + BP_PORT(bp)*4) &
AEU_INPUTS_ATTN_BITS_SPIO5);
}
/* gzip service functions */
static int bnx2x_gunzip_init(struct bnx2x *bp)
{
bp->gunzip_buf = dma_alloc_coherent(&bp->pdev->dev, FW_BUF_SIZE,
&bp->gunzip_mapping, GFP_KERNEL);
if (bp->gunzip_buf == NULL)
goto gunzip_nomem1;
bp->strm = kmalloc(sizeof(*bp->strm), GFP_KERNEL);
if (bp->strm == NULL)
goto gunzip_nomem2;
bp->strm->workspace = vmalloc(zlib_inflate_workspacesize());
if (bp->strm->workspace == NULL)
goto gunzip_nomem3;
return 0;
gunzip_nomem3:
kfree(bp->strm);
bp->strm = NULL;
gunzip_nomem2:
dma_free_coherent(&bp->pdev->dev, FW_BUF_SIZE, bp->gunzip_buf,
bp->gunzip_mapping);
bp->gunzip_buf = NULL;
gunzip_nomem1:
BNX2X_ERR("Cannot allocate firmware buffer for un-compression\n");
return -ENOMEM;
}
static void bnx2x_gunzip_end(struct bnx2x *bp)
{
if (bp->strm) {
vfree(bp->strm->workspace);
kfree(bp->strm);
bp->strm = NULL;
}
if (bp->gunzip_buf) {
dma_free_coherent(&bp->pdev->dev, FW_BUF_SIZE, bp->gunzip_buf,
bp->gunzip_mapping);
bp->gunzip_buf = NULL;
}
}
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
static int bnx2x_gunzip(struct bnx2x *bp, const u8 *zbuf, int len)
{
int n, rc;
/* check gzip header */
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
if ((zbuf[0] != 0x1f) || (zbuf[1] != 0x8b) || (zbuf[2] != Z_DEFLATED)) {
BNX2X_ERR("Bad gzip header\n");
return -EINVAL;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
}
n = 10;
#define FNAME 0x8
if (zbuf[3] & FNAME)
while ((zbuf[n++] != 0) && (n < len));
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
bp->strm->next_in = (typeof(bp->strm->next_in))zbuf + n;
bp->strm->avail_in = len - n;
bp->strm->next_out = bp->gunzip_buf;
bp->strm->avail_out = FW_BUF_SIZE;
rc = zlib_inflateInit2(bp->strm, -MAX_WBITS);
if (rc != Z_OK)
return rc;
rc = zlib_inflate(bp->strm, Z_FINISH);
if ((rc != Z_OK) && (rc != Z_STREAM_END))
netdev_err(bp->dev, "Firmware decompression error: %s\n",
bp->strm->msg);
bp->gunzip_outlen = (FW_BUF_SIZE - bp->strm->avail_out);
if (bp->gunzip_outlen & 0x3)
netdev_err(bp->dev,
"Firmware decompression error: gunzip_outlen (%d) not aligned\n",
bp->gunzip_outlen);
bp->gunzip_outlen >>= 2;
zlib_inflateEnd(bp->strm);
if (rc == Z_STREAM_END)
return 0;
return rc;
}
/* nic load/unload */
/*
* General service functions
*/
/* send a NIG loopback debug packet */
static void bnx2x_lb_pckt(struct bnx2x *bp)
{
u32 wb_write[3];
/* Ethernet source and destination addresses */
wb_write[0] = 0x55555555;
wb_write[1] = 0x55555555;
wb_write[2] = 0x20; /* SOP */
REG_WR_DMAE(bp, NIG_REG_DEBUG_PACKET_LB, wb_write, 3);
/* NON-IP protocol */
wb_write[0] = 0x09000000;
wb_write[1] = 0x55555555;
wb_write[2] = 0x10; /* EOP, eop_bvalid = 0 */
REG_WR_DMAE(bp, NIG_REG_DEBUG_PACKET_LB, wb_write, 3);
}
/* some of the internal memories
* are not directly readable from the driver
* to test them we send debug packets
*/
static int bnx2x_int_mem_test(struct bnx2x *bp)
{
int factor;
int count, i;
u32 val = 0;
if (CHIP_REV_IS_FPGA(bp))
factor = 120;
else if (CHIP_REV_IS_EMUL(bp))
factor = 200;
else
factor = 1;
/* Disable inputs of parser neighbor blocks */
REG_WR(bp, TSDM_REG_ENABLE_IN1, 0x0);
REG_WR(bp, TCM_REG_PRS_IFEN, 0x0);
REG_WR(bp, CFC_REG_DEBUG0, 0x1);
REG_WR(bp, NIG_REG_PRS_REQ_IN_EN, 0x0);
/* Write 0 to parser credits for CFC search request */
REG_WR(bp, PRS_REG_CFC_SEARCH_INITIAL_CREDIT, 0x0);
/* send Ethernet packet */
bnx2x_lb_pckt(bp);
/* TODO do i reset NIG statistic? */
/* Wait until NIG register shows 1 packet of size 0x10 */
count = 1000 * factor;
while (count) {
bnx2x_read_dmae(bp, NIG_REG_STAT2_BRB_OCTET, 2);
val = *bnx2x_sp(bp, wb_data[0]);
if (val == 0x10)
break;
usleep_range(10000, 20000);
count--;
}
if (val != 0x10) {
BNX2X_ERR("NIG timeout val = 0x%x\n", val);
return -1;
}
/* Wait until PRS register shows 1 packet */
count = 1000 * factor;
while (count) {
val = REG_RD(bp, PRS_REG_NUM_OF_PACKETS);
if (val == 1)
break;
usleep_range(10000, 20000);
count--;
}
if (val != 0x1) {
BNX2X_ERR("PRS timeout val = 0x%x\n", val);
return -2;
}
/* Reset and init BRB, PRS */
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_CLEAR, 0x03);
msleep(50);
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_SET, 0x03);
msleep(50);
bnx2x_init_block(bp, BLOCK_BRB1, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_PRS, PHASE_COMMON);
DP(NETIF_MSG_HW, "part2\n");
/* Disable inputs of parser neighbor blocks */
REG_WR(bp, TSDM_REG_ENABLE_IN1, 0x0);
REG_WR(bp, TCM_REG_PRS_IFEN, 0x0);
REG_WR(bp, CFC_REG_DEBUG0, 0x1);
REG_WR(bp, NIG_REG_PRS_REQ_IN_EN, 0x0);
/* Write 0 to parser credits for CFC search request */
REG_WR(bp, PRS_REG_CFC_SEARCH_INITIAL_CREDIT, 0x0);
/* send 10 Ethernet packets */
for (i = 0; i < 10; i++)
bnx2x_lb_pckt(bp);
/* Wait until NIG register shows 10 + 1
packets of size 11*0x10 = 0xb0 */
count = 1000 * factor;
while (count) {
bnx2x_read_dmae(bp, NIG_REG_STAT2_BRB_OCTET, 2);
val = *bnx2x_sp(bp, wb_data[0]);
if (val == 0xb0)
break;
usleep_range(10000, 20000);
count--;
}
if (val != 0xb0) {
BNX2X_ERR("NIG timeout val = 0x%x\n", val);
return -3;
}
/* Wait until PRS register shows 2 packets */
val = REG_RD(bp, PRS_REG_NUM_OF_PACKETS);
if (val != 2)
BNX2X_ERR("PRS timeout val = 0x%x\n", val);
/* Write 1 to parser credits for CFC search request */
REG_WR(bp, PRS_REG_CFC_SEARCH_INITIAL_CREDIT, 0x1);
/* Wait until PRS register shows 3 packets */
msleep(10 * factor);
/* Wait until NIG register shows 1 packet of size 0x10 */
val = REG_RD(bp, PRS_REG_NUM_OF_PACKETS);
if (val != 3)
BNX2X_ERR("PRS timeout val = 0x%x\n", val);
/* clear NIG EOP FIFO */
for (i = 0; i < 11; i++)
REG_RD(bp, NIG_REG_INGRESS_EOP_LB_FIFO);
val = REG_RD(bp, NIG_REG_INGRESS_EOP_LB_EMPTY);
if (val != 1) {
BNX2X_ERR("clear of NIG failed\n");
return -4;
}
/* Reset and init BRB, PRS, NIG */
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_CLEAR, 0x03);
msleep(50);
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_SET, 0x03);
msleep(50);
bnx2x_init_block(bp, BLOCK_BRB1, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_PRS, PHASE_COMMON);
if (!CNIC_SUPPORT(bp))
/* set NIC mode */
REG_WR(bp, PRS_REG_NIC_MODE, 1);
/* Enable inputs of parser neighbor blocks */
REG_WR(bp, TSDM_REG_ENABLE_IN1, 0x7fffffff);
REG_WR(bp, TCM_REG_PRS_IFEN, 0x1);
REG_WR(bp, CFC_REG_DEBUG0, 0x0);
REG_WR(bp, NIG_REG_PRS_REQ_IN_EN, 0x1);
DP(NETIF_MSG_HW, "done\n");
return 0; /* OK */
}
static void bnx2x_enable_blocks_attention(struct bnx2x *bp)
{
u32 val;
REG_WR(bp, PXP_REG_PXP_INT_MASK_0, 0);
if (!CHIP_IS_E1x(bp))
REG_WR(bp, PXP_REG_PXP_INT_MASK_1, 0x40);
else
REG_WR(bp, PXP_REG_PXP_INT_MASK_1, 0);
REG_WR(bp, DORQ_REG_DORQ_INT_MASK, 0);
REG_WR(bp, CFC_REG_CFC_INT_MASK, 0);
/*
* mask read length error interrupts in brb for parser
* (parsing unit and 'checksum and crc' unit)
* these errors are legal (PU reads fixed length and CAC can cause
* read length error on truncated packets)
*/
REG_WR(bp, BRB1_REG_BRB1_INT_MASK, 0xFC00);
REG_WR(bp, QM_REG_QM_INT_MASK, 0);
REG_WR(bp, TM_REG_TM_INT_MASK, 0);
REG_WR(bp, XSDM_REG_XSDM_INT_MASK_0, 0);
REG_WR(bp, XSDM_REG_XSDM_INT_MASK_1, 0);
REG_WR(bp, XCM_REG_XCM_INT_MASK, 0);
/* REG_WR(bp, XSEM_REG_XSEM_INT_MASK_0, 0); */
/* REG_WR(bp, XSEM_REG_XSEM_INT_MASK_1, 0); */
REG_WR(bp, USDM_REG_USDM_INT_MASK_0, 0);
REG_WR(bp, USDM_REG_USDM_INT_MASK_1, 0);
REG_WR(bp, UCM_REG_UCM_INT_MASK, 0);
/* REG_WR(bp, USEM_REG_USEM_INT_MASK_0, 0); */
/* REG_WR(bp, USEM_REG_USEM_INT_MASK_1, 0); */
REG_WR(bp, GRCBASE_UPB + PB_REG_PB_INT_MASK, 0);
REG_WR(bp, CSDM_REG_CSDM_INT_MASK_0, 0);
REG_WR(bp, CSDM_REG_CSDM_INT_MASK_1, 0);
REG_WR(bp, CCM_REG_CCM_INT_MASK, 0);
/* REG_WR(bp, CSEM_REG_CSEM_INT_MASK_0, 0); */
/* REG_WR(bp, CSEM_REG_CSEM_INT_MASK_1, 0); */
val = PXP2_PXP2_INT_MASK_0_REG_PGL_CPL_AFT |
PXP2_PXP2_INT_MASK_0_REG_PGL_CPL_OF |
PXP2_PXP2_INT_MASK_0_REG_PGL_PCIE_ATTN;
if (!CHIP_IS_E1x(bp))
val |= PXP2_PXP2_INT_MASK_0_REG_PGL_READ_BLOCKED |
PXP2_PXP2_INT_MASK_0_REG_PGL_WRITE_BLOCKED;
REG_WR(bp, PXP2_REG_PXP2_INT_MASK_0, val);
REG_WR(bp, TSDM_REG_TSDM_INT_MASK_0, 0);
REG_WR(bp, TSDM_REG_TSDM_INT_MASK_1, 0);
REG_WR(bp, TCM_REG_TCM_INT_MASK, 0);
/* REG_WR(bp, TSEM_REG_TSEM_INT_MASK_0, 0); */
if (!CHIP_IS_E1x(bp))
/* enable VFC attentions: bits 11 and 12, bits 31:13 reserved */
REG_WR(bp, TSEM_REG_TSEM_INT_MASK_1, 0x07ff);
REG_WR(bp, CDU_REG_CDU_INT_MASK, 0);
REG_WR(bp, DMAE_REG_DMAE_INT_MASK, 0);
/* REG_WR(bp, MISC_REG_MISC_INT_MASK, 0); */
REG_WR(bp, PBF_REG_PBF_INT_MASK, 0x18); /* bit 3,4 masked */
}
static void bnx2x_reset_common(struct bnx2x *bp)
{
u32 val = 0x1400;
/* reset_common */
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_CLEAR,
0xd3ffff7f);
if (CHIP_IS_E3(bp)) {
val |= MISC_REGISTERS_RESET_REG_2_MSTAT0;
val |= MISC_REGISTERS_RESET_REG_2_MSTAT1;
}
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_2_CLEAR, val);
}
static void bnx2x_setup_dmae(struct bnx2x *bp)
{
bp->dmae_ready = 0;
spin_lock_init(&bp->dmae_lock);
}
static void bnx2x_init_pxp(struct bnx2x *bp)
{
u16 devctl;
int r_order, w_order;
pcie_capability_read_word(bp->pdev, PCI_EXP_DEVCTL, &devctl);
DP(NETIF_MSG_HW, "read 0x%x from devctl\n", devctl);
w_order = ((devctl & PCI_EXP_DEVCTL_PAYLOAD) >> 5);
if (bp->mrrs == -1)
r_order = ((devctl & PCI_EXP_DEVCTL_READRQ) >> 12);
else {
DP(NETIF_MSG_HW, "force read order to %d\n", bp->mrrs);
r_order = bp->mrrs;
}
bnx2x_init_pxp_arb(bp, r_order, w_order);
}
static void bnx2x_setup_fan_failure_detection(struct bnx2x *bp)
{
int is_required;
u32 val;
int port;
if (BP_NOMCP(bp))
return;
is_required = 0;
val = SHMEM_RD(bp, dev_info.shared_hw_config.config2) &
SHARED_HW_CFG_FAN_FAILURE_MASK;
if (val == SHARED_HW_CFG_FAN_FAILURE_ENABLED)
is_required = 1;
/*
* The fan failure mechanism is usually related to the PHY type since
* the power consumption of the board is affected by the PHY. Currently,
* fan is required for most designs with SFX7101, BCM8727 and BCM8481.
*/
else if (val == SHARED_HW_CFG_FAN_FAILURE_PHY_TYPE)
for (port = PORT_0; port < PORT_MAX; port++) {
is_required |=
bnx2x_fan_failure_det_req(
bp,
bp->common.shmem_base,
bp->common.shmem2_base,
port);
}
DP(NETIF_MSG_HW, "fan detection setting: %d\n", is_required);
if (is_required == 0)
return;
/* Fan failure is indicated by SPIO 5 */
bnx2x_set_spio(bp, MISC_SPIO_SPIO5, MISC_SPIO_INPUT_HI_Z);
/* set to active low mode */
val = REG_RD(bp, MISC_REG_SPIO_INT);
val |= (MISC_SPIO_SPIO5 << MISC_SPIO_INT_OLD_SET_POS);
REG_WR(bp, MISC_REG_SPIO_INT, val);
/* enable interrupt to signal the IGU */
val = REG_RD(bp, MISC_REG_SPIO_EVENT_EN);
val |= MISC_SPIO_SPIO5;
REG_WR(bp, MISC_REG_SPIO_EVENT_EN, val);
}
void bnx2x_pf_disable(struct bnx2x *bp)
{
u32 val = REG_RD(bp, IGU_REG_PF_CONFIGURATION);
val &= ~IGU_PF_CONF_FUNC_EN;
REG_WR(bp, IGU_REG_PF_CONFIGURATION, val);
REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, 0);
REG_WR(bp, CFC_REG_WEAK_ENABLE_PF, 0);
}
static void bnx2x__common_init_phy(struct bnx2x *bp)
{
u32 shmem_base[2], shmem2_base[2];
/* Avoid common init in case MFW supports LFA */
if (SHMEM2_RD(bp, size) >
(u32)offsetof(struct shmem2_region, lfa_host_addr[BP_PORT(bp)]))
return;
shmem_base[0] = bp->common.shmem_base;
shmem2_base[0] = bp->common.shmem2_base;
if (!CHIP_IS_E1x(bp)) {
shmem_base[1] =
SHMEM2_RD(bp, other_shmem_base_addr);
shmem2_base[1] =
SHMEM2_RD(bp, other_shmem2_base_addr);
}
bnx2x_acquire_phy_lock(bp);
bnx2x_common_init_phy(bp, shmem_base, shmem2_base,
bp->common.chip_id);
bnx2x_release_phy_lock(bp);
}
static void bnx2x_config_endianity(struct bnx2x *bp, u32 val)
{
REG_WR(bp, PXP2_REG_RQ_QM_ENDIAN_M, val);
REG_WR(bp, PXP2_REG_RQ_TM_ENDIAN_M, val);
REG_WR(bp, PXP2_REG_RQ_SRC_ENDIAN_M, val);
REG_WR(bp, PXP2_REG_RQ_CDU_ENDIAN_M, val);
REG_WR(bp, PXP2_REG_RQ_DBG_ENDIAN_M, val);
/* make sure this value is 0 */
REG_WR(bp, PXP2_REG_RQ_HC_ENDIAN_M, 0);
REG_WR(bp, PXP2_REG_RD_QM_SWAP_MODE, val);
REG_WR(bp, PXP2_REG_RD_TM_SWAP_MODE, val);
REG_WR(bp, PXP2_REG_RD_SRC_SWAP_MODE, val);
REG_WR(bp, PXP2_REG_RD_CDURD_SWAP_MODE, val);
}
static void bnx2x_set_endianity(struct bnx2x *bp)
{
#ifdef __BIG_ENDIAN
bnx2x_config_endianity(bp, 1);
#else
bnx2x_config_endianity(bp, 0);
#endif
}
static void bnx2x_reset_endianity(struct bnx2x *bp)
{
bnx2x_config_endianity(bp, 0);
}
/**
* bnx2x_init_hw_common - initialize the HW at the COMMON phase.
*
* @bp: driver handle
*/
static int bnx2x_init_hw_common(struct bnx2x *bp)
{
u32 val;
DP(NETIF_MSG_HW, "starting common init func %d\n", BP_ABS_FUNC(bp));
/*
* take the RESET lock to protect undi_unload flow from accessing
* registers while we're resetting the chip
*/
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RESET);
bnx2x_reset_common(bp);
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_SET, 0xffffffff);
val = 0xfffc;
if (CHIP_IS_E3(bp)) {
val |= MISC_REGISTERS_RESET_REG_2_MSTAT0;
val |= MISC_REGISTERS_RESET_REG_2_MSTAT1;
}
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_2_SET, val);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RESET);
bnx2x_init_block(bp, BLOCK_MISC, PHASE_COMMON);
if (!CHIP_IS_E1x(bp)) {
u8 abs_func_id;
/**
* 4-port mode or 2-port mode we need to turn of master-enable
* for everyone, after that, turn it back on for self.
* so, we disregard multi-function or not, and always disable
* for all functions on the given path, this means 0,2,4,6 for
* path 0 and 1,3,5,7 for path 1
*/
for (abs_func_id = BP_PATH(bp);
abs_func_id < E2_FUNC_MAX*2; abs_func_id += 2) {
if (abs_func_id == BP_ABS_FUNC(bp)) {
REG_WR(bp,
PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER,
1);
continue;
}
bnx2x_pretend_func(bp, abs_func_id);
/* clear pf enable */
bnx2x_pf_disable(bp);
bnx2x_pretend_func(bp, BP_ABS_FUNC(bp));
}
}
bnx2x_init_block(bp, BLOCK_PXP, PHASE_COMMON);
if (CHIP_IS_E1(bp)) {
/* enable HW interrupt from PXP on USDM overflow
bit 16 on INT_MASK_0 */
REG_WR(bp, PXP_REG_PXP_INT_MASK_0, 0);
}
bnx2x_init_block(bp, BLOCK_PXP2, PHASE_COMMON);
bnx2x_init_pxp(bp);
bnx2x_set_endianity(bp);
bnx2x_ilt_init_page_size(bp, INITOP_SET);
if (CHIP_REV_IS_FPGA(bp) && CHIP_IS_E1H(bp))
REG_WR(bp, PXP2_REG_PGL_TAGS_LIMIT, 0x1);
/* let the HW do it's magic ... */
msleep(100);
/* finish PXP init */
val = REG_RD(bp, PXP2_REG_RQ_CFG_DONE);
if (val != 1) {
BNX2X_ERR("PXP2 CFG failed\n");
return -EBUSY;
}
val = REG_RD(bp, PXP2_REG_RD_INIT_DONE);
if (val != 1) {
BNX2X_ERR("PXP2 RD_INIT failed\n");
return -EBUSY;
}
/* Timers bug workaround E2 only. We need to set the entire ILT to
* have entries with value "0" and valid bit on.
* This needs to be done by the first PF that is loaded in a path
* (i.e. common phase)
*/
if (!CHIP_IS_E1x(bp)) {
/* In E2 there is a bug in the timers block that can cause function 6 / 7
* (i.e. vnic3) to start even if it is marked as "scan-off".
* This occurs when a different function (func2,3) is being marked
* as "scan-off". Real-life scenario for example: if a driver is being
* load-unloaded while func6,7 are down. This will cause the timer to access
* the ilt, translate to a logical address and send a request to read/write.
* Since the ilt for the function that is down is not valid, this will cause
* a translation error which is unrecoverable.
* The Workaround is intended to make sure that when this happens nothing fatal
* will occur. The workaround:
* 1. First PF driver which loads on a path will:
* a. After taking the chip out of reset, by using pretend,
* it will write "0" to the following registers of
* the other vnics.
* REG_WR(pdev, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, 0);
* REG_WR(pdev, CFC_REG_WEAK_ENABLE_PF,0);
* REG_WR(pdev, CFC_REG_STRONG_ENABLE_PF,0);
* And for itself it will write '1' to
* PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER to enable
* dmae-operations (writing to pram for example.)
* note: can be done for only function 6,7 but cleaner this
* way.
* b. Write zero+valid to the entire ILT.
* c. Init the first_timers_ilt_entry, last_timers_ilt_entry of
* VNIC3 (of that port). The range allocated will be the
* entire ILT. This is needed to prevent ILT range error.
* 2. Any PF driver load flow:
* a. ILT update with the physical addresses of the allocated
* logical pages.
* b. Wait 20msec. - note that this timeout is needed to make
* sure there are no requests in one of the PXP internal
* queues with "old" ILT addresses.
* c. PF enable in the PGLC.
* d. Clear the was_error of the PF in the PGLC. (could have
* occurred while driver was down)
* e. PF enable in the CFC (WEAK + STRONG)
* f. Timers scan enable
* 3. PF driver unload flow:
* a. Clear the Timers scan_en.
* b. Polling for scan_on=0 for that PF.
* c. Clear the PF enable bit in the PXP.
* d. Clear the PF enable in the CFC (WEAK + STRONG)
* e. Write zero+valid to all ILT entries (The valid bit must
* stay set)
* f. If this is VNIC 3 of a port then also init
* first_timers_ilt_entry to zero and last_timers_ilt_entry
* to the last entry in the ILT.
*
* Notes:
* Currently the PF error in the PGLC is non recoverable.
* In the future the there will be a recovery routine for this error.
* Currently attention is masked.
* Having an MCP lock on the load/unload process does not guarantee that
* there is no Timer disable during Func6/7 enable. This is because the
* Timers scan is currently being cleared by the MCP on FLR.
* Step 2.d can be done only for PF6/7 and the driver can also check if
* there is error before clearing it. But the flow above is simpler and
* more general.
* All ILT entries are written by zero+valid and not just PF6/7
* ILT entries since in the future the ILT entries allocation for
* PF-s might be dynamic.
*/
struct ilt_client_info ilt_cli;
struct bnx2x_ilt ilt;
memset(&ilt_cli, 0, sizeof(struct ilt_client_info));
memset(&ilt, 0, sizeof(struct bnx2x_ilt));
/* initialize dummy TM client */
ilt_cli.start = 0;
ilt_cli.end = ILT_NUM_PAGE_ENTRIES - 1;
ilt_cli.client_num = ILT_CLIENT_TM;
/* Step 1: set zeroes to all ilt page entries with valid bit on
* Step 2: set the timers first/last ilt entry to point
* to the entire range to prevent ILT range error for 3rd/4th
* vnic (this code assumes existence of the vnic)
*
* both steps performed by call to bnx2x_ilt_client_init_op()
* with dummy TM client
*
* we must use pretend since PXP2_REG_RQ_##blk##_FIRST_ILT
* and his brother are split registers
*/
bnx2x_pretend_func(bp, (BP_PATH(bp) + 6));
bnx2x_ilt_client_init_op_ilt(bp, &ilt, &ilt_cli, INITOP_CLEAR);
bnx2x_pretend_func(bp, BP_ABS_FUNC(bp));
REG_WR(bp, PXP2_REG_RQ_DRAM_ALIGN, BNX2X_PXP_DRAM_ALIGN);
REG_WR(bp, PXP2_REG_RQ_DRAM_ALIGN_RD, BNX2X_PXP_DRAM_ALIGN);
REG_WR(bp, PXP2_REG_RQ_DRAM_ALIGN_SEL, 1);
}
REG_WR(bp, PXP2_REG_RQ_DISABLE_INPUTS, 0);
REG_WR(bp, PXP2_REG_RD_DISABLE_INPUTS, 0);
if (!CHIP_IS_E1x(bp)) {
int factor = CHIP_REV_IS_EMUL(bp) ? 1000 :
(CHIP_REV_IS_FPGA(bp) ? 400 : 0);
bnx2x_init_block(bp, BLOCK_PGLUE_B, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_ATC, PHASE_COMMON);
/* let the HW do it's magic ... */
do {
msleep(200);
val = REG_RD(bp, ATC_REG_ATC_INIT_DONE);
} while (factor-- && (val != 1));
if (val != 1) {
BNX2X_ERR("ATC_INIT failed\n");
return -EBUSY;
}
}
bnx2x_init_block(bp, BLOCK_DMAE, PHASE_COMMON);
bnx2x_iov_init_dmae(bp);
/* clean the DMAE memory */
bp->dmae_ready = 1;
bnx2x_init_fill(bp, TSEM_REG_PRAM, 0, 8, 1);
bnx2x_init_block(bp, BLOCK_TCM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_UCM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_CCM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_XCM, PHASE_COMMON);
bnx2x_read_dmae(bp, XSEM_REG_PASSIVE_BUFFER, 3);
bnx2x_read_dmae(bp, CSEM_REG_PASSIVE_BUFFER, 3);
bnx2x_read_dmae(bp, TSEM_REG_PASSIVE_BUFFER, 3);
bnx2x_read_dmae(bp, USEM_REG_PASSIVE_BUFFER, 3);
bnx2x_init_block(bp, BLOCK_QM, PHASE_COMMON);
/* QM queues pointers table */
bnx2x_qm_init_ptr_table(bp, bp->qm_cid_count, INITOP_SET);
/* soft reset pulse */
REG_WR(bp, QM_REG_SOFT_RESET, 1);
REG_WR(bp, QM_REG_SOFT_RESET, 0);
if (CNIC_SUPPORT(bp))
bnx2x_init_block(bp, BLOCK_TM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_DORQ, PHASE_COMMON);
if (!CHIP_REV_IS_SLOW(bp))
/* enable hw interrupt from doorbell Q */
REG_WR(bp, DORQ_REG_DORQ_INT_MASK, 0);
bnx2x_init_block(bp, BLOCK_BRB1, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_PRS, PHASE_COMMON);
REG_WR(bp, PRS_REG_A_PRSU_20, 0xf);
if (!CHIP_IS_E1(bp))
REG_WR(bp, PRS_REG_E1HOV_MODE, bp->path_has_ovlan);
if (!CHIP_IS_E1x(bp) && !CHIP_IS_E3B0(bp)) {
if (IS_MF_AFEX(bp)) {
/* configure that VNTag and VLAN headers must be
* received in afex mode
*/
REG_WR(bp, PRS_REG_HDRS_AFTER_BASIC, 0xE);
REG_WR(bp, PRS_REG_MUST_HAVE_HDRS, 0xA);
REG_WR(bp, PRS_REG_HDRS_AFTER_TAG_0, 0x6);
REG_WR(bp, PRS_REG_TAG_ETHERTYPE_0, 0x8926);
REG_WR(bp, PRS_REG_TAG_LEN_0, 0x4);
} else {
/* Bit-map indicating which L2 hdrs may appear
* after the basic Ethernet header
*/
REG_WR(bp, PRS_REG_HDRS_AFTER_BASIC,
bp->path_has_ovlan ? 7 : 6);
}
}
bnx2x_init_block(bp, BLOCK_TSDM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_CSDM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_USDM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_XSDM, PHASE_COMMON);
if (!CHIP_IS_E1x(bp)) {
/* reset VFC memories */
REG_WR(bp, TSEM_REG_FAST_MEMORY + VFC_REG_MEMORIES_RST,
VFC_MEMORIES_RST_REG_CAM_RST |
VFC_MEMORIES_RST_REG_RAM_RST);
REG_WR(bp, XSEM_REG_FAST_MEMORY + VFC_REG_MEMORIES_RST,
VFC_MEMORIES_RST_REG_CAM_RST |
VFC_MEMORIES_RST_REG_RAM_RST);
msleep(20);
}
bnx2x_init_block(bp, BLOCK_TSEM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_USEM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_CSEM, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_XSEM, PHASE_COMMON);
/* sync semi rtc */
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_CLEAR,
0x80000000);
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_SET,
0x80000000);
bnx2x_init_block(bp, BLOCK_UPB, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_XPB, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_PBF, PHASE_COMMON);
if (!CHIP_IS_E1x(bp)) {
if (IS_MF_AFEX(bp)) {
/* configure that VNTag and VLAN headers must be
* sent in afex mode
*/
REG_WR(bp, PBF_REG_HDRS_AFTER_BASIC, 0xE);
REG_WR(bp, PBF_REG_MUST_HAVE_HDRS, 0xA);
REG_WR(bp, PBF_REG_HDRS_AFTER_TAG_0, 0x6);
REG_WR(bp, PBF_REG_TAG_ETHERTYPE_0, 0x8926);
REG_WR(bp, PBF_REG_TAG_LEN_0, 0x4);
} else {
REG_WR(bp, PBF_REG_HDRS_AFTER_BASIC,
bp->path_has_ovlan ? 7 : 6);
}
}
REG_WR(bp, SRC_REG_SOFT_RST, 1);
bnx2x_init_block(bp, BLOCK_SRC, PHASE_COMMON);
if (CNIC_SUPPORT(bp)) {
REG_WR(bp, SRC_REG_KEYSEARCH_0, 0x63285672);
REG_WR(bp, SRC_REG_KEYSEARCH_1, 0x24b8f2cc);
REG_WR(bp, SRC_REG_KEYSEARCH_2, 0x223aef9b);
REG_WR(bp, SRC_REG_KEYSEARCH_3, 0x26001e3a);
REG_WR(bp, SRC_REG_KEYSEARCH_4, 0x7ae91116);
REG_WR(bp, SRC_REG_KEYSEARCH_5, 0x5ce5230b);
REG_WR(bp, SRC_REG_KEYSEARCH_6, 0x298d8adf);
REG_WR(bp, SRC_REG_KEYSEARCH_7, 0x6eb0ff09);
REG_WR(bp, SRC_REG_KEYSEARCH_8, 0x1830f82f);
REG_WR(bp, SRC_REG_KEYSEARCH_9, 0x01e46be7);
}
REG_WR(bp, SRC_REG_SOFT_RST, 0);
if (sizeof(union cdu_context) != 1024)
/* we currently assume that a context is 1024 bytes */
dev_alert(&bp->pdev->dev,
"please adjust the size of cdu_context(%ld)\n",
(long)sizeof(union cdu_context));
bnx2x_init_block(bp, BLOCK_CDU, PHASE_COMMON);
val = (4 << 24) + (0 << 12) + 1024;
REG_WR(bp, CDU_REG_CDU_GLOBAL_PARAMS, val);
bnx2x_init_block(bp, BLOCK_CFC, PHASE_COMMON);
REG_WR(bp, CFC_REG_INIT_REG, 0x7FF);
/* enable context validation interrupt from CFC */
REG_WR(bp, CFC_REG_CFC_INT_MASK, 0);
/* set the thresholds to prevent CFC/CDU race */
REG_WR(bp, CFC_REG_DEBUG0, 0x20020000);
bnx2x_init_block(bp, BLOCK_HC, PHASE_COMMON);
if (!CHIP_IS_E1x(bp) && BP_NOMCP(bp))
REG_WR(bp, IGU_REG_RESET_MEMORIES, 0x36);
bnx2x_init_block(bp, BLOCK_IGU, PHASE_COMMON);
bnx2x_init_block(bp, BLOCK_MISC_AEU, PHASE_COMMON);
/* Reset PCIE errors for debug */
REG_WR(bp, 0x2814, 0xffffffff);
REG_WR(bp, 0x3820, 0xffffffff);
if (!CHIP_IS_E1x(bp)) {
REG_WR(bp, PCICFG_OFFSET + PXPCS_TL_CONTROL_5,
(PXPCS_TL_CONTROL_5_ERR_UNSPPORT1 |
PXPCS_TL_CONTROL_5_ERR_UNSPPORT));
REG_WR(bp, PCICFG_OFFSET + PXPCS_TL_FUNC345_STAT,
(PXPCS_TL_FUNC345_STAT_ERR_UNSPPORT4 |
PXPCS_TL_FUNC345_STAT_ERR_UNSPPORT3 |
PXPCS_TL_FUNC345_STAT_ERR_UNSPPORT2));
REG_WR(bp, PCICFG_OFFSET + PXPCS_TL_FUNC678_STAT,
(PXPCS_TL_FUNC678_STAT_ERR_UNSPPORT7 |
PXPCS_TL_FUNC678_STAT_ERR_UNSPPORT6 |
PXPCS_TL_FUNC678_STAT_ERR_UNSPPORT5));
}
bnx2x_init_block(bp, BLOCK_NIG, PHASE_COMMON);
if (!CHIP_IS_E1(bp)) {
/* in E3 this done in per-port section */
if (!CHIP_IS_E3(bp))
REG_WR(bp, NIG_REG_LLH_MF_MODE, IS_MF(bp));
}
if (CHIP_IS_E1H(bp))
/* not applicable for E2 (and above ...) */
REG_WR(bp, NIG_REG_LLH_E1HOV_MODE, IS_MF_SD(bp));
if (CHIP_REV_IS_SLOW(bp))
msleep(200);
/* finish CFC init */
val = reg_poll(bp, CFC_REG_LL_INIT_DONE, 1, 100, 10);
if (val != 1) {
BNX2X_ERR("CFC LL_INIT failed\n");
return -EBUSY;
}
val = reg_poll(bp, CFC_REG_AC_INIT_DONE, 1, 100, 10);
if (val != 1) {
BNX2X_ERR("CFC AC_INIT failed\n");
return -EBUSY;
}
val = reg_poll(bp, CFC_REG_CAM_INIT_DONE, 1, 100, 10);
if (val != 1) {
BNX2X_ERR("CFC CAM_INIT failed\n");
return -EBUSY;
}
REG_WR(bp, CFC_REG_DEBUG0, 0);
if (CHIP_IS_E1(bp)) {
/* read NIG statistic
to see if this is our first up since powerup */
bnx2x_read_dmae(bp, NIG_REG_STAT2_BRB_OCTET, 2);
val = *bnx2x_sp(bp, wb_data[0]);
/* do internal memory self test */
if ((val == 0) && bnx2x_int_mem_test(bp)) {
BNX2X_ERR("internal mem self test failed\n");
return -EBUSY;
}
}
bnx2x_setup_fan_failure_detection(bp);
/* clear PXP2 attentions */
REG_RD(bp, PXP2_REG_PXP2_INT_STS_CLR_0);
bnx2x_enable_blocks_attention(bp);
bnx2x_enable_blocks_parity(bp);
if (!BP_NOMCP(bp)) {
if (CHIP_IS_E1x(bp))
bnx2x__common_init_phy(bp);
} else
BNX2X_ERR("Bootcode is missing - can not initialize link\n");
if (SHMEM2_HAS(bp, netproc_fw_ver))
SHMEM2_WR(bp, netproc_fw_ver, REG_RD(bp, XSEM_REG_PRAM));
return 0;
}
/**
* bnx2x_init_hw_common_chip - init HW at the COMMON_CHIP phase.
*
* @bp: driver handle
*/
static int bnx2x_init_hw_common_chip(struct bnx2x *bp)
{
int rc = bnx2x_init_hw_common(bp);
if (rc)
return rc;
/* In E2 2-PORT mode, same ext phy is used for the two paths */
if (!BP_NOMCP(bp))
bnx2x__common_init_phy(bp);
return 0;
}
static int bnx2x_init_hw_port(struct bnx2x *bp)
{
int port = BP_PORT(bp);
int init_phase = port ? PHASE_PORT1 : PHASE_PORT0;
u32 low, high;
u32 val, reg;
DP(NETIF_MSG_HW, "starting port init port %d\n", port);
REG_WR(bp, NIG_REG_MASK_INTERRUPT_PORT0 + port*4, 0);
bnx2x_init_block(bp, BLOCK_MISC, init_phase);
bnx2x_init_block(bp, BLOCK_PXP, init_phase);
bnx2x_init_block(bp, BLOCK_PXP2, init_phase);
/* Timers bug workaround: disables the pf_master bit in pglue at
* common phase, we need to enable it here before any dmae access are
* attempted. Therefore we manually added the enable-master to the
* port phase (it also happens in the function phase)
*/
if (!CHIP_IS_E1x(bp))
REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, 1);
bnx2x_init_block(bp, BLOCK_ATC, init_phase);
bnx2x_init_block(bp, BLOCK_DMAE, init_phase);
bnx2x_init_block(bp, BLOCK_PGLUE_B, init_phase);
bnx2x_init_block(bp, BLOCK_QM, init_phase);
bnx2x_init_block(bp, BLOCK_TCM, init_phase);
bnx2x_init_block(bp, BLOCK_UCM, init_phase);
bnx2x_init_block(bp, BLOCK_CCM, init_phase);
bnx2x_init_block(bp, BLOCK_XCM, init_phase);
/* QM cid (connection) count */
bnx2x_qm_init_cid_count(bp, bp->qm_cid_count, INITOP_SET);
if (CNIC_SUPPORT(bp)) {
bnx2x_init_block(bp, BLOCK_TM, init_phase);
REG_WR(bp, TM_REG_LIN0_SCAN_TIME + port*4, 20);
REG_WR(bp, TM_REG_LIN0_MAX_ACTIVE_CID + port*4, 31);
}
bnx2x_init_block(bp, BLOCK_DORQ, init_phase);
bnx2x_init_block(bp, BLOCK_BRB1, init_phase);
if (CHIP_IS_E1(bp) || CHIP_IS_E1H(bp)) {
if (IS_MF(bp))
low = ((bp->flags & ONE_PORT_FLAG) ? 160 : 246);
else if (bp->dev->mtu > 4096) {
if (bp->flags & ONE_PORT_FLAG)
low = 160;
else {
val = bp->dev->mtu;
/* (24*1024 + val*4)/256 */
low = 96 + (val/64) +
((val % 64) ? 1 : 0);
}
} else
low = ((bp->flags & ONE_PORT_FLAG) ? 80 : 160);
high = low + 56; /* 14*1024/256 */
REG_WR(bp, BRB1_REG_PAUSE_LOW_THRESHOLD_0 + port*4, low);
REG_WR(bp, BRB1_REG_PAUSE_HIGH_THRESHOLD_0 + port*4, high);
}
if (CHIP_MODE_IS_4_PORT(bp))
REG_WR(bp, (BP_PORT(bp) ?
BRB1_REG_MAC_GUARANTIED_1 :
BRB1_REG_MAC_GUARANTIED_0), 40);
bnx2x_init_block(bp, BLOCK_PRS, init_phase);
if (CHIP_IS_E3B0(bp)) {
if (IS_MF_AFEX(bp)) {
/* configure headers for AFEX mode */
REG_WR(bp, BP_PORT(bp) ?
PRS_REG_HDRS_AFTER_BASIC_PORT_1 :
PRS_REG_HDRS_AFTER_BASIC_PORT_0, 0xE);
REG_WR(bp, BP_PORT(bp) ?
PRS_REG_HDRS_AFTER_TAG_0_PORT_1 :
PRS_REG_HDRS_AFTER_TAG_0_PORT_0, 0x6);
REG_WR(bp, BP_PORT(bp) ?
PRS_REG_MUST_HAVE_HDRS_PORT_1 :
PRS_REG_MUST_HAVE_HDRS_PORT_0, 0xA);
} else {
/* Ovlan exists only if we are in multi-function +
* switch-dependent mode, in switch-independent there
* is no ovlan headers
*/
REG_WR(bp, BP_PORT(bp) ?
PRS_REG_HDRS_AFTER_BASIC_PORT_1 :
PRS_REG_HDRS_AFTER_BASIC_PORT_0,
(bp->path_has_ovlan ? 7 : 6));
}
}
bnx2x_init_block(bp, BLOCK_TSDM, init_phase);
bnx2x_init_block(bp, BLOCK_CSDM, init_phase);
bnx2x_init_block(bp, BLOCK_USDM, init_phase);
bnx2x_init_block(bp, BLOCK_XSDM, init_phase);
bnx2x_init_block(bp, BLOCK_TSEM, init_phase);
bnx2x_init_block(bp, BLOCK_USEM, init_phase);
bnx2x_init_block(bp, BLOCK_CSEM, init_phase);
bnx2x_init_block(bp, BLOCK_XSEM, init_phase);
bnx2x_init_block(bp, BLOCK_UPB, init_phase);
bnx2x_init_block(bp, BLOCK_XPB, init_phase);
bnx2x_init_block(bp, BLOCK_PBF, init_phase);
if (CHIP_IS_E1x(bp)) {
/* configure PBF to work without PAUSE mtu 9000 */
REG_WR(bp, PBF_REG_P0_PAUSE_ENABLE + port*4, 0);
/* update threshold */
REG_WR(bp, PBF_REG_P0_ARB_THRSH + port*4, (9040/16));
/* update init credit */
REG_WR(bp, PBF_REG_P0_INIT_CRD + port*4, (9040/16) + 553 - 22);
/* probe changes */
REG_WR(bp, PBF_REG_INIT_P0 + port*4, 1);
udelay(50);
REG_WR(bp, PBF_REG_INIT_P0 + port*4, 0);
}
if (CNIC_SUPPORT(bp))
bnx2x_init_block(bp, BLOCK_SRC, init_phase);
bnx2x_init_block(bp, BLOCK_CDU, init_phase);
bnx2x_init_block(bp, BLOCK_CFC, init_phase);
if (CHIP_IS_E1(bp)) {
REG_WR(bp, HC_REG_LEADING_EDGE_0 + port*8, 0);
REG_WR(bp, HC_REG_TRAILING_EDGE_0 + port*8, 0);
}
bnx2x_init_block(bp, BLOCK_HC, init_phase);
bnx2x_init_block(bp, BLOCK_IGU, init_phase);
bnx2x_init_block(bp, BLOCK_MISC_AEU, init_phase);
/* init aeu_mask_attn_func_0/1:
* - SF mode: bits 3-7 are masked. Only bits 0-2 are in use
* - MF mode: bit 3 is masked. Bits 0-2 are in use as in SF
* bits 4-7 are used for "per vn group attention" */
val = IS_MF(bp) ? 0xF7 : 0x7;
/* Enable DCBX attention for all but E1 */
val |= CHIP_IS_E1(bp) ? 0 : 0x10;
REG_WR(bp, MISC_REG_AEU_MASK_ATTN_FUNC_0 + port*4, val);
/* SCPAD_PARITY should NOT trigger close the gates */
reg = port ? MISC_REG_AEU_ENABLE4_NIG_1 : MISC_REG_AEU_ENABLE4_NIG_0;
REG_WR(bp, reg,
REG_RD(bp, reg) &
~AEU_INPUTS_ATTN_BITS_MCP_LATCHED_SCPAD_PARITY);
reg = port ? MISC_REG_AEU_ENABLE4_PXP_1 : MISC_REG_AEU_ENABLE4_PXP_0;
REG_WR(bp, reg,
REG_RD(bp, reg) &
~AEU_INPUTS_ATTN_BITS_MCP_LATCHED_SCPAD_PARITY);
bnx2x_init_block(bp, BLOCK_NIG, init_phase);
if (!CHIP_IS_E1x(bp)) {
/* Bit-map indicating which L2 hdrs may appear after the
* basic Ethernet header
*/
if (IS_MF_AFEX(bp))
REG_WR(bp, BP_PORT(bp) ?
NIG_REG_P1_HDRS_AFTER_BASIC :
NIG_REG_P0_HDRS_AFTER_BASIC, 0xE);
else
REG_WR(bp, BP_PORT(bp) ?
NIG_REG_P1_HDRS_AFTER_BASIC :
NIG_REG_P0_HDRS_AFTER_BASIC,
IS_MF_SD(bp) ? 7 : 6);
if (CHIP_IS_E3(bp))
REG_WR(bp, BP_PORT(bp) ?
NIG_REG_LLH1_MF_MODE :
NIG_REG_LLH_MF_MODE, IS_MF(bp));
}
if (!CHIP_IS_E3(bp))
REG_WR(bp, NIG_REG_XGXS_SERDES0_MODE_SEL + port*4, 1);
if (!CHIP_IS_E1(bp)) {
/* 0x2 disable mf_ov, 0x1 enable */
REG_WR(bp, NIG_REG_LLH0_BRB1_DRV_MASK_MF + port*4,
(IS_MF_SD(bp) ? 0x1 : 0x2));
if (!CHIP_IS_E1x(bp)) {
val = 0;
switch (bp->mf_mode) {
case MULTI_FUNCTION_SD:
val = 1;
break;
case MULTI_FUNCTION_SI:
case MULTI_FUNCTION_AFEX:
val = 2;
break;
}
REG_WR(bp, (BP_PORT(bp) ? NIG_REG_LLH1_CLS_TYPE :
NIG_REG_LLH0_CLS_TYPE), val);
}
{
REG_WR(bp, NIG_REG_LLFC_ENABLE_0 + port*4, 0);
REG_WR(bp, NIG_REG_LLFC_OUT_EN_0 + port*4, 0);
REG_WR(bp, NIG_REG_PAUSE_ENABLE_0 + port*4, 1);
}
}
/* If SPIO5 is set to generate interrupts, enable it for this port */
val = REG_RD(bp, MISC_REG_SPIO_EVENT_EN);
if (val & MISC_SPIO_SPIO5) {
u32 reg_addr = (port ? MISC_REG_AEU_ENABLE1_FUNC_1_OUT_0 :
MISC_REG_AEU_ENABLE1_FUNC_0_OUT_0);
val = REG_RD(bp, reg_addr);
val |= AEU_INPUTS_ATTN_BITS_SPIO5;
REG_WR(bp, reg_addr, val);
}
return 0;
}
static void bnx2x_ilt_wr(struct bnx2x *bp, u32 index, dma_addr_t addr)
{
int reg;
u32 wb_write[2];
if (CHIP_IS_E1(bp))
reg = PXP2_REG_RQ_ONCHIP_AT + index*8;
else
reg = PXP2_REG_RQ_ONCHIP_AT_B0 + index*8;
wb_write[0] = ONCHIP_ADDR1(addr);
wb_write[1] = ONCHIP_ADDR2(addr);
REG_WR_DMAE(bp, reg, wb_write, 2);
}
void bnx2x_igu_clear_sb_gen(struct bnx2x *bp, u8 func, u8 idu_sb_id, bool is_pf)
{
u32 data, ctl, cnt = 100;
u32 igu_addr_data = IGU_REG_COMMAND_REG_32LSB_DATA;
u32 igu_addr_ctl = IGU_REG_COMMAND_REG_CTRL;
u32 igu_addr_ack = IGU_REG_CSTORM_TYPE_0_SB_CLEANUP + (idu_sb_id/32)*4;
u32 sb_bit = 1 << (idu_sb_id%32);
u32 func_encode = func | (is_pf ? 1 : 0) << IGU_FID_ENCODE_IS_PF_SHIFT;
u32 addr_encode = IGU_CMD_E2_PROD_UPD_BASE + idu_sb_id;
/* Not supported in BC mode */
if (CHIP_INT_MODE_IS_BC(bp))
return;
data = (IGU_USE_REGISTER_cstorm_type_0_sb_cleanup
<< IGU_REGULAR_CLEANUP_TYPE_SHIFT) |
IGU_REGULAR_CLEANUP_SET |
IGU_REGULAR_BCLEANUP;
ctl = addr_encode << IGU_CTRL_REG_ADDRESS_SHIFT |
func_encode << IGU_CTRL_REG_FID_SHIFT |
IGU_CTRL_CMD_TYPE_WR << IGU_CTRL_REG_TYPE_SHIFT;
DP(NETIF_MSG_HW, "write 0x%08x to IGU(via GRC) addr 0x%x\n",
data, igu_addr_data);
REG_WR(bp, igu_addr_data, data);
mmiowb();
barrier();
DP(NETIF_MSG_HW, "write 0x%08x to IGU(via GRC) addr 0x%x\n",
ctl, igu_addr_ctl);
REG_WR(bp, igu_addr_ctl, ctl);
mmiowb();
barrier();
/* wait for clean up to finish */
while (!(REG_RD(bp, igu_addr_ack) & sb_bit) && --cnt)
msleep(20);
if (!(REG_RD(bp, igu_addr_ack) & sb_bit)) {
DP(NETIF_MSG_HW,
"Unable to finish IGU cleanup: idu_sb_id %d offset %d bit %d (cnt %d)\n",
idu_sb_id, idu_sb_id/32, idu_sb_id%32, cnt);
}
}
static void bnx2x_igu_clear_sb(struct bnx2x *bp, u8 idu_sb_id)
{
bnx2x_igu_clear_sb_gen(bp, BP_FUNC(bp), idu_sb_id, true /*PF*/);
}
static void bnx2x_clear_func_ilt(struct bnx2x *bp, u32 func)
{
u32 i, base = FUNC_ILT_BASE(func);
for (i = base; i < base + ILT_PER_FUNC; i++)
bnx2x_ilt_wr(bp, i, 0);
}
static void bnx2x_init_searcher(struct bnx2x *bp)
{
int port = BP_PORT(bp);
bnx2x_src_init_t2(bp, bp->t2, bp->t2_mapping, SRC_CONN_NUM);
/* T1 hash bits value determines the T1 number of entries */
REG_WR(bp, SRC_REG_NUMBER_HASH_BITS0 + port*4, SRC_HASH_BITS);
}
static inline int bnx2x_func_switch_update(struct bnx2x *bp, int suspend)
{
int rc;
struct bnx2x_func_state_params func_params = {NULL};
struct bnx2x_func_switch_update_params *switch_update_params =
&func_params.params.switch_update;
/* Prepare parameters for function state transitions */
__set_bit(RAMROD_COMP_WAIT, &func_params.ramrod_flags);
__set_bit(RAMROD_RETRY, &func_params.ramrod_flags);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_SWITCH_UPDATE;
/* Function parameters */
__set_bit(BNX2X_F_UPDATE_TX_SWITCH_SUSPEND_CHNG,
&switch_update_params->changes);
if (suspend)
__set_bit(BNX2X_F_UPDATE_TX_SWITCH_SUSPEND,
&switch_update_params->changes);
rc = bnx2x_func_state_change(bp, &func_params);
return rc;
}
static int bnx2x_reset_nic_mode(struct bnx2x *bp)
{
int rc, i, port = BP_PORT(bp);
int vlan_en = 0, mac_en[NUM_MACS];
/* Close input from network */
if (bp->mf_mode == SINGLE_FUNCTION) {
bnx2x_set_rx_filter(&bp->link_params, 0);
} else {
vlan_en = REG_RD(bp, port ? NIG_REG_LLH1_FUNC_EN :
NIG_REG_LLH0_FUNC_EN);
REG_WR(bp, port ? NIG_REG_LLH1_FUNC_EN :
NIG_REG_LLH0_FUNC_EN, 0);
for (i = 0; i < NUM_MACS; i++) {
mac_en[i] = REG_RD(bp, port ?
(NIG_REG_LLH1_FUNC_MEM_ENABLE +
4 * i) :
(NIG_REG_LLH0_FUNC_MEM_ENABLE +
4 * i));
REG_WR(bp, port ? (NIG_REG_LLH1_FUNC_MEM_ENABLE +
4 * i) :
(NIG_REG_LLH0_FUNC_MEM_ENABLE + 4 * i), 0);
}
}
/* Close BMC to host */
REG_WR(bp, port ? NIG_REG_P0_TX_MNG_HOST_ENABLE :
NIG_REG_P1_TX_MNG_HOST_ENABLE, 0);
/* Suspend Tx switching to the PF. Completion of this ramrod
* further guarantees that all the packets of that PF / child
* VFs in BRB were processed by the Parser, so it is safe to
* change the NIC_MODE register.
*/
rc = bnx2x_func_switch_update(bp, 1);
if (rc) {
BNX2X_ERR("Can't suspend tx-switching!\n");
return rc;
}
/* Change NIC_MODE register */
REG_WR(bp, PRS_REG_NIC_MODE, 0);
/* Open input from network */
if (bp->mf_mode == SINGLE_FUNCTION) {
bnx2x_set_rx_filter(&bp->link_params, 1);
} else {
REG_WR(bp, port ? NIG_REG_LLH1_FUNC_EN :
NIG_REG_LLH0_FUNC_EN, vlan_en);
for (i = 0; i < NUM_MACS; i++) {
REG_WR(bp, port ? (NIG_REG_LLH1_FUNC_MEM_ENABLE +
4 * i) :
(NIG_REG_LLH0_FUNC_MEM_ENABLE + 4 * i),
mac_en[i]);
}
}
/* Enable BMC to host */
REG_WR(bp, port ? NIG_REG_P0_TX_MNG_HOST_ENABLE :
NIG_REG_P1_TX_MNG_HOST_ENABLE, 1);
/* Resume Tx switching to the PF */
rc = bnx2x_func_switch_update(bp, 0);
if (rc) {
BNX2X_ERR("Can't resume tx-switching!\n");
return rc;
}
DP(NETIF_MSG_IFUP, "NIC MODE disabled\n");
return 0;
}
int bnx2x_init_hw_func_cnic(struct bnx2x *bp)
{
int rc;
bnx2x_ilt_init_op_cnic(bp, INITOP_SET);
if (CONFIGURE_NIC_MODE(bp)) {
/* Configure searcher as part of function hw init */
bnx2x_init_searcher(bp);
/* Reset NIC mode */
rc = bnx2x_reset_nic_mode(bp);
if (rc)
BNX2X_ERR("Can't change NIC mode!\n");
return rc;
}
return 0;
}
/* previous driver DMAE transaction may have occurred when pre-boot stage ended
* and boot began, or when kdump kernel was loaded. Either case would invalidate
* the addresses of the transaction, resulting in was-error bit set in the pci
* causing all hw-to-host pcie transactions to timeout. If this happened we want
* to clear the interrupt which detected this from the pglueb and the was done
* bit
*/
static void bnx2x_clean_pglue_errors(struct bnx2x *bp)
{
if (!CHIP_IS_E1x(bp))
REG_WR(bp, PGLUE_B_REG_WAS_ERROR_PF_7_0_CLR,
1 << BP_ABS_FUNC(bp));
}
static int bnx2x_init_hw_func(struct bnx2x *bp)
{
int port = BP_PORT(bp);
int func = BP_FUNC(bp);
int init_phase = PHASE_PF0 + func;
struct bnx2x_ilt *ilt = BP_ILT(bp);
u16 cdu_ilt_start;
u32 addr, val;
u32 main_mem_base, main_mem_size, main_mem_prty_clr;
int i, main_mem_width, rc;
DP(NETIF_MSG_HW, "starting func init func %d\n", func);
/* FLR cleanup - hmmm */
if (!CHIP_IS_E1x(bp)) {
rc = bnx2x_pf_flr_clnup(bp);
if (rc) {
bnx2x_fw_dump(bp);
return rc;
}
}
/* set MSI reconfigure capability */
if (bp->common.int_block == INT_BLOCK_HC) {
addr = (port ? HC_REG_CONFIG_1 : HC_REG_CONFIG_0);
val = REG_RD(bp, addr);
val |= HC_CONFIG_0_REG_MSI_ATTN_EN_0;
REG_WR(bp, addr, val);
}
bnx2x_init_block(bp, BLOCK_PXP, init_phase);
bnx2x_init_block(bp, BLOCK_PXP2, init_phase);
ilt = BP_ILT(bp);
cdu_ilt_start = ilt->clients[ILT_CLIENT_CDU].start;
if (IS_SRIOV(bp))
cdu_ilt_start += BNX2X_FIRST_VF_CID/ILT_PAGE_CIDS;
cdu_ilt_start = bnx2x_iov_init_ilt(bp, cdu_ilt_start);
/* since BNX2X_FIRST_VF_CID > 0 the PF L2 cids precedes
* those of the VFs, so start line should be reset
*/
cdu_ilt_start = ilt->clients[ILT_CLIENT_CDU].start;
for (i = 0; i < L2_ILT_LINES(bp); i++) {
ilt->lines[cdu_ilt_start + i].page = bp->context[i].vcxt;
ilt->lines[cdu_ilt_start + i].page_mapping =
bp->context[i].cxt_mapping;
ilt->lines[cdu_ilt_start + i].size = bp->context[i].size;
}
bnx2x_ilt_init_op(bp, INITOP_SET);
if (!CONFIGURE_NIC_MODE(bp)) {
bnx2x_init_searcher(bp);
REG_WR(bp, PRS_REG_NIC_MODE, 0);
DP(NETIF_MSG_IFUP, "NIC MODE disabled\n");
} else {
/* Set NIC mode */
REG_WR(bp, PRS_REG_NIC_MODE, 1);
DP(NETIF_MSG_IFUP, "NIC MODE configured\n");
}
if (!CHIP_IS_E1x(bp)) {
u32 pf_conf = IGU_PF_CONF_FUNC_EN;
/* Turn on a single ISR mode in IGU if driver is going to use
* INT#x or MSI
*/
if (!(bp->flags & USING_MSIX_FLAG))
pf_conf |= IGU_PF_CONF_SINGLE_ISR_EN;
/*
* Timers workaround bug: function init part.
* Need to wait 20msec after initializing ILT,
* needed to make sure there are no requests in
* one of the PXP internal queues with "old" ILT addresses
*/
msleep(20);
/*
* Master enable - Due to WB DMAE writes performed before this
* register is re-initialized as part of the regular function
* init
*/
REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, 1);
/* Enable the function in IGU */
REG_WR(bp, IGU_REG_PF_CONFIGURATION, pf_conf);
}
bp->dmae_ready = 1;
bnx2x_init_block(bp, BLOCK_PGLUE_B, init_phase);
bnx2x_clean_pglue_errors(bp);
bnx2x_init_block(bp, BLOCK_ATC, init_phase);
bnx2x_init_block(bp, BLOCK_DMAE, init_phase);
bnx2x_init_block(bp, BLOCK_NIG, init_phase);
bnx2x_init_block(bp, BLOCK_SRC, init_phase);
bnx2x_init_block(bp, BLOCK_MISC, init_phase);
bnx2x_init_block(bp, BLOCK_TCM, init_phase);
bnx2x_init_block(bp, BLOCK_UCM, init_phase);
bnx2x_init_block(bp, BLOCK_CCM, init_phase);
bnx2x_init_block(bp, BLOCK_XCM, init_phase);
bnx2x_init_block(bp, BLOCK_TSEM, init_phase);
bnx2x_init_block(bp, BLOCK_USEM, init_phase);
bnx2x_init_block(bp, BLOCK_CSEM, init_phase);
bnx2x_init_block(bp, BLOCK_XSEM, init_phase);
if (!CHIP_IS_E1x(bp))
REG_WR(bp, QM_REG_PF_EN, 1);
if (!CHIP_IS_E1x(bp)) {
REG_WR(bp, TSEM_REG_VFPF_ERR_NUM, BNX2X_MAX_NUM_OF_VFS + func);
REG_WR(bp, USEM_REG_VFPF_ERR_NUM, BNX2X_MAX_NUM_OF_VFS + func);
REG_WR(bp, CSEM_REG_VFPF_ERR_NUM, BNX2X_MAX_NUM_OF_VFS + func);
REG_WR(bp, XSEM_REG_VFPF_ERR_NUM, BNX2X_MAX_NUM_OF_VFS + func);
}
bnx2x_init_block(bp, BLOCK_QM, init_phase);
bnx2x_init_block(bp, BLOCK_TM, init_phase);
bnx2x_init_block(bp, BLOCK_DORQ, init_phase);
REG_WR(bp, DORQ_REG_MODE_ACT, 1); /* no dpm */
bnx2x_iov_init_dq(bp);
bnx2x_init_block(bp, BLOCK_BRB1, init_phase);
bnx2x_init_block(bp, BLOCK_PRS, init_phase);
bnx2x_init_block(bp, BLOCK_TSDM, init_phase);
bnx2x_init_block(bp, BLOCK_CSDM, init_phase);
bnx2x_init_block(bp, BLOCK_USDM, init_phase);
bnx2x_init_block(bp, BLOCK_XSDM, init_phase);
bnx2x_init_block(bp, BLOCK_UPB, init_phase);
bnx2x_init_block(bp, BLOCK_XPB, init_phase);
bnx2x_init_block(bp, BLOCK_PBF, init_phase);
if (!CHIP_IS_E1x(bp))
REG_WR(bp, PBF_REG_DISABLE_PF, 0);
bnx2x_init_block(bp, BLOCK_CDU, init_phase);
bnx2x_init_block(bp, BLOCK_CFC, init_phase);
if (!CHIP_IS_E1x(bp))
REG_WR(bp, CFC_REG_WEAK_ENABLE_PF, 1);
if (IS_MF(bp)) {
if (!(IS_MF_UFP(bp) && BNX2X_IS_MF_SD_PROTOCOL_FCOE(bp))) {
REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port * 8, 1);
REG_WR(bp, NIG_REG_LLH0_FUNC_VLAN_ID + port * 8,
bp->mf_ov);
}
}
bnx2x_init_block(bp, BLOCK_MISC_AEU, init_phase);
/* HC init per function */
if (bp->common.int_block == INT_BLOCK_HC) {
if (CHIP_IS_E1H(bp)) {
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_12 + func*4, 0);
REG_WR(bp, HC_REG_LEADING_EDGE_0 + port*8, 0);
REG_WR(bp, HC_REG_TRAILING_EDGE_0 + port*8, 0);
}
bnx2x_init_block(bp, BLOCK_HC, init_phase);
} else {
int num_segs, sb_idx, prod_offset;
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_12 + func*4, 0);
if (!CHIP_IS_E1x(bp)) {
REG_WR(bp, IGU_REG_LEADING_EDGE_LATCH, 0);
REG_WR(bp, IGU_REG_TRAILING_EDGE_LATCH, 0);
}
bnx2x_init_block(bp, BLOCK_IGU, init_phase);
if (!CHIP_IS_E1x(bp)) {
int dsb_idx = 0;
/**
* Producer memory:
* E2 mode: address 0-135 match to the mapping memory;
* 136 - PF0 default prod; 137 - PF1 default prod;
* 138 - PF2 default prod; 139 - PF3 default prod;
* 140 - PF0 attn prod; 141 - PF1 attn prod;
* 142 - PF2 attn prod; 143 - PF3 attn prod;
* 144-147 reserved.
*
* E1.5 mode - In backward compatible mode;
* for non default SB; each even line in the memory
* holds the U producer and each odd line hold
* the C producer. The first 128 producers are for
* NDSB (PF0 - 0-31; PF1 - 32-63 and so on). The last 20
* producers are for the DSB for each PF.
* Each PF has five segments: (the order inside each
* segment is PF0; PF1; PF2; PF3) - 128-131 U prods;
* 132-135 C prods; 136-139 X prods; 140-143 T prods;
* 144-147 attn prods;
*/
/* non-default-status-blocks */
num_segs = CHIP_INT_MODE_IS_BC(bp) ?
IGU_BC_NDSB_NUM_SEGS : IGU_NORM_NDSB_NUM_SEGS;
for (sb_idx = 0; sb_idx < bp->igu_sb_cnt; sb_idx++) {
prod_offset = (bp->igu_base_sb + sb_idx) *
num_segs;
for (i = 0; i < num_segs; i++) {
addr = IGU_REG_PROD_CONS_MEMORY +
(prod_offset + i) * 4;
REG_WR(bp, addr, 0);
}
/* send consumer update with value 0 */
bnx2x_ack_sb(bp, bp->igu_base_sb + sb_idx,
USTORM_ID, 0, IGU_INT_NOP, 1);
bnx2x_igu_clear_sb(bp,
bp->igu_base_sb + sb_idx);
}
/* default-status-blocks */
num_segs = CHIP_INT_MODE_IS_BC(bp) ?
IGU_BC_DSB_NUM_SEGS : IGU_NORM_DSB_NUM_SEGS;
if (CHIP_MODE_IS_4_PORT(bp))
dsb_idx = BP_FUNC(bp);
else
dsb_idx = BP_VN(bp);
prod_offset = (CHIP_INT_MODE_IS_BC(bp) ?
IGU_BC_BASE_DSB_PROD + dsb_idx :
IGU_NORM_BASE_DSB_PROD + dsb_idx);
/*
* igu prods come in chunks of E1HVN_MAX (4) -
* does not matters what is the current chip mode
*/
for (i = 0; i < (num_segs * E1HVN_MAX);
i += E1HVN_MAX) {
addr = IGU_REG_PROD_CONS_MEMORY +
(prod_offset + i)*4;
REG_WR(bp, addr, 0);
}
/* send consumer update with 0 */
if (CHIP_INT_MODE_IS_BC(bp)) {
bnx2x_ack_sb(bp, bp->igu_dsb_id,
USTORM_ID, 0, IGU_INT_NOP, 1);
bnx2x_ack_sb(bp, bp->igu_dsb_id,
CSTORM_ID, 0, IGU_INT_NOP, 1);
bnx2x_ack_sb(bp, bp->igu_dsb_id,
XSTORM_ID, 0, IGU_INT_NOP, 1);
bnx2x_ack_sb(bp, bp->igu_dsb_id,
TSTORM_ID, 0, IGU_INT_NOP, 1);
bnx2x_ack_sb(bp, bp->igu_dsb_id,
ATTENTION_ID, 0, IGU_INT_NOP, 1);
} else {
bnx2x_ack_sb(bp, bp->igu_dsb_id,
USTORM_ID, 0, IGU_INT_NOP, 1);
bnx2x_ack_sb(bp, bp->igu_dsb_id,
ATTENTION_ID, 0, IGU_INT_NOP, 1);
}
bnx2x_igu_clear_sb(bp, bp->igu_dsb_id);
/* !!! These should become driver const once
rf-tool supports split-68 const */
REG_WR(bp, IGU_REG_SB_INT_BEFORE_MASK_LSB, 0);
REG_WR(bp, IGU_REG_SB_INT_BEFORE_MASK_MSB, 0);
REG_WR(bp, IGU_REG_SB_MASK_LSB, 0);
REG_WR(bp, IGU_REG_SB_MASK_MSB, 0);
REG_WR(bp, IGU_REG_PBA_STATUS_LSB, 0);
REG_WR(bp, IGU_REG_PBA_STATUS_MSB, 0);
}
}
/* Reset PCIE errors for debug */
REG_WR(bp, 0x2114, 0xffffffff);
REG_WR(bp, 0x2120, 0xffffffff);
if (CHIP_IS_E1x(bp)) {
main_mem_size = HC_REG_MAIN_MEMORY_SIZE / 2; /*dwords*/
main_mem_base = HC_REG_MAIN_MEMORY +
BP_PORT(bp) * (main_mem_size * 4);
main_mem_prty_clr = HC_REG_HC_PRTY_STS_CLR;
main_mem_width = 8;
val = REG_RD(bp, main_mem_prty_clr);
if (val)
DP(NETIF_MSG_HW,
"Hmmm... Parity errors in HC block during function init (0x%x)!\n",
val);
/* Clear "false" parity errors in MSI-X table */
for (i = main_mem_base;
i < main_mem_base + main_mem_size * 4;
i += main_mem_width) {
bnx2x_read_dmae(bp, i, main_mem_width / 4);
bnx2x_write_dmae(bp, bnx2x_sp_mapping(bp, wb_data),
i, main_mem_width / 4);
}
/* Clear HC parity attention */
REG_RD(bp, main_mem_prty_clr);
}
#ifdef BNX2X_STOP_ON_ERROR
/* Enable STORMs SP logging */
REG_WR8(bp, BAR_USTRORM_INTMEM +
USTORM_RECORD_SLOW_PATH_OFFSET(BP_FUNC(bp)), 1);
REG_WR8(bp, BAR_TSTRORM_INTMEM +
TSTORM_RECORD_SLOW_PATH_OFFSET(BP_FUNC(bp)), 1);
REG_WR8(bp, BAR_CSTRORM_INTMEM +
CSTORM_RECORD_SLOW_PATH_OFFSET(BP_FUNC(bp)), 1);
REG_WR8(bp, BAR_XSTRORM_INTMEM +
XSTORM_RECORD_SLOW_PATH_OFFSET(BP_FUNC(bp)), 1);
#endif
bnx2x_phy_probe(&bp->link_params);
return 0;
}
void bnx2x_free_mem_cnic(struct bnx2x *bp)
{
bnx2x_ilt_mem_op_cnic(bp, ILT_MEMOP_FREE);
if (!CHIP_IS_E1x(bp))
BNX2X_PCI_FREE(bp->cnic_sb.e2_sb, bp->cnic_sb_mapping,
sizeof(struct host_hc_status_block_e2));
else
BNX2X_PCI_FREE(bp->cnic_sb.e1x_sb, bp->cnic_sb_mapping,
sizeof(struct host_hc_status_block_e1x));
BNX2X_PCI_FREE(bp->t2, bp->t2_mapping, SRC_T2_SZ);
}
void bnx2x_free_mem(struct bnx2x *bp)
{
int i;
BNX2X_PCI_FREE(bp->fw_stats, bp->fw_stats_mapping,
bp->fw_stats_data_sz + bp->fw_stats_req_sz);
if (IS_VF(bp))
return;
BNX2X_PCI_FREE(bp->def_status_blk, bp->def_status_blk_mapping,
sizeof(struct host_sp_status_block));
BNX2X_PCI_FREE(bp->slowpath, bp->slowpath_mapping,
sizeof(struct bnx2x_slowpath));
for (i = 0; i < L2_ILT_LINES(bp); i++)
BNX2X_PCI_FREE(bp->context[i].vcxt, bp->context[i].cxt_mapping,
bp->context[i].size);
bnx2x_ilt_mem_op(bp, ILT_MEMOP_FREE);
BNX2X_FREE(bp->ilt->lines);
BNX2X_PCI_FREE(bp->spq, bp->spq_mapping, BCM_PAGE_SIZE);
BNX2X_PCI_FREE(bp->eq_ring, bp->eq_mapping,
BCM_PAGE_SIZE * NUM_EQ_PAGES);
BNX2X_PCI_FREE(bp->t2, bp->t2_mapping, SRC_T2_SZ);
bnx2x_iov_free_mem(bp);
}
int bnx2x_alloc_mem_cnic(struct bnx2x *bp)
{
if (!CHIP_IS_E1x(bp)) {
/* size = the status block + ramrod buffers */
bp->cnic_sb.e2_sb = BNX2X_PCI_ALLOC(&bp->cnic_sb_mapping,
sizeof(struct host_hc_status_block_e2));
if (!bp->cnic_sb.e2_sb)
goto alloc_mem_err;
} else {
bp->cnic_sb.e1x_sb = BNX2X_PCI_ALLOC(&bp->cnic_sb_mapping,
sizeof(struct host_hc_status_block_e1x));
if (!bp->cnic_sb.e1x_sb)
goto alloc_mem_err;
}
if (CONFIGURE_NIC_MODE(bp) && !bp->t2) {
/* allocate searcher T2 table, as it wasn't allocated before */
bp->t2 = BNX2X_PCI_ALLOC(&bp->t2_mapping, SRC_T2_SZ);
if (!bp->t2)
goto alloc_mem_err;
}
/* write address to which L5 should insert its values */
bp->cnic_eth_dev.addr_drv_info_to_mcp =
&bp->slowpath->drv_info_to_mcp;
if (bnx2x_ilt_mem_op_cnic(bp, ILT_MEMOP_ALLOC))
goto alloc_mem_err;
return 0;
alloc_mem_err:
bnx2x_free_mem_cnic(bp);
BNX2X_ERR("Can't allocate memory\n");
return -ENOMEM;
}
int bnx2x_alloc_mem(struct bnx2x *bp)
{
int i, allocated, context_size;
if (!CONFIGURE_NIC_MODE(bp) && !bp->t2) {
/* allocate searcher T2 table */
bp->t2 = BNX2X_PCI_ALLOC(&bp->t2_mapping, SRC_T2_SZ);
if (!bp->t2)
goto alloc_mem_err;
}
bp->def_status_blk = BNX2X_PCI_ALLOC(&bp->def_status_blk_mapping,
sizeof(struct host_sp_status_block));
if (!bp->def_status_blk)
goto alloc_mem_err;
bp->slowpath = BNX2X_PCI_ALLOC(&bp->slowpath_mapping,
sizeof(struct bnx2x_slowpath));
if (!bp->slowpath)
goto alloc_mem_err;
/* Allocate memory for CDU context:
* This memory is allocated separately and not in the generic ILT
* functions because CDU differs in few aspects:
* 1. There are multiple entities allocating memory for context -
* 'regular' driver, CNIC and SRIOV driver. Each separately controls
* its own ILT lines.
* 2. Since CDU page-size is not a single 4KB page (which is the case
* for the other ILT clients), to be efficient we want to support
* allocation of sub-page-size in the last entry.
* 3. Context pointers are used by the driver to pass to FW / update
* the context (for the other ILT clients the pointers are used just to
* free the memory during unload).
*/
context_size = sizeof(union cdu_context) * BNX2X_L2_CID_COUNT(bp);
for (i = 0, allocated = 0; allocated < context_size; i++) {
bp->context[i].size = min(CDU_ILT_PAGE_SZ,
(context_size - allocated));
bp->context[i].vcxt = BNX2X_PCI_ALLOC(&bp->context[i].cxt_mapping,
bp->context[i].size);
if (!bp->context[i].vcxt)
goto alloc_mem_err;
allocated += bp->context[i].size;
}
bp->ilt->lines = kcalloc(ILT_MAX_LINES, sizeof(struct ilt_line),
GFP_KERNEL);
if (!bp->ilt->lines)
goto alloc_mem_err;
if (bnx2x_ilt_mem_op(bp, ILT_MEMOP_ALLOC))
goto alloc_mem_err;
if (bnx2x_iov_alloc_mem(bp))
goto alloc_mem_err;
/* Slow path ring */
bp->spq = BNX2X_PCI_ALLOC(&bp->spq_mapping, BCM_PAGE_SIZE);
if (!bp->spq)
goto alloc_mem_err;
/* EQ */
bp->eq_ring = BNX2X_PCI_ALLOC(&bp->eq_mapping,
BCM_PAGE_SIZE * NUM_EQ_PAGES);
if (!bp->eq_ring)
goto alloc_mem_err;
return 0;
alloc_mem_err:
bnx2x_free_mem(bp);
BNX2X_ERR("Can't allocate memory\n");
return -ENOMEM;
}
/*
* Init service functions
*/
int bnx2x_set_mac_one(struct bnx2x *bp, u8 *mac,
struct bnx2x_vlan_mac_obj *obj, bool set,
int mac_type, unsigned long *ramrod_flags)
{
int rc;
struct bnx2x_vlan_mac_ramrod_params ramrod_param;
memset(&ramrod_param, 0, sizeof(ramrod_param));
/* Fill general parameters */
ramrod_param.vlan_mac_obj = obj;
ramrod_param.ramrod_flags = *ramrod_flags;
/* Fill a user request section if needed */
if (!test_bit(RAMROD_CONT, ramrod_flags)) {
memcpy(ramrod_param.user_req.u.mac.mac, mac, ETH_ALEN);
__set_bit(mac_type, &ramrod_param.user_req.vlan_mac_flags);
/* Set the command: ADD or DEL */
if (set)
ramrod_param.user_req.cmd = BNX2X_VLAN_MAC_ADD;
else
ramrod_param.user_req.cmd = BNX2X_VLAN_MAC_DEL;
}
rc = bnx2x_config_vlan_mac(bp, &ramrod_param);
if (rc == -EEXIST) {
DP(BNX2X_MSG_SP, "Failed to schedule ADD operations: %d\n", rc);
/* do not treat adding same MAC as error */
rc = 0;
} else if (rc < 0)
BNX2X_ERR("%s MAC failed\n", (set ? "Set" : "Del"));
return rc;
}
int bnx2x_set_vlan_one(struct bnx2x *bp, u16 vlan,
struct bnx2x_vlan_mac_obj *obj, bool set,
unsigned long *ramrod_flags)
{
int rc;
struct bnx2x_vlan_mac_ramrod_params ramrod_param;
memset(&ramrod_param, 0, sizeof(ramrod_param));
/* Fill general parameters */
ramrod_param.vlan_mac_obj = obj;
ramrod_param.ramrod_flags = *ramrod_flags;
/* Fill a user request section if needed */
if (!test_bit(RAMROD_CONT, ramrod_flags)) {
ramrod_param.user_req.u.vlan.vlan = vlan;
/* Set the command: ADD or DEL */
if (set)
ramrod_param.user_req.cmd = BNX2X_VLAN_MAC_ADD;
else
ramrod_param.user_req.cmd = BNX2X_VLAN_MAC_DEL;
}
rc = bnx2x_config_vlan_mac(bp, &ramrod_param);
if (rc == -EEXIST) {
/* Do not treat adding same vlan as error. */
DP(BNX2X_MSG_SP, "Failed to schedule ADD operations: %d\n", rc);
rc = 0;
} else if (rc < 0) {
BNX2X_ERR("%s VLAN failed\n", (set ? "Set" : "Del"));
}
return rc;
}
int bnx2x_del_all_macs(struct bnx2x *bp,
struct bnx2x_vlan_mac_obj *mac_obj,
int mac_type, bool wait_for_comp)
{
int rc;
unsigned long ramrod_flags = 0, vlan_mac_flags = 0;
/* Wait for completion of requested */
if (wait_for_comp)
__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
/* Set the mac type of addresses we want to clear */
__set_bit(mac_type, &vlan_mac_flags);
rc = mac_obj->delete_all(bp, mac_obj, &vlan_mac_flags, &ramrod_flags);
if (rc < 0)
BNX2X_ERR("Failed to delete MACs: %d\n", rc);
return rc;
}
int bnx2x_set_eth_mac(struct bnx2x *bp, bool set)
{
if (IS_PF(bp)) {
unsigned long ramrod_flags = 0;
DP(NETIF_MSG_IFUP, "Adding Eth MAC\n");
__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
return bnx2x_set_mac_one(bp, bp->dev->dev_addr,
&bp->sp_objs->mac_obj, set,
BNX2X_ETH_MAC, &ramrod_flags);
} else { /* vf */
return bnx2x_vfpf_config_mac(bp, bp->dev->dev_addr,
bp->fp->index, set);
}
}
int bnx2x_setup_leading(struct bnx2x *bp)
{
if (IS_PF(bp))
return bnx2x_setup_queue(bp, &bp->fp[0], true);
else /* VF */
return bnx2x_vfpf_setup_q(bp, &bp->fp[0], true);
}
/**
* bnx2x_set_int_mode - configure interrupt mode
*
* @bp: driver handle
*
* In case of MSI-X it will also try to enable MSI-X.
*/
int bnx2x_set_int_mode(struct bnx2x *bp)
{
int rc = 0;
if (IS_VF(bp) && int_mode != BNX2X_INT_MODE_MSIX) {
BNX2X_ERR("VF not loaded since interrupt mode not msix\n");
return -EINVAL;
}
switch (int_mode) {
case BNX2X_INT_MODE_MSIX:
/* attempt to enable msix */
rc = bnx2x_enable_msix(bp);
/* msix attained */
if (!rc)
return 0;
/* vfs use only msix */
if (rc && IS_VF(bp))
return rc;
/* failed to enable multiple MSI-X */
BNX2X_DEV_INFO("Failed to enable multiple MSI-X (%d), set number of queues to %d\n",
bp->num_queues,
1 + bp->num_cnic_queues);
/* falling through... */
case BNX2X_INT_MODE_MSI:
bnx2x_enable_msi(bp);
/* falling through... */
case BNX2X_INT_MODE_INTX:
bp->num_ethernet_queues = 1;
bp->num_queues = bp->num_ethernet_queues + bp->num_cnic_queues;
BNX2X_DEV_INFO("set number of queues to 1\n");
break;
default:
BNX2X_DEV_INFO("unknown value in int_mode module parameter\n");
return -EINVAL;
}
return 0;
}
/* must be called prior to any HW initializations */
static inline u16 bnx2x_cid_ilt_lines(struct bnx2x *bp)
{
if (IS_SRIOV(bp))
return (BNX2X_FIRST_VF_CID + BNX2X_VF_CIDS)/ILT_PAGE_CIDS;
return L2_ILT_LINES(bp);
}
void bnx2x_ilt_set_info(struct bnx2x *bp)
{
struct ilt_client_info *ilt_client;
struct bnx2x_ilt *ilt = BP_ILT(bp);
u16 line = 0;
ilt->start_line = FUNC_ILT_BASE(BP_FUNC(bp));
DP(BNX2X_MSG_SP, "ilt starts at line %d\n", ilt->start_line);
/* CDU */
ilt_client = &ilt->clients[ILT_CLIENT_CDU];
ilt_client->client_num = ILT_CLIENT_CDU;
ilt_client->page_size = CDU_ILT_PAGE_SZ;
ilt_client->flags = ILT_CLIENT_SKIP_MEM;
ilt_client->start = line;
line += bnx2x_cid_ilt_lines(bp);
if (CNIC_SUPPORT(bp))
line += CNIC_ILT_LINES;
ilt_client->end = line - 1;
DP(NETIF_MSG_IFUP, "ilt client[CDU]: start %d, end %d, psz 0x%x, flags 0x%x, hw psz %d\n",
ilt_client->start,
ilt_client->end,
ilt_client->page_size,
ilt_client->flags,
ilog2(ilt_client->page_size >> 12));
/* QM */
if (QM_INIT(bp->qm_cid_count)) {
ilt_client = &ilt->clients[ILT_CLIENT_QM];
ilt_client->client_num = ILT_CLIENT_QM;
ilt_client->page_size = QM_ILT_PAGE_SZ;
ilt_client->flags = 0;
ilt_client->start = line;
/* 4 bytes for each cid */
line += DIV_ROUND_UP(bp->qm_cid_count * QM_QUEUES_PER_FUNC * 4,
QM_ILT_PAGE_SZ);
ilt_client->end = line - 1;
DP(NETIF_MSG_IFUP,
"ilt client[QM]: start %d, end %d, psz 0x%x, flags 0x%x, hw psz %d\n",
ilt_client->start,
ilt_client->end,
ilt_client->page_size,
ilt_client->flags,
ilog2(ilt_client->page_size >> 12));
}
if (CNIC_SUPPORT(bp)) {
/* SRC */
ilt_client = &ilt->clients[ILT_CLIENT_SRC];
ilt_client->client_num = ILT_CLIENT_SRC;
ilt_client->page_size = SRC_ILT_PAGE_SZ;
ilt_client->flags = 0;
ilt_client->start = line;
line += SRC_ILT_LINES;
ilt_client->end = line - 1;
DP(NETIF_MSG_IFUP,
"ilt client[SRC]: start %d, end %d, psz 0x%x, flags 0x%x, hw psz %d\n",
ilt_client->start,
ilt_client->end,
ilt_client->page_size,
ilt_client->flags,
ilog2(ilt_client->page_size >> 12));
/* TM */
ilt_client = &ilt->clients[ILT_CLIENT_TM];
ilt_client->client_num = ILT_CLIENT_TM;
ilt_client->page_size = TM_ILT_PAGE_SZ;
ilt_client->flags = 0;
ilt_client->start = line;
line += TM_ILT_LINES;
ilt_client->end = line - 1;
DP(NETIF_MSG_IFUP,
"ilt client[TM]: start %d, end %d, psz 0x%x, flags 0x%x, hw psz %d\n",
ilt_client->start,
ilt_client->end,
ilt_client->page_size,
ilt_client->flags,
ilog2(ilt_client->page_size >> 12));
}
BUG_ON(line > ILT_MAX_LINES);
}
/**
* bnx2x_pf_q_prep_init - prepare INIT transition parameters
*
* @bp: driver handle
* @fp: pointer to fastpath
* @init_params: pointer to parameters structure
*
* parameters configured:
* - HC configuration
* - Queue's CDU context
*/
static void bnx2x_pf_q_prep_init(struct bnx2x *bp,
struct bnx2x_fastpath *fp, struct bnx2x_queue_init_params *init_params)
{
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 cos;
int cxt_index, cxt_offset;
/* FCoE Queue uses Default SB, thus has no HC capabilities */
if (!IS_FCOE_FP(fp)) {
__set_bit(BNX2X_Q_FLG_HC, &init_params->rx.flags);
__set_bit(BNX2X_Q_FLG_HC, &init_params->tx.flags);
/* If HC is supported, enable host coalescing in the transition
* to INIT state.
*/
__set_bit(BNX2X_Q_FLG_HC_EN, &init_params->rx.flags);
__set_bit(BNX2X_Q_FLG_HC_EN, &init_params->tx.flags);
/* HC rate */
init_params->rx.hc_rate = bp->rx_ticks ?
(1000000 / bp->rx_ticks) : 0;
init_params->tx.hc_rate = bp->tx_ticks ?
(1000000 / bp->tx_ticks) : 0;
/* FW SB ID */
init_params->rx.fw_sb_id = init_params->tx.fw_sb_id =
fp->fw_sb_id;
/*
* CQ index among the SB indices: FCoE clients uses the default
* SB, therefore it's different.
*/
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
init_params->rx.sb_cq_index = HC_INDEX_ETH_RX_CQ_CONS;
init_params->tx.sb_cq_index = HC_INDEX_ETH_FIRST_TX_CQ_CONS;
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* set maximum number of COSs supported by this queue */
init_params->max_cos = fp->max_cos;
DP(NETIF_MSG_IFUP, "fp: %d setting queue params max cos to: %d\n",
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
fp->index, init_params->max_cos);
/* set the context pointers queue object */
for (cos = FIRST_TX_COS_INDEX; cos < init_params->max_cos; cos++) {
cxt_index = fp->txdata_ptr[cos]->cid / ILT_PAGE_CIDS;
cxt_offset = fp->txdata_ptr[cos]->cid - (cxt_index *
ILT_PAGE_CIDS);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
init_params->cxts[cos] =
&bp->context[cxt_index].vcxt[cxt_offset].eth;
}
}
static int bnx2x_setup_tx_only(struct bnx2x *bp, struct bnx2x_fastpath *fp,
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
struct bnx2x_queue_state_params *q_params,
struct bnx2x_queue_setup_tx_only_params *tx_only_params,
int tx_index, bool leading)
{
memset(tx_only_params, 0, sizeof(*tx_only_params));
/* Set the command */
q_params->cmd = BNX2X_Q_CMD_SETUP_TX_ONLY;
/* Set tx-only QUEUE flags: don't zero statistics */
tx_only_params->flags = bnx2x_get_common_flags(bp, fp, false);
/* choose the index of the cid to send the slow path on */
tx_only_params->cid_index = tx_index;
/* Set general TX_ONLY_SETUP parameters */
bnx2x_pf_q_prep_general(bp, fp, &tx_only_params->gen_params, tx_index);
/* Set Tx TX_ONLY_SETUP parameters */
bnx2x_pf_tx_q_prep(bp, fp, &tx_only_params->txq_params, tx_index);
DP(NETIF_MSG_IFUP,
"preparing to send tx-only ramrod for connection: cos %d, primary cid %d, cid %d, client id %d, sp-client id %d, flags %lx\n",
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
tx_index, q_params->q_obj->cids[FIRST_TX_COS_INDEX],
q_params->q_obj->cids[tx_index], q_params->q_obj->cl_id,
tx_only_params->gen_params.spcl_id, tx_only_params->flags);
/* send the ramrod */
return bnx2x_queue_state_change(bp, q_params);
}
/**
* bnx2x_setup_queue - setup queue
*
* @bp: driver handle
* @fp: pointer to fastpath
* @leading: is leading
*
* This function performs 2 steps in a Queue state machine
* actually: 1) RESET->INIT 2) INIT->SETUP
*/
int bnx2x_setup_queue(struct bnx2x *bp, struct bnx2x_fastpath *fp,
bool leading)
{
struct bnx2x_queue_state_params q_params = {NULL};
struct bnx2x_queue_setup_params *setup_params =
&q_params.params.setup;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
struct bnx2x_queue_setup_tx_only_params *tx_only_params =
&q_params.params.tx_only;
int rc;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 tx_index;
DP(NETIF_MSG_IFUP, "setting up queue %d\n", fp->index);
/* reset IGU state skip FCoE L2 queue */
if (!IS_FCOE_FP(fp))
bnx2x_ack_sb(bp, fp->igu_sb_id, USTORM_ID, 0,
IGU_INT_ENABLE, 0);
q_params.q_obj = &bnx2x_sp_obj(bp, fp).q_obj;
/* We want to wait for completion in this context */
__set_bit(RAMROD_COMP_WAIT, &q_params.ramrod_flags);
/* Prepare the INIT parameters */
bnx2x_pf_q_prep_init(bp, fp, &q_params.params.init);
/* Set the command */
q_params.cmd = BNX2X_Q_CMD_INIT;
/* Change the state to INIT */
rc = bnx2x_queue_state_change(bp, &q_params);
if (rc) {
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
BNX2X_ERR("Queue(%d) INIT failed\n", fp->index);
return rc;
}
DP(NETIF_MSG_IFUP, "init complete\n");
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* Now move the Queue to the SETUP state... */
memset(setup_params, 0, sizeof(*setup_params));
/* Set QUEUE flags */
setup_params->flags = bnx2x_get_q_flags(bp, fp, leading);
/* Set general SETUP parameters */
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bnx2x_pf_q_prep_general(bp, fp, &setup_params->gen_params,
FIRST_TX_COS_INDEX);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bnx2x_pf_rx_q_prep(bp, fp, &setup_params->pause_params,
&setup_params->rxq_params);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bnx2x_pf_tx_q_prep(bp, fp, &setup_params->txq_params,
FIRST_TX_COS_INDEX);
/* Set the command */
q_params.cmd = BNX2X_Q_CMD_SETUP;
if (IS_FCOE_FP(fp))
bp->fcoe_init = true;
/* Change the state to SETUP */
rc = bnx2x_queue_state_change(bp, &q_params);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
if (rc) {
BNX2X_ERR("Queue(%d) SETUP failed\n", fp->index);
return rc;
}
/* loop through the relevant tx-only indices */
for (tx_index = FIRST_TX_ONLY_COS_INDEX;
tx_index < fp->max_cos;
tx_index++) {
/* prepare and send tx-only ramrod*/
rc = bnx2x_setup_tx_only(bp, fp, &q_params,
tx_only_params, tx_index, leading);
if (rc) {
BNX2X_ERR("Queue(%d.%d) TX_ONLY_SETUP failed\n",
fp->index, tx_index);
return rc;
}
}
return rc;
}
static int bnx2x_stop_queue(struct bnx2x *bp, int index)
{
struct bnx2x_fastpath *fp = &bp->fp[index];
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
struct bnx2x_fp_txdata *txdata;
struct bnx2x_queue_state_params q_params = {NULL};
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
int rc, tx_index;
DP(NETIF_MSG_IFDOWN, "stopping queue %d cid %d\n", index, fp->cid);
q_params.q_obj = &bnx2x_sp_obj(bp, fp).q_obj;
/* We want to wait for completion in this context */
__set_bit(RAMROD_COMP_WAIT, &q_params.ramrod_flags);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* close tx-only connections */
for (tx_index = FIRST_TX_ONLY_COS_INDEX;
tx_index < fp->max_cos;
tx_index++){
/* ascertain this is a normal queue*/
txdata = fp->txdata_ptr[tx_index];
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
DP(NETIF_MSG_IFDOWN, "stopping tx-only queue %d\n",
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
txdata->txq_index);
/* send halt terminate on tx-only connection */
q_params.cmd = BNX2X_Q_CMD_TERMINATE;
memset(&q_params.params.terminate, 0,
sizeof(q_params.params.terminate));
q_params.params.terminate.cid_index = tx_index;
rc = bnx2x_queue_state_change(bp, &q_params);
if (rc)
return rc;
/* send halt terminate on tx-only connection */
q_params.cmd = BNX2X_Q_CMD_CFC_DEL;
memset(&q_params.params.cfc_del, 0,
sizeof(q_params.params.cfc_del));
q_params.params.cfc_del.cid_index = tx_index;
rc = bnx2x_queue_state_change(bp, &q_params);
if (rc)
return rc;
}
/* Stop the primary connection: */
/* ...halt the connection */
q_params.cmd = BNX2X_Q_CMD_HALT;
rc = bnx2x_queue_state_change(bp, &q_params);
if (rc)
return rc;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* ...terminate the connection */
q_params.cmd = BNX2X_Q_CMD_TERMINATE;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
memset(&q_params.params.terminate, 0,
sizeof(q_params.params.terminate));
q_params.params.terminate.cid_index = FIRST_TX_COS_INDEX;
rc = bnx2x_queue_state_change(bp, &q_params);
if (rc)
return rc;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* ...delete cfc entry */
q_params.cmd = BNX2X_Q_CMD_CFC_DEL;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
memset(&q_params.params.cfc_del, 0,
sizeof(q_params.params.cfc_del));
q_params.params.cfc_del.cid_index = FIRST_TX_COS_INDEX;
return bnx2x_queue_state_change(bp, &q_params);
}
static void bnx2x_reset_func(struct bnx2x *bp)
{
int port = BP_PORT(bp);
int func = BP_FUNC(bp);
int i;
/* Disable the function in the FW */
REG_WR8(bp, BAR_XSTRORM_INTMEM + XSTORM_FUNC_EN_OFFSET(func), 0);
REG_WR8(bp, BAR_CSTRORM_INTMEM + CSTORM_FUNC_EN_OFFSET(func), 0);
REG_WR8(bp, BAR_TSTRORM_INTMEM + TSTORM_FUNC_EN_OFFSET(func), 0);
REG_WR8(bp, BAR_USTRORM_INTMEM + USTORM_FUNC_EN_OFFSET(func), 0);
/* FP SBs */
for_each_eth_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
REG_WR8(bp, BAR_CSTRORM_INTMEM +
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
CSTORM_STATUS_BLOCK_DATA_STATE_OFFSET(fp->fw_sb_id),
SB_DISABLED);
}
if (CNIC_LOADED(bp))
/* CNIC SB */
REG_WR8(bp, BAR_CSTRORM_INTMEM +
CSTORM_STATUS_BLOCK_DATA_STATE_OFFSET
(bnx2x_cnic_fw_sb_id(bp)), SB_DISABLED);
/* SP SB */
REG_WR8(bp, BAR_CSTRORM_INTMEM +
CSTORM_SP_STATUS_BLOCK_DATA_STATE_OFFSET(func),
SB_DISABLED);
for (i = 0; i < XSTORM_SPQ_DATA_SIZE / 4; i++)
REG_WR(bp, BAR_XSTRORM_INTMEM + XSTORM_SPQ_DATA_OFFSET(func),
0);
/* Configure IGU */
if (bp->common.int_block == INT_BLOCK_HC) {
REG_WR(bp, HC_REG_LEADING_EDGE_0 + port*8, 0);
REG_WR(bp, HC_REG_TRAILING_EDGE_0 + port*8, 0);
} else {
REG_WR(bp, IGU_REG_LEADING_EDGE_LATCH, 0);
REG_WR(bp, IGU_REG_TRAILING_EDGE_LATCH, 0);
}
if (CNIC_LOADED(bp)) {
/* Disable Timer scan */
REG_WR(bp, TM_REG_EN_LINEAR0_TIMER + port*4, 0);
/*
* Wait for at least 10ms and up to 2 second for the timers
* scan to complete
*/
for (i = 0; i < 200; i++) {
usleep_range(10000, 20000);
if (!REG_RD(bp, TM_REG_LIN0_SCAN_ON + port*4))
break;
}
}
/* Clear ILT */
bnx2x_clear_func_ilt(bp, func);
/* Timers workaround bug for E2: if this is vnic-3,
* we need to set the entire ilt range for this timers.
*/
if (!CHIP_IS_E1x(bp) && BP_VN(bp) == 3) {
struct ilt_client_info ilt_cli;
/* use dummy TM client */
memset(&ilt_cli, 0, sizeof(struct ilt_client_info));
ilt_cli.start = 0;
ilt_cli.end = ILT_NUM_PAGE_ENTRIES - 1;
ilt_cli.client_num = ILT_CLIENT_TM;
bnx2x_ilt_boundry_init_op(bp, &ilt_cli, 0, INITOP_CLEAR);
}
/* this assumes that reset_port() called before reset_func()*/
if (!CHIP_IS_E1x(bp))
bnx2x_pf_disable(bp);
bp->dmae_ready = 0;
}
static void bnx2x_reset_port(struct bnx2x *bp)
{
int port = BP_PORT(bp);
u32 val;
/* Reset physical Link */
bnx2x__link_reset(bp);
REG_WR(bp, NIG_REG_MASK_INTERRUPT_PORT0 + port*4, 0);
/* Do not rcv packets to BRB */
REG_WR(bp, NIG_REG_LLH0_BRB1_DRV_MASK + port*4, 0x0);
/* Do not direct rcv packets that are not for MCP to the BRB */
REG_WR(bp, (port ? NIG_REG_LLH1_BRB1_NOT_MCP :
NIG_REG_LLH0_BRB1_NOT_MCP), 0x0);
/* Configure AEU */
REG_WR(bp, MISC_REG_AEU_MASK_ATTN_FUNC_0 + port*4, 0);
msleep(100);
/* Check for BRB port occupancy */
val = REG_RD(bp, BRB1_REG_PORT_NUM_OCC_BLOCKS_0 + port*4);
if (val)
DP(NETIF_MSG_IFDOWN,
"BRB1 is not empty %d blocks are occupied\n", val);
/* TODO: Close Doorbell port? */
}
static int bnx2x_reset_hw(struct bnx2x *bp, u32 load_code)
{
struct bnx2x_func_state_params func_params = {NULL};
/* Prepare parameters for function state transitions */
__set_bit(RAMROD_COMP_WAIT, &func_params.ramrod_flags);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_HW_RESET;
func_params.params.hw_init.load_phase = load_code;
return bnx2x_func_state_change(bp, &func_params);
}
static int bnx2x_func_stop(struct bnx2x *bp)
{
struct bnx2x_func_state_params func_params = {NULL};
int rc;
/* Prepare parameters for function state transitions */
__set_bit(RAMROD_COMP_WAIT, &func_params.ramrod_flags);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_STOP;
/*
* Try to stop the function the 'good way'. If fails (in case
* of a parity error during bnx2x_chip_cleanup()) and we are
* not in a debug mode, perform a state transaction in order to
* enable further HW_RESET transaction.
*/
rc = bnx2x_func_state_change(bp, &func_params);
if (rc) {
#ifdef BNX2X_STOP_ON_ERROR
return rc;
#else
BNX2X_ERR("FUNC_STOP ramrod failed. Running a dry transaction\n");
__set_bit(RAMROD_DRV_CLR_ONLY, &func_params.ramrod_flags);
return bnx2x_func_state_change(bp, &func_params);
#endif
}
return 0;
}
/**
* bnx2x_send_unload_req - request unload mode from the MCP.
*
* @bp: driver handle
* @unload_mode: requested function's unload mode
*
* Return unload mode returned by the MCP: COMMON, PORT or FUNC.
*/
u32 bnx2x_send_unload_req(struct bnx2x *bp, int unload_mode)
{
u32 reset_code = 0;
int port = BP_PORT(bp);
/* Select the UNLOAD request mode */
if (unload_mode == UNLOAD_NORMAL)
reset_code = DRV_MSG_CODE_UNLOAD_REQ_WOL_DIS;
else if (bp->flags & NO_WOL_FLAG)
reset_code = DRV_MSG_CODE_UNLOAD_REQ_WOL_MCP;
else if (bp->wol) {
u32 emac_base = port ? GRCBASE_EMAC1 : GRCBASE_EMAC0;
u8 *mac_addr = bp->dev->dev_addr;
struct pci_dev *pdev = bp->pdev;
u32 val;
u16 pmc;
/* The mac address is written to entries 1-4 to
* preserve entry 0 which is used by the PMF
*/
u8 entry = (BP_VN(bp) + 1)*8;
val = (mac_addr[0] << 8) | mac_addr[1];
EMAC_WR(bp, EMAC_REG_EMAC_MAC_MATCH + entry, val);
val = (mac_addr[2] << 24) | (mac_addr[3] << 16) |
(mac_addr[4] << 8) | mac_addr[5];
EMAC_WR(bp, EMAC_REG_EMAC_MAC_MATCH + entry + 4, val);
/* Enable the PME and clear the status */
pci_read_config_word(pdev, pdev->pm_cap + PCI_PM_CTRL, &pmc);
pmc |= PCI_PM_CTRL_PME_ENABLE | PCI_PM_CTRL_PME_STATUS;
pci_write_config_word(pdev, pdev->pm_cap + PCI_PM_CTRL, pmc);
reset_code = DRV_MSG_CODE_UNLOAD_REQ_WOL_EN;
} else
reset_code = DRV_MSG_CODE_UNLOAD_REQ_WOL_DIS;
/* Send the request to the MCP */
if (!BP_NOMCP(bp))
reset_code = bnx2x_fw_command(bp, reset_code, 0);
else {
int path = BP_PATH(bp);
DP(NETIF_MSG_IFDOWN, "NO MCP - load counts[%d] %d, %d, %d\n",
path, bnx2x_load_count[path][0], bnx2x_load_count[path][1],
bnx2x_load_count[path][2]);
bnx2x_load_count[path][0]--;
bnx2x_load_count[path][1 + port]--;
DP(NETIF_MSG_IFDOWN, "NO MCP - new load counts[%d] %d, %d, %d\n",
path, bnx2x_load_count[path][0], bnx2x_load_count[path][1],
bnx2x_load_count[path][2]);
if (bnx2x_load_count[path][0] == 0)
reset_code = FW_MSG_CODE_DRV_UNLOAD_COMMON;
else if (bnx2x_load_count[path][1 + port] == 0)
reset_code = FW_MSG_CODE_DRV_UNLOAD_PORT;
else
reset_code = FW_MSG_CODE_DRV_UNLOAD_FUNCTION;
}
return reset_code;
}
/**
* bnx2x_send_unload_done - send UNLOAD_DONE command to the MCP.
*
* @bp: driver handle
* @keep_link: true iff link should be kept up
*/
void bnx2x_send_unload_done(struct bnx2x *bp, bool keep_link)
{
u32 reset_param = keep_link ? DRV_MSG_CODE_UNLOAD_SKIP_LINK_RESET : 0;
/* Report UNLOAD_DONE to MCP */
if (!BP_NOMCP(bp))
bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_DONE, reset_param);
}
static int bnx2x_func_wait_started(struct bnx2x *bp)
{
int tout = 50;
int msix = (bp->flags & USING_MSIX_FLAG) ? 1 : 0;
if (!bp->port.pmf)
return 0;
/*
* (assumption: No Attention from MCP at this stage)
* PMF probably in the middle of TX disable/enable transaction
* 1. Sync IRS for default SB
* 2. Sync SP queue - this guarantees us that attention handling started
* 3. Wait, that TX disable/enable transaction completes
*
* 1+2 guarantee that if DCBx attention was scheduled it already changed
* pending bit of transaction from STARTED-->TX_STOPPED, if we already
* received completion for the transaction the state is TX_STOPPED.
* State will return to STARTED after completion of TX_STOPPED-->STARTED
* transaction.
*/
/* make sure default SB ISR is done */
if (msix)
synchronize_irq(bp->msix_table[0].vector);
else
synchronize_irq(bp->pdev->irq);
flush_workqueue(bnx2x_wq);
flush_workqueue(bnx2x_iov_wq);
while (bnx2x_func_get_state(bp, &bp->func_obj) !=
BNX2X_F_STATE_STARTED && tout--)
msleep(20);
if (bnx2x_func_get_state(bp, &bp->func_obj) !=
BNX2X_F_STATE_STARTED) {
#ifdef BNX2X_STOP_ON_ERROR
BNX2X_ERR("Wrong function state\n");
return -EBUSY;
#else
/*
* Failed to complete the transaction in a "good way"
* Force both transactions with CLR bit
*/
struct bnx2x_func_state_params func_params = {NULL};
DP(NETIF_MSG_IFDOWN,
"Hmmm... Unexpected function state! Forcing STARTED-->TX_STOPPED-->STARTED\n");
func_params.f_obj = &bp->func_obj;
__set_bit(RAMROD_DRV_CLR_ONLY,
&func_params.ramrod_flags);
/* STARTED-->TX_ST0PPED */
func_params.cmd = BNX2X_F_CMD_TX_STOP;
bnx2x_func_state_change(bp, &func_params);
/* TX_ST0PPED-->STARTED */
func_params.cmd = BNX2X_F_CMD_TX_START;
return bnx2x_func_state_change(bp, &func_params);
#endif
}
return 0;
}
static void bnx2x_disable_ptp(struct bnx2x *bp)
{
int port = BP_PORT(bp);
/* Disable sending PTP packets to host */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_TO_HOST :
NIG_REG_P0_LLH_PTP_TO_HOST, 0x0);
/* Reset PTP event detection rules */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_PARAM_MASK :
NIG_REG_P0_LLH_PTP_PARAM_MASK, 0x7FF);
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_RULE_MASK :
NIG_REG_P0_LLH_PTP_RULE_MASK, 0x3FFF);
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_PARAM_MASK :
NIG_REG_P0_TLLH_PTP_PARAM_MASK, 0x7FF);
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_RULE_MASK :
NIG_REG_P0_TLLH_PTP_RULE_MASK, 0x3FFF);
/* Disable the PTP feature */
REG_WR(bp, port ? NIG_REG_P1_PTP_EN :
NIG_REG_P0_PTP_EN, 0x0);
}
/* Called during unload, to stop PTP-related stuff */
static void bnx2x_stop_ptp(struct bnx2x *bp)
{
/* Cancel PTP work queue. Should be done after the Tx queues are
* drained to prevent additional scheduling.
*/
cancel_work_sync(&bp->ptp_task);
if (bp->ptp_tx_skb) {
dev_kfree_skb_any(bp->ptp_tx_skb);
bp->ptp_tx_skb = NULL;
}
/* Disable PTP in HW */
bnx2x_disable_ptp(bp);
DP(BNX2X_MSG_PTP, "PTP stop ended successfully\n");
}
void bnx2x_chip_cleanup(struct bnx2x *bp, int unload_mode, bool keep_link)
{
int port = BP_PORT(bp);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
int i, rc = 0;
u8 cos;
struct bnx2x_mcast_ramrod_params rparam = {NULL};
u32 reset_code;
/* Wait until tx fastpath tasks complete */
for_each_tx_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
for_each_cos_in_tx_queue(fp, cos)
rc = bnx2x_clean_tx_queue(bp, fp->txdata_ptr[cos]);
#ifdef BNX2X_STOP_ON_ERROR
if (rc)
return;
#endif
}
/* Give HW time to discard old tx messages */
usleep_range(1000, 2000);
/* Clean all ETH MACs */
rc = bnx2x_del_all_macs(bp, &bp->sp_objs[0].mac_obj, BNX2X_ETH_MAC,
false);
if (rc < 0)
BNX2X_ERR("Failed to delete all ETH macs: %d\n", rc);
/* Clean up UC list */
rc = bnx2x_del_all_macs(bp, &bp->sp_objs[0].mac_obj, BNX2X_UC_LIST_MAC,
true);
if (rc < 0)
BNX2X_ERR("Failed to schedule DEL commands for UC MACs list: %d\n",
rc);
/* Disable LLH */
if (!CHIP_IS_E1(bp))
REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port*8, 0);
/* Set "drop all" (stop Rx).
* We need to take a netif_addr_lock() here in order to prevent
* a race between the completion code and this code.
*/
netif_addr_lock_bh(bp->dev);
/* Schedule the rx_mode command */
if (test_bit(BNX2X_FILTER_RX_MODE_PENDING, &bp->sp_state))
set_bit(BNX2X_FILTER_RX_MODE_SCHED, &bp->sp_state);
else
bnx2x_set_storm_rx_mode(bp);
/* Cleanup multicast configuration */
rparam.mcast_obj = &bp->mcast_obj;
rc = bnx2x_config_mcast(bp, &rparam, BNX2X_MCAST_CMD_DEL);
if (rc < 0)
BNX2X_ERR("Failed to send DEL multicast command: %d\n", rc);
netif_addr_unlock_bh(bp->dev);
bnx2x_iov_chip_cleanup(bp);
/*
* Send the UNLOAD_REQUEST to the MCP. This will return if
* this function should perform FUNC, PORT or COMMON HW
* reset.
*/
reset_code = bnx2x_send_unload_req(bp, unload_mode);
/*
* (assumption: No Attention from MCP at this stage)
* PMF probably in the middle of TX disable/enable transaction
*/
rc = bnx2x_func_wait_started(bp);
if (rc) {
BNX2X_ERR("bnx2x_func_wait_started failed\n");
#ifdef BNX2X_STOP_ON_ERROR
return;
#endif
}
/* Close multi and leading connections
* Completions for ramrods are collected in a synchronous way
*/
for_each_eth_queue(bp, i)
if (bnx2x_stop_queue(bp, i))
#ifdef BNX2X_STOP_ON_ERROR
return;
#else
goto unload_error;
#endif
if (CNIC_LOADED(bp)) {
for_each_cnic_queue(bp, i)
if (bnx2x_stop_queue(bp, i))
#ifdef BNX2X_STOP_ON_ERROR
return;
#else
goto unload_error;
#endif
}
/* If SP settings didn't get completed so far - something
* very wrong has happen.
*/
if (!bnx2x_wait_sp_comp(bp, ~0x0UL))
BNX2X_ERR("Hmmm... Common slow path ramrods got stuck!\n");
#ifndef BNX2X_STOP_ON_ERROR
unload_error:
#endif
rc = bnx2x_func_stop(bp);
if (rc) {
BNX2X_ERR("Function stop failed!\n");
#ifdef BNX2X_STOP_ON_ERROR
return;
#endif
}
/* stop_ptp should be after the Tx queues are drained to prevent
* scheduling to the cancelled PTP work queue. It should also be after
* function stop ramrod is sent, since as part of this ramrod FW access
* PTP registers.
*/
bnx2x: fix lockdep splat Michel reported following lockdep splat [ 44.718117] INFO: trying to register non-static key. [ 44.723081] the code is fine but needs lockdep annotation. [ 44.728559] turning off the locking correctness validator. [ 44.734036] CPU: 8 PID: 5483 Comm: ethtool Not tainted 4.1.0 [ 44.770289] Call Trace: [ 44.772741] [<ffffffff816eb1cd>] dump_stack+0x4c/0x65 [ 44.777879] [<ffffffff8111d921>] ? console_unlock+0x1f1/0x510 [ 44.783708] [<ffffffff811121f5>] __lock_acquire+0x1d05/0x1f10 [ 44.789538] [<ffffffff8111370a>] ? mark_held_locks+0x6a/0x90 [ 44.795276] [<ffffffff81113835>] ? trace_hardirqs_on_caller+0x105/0x1d0 [ 44.801967] [<ffffffff8111390d>] ? trace_hardirqs_on+0xd/0x10 [ 44.807793] [<ffffffff811330fa>] ? hrtimer_try_to_cancel+0x4a/0x250 [ 44.814142] [<ffffffff81112ba6>] lock_acquire+0xb6/0x290 [ 44.819537] [<ffffffff810d6675>] ? flush_work+0x5/0x280 [ 44.824844] [<ffffffff810d66ad>] flush_work+0x3d/0x280 [ 44.830061] [<ffffffff810d6675>] ? flush_work+0x5/0x280 [ 44.835366] [<ffffffff816f3c43>] ? schedule_hrtimeout_range+0x13/0x20 [ 44.841889] [<ffffffff8112ec9b>] ? usleep_range+0x4b/0x50 [ 44.847365] [<ffffffff8111370a>] ? mark_held_locks+0x6a/0x90 [ 44.853102] [<ffffffff810d8585>] ? __cancel_work_timer+0x105/0x1c0 [ 44.859359] [<ffffffff81113835>] ? trace_hardirqs_on_caller+0x105/0x1d0 [ 44.866045] [<ffffffff810d851f>] __cancel_work_timer+0x9f/0x1c0 [ 44.872048] [<ffffffffa0010982>] ? bnx2x_func_stop+0x42/0x90 [bnx2x] [ 44.878481] [<ffffffff810d8670>] cancel_work_sync+0x10/0x20 [ 44.884134] [<ffffffffa00259e5>] bnx2x_chip_cleanup+0x245/0x730 [bnx2x] [ 44.890829] [<ffffffff8110ce02>] ? up+0x32/0x50 [ 44.895439] [<ffffffff811306b5>] ? del_timer_sync+0x5/0xd0 [ 44.901005] [<ffffffffa005596d>] bnx2x_nic_unload+0x20d/0x8e0 [bnx2x] [ 44.907527] [<ffffffff811f1aef>] ? might_fault+0x5f/0xb0 [ 44.912921] [<ffffffffa005851c>] bnx2x_reload_if_running+0x2c/0x50 [bnx2x] [ 44.919879] [<ffffffffa005a3c5>] bnx2x_set_ringparam+0x2b5/0x460 [bnx2x] [ 44.926664] [<ffffffff815d498b>] dev_ethtool+0x55b/0x1c40 [ 44.932148] [<ffffffff815dfdc7>] ? rtnl_lock+0x17/0x20 [ 44.937364] [<ffffffff815e7f8b>] dev_ioctl+0x17b/0x630 [ 44.942582] [<ffffffff815abf8d>] sock_do_ioctl+0x5d/0x70 [ 44.947972] [<ffffffff815ac013>] sock_ioctl+0x73/0x280 [ 44.953192] [<ffffffff8124c1c8>] do_vfs_ioctl+0x88/0x5b0 [ 44.958587] [<ffffffff8110d0b3>] ? up_read+0x23/0x40 [ 44.963631] [<ffffffff812584cc>] ? __fget_light+0x6c/0xa0 [ 44.969105] [<ffffffff8124c781>] SyS_ioctl+0x91/0xb0 [ 44.974149] [<ffffffff816f4dd7>] system_call_fastpath+0x12/0x6f As bnx2x_init_ptp() is only called if bp->flags contains PTP_SUPPORTED, we also need to guard bnx2x_stop_ptp() with same condition, otherwise ptp_task workqueue is not initialized and kernel barfs on cancel_work_sync() Fixes: eeed018cbfa30 ("bnx2x: Add timestamping and PTP hardware clock support") Reported-by: Michel Lespinasse <walken@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Michal Kalderon <Michal.Kalderon@qlogic.com> Cc: Ariel Elior <Ariel.Elior@qlogic.com> Cc: Yuval Mintz <Yuval.Mintz@qlogic.com> Cc: David Decotigny <decot@google.com> Acked-by: Sony Chacko <sony.chacko@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-06-26 14:32:29 +09:00
if (bp->flags & PTP_SUPPORTED)
bnx2x_stop_ptp(bp);
/* Disable HW interrupts, NAPI */
bnx2x_netif_stop(bp, 1);
/* Delete all NAPI objects */
bnx2x_del_all_napi(bp);
if (CNIC_LOADED(bp))
bnx2x_del_all_napi_cnic(bp);
/* Release IRQs */
bnx2x_free_irq(bp);
bnx2x: don't reset chip on cleanup if PCI function is offline When PCI error is detected, in some architectures (like PowerPC) a slot reset is performed - the driver's error handlers are in charge of "disable" device before the reset, and re-enable it after a successful slot reset. There are two cases though that another path is taken on the code: if the slot reset is not successful or if too many errors already happened in the specific adapter (meaning that possibly the device is experiencing a HW failure that slot reset is not able to solve), the core PCI error mechanism (called EEH in PowerPC) will remove the adapter from the system, since it will consider this as a permanent failure on device. In this case, a path is taken that leads to bnx2x_chip_cleanup() calling bnx2x_reset_hw(), which then tries to perform a HW reset on chip. This reset won't succeed since the HW is in a fault state, which can be seen by multiple messages on kernel log like below: bnx2x: [bnx2x_issue_dmae_with_comp:552(eth1)]DMAE timeout! bnx2x: [bnx2x_write_dmae:600(eth1)]DMAE returned failure -1 After some time, the PCI error mechanism gives up on waiting the driver's correct removal procedure and forcibly remove the adapter from the system. We can see soft lockup while core PCI error mechanism is waiting for driver to accomplish the right removal process. This patch adds a verification to avoid a chip reset whenever the function is in PCI error state - since this case is only reached when we have a device being removed because of a permanent failure, the HW chip reset is not expected to work fine neither is necessary. Also, as a minor improvement in error path, we avoid the MCP information dump in case of non-recoverable PCI error (when adapter is about to be removed), since it will certainly fail. Reported-by: Harsha Thyagaraja <hathyaga@in.ibm.com> Signed-off-by: Guilherme G. Piccoli <gpiccoli@linux.vnet.ibm.com> Acked-By: Yuval Mintz <Yuval.Mintz@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-09-01 00:11:57 +09:00
/* Reset the chip, unless PCI function is offline. If we reach this
* point following a PCI error handling, it means device is really
* in a bad state and we're about to remove it, so reset the chip
* is not a good idea.
*/
if (!pci_channel_offline(bp->pdev)) {
rc = bnx2x_reset_hw(bp, reset_code);
if (rc)
BNX2X_ERR("HW_RESET failed\n");
}
/* Report UNLOAD_DONE to MCP */
bnx2x_send_unload_done(bp, keep_link);
}
void bnx2x_disable_close_the_gate(struct bnx2x *bp)
{
u32 val;
DP(NETIF_MSG_IFDOWN, "Disabling \"close the gates\"\n");
if (CHIP_IS_E1(bp)) {
int port = BP_PORT(bp);
u32 addr = port ? MISC_REG_AEU_MASK_ATTN_FUNC_1 :
MISC_REG_AEU_MASK_ATTN_FUNC_0;
val = REG_RD(bp, addr);
val &= ~(0x300);
REG_WR(bp, addr, val);
} else {
val = REG_RD(bp, MISC_REG_AEU_GENERAL_MASK);
val &= ~(MISC_AEU_GENERAL_MASK_REG_AEU_PXP_CLOSE_MASK |
MISC_AEU_GENERAL_MASK_REG_AEU_NIG_CLOSE_MASK);
REG_WR(bp, MISC_REG_AEU_GENERAL_MASK, val);
}
}
/* Close gates #2, #3 and #4: */
static void bnx2x_set_234_gates(struct bnx2x *bp, bool close)
{
u32 val;
/* Gates #2 and #4a are closed/opened for "not E1" only */
if (!CHIP_IS_E1(bp)) {
/* #4 */
REG_WR(bp, PXP_REG_HST_DISCARD_DOORBELLS, !!close);
/* #2 */
REG_WR(bp, PXP_REG_HST_DISCARD_INTERNAL_WRITES, !!close);
}
/* #3 */
if (CHIP_IS_E1x(bp)) {
/* Prevent interrupts from HC on both ports */
val = REG_RD(bp, HC_REG_CONFIG_1);
REG_WR(bp, HC_REG_CONFIG_1,
(!close) ? (val | HC_CONFIG_1_REG_BLOCK_DISABLE_1) :
(val & ~(u32)HC_CONFIG_1_REG_BLOCK_DISABLE_1));
val = REG_RD(bp, HC_REG_CONFIG_0);
REG_WR(bp, HC_REG_CONFIG_0,
(!close) ? (val | HC_CONFIG_0_REG_BLOCK_DISABLE_0) :
(val & ~(u32)HC_CONFIG_0_REG_BLOCK_DISABLE_0));
} else {
/* Prevent incoming interrupts in IGU */
val = REG_RD(bp, IGU_REG_BLOCK_CONFIGURATION);
REG_WR(bp, IGU_REG_BLOCK_CONFIGURATION,
(!close) ?
(val | IGU_BLOCK_CONFIGURATION_REG_BLOCK_ENABLE) :
(val & ~(u32)IGU_BLOCK_CONFIGURATION_REG_BLOCK_ENABLE));
}
DP(NETIF_MSG_HW | NETIF_MSG_IFUP, "%s gates #2, #3 and #4\n",
close ? "closing" : "opening");
mmiowb();
}
#define SHARED_MF_CLP_MAGIC 0x80000000 /* `magic' bit */
static void bnx2x_clp_reset_prep(struct bnx2x *bp, u32 *magic_val)
{
/* Do some magic... */
u32 val = MF_CFG_RD(bp, shared_mf_config.clp_mb);
*magic_val = val & SHARED_MF_CLP_MAGIC;
MF_CFG_WR(bp, shared_mf_config.clp_mb, val | SHARED_MF_CLP_MAGIC);
}
/**
* bnx2x_clp_reset_done - restore the value of the `magic' bit.
*
* @bp: driver handle
* @magic_val: old value of the `magic' bit.
*/
static void bnx2x_clp_reset_done(struct bnx2x *bp, u32 magic_val)
{
/* Restore the `magic' bit value... */
u32 val = MF_CFG_RD(bp, shared_mf_config.clp_mb);
MF_CFG_WR(bp, shared_mf_config.clp_mb,
(val & (~SHARED_MF_CLP_MAGIC)) | magic_val);
}
/**
* bnx2x_reset_mcp_prep - prepare for MCP reset.
*
* @bp: driver handle
* @magic_val: old value of 'magic' bit.
*
* Takes care of CLP configurations.
*/
static void bnx2x_reset_mcp_prep(struct bnx2x *bp, u32 *magic_val)
{
u32 shmem;
u32 validity_offset;
DP(NETIF_MSG_HW | NETIF_MSG_IFUP, "Starting\n");
/* Set `magic' bit in order to save MF config */
if (!CHIP_IS_E1(bp))
bnx2x_clp_reset_prep(bp, magic_val);
/* Get shmem offset */
shmem = REG_RD(bp, MISC_REG_SHARED_MEM_ADDR);
validity_offset =
offsetof(struct shmem_region, validity_map[BP_PORT(bp)]);
/* Clear validity map flags */
if (shmem > 0)
REG_WR(bp, shmem + validity_offset, 0);
}
#define MCP_TIMEOUT 5000 /* 5 seconds (in ms) */
#define MCP_ONE_TIMEOUT 100 /* 100 ms */
/**
* bnx2x_mcp_wait_one - wait for MCP_ONE_TIMEOUT
*
* @bp: driver handle
*/
static void bnx2x_mcp_wait_one(struct bnx2x *bp)
{
/* special handling for emulation and FPGA,
wait 10 times longer */
if (CHIP_REV_IS_SLOW(bp))
msleep(MCP_ONE_TIMEOUT*10);
else
msleep(MCP_ONE_TIMEOUT);
}
/*
* initializes bp->common.shmem_base and waits for validity signature to appear
*/
static int bnx2x_init_shmem(struct bnx2x *bp)
{
int cnt = 0;
u32 val = 0;
do {
bp->common.shmem_base = REG_RD(bp, MISC_REG_SHARED_MEM_ADDR);
/* If we read all 0xFFs, means we are in PCI error state and
* should bail out to avoid crashes on adapter's FW reads.
*/
if (bp->common.shmem_base == 0xFFFFFFFF) {
bp->flags |= NO_MCP_FLAG;
return -ENODEV;
}
if (bp->common.shmem_base) {
val = SHMEM_RD(bp, validity_map[BP_PORT(bp)]);
if (val & SHR_MEM_VALIDITY_MB)
return 0;
}
bnx2x_mcp_wait_one(bp);
} while (cnt++ < (MCP_TIMEOUT / MCP_ONE_TIMEOUT));
BNX2X_ERR("BAD MCP validity signature\n");
return -ENODEV;
}
static int bnx2x_reset_mcp_comp(struct bnx2x *bp, u32 magic_val)
{
int rc = bnx2x_init_shmem(bp);
/* Restore the `magic' bit value */
if (!CHIP_IS_E1(bp))
bnx2x_clp_reset_done(bp, magic_val);
return rc;
}
static void bnx2x_pxp_prep(struct bnx2x *bp)
{
if (!CHIP_IS_E1(bp)) {
REG_WR(bp, PXP2_REG_RD_START_INIT, 0);
REG_WR(bp, PXP2_REG_RQ_RBC_DONE, 0);
mmiowb();
}
}
/*
* Reset the whole chip except for:
* - PCIE core
* - PCI Glue, PSWHST, PXP/PXP2 RF (all controlled by
* one reset bit)
* - IGU
* - MISC (including AEU)
* - GRC
* - RBCN, RBCP
*/
static void bnx2x_process_kill_chip_reset(struct bnx2x *bp, bool global)
{
u32 not_reset_mask1, reset_mask1, not_reset_mask2, reset_mask2;
u32 global_bits2, stay_reset2;
/*
* Bits that have to be set in reset_mask2 if we want to reset 'global'
* (per chip) blocks.
*/
global_bits2 =
MISC_REGISTERS_RESET_REG_2_RST_MCP_N_RESET_CMN_CPU |
MISC_REGISTERS_RESET_REG_2_RST_MCP_N_RESET_CMN_CORE;
/* Don't reset the following blocks.
* Important: per port blocks (such as EMAC, BMAC, UMAC) can't be
* reset, as in 4 port device they might still be owned
* by the MCP (there is only one leader per path).
*/
not_reset_mask1 =
MISC_REGISTERS_RESET_REG_1_RST_HC |
MISC_REGISTERS_RESET_REG_1_RST_PXPV |
MISC_REGISTERS_RESET_REG_1_RST_PXP;
not_reset_mask2 =
MISC_REGISTERS_RESET_REG_2_RST_PCI_MDIO |
MISC_REGISTERS_RESET_REG_2_RST_EMAC0_HARD_CORE |
MISC_REGISTERS_RESET_REG_2_RST_EMAC1_HARD_CORE |
MISC_REGISTERS_RESET_REG_2_RST_MISC_CORE |
MISC_REGISTERS_RESET_REG_2_RST_RBCN |
MISC_REGISTERS_RESET_REG_2_RST_GRC |
MISC_REGISTERS_RESET_REG_2_RST_MCP_N_RESET_REG_HARD_CORE |
MISC_REGISTERS_RESET_REG_2_RST_MCP_N_HARD_CORE_RST_B |
MISC_REGISTERS_RESET_REG_2_RST_ATC |
MISC_REGISTERS_RESET_REG_2_PGLC |
MISC_REGISTERS_RESET_REG_2_RST_BMAC0 |
MISC_REGISTERS_RESET_REG_2_RST_BMAC1 |
MISC_REGISTERS_RESET_REG_2_RST_EMAC0 |
MISC_REGISTERS_RESET_REG_2_RST_EMAC1 |
MISC_REGISTERS_RESET_REG_2_UMAC0 |
MISC_REGISTERS_RESET_REG_2_UMAC1;
/*
* Keep the following blocks in reset:
* - all xxMACs are handled by the bnx2x_link code.
*/
stay_reset2 =
MISC_REGISTERS_RESET_REG_2_XMAC |
MISC_REGISTERS_RESET_REG_2_XMAC_SOFT;
/* Full reset masks according to the chip */
reset_mask1 = 0xffffffff;
if (CHIP_IS_E1(bp))
reset_mask2 = 0xffff;
else if (CHIP_IS_E1H(bp))
reset_mask2 = 0x1ffff;
else if (CHIP_IS_E2(bp))
reset_mask2 = 0xfffff;
else /* CHIP_IS_E3 */
reset_mask2 = 0x3ffffff;
/* Don't reset global blocks unless we need to */
if (!global)
reset_mask2 &= ~global_bits2;
/*
* In case of attention in the QM, we need to reset PXP
* (MISC_REGISTERS_RESET_REG_2_RST_PXP_RQ_RD_WR) before QM
* because otherwise QM reset would release 'close the gates' shortly
* before resetting the PXP, then the PSWRQ would send a write
* request to PGLUE. Then when PXP is reset, PGLUE would try to
* read the payload data from PSWWR, but PSWWR would not
* respond. The write queue in PGLUE would stuck, dmae commands
* would not return. Therefore it's important to reset the second
* reset register (containing the
* MISC_REGISTERS_RESET_REG_2_RST_PXP_RQ_RD_WR bit) before the
* first one (containing the MISC_REGISTERS_RESET_REG_1_RST_QM
* bit).
*/
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_2_CLEAR,
reset_mask2 & (~not_reset_mask2));
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_CLEAR,
reset_mask1 & (~not_reset_mask1));
barrier();
mmiowb();
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_2_SET,
reset_mask2 & (~stay_reset2));
barrier();
mmiowb();
REG_WR(bp, GRCBASE_MISC + MISC_REGISTERS_RESET_REG_1_SET, reset_mask1);
mmiowb();
}
/**
* bnx2x_er_poll_igu_vq - poll for pending writes bit.
* It should get cleared in no more than 1s.
*
* @bp: driver handle
*
* It should get cleared in no more than 1s. Returns 0 if
* pending writes bit gets cleared.
*/
static int bnx2x_er_poll_igu_vq(struct bnx2x *bp)
{
u32 cnt = 1000;
u32 pend_bits = 0;
do {
pend_bits = REG_RD(bp, IGU_REG_PENDING_BITS_STATUS);
if (pend_bits == 0)
break;
usleep_range(1000, 2000);
} while (cnt-- > 0);
if (cnt <= 0) {
BNX2X_ERR("Still pending IGU requests pend_bits=%x!\n",
pend_bits);
return -EBUSY;
}
return 0;
}
static int bnx2x_process_kill(struct bnx2x *bp, bool global)
{
int cnt = 1000;
u32 val = 0;
u32 sr_cnt, blk_cnt, port_is_idle_0, port_is_idle_1, pgl_exp_rom2;
u32 tags_63_32 = 0;
/* Empty the Tetris buffer, wait for 1s */
do {
sr_cnt = REG_RD(bp, PXP2_REG_RD_SR_CNT);
blk_cnt = REG_RD(bp, PXP2_REG_RD_BLK_CNT);
port_is_idle_0 = REG_RD(bp, PXP2_REG_RD_PORT_IS_IDLE_0);
port_is_idle_1 = REG_RD(bp, PXP2_REG_RD_PORT_IS_IDLE_1);
pgl_exp_rom2 = REG_RD(bp, PXP2_REG_PGL_EXP_ROM2);
if (CHIP_IS_E3(bp))
tags_63_32 = REG_RD(bp, PGLUE_B_REG_TAGS_63_32);
if ((sr_cnt == 0x7e) && (blk_cnt == 0xa0) &&
((port_is_idle_0 & 0x1) == 0x1) &&
((port_is_idle_1 & 0x1) == 0x1) &&
(pgl_exp_rom2 == 0xffffffff) &&
(!CHIP_IS_E3(bp) || (tags_63_32 == 0xffffffff)))
break;
usleep_range(1000, 2000);
} while (cnt-- > 0);
if (cnt <= 0) {
BNX2X_ERR("Tetris buffer didn't get empty or there are still outstanding read requests after 1s!\n");
BNX2X_ERR("sr_cnt=0x%08x, blk_cnt=0x%08x, port_is_idle_0=0x%08x, port_is_idle_1=0x%08x, pgl_exp_rom2=0x%08x\n",
sr_cnt, blk_cnt, port_is_idle_0, port_is_idle_1,
pgl_exp_rom2);
return -EAGAIN;
}
barrier();
/* Close gates #2, #3 and #4 */
bnx2x_set_234_gates(bp, true);
/* Poll for IGU VQs for 57712 and newer chips */
if (!CHIP_IS_E1x(bp) && bnx2x_er_poll_igu_vq(bp))
return -EAGAIN;
/* TBD: Indicate that "process kill" is in progress to MCP */
/* Clear "unprepared" bit */
REG_WR(bp, MISC_REG_UNPREPARED, 0);
barrier();
/* Make sure all is written to the chip before the reset */
mmiowb();
/* Wait for 1ms to empty GLUE and PCI-E core queues,
* PSWHST, GRC and PSWRD Tetris buffer.
*/
usleep_range(1000, 2000);
/* Prepare to chip reset: */
/* MCP */
if (global)
bnx2x_reset_mcp_prep(bp, &val);
/* PXP */
bnx2x_pxp_prep(bp);
barrier();
/* reset the chip */
bnx2x_process_kill_chip_reset(bp, global);
barrier();
/* clear errors in PGB */
if (!CHIP_IS_E1x(bp))
REG_WR(bp, PGLUE_B_REG_LATCHED_ERRORS_CLR, 0x7f);
/* Recover after reset: */
/* MCP */
if (global && bnx2x_reset_mcp_comp(bp, val))
return -EAGAIN;
/* TBD: Add resetting the NO_MCP mode DB here */
/* Open the gates #2, #3 and #4 */
bnx2x_set_234_gates(bp, false);
/* TBD: IGU/AEU preparation bring back the AEU/IGU to a
* reset state, re-enable attentions. */
return 0;
}
static int bnx2x_leader_reset(struct bnx2x *bp)
{
int rc = 0;
bool global = bnx2x_reset_is_global(bp);
u32 load_code;
/* if not going to reset MCP - load "fake" driver to reset HW while
* driver is owner of the HW
*/
if (!global && !BP_NOMCP(bp)) {
load_code = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_REQ,
DRV_MSG_CODE_LOAD_REQ_WITH_LFA);
if (!load_code) {
BNX2X_ERR("MCP response failure, aborting\n");
rc = -EAGAIN;
goto exit_leader_reset;
}
if ((load_code != FW_MSG_CODE_DRV_LOAD_COMMON_CHIP) &&
(load_code != FW_MSG_CODE_DRV_LOAD_COMMON)) {
BNX2X_ERR("MCP unexpected resp, aborting\n");
rc = -EAGAIN;
goto exit_leader_reset2;
}
load_code = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_DONE, 0);
if (!load_code) {
BNX2X_ERR("MCP response failure, aborting\n");
rc = -EAGAIN;
goto exit_leader_reset2;
}
}
/* Try to recover after the failure */
if (bnx2x_process_kill(bp, global)) {
BNX2X_ERR("Something bad had happen on engine %d! Aii!\n",
BP_PATH(bp));
rc = -EAGAIN;
goto exit_leader_reset2;
}
/*
* Clear RESET_IN_PROGRES and RESET_GLOBAL bits and update the driver
* state.
*/
bnx2x_set_reset_done(bp);
if (global)
bnx2x_clear_reset_global(bp);
exit_leader_reset2:
/* unload "fake driver" if it was loaded */
if (!global && !BP_NOMCP(bp)) {
bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_REQ_WOL_MCP, 0);
bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_DONE, 0);
}
exit_leader_reset:
bp->is_leader = 0;
bnx2x_release_leader_lock(bp);
smp_mb();
return rc;
}
static void bnx2x_recovery_failed(struct bnx2x *bp)
{
netdev_err(bp->dev, "Recovery has failed. Power cycle is needed.\n");
/* Disconnect this device */
netif_device_detach(bp->dev);
/*
* Block ifup for all function on this engine until "process kill"
* or power cycle.
*/
bnx2x_set_reset_in_progress(bp);
/* Shut down the power */
bnx2x_set_power_state(bp, PCI_D3hot);
bp->recovery_state = BNX2X_RECOVERY_FAILED;
smp_mb();
}
/*
* Assumption: runs under rtnl lock. This together with the fact
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
* that it's called only from bnx2x_sp_rtnl() ensure that it
* will never be called when netif_running(bp->dev) is false.
*/
static void bnx2x_parity_recover(struct bnx2x *bp)
{
bool global = false;
u32 error_recovered, error_unrecovered;
bool is_parity;
DP(NETIF_MSG_HW, "Handling parity\n");
while (1) {
switch (bp->recovery_state) {
case BNX2X_RECOVERY_INIT:
DP(NETIF_MSG_HW, "State is BNX2X_RECOVERY_INIT\n");
is_parity = bnx2x_chk_parity_attn(bp, &global, false);
WARN_ON(!is_parity);
/* Try to get a LEADER_LOCK HW lock */
if (bnx2x_trylock_leader_lock(bp)) {
bnx2x_set_reset_in_progress(bp);
/*
* Check if there is a global attention and if
* there was a global attention, set the global
* reset bit.
*/
if (global)
bnx2x_set_reset_global(bp);
bp->is_leader = 1;
}
/* Stop the driver */
/* If interface has been removed - break */
if (bnx2x_nic_unload(bp, UNLOAD_RECOVERY, false))
return;
bp->recovery_state = BNX2X_RECOVERY_WAIT;
/* Ensure "is_leader", MCP command sequence and
* "recovery_state" update values are seen on other
* CPUs.
*/
smp_mb();
break;
case BNX2X_RECOVERY_WAIT:
DP(NETIF_MSG_HW, "State is BNX2X_RECOVERY_WAIT\n");
if (bp->is_leader) {
int other_engine = BP_PATH(bp) ? 0 : 1;
bool other_load_status =
bnx2x_get_load_status(bp, other_engine);
bool load_status =
bnx2x_get_load_status(bp, BP_PATH(bp));
global = bnx2x_reset_is_global(bp);
/*
* In case of a parity in a global block, let
* the first leader that performs a
* leader_reset() reset the global blocks in
* order to clear global attentions. Otherwise
* the gates will remain closed for that
* engine.
*/
if (load_status ||
(global && other_load_status)) {
/* Wait until all other functions get
* down.
*/
schedule_delayed_work(&bp->sp_rtnl_task,
HZ/10);
return;
} else {
/* If all other functions got down -
* try to bring the chip back to
* normal. In any case it's an exit
* point for a leader.
*/
if (bnx2x_leader_reset(bp)) {
bnx2x_recovery_failed(bp);
return;
}
/* If we are here, means that the
* leader has succeeded and doesn't
* want to be a leader any more. Try
* to continue as a none-leader.
*/
break;
}
} else { /* non-leader */
if (!bnx2x_reset_is_done(bp, BP_PATH(bp))) {
/* Try to get a LEADER_LOCK HW lock as
* long as a former leader may have
* been unloaded by the user or
* released a leadership by another
* reason.
*/
if (bnx2x_trylock_leader_lock(bp)) {
/* I'm a leader now! Restart a
* switch case.
*/
bp->is_leader = 1;
break;
}
schedule_delayed_work(&bp->sp_rtnl_task,
HZ/10);
return;
} else {
/*
* If there was a global attention, wait
* for it to be cleared.
*/
if (bnx2x_reset_is_global(bp)) {
schedule_delayed_work(
&bp->sp_rtnl_task,
HZ/10);
return;
}
error_recovered =
bp->eth_stats.recoverable_error;
error_unrecovered =
bp->eth_stats.unrecoverable_error;
bp->recovery_state =
BNX2X_RECOVERY_NIC_LOADING;
if (bnx2x_nic_load(bp, LOAD_NORMAL)) {
error_unrecovered++;
netdev_err(bp->dev,
"Recovery failed. Power cycle needed\n");
/* Disconnect this device */
netif_device_detach(bp->dev);
/* Shut down the power */
bnx2x_set_power_state(
bp, PCI_D3hot);
smp_mb();
} else {
bp->recovery_state =
BNX2X_RECOVERY_DONE;
error_recovered++;
smp_mb();
}
bp->eth_stats.recoverable_error =
error_recovered;
bp->eth_stats.unrecoverable_error =
error_unrecovered;
return;
}
}
default:
return;
}
}
}
static int bnx2x_udp_port_update(struct bnx2x *bp)
{
struct bnx2x_func_switch_update_params *switch_update_params;
struct bnx2x_func_state_params func_params = {NULL};
struct bnx2x_udp_tunnel *udp_tunnel;
u16 vxlan_port = 0, geneve_port = 0;
int rc;
switch_update_params = &func_params.params.switch_update;
/* Prepare parameters for function state transitions */
__set_bit(RAMROD_COMP_WAIT, &func_params.ramrod_flags);
__set_bit(RAMROD_RETRY, &func_params.ramrod_flags);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_SWITCH_UPDATE;
/* Function parameters */
__set_bit(BNX2X_F_UPDATE_TUNNEL_CFG_CHNG,
&switch_update_params->changes);
if (bp->udp_tunnel_ports[BNX2X_UDP_PORT_GENEVE].count) {
udp_tunnel = &bp->udp_tunnel_ports[BNX2X_UDP_PORT_GENEVE];
geneve_port = udp_tunnel->dst_port;
switch_update_params->geneve_dst_port = geneve_port;
}
if (bp->udp_tunnel_ports[BNX2X_UDP_PORT_VXLAN].count) {
udp_tunnel = &bp->udp_tunnel_ports[BNX2X_UDP_PORT_VXLAN];
vxlan_port = udp_tunnel->dst_port;
switch_update_params->vxlan_dst_port = vxlan_port;
}
/* Re-enable inner-rss for the offloaded UDP tunnels */
__set_bit(BNX2X_F_UPDATE_TUNNEL_INNER_RSS,
&switch_update_params->changes);
rc = bnx2x_func_state_change(bp, &func_params);
if (rc)
BNX2X_ERR("failed to set UDP dst port to %04x %04x (rc = 0x%x)\n",
vxlan_port, geneve_port, rc);
else
DP(BNX2X_MSG_SP,
"Configured UDP ports: Vxlan [%04x] Geneve [%04x]\n",
vxlan_port, geneve_port);
return rc;
}
static void __bnx2x_add_udp_port(struct bnx2x *bp, u16 port,
enum bnx2x_udp_port_type type)
{
struct bnx2x_udp_tunnel *udp_port = &bp->udp_tunnel_ports[type];
if (!netif_running(bp->dev) || !IS_PF(bp) || CHIP_IS_E1x(bp))
return;
if (udp_port->count && udp_port->dst_port == port) {
udp_port->count++;
return;
}
if (udp_port->count) {
DP(BNX2X_MSG_SP,
"UDP tunnel [%d] - destination port limit reached\n",
type);
return;
}
udp_port->dst_port = port;
udp_port->count = 1;
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_CHANGE_UDP_PORT, 0);
}
static void __bnx2x_del_udp_port(struct bnx2x *bp, u16 port,
enum bnx2x_udp_port_type type)
{
struct bnx2x_udp_tunnel *udp_port = &bp->udp_tunnel_ports[type];
if (!IS_PF(bp) || CHIP_IS_E1x(bp))
return;
if (!udp_port->count || udp_port->dst_port != port) {
DP(BNX2X_MSG_SP, "Invalid UDP tunnel [%d] port\n",
type);
return;
}
/* Remove reference, and make certain it's no longer in use */
udp_port->count--;
if (udp_port->count)
return;
udp_port->dst_port = 0;
if (netif_running(bp->dev))
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_CHANGE_UDP_PORT, 0);
else
DP(BNX2X_MSG_SP, "Deleted UDP tunnel [%d] port %d\n",
type, port);
}
static void bnx2x_udp_tunnel_add(struct net_device *netdev,
struct udp_tunnel_info *ti)
{
struct bnx2x *bp = netdev_priv(netdev);
u16 t_port = ntohs(ti->port);
switch (ti->type) {
case UDP_TUNNEL_TYPE_VXLAN:
__bnx2x_add_udp_port(bp, t_port, BNX2X_UDP_PORT_VXLAN);
break;
case UDP_TUNNEL_TYPE_GENEVE:
__bnx2x_add_udp_port(bp, t_port, BNX2X_UDP_PORT_GENEVE);
break;
default:
break;
}
}
static void bnx2x_udp_tunnel_del(struct net_device *netdev,
struct udp_tunnel_info *ti)
{
struct bnx2x *bp = netdev_priv(netdev);
u16 t_port = ntohs(ti->port);
switch (ti->type) {
case UDP_TUNNEL_TYPE_VXLAN:
__bnx2x_del_udp_port(bp, t_port, BNX2X_UDP_PORT_VXLAN);
break;
case UDP_TUNNEL_TYPE_GENEVE:
__bnx2x_del_udp_port(bp, t_port, BNX2X_UDP_PORT_GENEVE);
break;
default:
break;
}
}
static int bnx2x_close(struct net_device *dev);
/* bnx2x_nic_unload() flushes the bnx2x_wq, thus reset task is
* scheduled on a general queue in order to prevent a dead lock.
*/
static void bnx2x_sp_rtnl_task(struct work_struct *work)
{
struct bnx2x *bp = container_of(work, struct bnx2x, sp_rtnl_task.work);
rtnl_lock();
if (!netif_running(bp->dev)) {
rtnl_unlock();
return;
}
if (unlikely(bp->recovery_state != BNX2X_RECOVERY_DONE)) {
#ifdef BNX2X_STOP_ON_ERROR
BNX2X_ERR("recovery flow called but STOP_ON_ERROR defined so reset not done to allow debug dump,\n"
"you will need to reboot when done\n");
goto sp_rtnl_not_reset;
#endif
/*
* Clear all pending SP commands as we are going to reset the
* function anyway.
*/
bp->sp_rtnl_state = 0;
smp_mb();
bnx2x_parity_recover(bp);
rtnl_unlock();
return;
}
if (test_and_clear_bit(BNX2X_SP_RTNL_TX_TIMEOUT, &bp->sp_rtnl_state)) {
#ifdef BNX2X_STOP_ON_ERROR
BNX2X_ERR("recovery flow called but STOP_ON_ERROR defined so reset not done to allow debug dump,\n"
"you will need to reboot when done\n");
goto sp_rtnl_not_reset;
#endif
/*
* Clear all pending SP commands as we are going to reset the
* function anyway.
*/
bp->sp_rtnl_state = 0;
smp_mb();
bnx2x_nic_unload(bp, UNLOAD_NORMAL, true);
bnx2x_nic_load(bp, LOAD_NORMAL);
rtnl_unlock();
return;
}
#ifdef BNX2X_STOP_ON_ERROR
sp_rtnl_not_reset:
#endif
if (test_and_clear_bit(BNX2X_SP_RTNL_SETUP_TC, &bp->sp_rtnl_state))
bnx2x_setup_tc(bp->dev, bp->dcbx_port_params.ets.num_of_cos);
if (test_and_clear_bit(BNX2X_SP_RTNL_AFEX_F_UPDATE, &bp->sp_rtnl_state))
bnx2x_after_function_update(bp);
/*
* in case of fan failure we need to reset id if the "stop on error"
* debug flag is set, since we trying to prevent permanent overheating
* damage
*/
if (test_and_clear_bit(BNX2X_SP_RTNL_FAN_FAILURE, &bp->sp_rtnl_state)) {
DP(NETIF_MSG_HW, "fan failure detected. Unloading driver\n");
netif_device_detach(bp->dev);
bnx2x_close(bp->dev);
rtnl_unlock();
return;
}
if (test_and_clear_bit(BNX2X_SP_RTNL_VFPF_MCAST, &bp->sp_rtnl_state)) {
DP(BNX2X_MSG_SP,
"sending set mcast vf pf channel message from rtnl sp-task\n");
bnx2x_vfpf_set_mcast(bp->dev);
}
if (test_and_clear_bit(BNX2X_SP_RTNL_VFPF_CHANNEL_DOWN,
&bp->sp_rtnl_state)){
if (!test_bit(__LINK_STATE_NOCARRIER, &bp->dev->state)) {
bnx2x_tx_disable(bp);
BNX2X_ERR("PF indicated channel is not servicable anymore. This means this VF device is no longer operational\n");
}
}
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
if (test_and_clear_bit(BNX2X_SP_RTNL_RX_MODE, &bp->sp_rtnl_state)) {
DP(BNX2X_MSG_SP, "Handling Rx Mode setting\n");
bnx2x_set_rx_mode_inner(bp);
}
if (test_and_clear_bit(BNX2X_SP_RTNL_HYPERVISOR_VLAN,
&bp->sp_rtnl_state))
bnx2x_pf_set_vfs_vlan(bp);
if (test_and_clear_bit(BNX2X_SP_RTNL_TX_STOP, &bp->sp_rtnl_state)) {
bnx2x_dcbx_stop_hw_tx(bp);
bnx2x_dcbx_resume_hw_tx(bp);
}
if (test_and_clear_bit(BNX2X_SP_RTNL_GET_DRV_VERSION,
&bp->sp_rtnl_state))
bnx2x_update_mng_version(bp);
if (test_and_clear_bit(BNX2X_SP_RTNL_CHANGE_UDP_PORT,
&bp->sp_rtnl_state)) {
if (bnx2x_udp_port_update(bp)) {
/* On error, forget configuration */
memset(bp->udp_tunnel_ports, 0,
sizeof(struct bnx2x_udp_tunnel) *
BNX2X_UDP_PORT_MAX);
} else {
/* Since we don't store additional port information,
* if no ports are configured for any feature ask for
* information about currently configured ports.
*/
if (!bp->udp_tunnel_ports[BNX2X_UDP_PORT_VXLAN].count &&
!bp->udp_tunnel_ports[BNX2X_UDP_PORT_GENEVE].count)
udp_tunnel_get_rx_info(bp->dev);
}
}
/* work which needs rtnl lock not-taken (as it takes the lock itself and
* can be called from other contexts as well)
*/
rtnl_unlock();
/* enable SR-IOV if applicable */
if (IS_SRIOV(bp) && test_and_clear_bit(BNX2X_SP_RTNL_ENABLE_SRIOV,
&bp->sp_rtnl_state)) {
bnx2x_disable_sriov(bp);
bnx2x_enable_sriov(bp);
}
}
static void bnx2x_period_task(struct work_struct *work)
{
struct bnx2x *bp = container_of(work, struct bnx2x, period_task.work);
if (!netif_running(bp->dev))
goto period_task_exit;
if (CHIP_REV_IS_SLOW(bp)) {
BNX2X_ERR("period task called on emulation, ignoring\n");
goto period_task_exit;
}
bnx2x_acquire_phy_lock(bp);
/*
* The barrier is needed to ensure the ordering between the writing to
* the bp->port.pmf in the bnx2x_nic_load() or bnx2x_pmf_update() and
* the reading here.
*/
smp_mb();
if (bp->port.pmf) {
bnx2x_period_func(&bp->link_params, &bp->link_vars);
/* Re-queue task in 1 sec */
queue_delayed_work(bnx2x_wq, &bp->period_task, 1*HZ);
}
bnx2x_release_phy_lock(bp);
period_task_exit:
return;
}
/*
* Init service functions
*/
static u32 bnx2x_get_pretend_reg(struct bnx2x *bp)
{
u32 base = PXP2_REG_PGL_PRETEND_FUNC_F0;
u32 stride = PXP2_REG_PGL_PRETEND_FUNC_F1 - base;
return base + (BP_ABS_FUNC(bp)) * stride;
}
static bool bnx2x_prev_unload_close_umac(struct bnx2x *bp,
u8 port, u32 reset_reg,
struct bnx2x_mac_vals *vals)
{
u32 mask = MISC_REGISTERS_RESET_REG_2_UMAC0 << port;
u32 base_addr;
if (!(mask & reset_reg))
return false;
BNX2X_DEV_INFO("Disable umac Rx %02x\n", port);
base_addr = port ? GRCBASE_UMAC1 : GRCBASE_UMAC0;
vals->umac_addr[port] = base_addr + UMAC_REG_COMMAND_CONFIG;
vals->umac_val[port] = REG_RD(bp, vals->umac_addr[port]);
REG_WR(bp, vals->umac_addr[port], 0);
return true;
}
static void bnx2x_prev_unload_close_mac(struct bnx2x *bp,
struct bnx2x_mac_vals *vals)
{
u32 val, base_addr, offset, mask, reset_reg;
bool mac_stopped = false;
u8 port = BP_PORT(bp);
/* reset addresses as they also mark which values were changed */
memset(vals, 0, sizeof(*vals));
reset_reg = REG_RD(bp, MISC_REG_RESET_REG_2);
if (!CHIP_IS_E3(bp)) {
val = REG_RD(bp, NIG_REG_BMAC0_REGS_OUT_EN + port * 4);
mask = MISC_REGISTERS_RESET_REG_2_RST_BMAC0 << port;
if ((mask & reset_reg) && val) {
u32 wb_data[2];
BNX2X_DEV_INFO("Disable bmac Rx\n");
base_addr = BP_PORT(bp) ? NIG_REG_INGRESS_BMAC1_MEM
: NIG_REG_INGRESS_BMAC0_MEM;
offset = CHIP_IS_E2(bp) ? BIGMAC2_REGISTER_BMAC_CONTROL
: BIGMAC_REGISTER_BMAC_CONTROL;
/*
* use rd/wr since we cannot use dmae. This is safe
* since MCP won't access the bus due to the request
* to unload, and no function on the path can be
* loaded at this time.
*/
wb_data[0] = REG_RD(bp, base_addr + offset);
wb_data[1] = REG_RD(bp, base_addr + offset + 0x4);
vals->bmac_addr = base_addr + offset;
vals->bmac_val[0] = wb_data[0];
vals->bmac_val[1] = wb_data[1];
wb_data[0] &= ~BMAC_CONTROL_RX_ENABLE;
REG_WR(bp, vals->bmac_addr, wb_data[0]);
REG_WR(bp, vals->bmac_addr + 0x4, wb_data[1]);
}
BNX2X_DEV_INFO("Disable emac Rx\n");
vals->emac_addr = NIG_REG_NIG_EMAC0_EN + BP_PORT(bp)*4;
vals->emac_val = REG_RD(bp, vals->emac_addr);
REG_WR(bp, vals->emac_addr, 0);
mac_stopped = true;
} else {
if (reset_reg & MISC_REGISTERS_RESET_REG_2_XMAC) {
BNX2X_DEV_INFO("Disable xmac Rx\n");
base_addr = BP_PORT(bp) ? GRCBASE_XMAC1 : GRCBASE_XMAC0;
val = REG_RD(bp, base_addr + XMAC_REG_PFC_CTRL_HI);
REG_WR(bp, base_addr + XMAC_REG_PFC_CTRL_HI,
val & ~(1 << 1));
REG_WR(bp, base_addr + XMAC_REG_PFC_CTRL_HI,
val | (1 << 1));
vals->xmac_addr = base_addr + XMAC_REG_CTRL;
vals->xmac_val = REG_RD(bp, vals->xmac_addr);
REG_WR(bp, vals->xmac_addr, 0);
mac_stopped = true;
}
mac_stopped |= bnx2x_prev_unload_close_umac(bp, 0,
reset_reg, vals);
mac_stopped |= bnx2x_prev_unload_close_umac(bp, 1,
reset_reg, vals);
}
if (mac_stopped)
msleep(20);
}
#define BNX2X_PREV_UNDI_PROD_ADDR(p) (BAR_TSTRORM_INTMEM + 0x1508 + ((p) << 4))
#define BNX2X_PREV_UNDI_PROD_ADDR_H(f) (BAR_TSTRORM_INTMEM + \
0x1848 + ((f) << 4))
#define BNX2X_PREV_UNDI_RCQ(val) ((val) & 0xffff)
#define BNX2X_PREV_UNDI_BD(val) ((val) >> 16 & 0xffff)
#define BNX2X_PREV_UNDI_PROD(rcq, bd) ((bd) << 16 | (rcq))
#define BCM_5710_UNDI_FW_MF_MAJOR (0x07)
#define BCM_5710_UNDI_FW_MF_MINOR (0x08)
#define BCM_5710_UNDI_FW_MF_VERS (0x05)
static bool bnx2x_prev_is_after_undi(struct bnx2x *bp)
{
/* UNDI marks its presence in DORQ -
* it initializes CID offset for normal bell to 0x7
*/
if (!(REG_RD(bp, MISC_REG_RESET_REG_1) &
MISC_REGISTERS_RESET_REG_1_RST_DORQ))
return false;
if (REG_RD(bp, DORQ_REG_NORM_CID_OFST) == 0x7) {
BNX2X_DEV_INFO("UNDI previously loaded\n");
return true;
}
return false;
}
static void bnx2x_prev_unload_undi_inc(struct bnx2x *bp, u8 inc)
{
u16 rcq, bd;
u32 addr, tmp_reg;
if (BP_FUNC(bp) < 2)
addr = BNX2X_PREV_UNDI_PROD_ADDR(BP_PORT(bp));
else
addr = BNX2X_PREV_UNDI_PROD_ADDR_H(BP_FUNC(bp) - 2);
tmp_reg = REG_RD(bp, addr);
rcq = BNX2X_PREV_UNDI_RCQ(tmp_reg) + inc;
bd = BNX2X_PREV_UNDI_BD(tmp_reg) + inc;
tmp_reg = BNX2X_PREV_UNDI_PROD(rcq, bd);
REG_WR(bp, addr, tmp_reg);
BNX2X_DEV_INFO("UNDI producer [%d/%d][%08x] rings bd -> 0x%04x, rcq -> 0x%04x\n",
BP_PORT(bp), BP_FUNC(bp), addr, bd, rcq);
}
static int bnx2x_prev_mcp_done(struct bnx2x *bp)
{
u32 rc = bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_DONE,
DRV_MSG_CODE_UNLOAD_SKIP_LINK_RESET);
if (!rc) {
BNX2X_ERR("MCP response failure, aborting\n");
return -EBUSY;
}
return 0;
}
static struct bnx2x_prev_path_list *
bnx2x_prev_path_get_entry(struct bnx2x *bp)
{
struct bnx2x_prev_path_list *tmp_list;
list_for_each_entry(tmp_list, &bnx2x_prev_list, list)
if (PCI_SLOT(bp->pdev->devfn) == tmp_list->slot &&
bp->pdev->bus->number == tmp_list->bus &&
BP_PATH(bp) == tmp_list->path)
return tmp_list;
return NULL;
}
static int bnx2x_prev_path_mark_eeh(struct bnx2x *bp)
{
struct bnx2x_prev_path_list *tmp_list;
int rc;
rc = down_interruptible(&bnx2x_prev_sem);
if (rc) {
BNX2X_ERR("Received %d when tried to take lock\n", rc);
return rc;
}
tmp_list = bnx2x_prev_path_get_entry(bp);
if (tmp_list) {
tmp_list->aer = 1;
rc = 0;
} else {
BNX2X_ERR("path %d: Entry does not exist for eeh; Flow occurs before initial insmod is over ?\n",
BP_PATH(bp));
}
up(&bnx2x_prev_sem);
return rc;
}
static bool bnx2x_prev_is_path_marked(struct bnx2x *bp)
{
struct bnx2x_prev_path_list *tmp_list;
bool rc = false;
if (down_trylock(&bnx2x_prev_sem))
return false;
tmp_list = bnx2x_prev_path_get_entry(bp);
if (tmp_list) {
if (tmp_list->aer) {
DP(NETIF_MSG_HW, "Path %d was marked by AER\n",
BP_PATH(bp));
} else {
rc = true;
BNX2X_DEV_INFO("Path %d was already cleaned from previous drivers\n",
BP_PATH(bp));
}
}
up(&bnx2x_prev_sem);
return rc;
}
bool bnx2x_port_after_undi(struct bnx2x *bp)
{
struct bnx2x_prev_path_list *entry;
bool val;
down(&bnx2x_prev_sem);
entry = bnx2x_prev_path_get_entry(bp);
val = !!(entry && (entry->undi & (1 << BP_PORT(bp))));
up(&bnx2x_prev_sem);
return val;
}
static int bnx2x_prev_mark_path(struct bnx2x *bp, bool after_undi)
{
struct bnx2x_prev_path_list *tmp_list;
int rc;
rc = down_interruptible(&bnx2x_prev_sem);
if (rc) {
BNX2X_ERR("Received %d when tried to take lock\n", rc);
return rc;
}
/* Check whether the entry for this path already exists */
tmp_list = bnx2x_prev_path_get_entry(bp);
if (tmp_list) {
if (!tmp_list->aer) {
BNX2X_ERR("Re-Marking the path.\n");
} else {
DP(NETIF_MSG_HW, "Removing AER indication from path %d\n",
BP_PATH(bp));
tmp_list->aer = 0;
}
up(&bnx2x_prev_sem);
return 0;
}
up(&bnx2x_prev_sem);
/* Create an entry for this path and add it */
tmp_list = kmalloc(sizeof(struct bnx2x_prev_path_list), GFP_KERNEL);
if (!tmp_list) {
BNX2X_ERR("Failed to allocate 'bnx2x_prev_path_list'\n");
return -ENOMEM;
}
tmp_list->bus = bp->pdev->bus->number;
tmp_list->slot = PCI_SLOT(bp->pdev->devfn);
tmp_list->path = BP_PATH(bp);
tmp_list->aer = 0;
tmp_list->undi = after_undi ? (1 << BP_PORT(bp)) : 0;
rc = down_interruptible(&bnx2x_prev_sem);
if (rc) {
BNX2X_ERR("Received %d when tried to take lock\n", rc);
kfree(tmp_list);
} else {
DP(NETIF_MSG_HW, "Marked path [%d] - finished previous unload\n",
BP_PATH(bp));
list_add(&tmp_list->list, &bnx2x_prev_list);
up(&bnx2x_prev_sem);
}
return rc;
}
static int bnx2x_do_flr(struct bnx2x *bp)
{
struct pci_dev *dev = bp->pdev;
if (CHIP_IS_E1x(bp)) {
BNX2X_DEV_INFO("FLR not supported in E1/E1H\n");
return -EINVAL;
}
/* only bootcode REQ_BC_VER_4_INITIATE_FLR and onwards support flr */
if (bp->common.bc_ver < REQ_BC_VER_4_INITIATE_FLR) {
BNX2X_ERR("FLR not supported by BC_VER: 0x%x\n",
bp->common.bc_ver);
return -EINVAL;
}
if (!pci_wait_for_pending_transaction(dev))
dev_err(&dev->dev, "transaction is not cleared; proceeding with reset anyway\n");
BNX2X_DEV_INFO("Initiating FLR\n");
bnx2x_fw_command(bp, DRV_MSG_CODE_INITIATE_FLR, 0);
return 0;
}
static int bnx2x_prev_unload_uncommon(struct bnx2x *bp)
{
int rc;
BNX2X_DEV_INFO("Uncommon unload Flow\n");
/* Test if previous unload process was already finished for this path */
if (bnx2x_prev_is_path_marked(bp))
return bnx2x_prev_mcp_done(bp);
BNX2X_DEV_INFO("Path is unmarked\n");
/* Cannot proceed with FLR if UNDI is loaded, since FW does not match */
if (bnx2x_prev_is_after_undi(bp))
goto out;
/* If function has FLR capabilities, and existing FW version matches
* the one required, then FLR will be sufficient to clean any residue
* left by previous driver
*/
rc = bnx2x_compare_fw_ver(bp, FW_MSG_CODE_DRV_LOAD_FUNCTION, false);
if (!rc) {
/* fw version is good */
BNX2X_DEV_INFO("FW version matches our own. Attempting FLR\n");
rc = bnx2x_do_flr(bp);
}
if (!rc) {
/* FLR was performed */
BNX2X_DEV_INFO("FLR successful\n");
return 0;
}
BNX2X_DEV_INFO("Could not FLR\n");
out:
/* Close the MCP request, return failure*/
rc = bnx2x_prev_mcp_done(bp);
if (!rc)
rc = BNX2X_PREV_WAIT_NEEDED;
return rc;
}
static int bnx2x_prev_unload_common(struct bnx2x *bp)
{
u32 reset_reg, tmp_reg = 0, rc;
bool prev_undi = false;
struct bnx2x_mac_vals mac_vals;
/* It is possible a previous function received 'common' answer,
* but hasn't loaded yet, therefore creating a scenario of
* multiple functions receiving 'common' on the same path.
*/
BNX2X_DEV_INFO("Common unload Flow\n");
memset(&mac_vals, 0, sizeof(mac_vals));
if (bnx2x_prev_is_path_marked(bp))
return bnx2x_prev_mcp_done(bp);
reset_reg = REG_RD(bp, MISC_REG_RESET_REG_1);
/* Reset should be performed after BRB is emptied */
if (reset_reg & MISC_REGISTERS_RESET_REG_1_RST_BRB1) {
u32 timer_count = 1000;
/* Close the MAC Rx to prevent BRB from filling up */
bnx2x_prev_unload_close_mac(bp, &mac_vals);
/* close LLH filters for both ports towards the BRB */
bnx2x_set_rx_filter(&bp->link_params, 0);
bp->link_params.port ^= 1;
bnx2x_set_rx_filter(&bp->link_params, 0);
bp->link_params.port ^= 1;
/* Check if the UNDI driver was previously loaded */
if (bnx2x_prev_is_after_undi(bp)) {
prev_undi = true;
/* clear the UNDI indication */
REG_WR(bp, DORQ_REG_NORM_CID_OFST, 0);
/* clear possible idle check errors */
REG_RD(bp, NIG_REG_NIG_INT_STS_CLR_0);
}
if (!CHIP_IS_E1x(bp))
/* block FW from writing to host */
REG_WR(bp, PGLUE_B_REG_INTERNAL_PFID_ENABLE_MASTER, 0);
/* wait until BRB is empty */
tmp_reg = REG_RD(bp, BRB1_REG_NUM_OF_FULL_BLOCKS);
while (timer_count) {
u32 prev_brb = tmp_reg;
tmp_reg = REG_RD(bp, BRB1_REG_NUM_OF_FULL_BLOCKS);
if (!tmp_reg)
break;
BNX2X_DEV_INFO("BRB still has 0x%08x\n", tmp_reg);
/* reset timer as long as BRB actually gets emptied */
if (prev_brb > tmp_reg)
timer_count = 1000;
else
timer_count--;
/* If UNDI resides in memory, manually increment it */
if (prev_undi)
bnx2x_prev_unload_undi_inc(bp, 1);
udelay(10);
}
if (!timer_count)
BNX2X_ERR("Failed to empty BRB, hope for the best\n");
}
/* No packets are in the pipeline, path is ready for reset */
bnx2x_reset_common(bp);
if (mac_vals.xmac_addr)
REG_WR(bp, mac_vals.xmac_addr, mac_vals.xmac_val);
if (mac_vals.umac_addr[0])
REG_WR(bp, mac_vals.umac_addr[0], mac_vals.umac_val[0]);
if (mac_vals.umac_addr[1])
REG_WR(bp, mac_vals.umac_addr[1], mac_vals.umac_val[1]);
if (mac_vals.emac_addr)
REG_WR(bp, mac_vals.emac_addr, mac_vals.emac_val);
if (mac_vals.bmac_addr) {
REG_WR(bp, mac_vals.bmac_addr, mac_vals.bmac_val[0]);
REG_WR(bp, mac_vals.bmac_addr + 4, mac_vals.bmac_val[1]);
}
rc = bnx2x_prev_mark_path(bp, prev_undi);
if (rc) {
bnx2x_prev_mcp_done(bp);
return rc;
}
return bnx2x_prev_mcp_done(bp);
}
static int bnx2x_prev_unload(struct bnx2x *bp)
{
int time_counter = 10;
u32 rc, fw, hw_lock_reg, hw_lock_val;
BNX2X_DEV_INFO("Entering Previous Unload Flow\n");
/* clear hw from errors which may have resulted from an interrupted
* dmae transaction.
*/
bnx2x_clean_pglue_errors(bp);
/* Release previously held locks */
hw_lock_reg = (BP_FUNC(bp) <= 5) ?
(MISC_REG_DRIVER_CONTROL_1 + BP_FUNC(bp) * 8) :
(MISC_REG_DRIVER_CONTROL_7 + (BP_FUNC(bp) - 6) * 8);
hw_lock_val = REG_RD(bp, hw_lock_reg);
if (hw_lock_val) {
if (hw_lock_val & HW_LOCK_RESOURCE_NVRAM) {
BNX2X_DEV_INFO("Release Previously held NVRAM lock\n");
REG_WR(bp, MCP_REG_MCPR_NVM_SW_ARB,
(MCPR_NVM_SW_ARB_ARB_REQ_CLR1 << BP_PORT(bp)));
}
BNX2X_DEV_INFO("Release Previously held hw lock\n");
REG_WR(bp, hw_lock_reg, 0xffffffff);
} else
BNX2X_DEV_INFO("No need to release hw/nvram locks\n");
if (MCPR_ACCESS_LOCK_LOCK & REG_RD(bp, MCP_REG_MCPR_ACCESS_LOCK)) {
BNX2X_DEV_INFO("Release previously held alr\n");
bnx2x_release_alr(bp);
}
do {
int aer = 0;
/* Lock MCP using an unload request */
fw = bnx2x_fw_command(bp, DRV_MSG_CODE_UNLOAD_REQ_WOL_DIS, 0);
if (!fw) {
BNX2X_ERR("MCP response failure, aborting\n");
rc = -EBUSY;
break;
}
rc = down_interruptible(&bnx2x_prev_sem);
if (rc) {
BNX2X_ERR("Cannot check for AER; Received %d when tried to take lock\n",
rc);
} else {
/* If Path is marked by EEH, ignore unload status */
aer = !!(bnx2x_prev_path_get_entry(bp) &&
bnx2x_prev_path_get_entry(bp)->aer);
up(&bnx2x_prev_sem);
}
if (fw == FW_MSG_CODE_DRV_UNLOAD_COMMON || aer) {
rc = bnx2x_prev_unload_common(bp);
break;
}
/* non-common reply from MCP might require looping */
rc = bnx2x_prev_unload_uncommon(bp);
if (rc != BNX2X_PREV_WAIT_NEEDED)
break;
msleep(20);
} while (--time_counter);
if (!time_counter || rc) {
BNX2X_DEV_INFO("Unloading previous driver did not occur, Possibly due to MF UNDI\n");
rc = -EPROBE_DEFER;
}
/* Mark function if its port was used to boot from SAN */
if (bnx2x_port_after_undi(bp))
bp->link_params.feature_config_flags |=
FEATURE_CONFIG_BOOT_FROM_SAN;
BNX2X_DEV_INFO("Finished Previous Unload Flow [%d]\n", rc);
return rc;
}
static void bnx2x_get_common_hwinfo(struct bnx2x *bp)
{
u32 val, val2, val3, val4, id, boot_mode;
u16 pmc;
/* Get the chip revision id and number. */
/* chip num:16-31, rev:12-15, metal:4-11, bond_id:0-3 */
val = REG_RD(bp, MISC_REG_CHIP_NUM);
id = ((val & 0xffff) << 16);
val = REG_RD(bp, MISC_REG_CHIP_REV);
id |= ((val & 0xf) << 12);
/* Metal is read from PCI regs, but we can't access >=0x400 from
* the configuration space (so we need to reg_rd)
*/
val = REG_RD(bp, PCICFG_OFFSET + PCI_ID_VAL3);
id |= (((val >> 24) & 0xf) << 4);
val = REG_RD(bp, MISC_REG_BOND_ID);
id |= (val & 0xf);
bp->common.chip_id = id;
/* force 57811 according to MISC register */
if (REG_RD(bp, MISC_REG_CHIP_TYPE) & MISC_REG_CHIP_TYPE_57811_MASK) {
if (CHIP_IS_57810(bp))
bp->common.chip_id = (CHIP_NUM_57811 << 16) |
(bp->common.chip_id & 0x0000FFFF);
else if (CHIP_IS_57810_MF(bp))
bp->common.chip_id = (CHIP_NUM_57811_MF << 16) |
(bp->common.chip_id & 0x0000FFFF);
bp->common.chip_id |= 0x1;
}
/* Set doorbell size */
bp->db_size = (1 << BNX2X_DB_SHIFT);
if (!CHIP_IS_E1x(bp)) {
val = REG_RD(bp, MISC_REG_PORT4MODE_EN_OVWR);
if ((val & 1) == 0)
val = REG_RD(bp, MISC_REG_PORT4MODE_EN);
else
val = (val >> 1) & 1;
BNX2X_DEV_INFO("chip is in %s\n", val ? "4_PORT_MODE" :
"2_PORT_MODE");
bp->common.chip_port_mode = val ? CHIP_4_PORT_MODE :
CHIP_2_PORT_MODE;
if (CHIP_MODE_IS_4_PORT(bp))
bp->pfid = (bp->pf_num >> 1); /* 0..3 */
else
bp->pfid = (bp->pf_num & 0x6); /* 0, 2, 4, 6 */
} else {
bp->common.chip_port_mode = CHIP_PORT_MODE_NONE; /* N/A */
bp->pfid = bp->pf_num; /* 0..7 */
}
BNX2X_DEV_INFO("pf_id: %x", bp->pfid);
bp->link_params.chip_id = bp->common.chip_id;
BNX2X_DEV_INFO("chip ID is 0x%x\n", id);
val = (REG_RD(bp, 0x2874) & 0x55);
if ((bp->common.chip_id & 0x1) ||
(CHIP_IS_E1(bp) && val) || (CHIP_IS_E1H(bp) && (val == 0x55))) {
bp->flags |= ONE_PORT_FLAG;
BNX2X_DEV_INFO("single port device\n");
}
val = REG_RD(bp, MCP_REG_MCPR_NVM_CFG4);
bp->common.flash_size = (BNX2X_NVRAM_1MB_SIZE <<
(val & MCPR_NVM_CFG4_FLASH_SIZE));
BNX2X_DEV_INFO("flash_size 0x%x (%d)\n",
bp->common.flash_size, bp->common.flash_size);
bnx2x_init_shmem(bp);
bp->common.shmem2_base = REG_RD(bp, (BP_PATH(bp) ?
MISC_REG_GENERIC_CR_1 :
MISC_REG_GENERIC_CR_0));
bp->link_params.shmem_base = bp->common.shmem_base;
bp->link_params.shmem2_base = bp->common.shmem2_base;
if (SHMEM2_RD(bp, size) >
(u32)offsetof(struct shmem2_region, lfa_host_addr[BP_PORT(bp)]))
bp->link_params.lfa_base =
REG_RD(bp, bp->common.shmem2_base +
(u32)offsetof(struct shmem2_region,
lfa_host_addr[BP_PORT(bp)]));
else
bp->link_params.lfa_base = 0;
BNX2X_DEV_INFO("shmem offset 0x%x shmem2 offset 0x%x\n",
bp->common.shmem_base, bp->common.shmem2_base);
if (!bp->common.shmem_base) {
BNX2X_DEV_INFO("MCP not active\n");
bp->flags |= NO_MCP_FLAG;
return;
}
bp->common.hw_config = SHMEM_RD(bp, dev_info.shared_hw_config.config);
BNX2X_DEV_INFO("hw_config 0x%08x\n", bp->common.hw_config);
bp->link_params.hw_led_mode = ((bp->common.hw_config &
SHARED_HW_CFG_LED_MODE_MASK) >>
SHARED_HW_CFG_LED_MODE_SHIFT);
bp->link_params.feature_config_flags = 0;
val = SHMEM_RD(bp, dev_info.shared_feature_config.config);
if (val & SHARED_FEAT_CFG_OVERRIDE_PREEMPHASIS_CFG_ENABLED)
bp->link_params.feature_config_flags |=
FEATURE_CONFIG_OVERRIDE_PREEMPHASIS_ENABLED;
else
bp->link_params.feature_config_flags &=
~FEATURE_CONFIG_OVERRIDE_PREEMPHASIS_ENABLED;
val = SHMEM_RD(bp, dev_info.bc_rev) >> 8;
bp->common.bc_ver = val;
BNX2X_DEV_INFO("bc_ver %X\n", val);
if (val < BNX2X_BC_VER) {
/* for now only warn
* later we might need to enforce this */
BNX2X_ERR("This driver needs bc_ver %X but found %X, please upgrade BC\n",
BNX2X_BC_VER, val);
}
bp->link_params.feature_config_flags |=
(val >= REQ_BC_VER_4_VRFY_FIRST_PHY_OPT_MDL) ?
FEATURE_CONFIG_BC_SUPPORTS_OPT_MDL_VRFY : 0;
bp->link_params.feature_config_flags |=
(val >= REQ_BC_VER_4_VRFY_SPECIFIC_PHY_OPT_MDL) ?
FEATURE_CONFIG_BC_SUPPORTS_DUAL_PHY_OPT_MDL_VRFY : 0;
bp->link_params.feature_config_flags |=
(val >= REQ_BC_VER_4_VRFY_AFEX_SUPPORTED) ?
FEATURE_CONFIG_BC_SUPPORTS_AFEX : 0;
bp->link_params.feature_config_flags |=
(val >= REQ_BC_VER_4_SFP_TX_DISABLE_SUPPORTED) ?
FEATURE_CONFIG_BC_SUPPORTS_SFP_TX_DISABLED : 0;
bp->link_params.feature_config_flags |=
(val >= REQ_BC_VER_4_MT_SUPPORTED) ?
FEATURE_CONFIG_MT_SUPPORT : 0;
bp->flags |= (val >= REQ_BC_VER_4_PFC_STATS_SUPPORTED) ?
BC_SUPPORTS_PFC_STATS : 0;
bp->flags |= (val >= REQ_BC_VER_4_FCOE_FEATURES) ?
BC_SUPPORTS_FCOE_FEATURES : 0;
bp->flags |= (val >= REQ_BC_VER_4_DCBX_ADMIN_MSG_NON_PMF) ?
BC_SUPPORTS_DCBX_MSG_NON_PMF : 0;
bp->flags |= (val >= REQ_BC_VER_4_RMMOD_CMD) ?
BC_SUPPORTS_RMMOD_CMD : 0;
boot_mode = SHMEM_RD(bp,
dev_info.port_feature_config[BP_PORT(bp)].mba_config) &
PORT_FEATURE_MBA_BOOT_AGENT_TYPE_MASK;
switch (boot_mode) {
case PORT_FEATURE_MBA_BOOT_AGENT_TYPE_PXE:
bp->common.boot_mode = FEATURE_ETH_BOOTMODE_PXE;
break;
case PORT_FEATURE_MBA_BOOT_AGENT_TYPE_ISCSIB:
bp->common.boot_mode = FEATURE_ETH_BOOTMODE_ISCSI;
break;
case PORT_FEATURE_MBA_BOOT_AGENT_TYPE_FCOE_BOOT:
bp->common.boot_mode = FEATURE_ETH_BOOTMODE_FCOE;
break;
case PORT_FEATURE_MBA_BOOT_AGENT_TYPE_NONE:
bp->common.boot_mode = FEATURE_ETH_BOOTMODE_NONE;
break;
}
pci_read_config_word(bp->pdev, bp->pdev->pm_cap + PCI_PM_PMC, &pmc);
bp->flags |= (pmc & PCI_PM_CAP_PME_D3cold) ? 0 : NO_WOL_FLAG;
BNX2X_DEV_INFO("%sWoL capable\n",
(bp->flags & NO_WOL_FLAG) ? "not " : "");
val = SHMEM_RD(bp, dev_info.shared_hw_config.part_num);
val2 = SHMEM_RD(bp, dev_info.shared_hw_config.part_num[4]);
val3 = SHMEM_RD(bp, dev_info.shared_hw_config.part_num[8]);
val4 = SHMEM_RD(bp, dev_info.shared_hw_config.part_num[12]);
dev_info(&bp->pdev->dev, "part number %X-%X-%X-%X\n",
val, val2, val3, val4);
}
#define IGU_FID(val) GET_FIELD((val), IGU_REG_MAPPING_MEMORY_FID)
#define IGU_VEC(val) GET_FIELD((val), IGU_REG_MAPPING_MEMORY_VECTOR)
static int bnx2x_get_igu_cam_info(struct bnx2x *bp)
{
int pfid = BP_FUNC(bp);
int igu_sb_id;
u32 val;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
u8 fid, igu_sb_cnt = 0;
bp->igu_base_sb = 0xff;
if (CHIP_INT_MODE_IS_BC(bp)) {
int vn = BP_VN(bp);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
igu_sb_cnt = bp->igu_sb_cnt;
bp->igu_base_sb = (CHIP_MODE_IS_4_PORT(bp) ? pfid : vn) *
FP_SB_MAX_E1x;
bp->igu_dsb_id = E1HVN_MAX * FP_SB_MAX_E1x +
(CHIP_MODE_IS_4_PORT(bp) ? pfid : vn);
return 0;
}
/* IGU in normal mode - read CAM */
for (igu_sb_id = 0; igu_sb_id < IGU_REG_MAPPING_MEMORY_SIZE;
igu_sb_id++) {
val = REG_RD(bp, IGU_REG_MAPPING_MEMORY + igu_sb_id * 4);
if (!(val & IGU_REG_MAPPING_MEMORY_VALID))
continue;
fid = IGU_FID(val);
if ((fid & IGU_FID_ENCODE_IS_PF)) {
if ((fid & IGU_FID_PF_NUM_MASK) != pfid)
continue;
if (IGU_VEC(val) == 0)
/* default status block */
bp->igu_dsb_id = igu_sb_id;
else {
if (bp->igu_base_sb == 0xff)
bp->igu_base_sb = igu_sb_id;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
igu_sb_cnt++;
}
}
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
#ifdef CONFIG_PCI_MSI
/* Due to new PF resource allocation by MFW T7.4 and above, it's
* optional that number of CAM entries will not be equal to the value
* advertised in PCI.
* Driver should use the minimal value of both as the actual status
* block count
*/
bp->igu_sb_cnt = min_t(int, bp->igu_sb_cnt, igu_sb_cnt);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
#endif
if (igu_sb_cnt == 0) {
BNX2X_ERR("CAM configuration error\n");
return -EINVAL;
}
return 0;
}
static void bnx2x_link_settings_supported(struct bnx2x *bp, u32 switch_cfg)
{
int cfg_size = 0, idx, port = BP_PORT(bp);
/* Aggregation of supported attributes of all external phys */
bp->port.supported[0] = 0;
bp->port.supported[1] = 0;
switch (bp->link_params.num_phys) {
case 1:
bp->port.supported[0] = bp->link_params.phy[INT_PHY].supported;
cfg_size = 1;
break;
case 2:
bp->port.supported[0] = bp->link_params.phy[EXT_PHY1].supported;
cfg_size = 1;
break;
case 3:
if (bp->link_params.multi_phy_config &
PORT_HW_CFG_PHY_SWAPPED_ENABLED) {
bp->port.supported[1] =
bp->link_params.phy[EXT_PHY1].supported;
bp->port.supported[0] =
bp->link_params.phy[EXT_PHY2].supported;
} else {
bp->port.supported[0] =
bp->link_params.phy[EXT_PHY1].supported;
bp->port.supported[1] =
bp->link_params.phy[EXT_PHY2].supported;
}
cfg_size = 2;
break;
}
if (!(bp->port.supported[0] || bp->port.supported[1])) {
BNX2X_ERR("NVRAM config error. BAD phy config. PHY1 config 0x%x, PHY2 config 0x%x\n",
SHMEM_RD(bp,
dev_info.port_hw_config[port].external_phy_config),
SHMEM_RD(bp,
dev_info.port_hw_config[port].external_phy_config2));
return;
}
if (CHIP_IS_E3(bp))
bp->port.phy_addr = REG_RD(bp, MISC_REG_WC0_CTRL_PHY_ADDR);
else {
switch (switch_cfg) {
case SWITCH_CFG_1G:
bp->port.phy_addr = REG_RD(
bp, NIG_REG_SERDES0_CTRL_PHY_ADDR + port*0x10);
break;
case SWITCH_CFG_10G:
bp->port.phy_addr = REG_RD(
bp, NIG_REG_XGXS0_CTRL_PHY_ADDR + port*0x18);
break;
default:
BNX2X_ERR("BAD switch_cfg link_config 0x%x\n",
bp->port.link_config[0]);
return;
}
}
BNX2X_DEV_INFO("phy_addr 0x%x\n", bp->port.phy_addr);
/* mask what we support according to speed_cap_mask per configuration */
for (idx = 0; idx < cfg_size; idx++) {
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_10M_HALF))
bp->port.supported[idx] &= ~SUPPORTED_10baseT_Half;
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_10M_FULL))
bp->port.supported[idx] &= ~SUPPORTED_10baseT_Full;
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_100M_HALF))
bp->port.supported[idx] &= ~SUPPORTED_100baseT_Half;
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_100M_FULL))
bp->port.supported[idx] &= ~SUPPORTED_100baseT_Full;
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_1G))
bp->port.supported[idx] &= ~(SUPPORTED_1000baseT_Half |
SUPPORTED_1000baseT_Full);
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_2_5G))
bp->port.supported[idx] &= ~SUPPORTED_2500baseX_Full;
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_10G))
bp->port.supported[idx] &= ~SUPPORTED_10000baseT_Full;
if (!(bp->link_params.speed_cap_mask[idx] &
PORT_HW_CFG_SPEED_CAPABILITY_D0_20G))
bp->port.supported[idx] &= ~SUPPORTED_20000baseKR2_Full;
}
BNX2X_DEV_INFO("supported 0x%x 0x%x\n", bp->port.supported[0],
bp->port.supported[1]);
}
static void bnx2x_link_settings_requested(struct bnx2x *bp)
{
u32 link_config, idx, cfg_size = 0;
bp->port.advertising[0] = 0;
bp->port.advertising[1] = 0;
switch (bp->link_params.num_phys) {
case 1:
case 2:
cfg_size = 1;
break;
case 3:
cfg_size = 2;
break;
}
for (idx = 0; idx < cfg_size; idx++) {
bp->link_params.req_duplex[idx] = DUPLEX_FULL;
link_config = bp->port.link_config[idx];
switch (link_config & PORT_FEATURE_LINK_SPEED_MASK) {
case PORT_FEATURE_LINK_SPEED_AUTO:
if (bp->port.supported[idx] & SUPPORTED_Autoneg) {
bp->link_params.req_line_speed[idx] =
SPEED_AUTO_NEG;
bp->port.advertising[idx] |=
bp->port.supported[idx];
if (bp->link_params.phy[EXT_PHY1].type ==
PORT_HW_CFG_XGXS_EXT_PHY_TYPE_BCM84833)
bp->port.advertising[idx] |=
(SUPPORTED_100baseT_Half |
SUPPORTED_100baseT_Full);
} else {
/* force 10G, no AN */
bp->link_params.req_line_speed[idx] =
SPEED_10000;
bp->port.advertising[idx] |=
(ADVERTISED_10000baseT_Full |
ADVERTISED_FIBRE);
continue;
}
break;
case PORT_FEATURE_LINK_SPEED_10M_FULL:
if (bp->port.supported[idx] & SUPPORTED_10baseT_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_10;
bp->port.advertising[idx] |=
(ADVERTISED_10baseT_Full |
ADVERTISED_TP);
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_10M_HALF:
if (bp->port.supported[idx] & SUPPORTED_10baseT_Half) {
bp->link_params.req_line_speed[idx] =
SPEED_10;
bp->link_params.req_duplex[idx] =
DUPLEX_HALF;
bp->port.advertising[idx] |=
(ADVERTISED_10baseT_Half |
ADVERTISED_TP);
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_100M_FULL:
if (bp->port.supported[idx] &
SUPPORTED_100baseT_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_100;
bp->port.advertising[idx] |=
(ADVERTISED_100baseT_Full |
ADVERTISED_TP);
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_100M_HALF:
if (bp->port.supported[idx] &
SUPPORTED_100baseT_Half) {
bp->link_params.req_line_speed[idx] =
SPEED_100;
bp->link_params.req_duplex[idx] =
DUPLEX_HALF;
bp->port.advertising[idx] |=
(ADVERTISED_100baseT_Half |
ADVERTISED_TP);
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_1G:
if (bp->port.supported[idx] &
SUPPORTED_1000baseT_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_1000;
bp->port.advertising[idx] |=
(ADVERTISED_1000baseT_Full |
ADVERTISED_TP);
} else if (bp->port.supported[idx] &
SUPPORTED_1000baseKX_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_1000;
bp->port.advertising[idx] |=
ADVERTISED_1000baseKX_Full;
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_2_5G:
if (bp->port.supported[idx] &
SUPPORTED_2500baseX_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_2500;
bp->port.advertising[idx] |=
(ADVERTISED_2500baseX_Full |
ADVERTISED_TP);
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_10G_CX4:
if (bp->port.supported[idx] &
SUPPORTED_10000baseT_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_10000;
bp->port.advertising[idx] |=
(ADVERTISED_10000baseT_Full |
ADVERTISED_FIBRE);
} else if (bp->port.supported[idx] &
SUPPORTED_10000baseKR_Full) {
bp->link_params.req_line_speed[idx] =
SPEED_10000;
bp->port.advertising[idx] |=
(ADVERTISED_10000baseKR_Full |
ADVERTISED_FIBRE);
} else {
BNX2X_ERR("NVRAM config error. Invalid link_config 0x%x speed_cap_mask 0x%x\n",
link_config,
bp->link_params.speed_cap_mask[idx]);
return;
}
break;
case PORT_FEATURE_LINK_SPEED_20G:
bp->link_params.req_line_speed[idx] = SPEED_20000;
break;
default:
BNX2X_ERR("NVRAM config error. BAD link speed link_config 0x%x\n",
link_config);
bp->link_params.req_line_speed[idx] =
SPEED_AUTO_NEG;
bp->port.advertising[idx] =
bp->port.supported[idx];
break;
}
bp->link_params.req_flow_ctrl[idx] = (link_config &
PORT_FEATURE_FLOW_CONTROL_MASK);
if (bp->link_params.req_flow_ctrl[idx] ==
BNX2X_FLOW_CTRL_AUTO) {
if (!(bp->port.supported[idx] & SUPPORTED_Autoneg))
bp->link_params.req_flow_ctrl[idx] =
BNX2X_FLOW_CTRL_NONE;
else
bnx2x_set_requested_fc(bp);
}
BNX2X_DEV_INFO("req_line_speed %d req_duplex %d req_flow_ctrl 0x%x advertising 0x%x\n",
bp->link_params.req_line_speed[idx],
bp->link_params.req_duplex[idx],
bp->link_params.req_flow_ctrl[idx],
bp->port.advertising[idx]);
}
}
static void bnx2x_set_mac_buf(u8 *mac_buf, u32 mac_lo, u16 mac_hi)
{
__be16 mac_hi_be = cpu_to_be16(mac_hi);
__be32 mac_lo_be = cpu_to_be32(mac_lo);
memcpy(mac_buf, &mac_hi_be, sizeof(mac_hi_be));
memcpy(mac_buf + sizeof(mac_hi_be), &mac_lo_be, sizeof(mac_lo_be));
}
static void bnx2x_get_port_hwinfo(struct bnx2x *bp)
{
int port = BP_PORT(bp);
u32 config;
u32 ext_phy_type, ext_phy_config, eee_mode;
bp->link_params.bp = bp;
bp->link_params.port = port;
bp->link_params.lane_config =
SHMEM_RD(bp, dev_info.port_hw_config[port].lane_config);
bp->link_params.speed_cap_mask[0] =
SHMEM_RD(bp,
dev_info.port_hw_config[port].speed_capability_mask) &
PORT_HW_CFG_SPEED_CAPABILITY_D0_MASK;
bp->link_params.speed_cap_mask[1] =
SHMEM_RD(bp,
dev_info.port_hw_config[port].speed_capability_mask2) &
PORT_HW_CFG_SPEED_CAPABILITY_D0_MASK;
bp->port.link_config[0] =
SHMEM_RD(bp, dev_info.port_feature_config[port].link_config);
bp->port.link_config[1] =
SHMEM_RD(bp, dev_info.port_feature_config[port].link_config2);
bp->link_params.multi_phy_config =
SHMEM_RD(bp, dev_info.port_hw_config[port].multi_phy_config);
/* If the device is capable of WoL, set the default state according
* to the HW
*/
config = SHMEM_RD(bp, dev_info.port_feature_config[port].config);
bp->wol = (!(bp->flags & NO_WOL_FLAG) &&
(config & PORT_FEATURE_WOL_ENABLED));
if ((config & PORT_FEAT_CFG_STORAGE_PERSONALITY_MASK) ==
PORT_FEAT_CFG_STORAGE_PERSONALITY_FCOE && !IS_MF(bp))
bp->flags |= NO_ISCSI_FLAG;
if ((config & PORT_FEAT_CFG_STORAGE_PERSONALITY_MASK) ==
PORT_FEAT_CFG_STORAGE_PERSONALITY_ISCSI && !(IS_MF(bp)))
bp->flags |= NO_FCOE_FLAG;
BNX2X_DEV_INFO("lane_config 0x%08x speed_cap_mask0 0x%08x link_config0 0x%08x\n",
bp->link_params.lane_config,
bp->link_params.speed_cap_mask[0],
bp->port.link_config[0]);
bp->link_params.switch_cfg = (bp->port.link_config[0] &
PORT_FEATURE_CONNECTED_SWITCH_MASK);
bnx2x_phy_probe(&bp->link_params);
bnx2x_link_settings_supported(bp, bp->link_params.switch_cfg);
bnx2x_link_settings_requested(bp);
/*
* If connected directly, work with the internal PHY, otherwise, work
* with the external PHY
*/
ext_phy_config =
SHMEM_RD(bp,
dev_info.port_hw_config[port].external_phy_config);
ext_phy_type = XGXS_EXT_PHY_TYPE(ext_phy_config);
if (ext_phy_type == PORT_HW_CFG_XGXS_EXT_PHY_TYPE_DIRECT)
bp->mdio.prtad = bp->port.phy_addr;
else if ((ext_phy_type != PORT_HW_CFG_XGXS_EXT_PHY_TYPE_FAILURE) &&
(ext_phy_type != PORT_HW_CFG_XGXS_EXT_PHY_TYPE_NOT_CONN))
bp->mdio.prtad =
XGXS_EXT_PHY_ADDR(ext_phy_config);
/* Configure link feature according to nvram value */
eee_mode = (((SHMEM_RD(bp, dev_info.
port_feature_config[port].eee_power_mode)) &
PORT_FEAT_CFG_EEE_POWER_MODE_MASK) >>
PORT_FEAT_CFG_EEE_POWER_MODE_SHIFT);
if (eee_mode != PORT_FEAT_CFG_EEE_POWER_MODE_DISABLED) {
bp->link_params.eee_mode = EEE_MODE_ADV_LPI |
EEE_MODE_ENABLE_LPI |
EEE_MODE_OUTPUT_TIME;
} else {
bp->link_params.eee_mode = 0;
}
}
void bnx2x_get_iscsi_info(struct bnx2x *bp)
{
u32 no_flags = NO_ISCSI_FLAG;
int port = BP_PORT(bp);
u32 max_iscsi_conn = FW_ENCODE_32BIT_PATTERN ^ SHMEM_RD(bp,
drv_lic_key[port].max_iscsi_conn);
if (!CNIC_SUPPORT(bp)) {
bp->flags |= no_flags;
return;
}
/* Get the number of maximum allowed iSCSI connections */
bp->cnic_eth_dev.max_iscsi_conn =
(max_iscsi_conn & BNX2X_MAX_ISCSI_INIT_CONN_MASK) >>
BNX2X_MAX_ISCSI_INIT_CONN_SHIFT;
BNX2X_DEV_INFO("max_iscsi_conn 0x%x\n",
bp->cnic_eth_dev.max_iscsi_conn);
/*
* If maximum allowed number of connections is zero -
* disable the feature.
*/
if (!bp->cnic_eth_dev.max_iscsi_conn)
bp->flags |= no_flags;
}
static void bnx2x_get_ext_wwn_info(struct bnx2x *bp, int func)
{
/* Port info */
bp->cnic_eth_dev.fcoe_wwn_port_name_hi =
MF_CFG_RD(bp, func_ext_config[func].fcoe_wwn_port_name_upper);
bp->cnic_eth_dev.fcoe_wwn_port_name_lo =
MF_CFG_RD(bp, func_ext_config[func].fcoe_wwn_port_name_lower);
/* Node info */
bp->cnic_eth_dev.fcoe_wwn_node_name_hi =
MF_CFG_RD(bp, func_ext_config[func].fcoe_wwn_node_name_upper);
bp->cnic_eth_dev.fcoe_wwn_node_name_lo =
MF_CFG_RD(bp, func_ext_config[func].fcoe_wwn_node_name_lower);
}
static int bnx2x_shared_fcoe_funcs(struct bnx2x *bp)
{
u8 count = 0;
if (IS_MF(bp)) {
u8 fid;
/* iterate over absolute function ids for this path: */
for (fid = BP_PATH(bp); fid < E2_FUNC_MAX * 2; fid += 2) {
if (IS_MF_SD(bp)) {
u32 cfg = MF_CFG_RD(bp,
func_mf_config[fid].config);
if (!(cfg & FUNC_MF_CFG_FUNC_HIDE) &&
((cfg & FUNC_MF_CFG_PROTOCOL_MASK) ==
FUNC_MF_CFG_PROTOCOL_FCOE))
count++;
} else {
u32 cfg = MF_CFG_RD(bp,
func_ext_config[fid].
func_cfg);
if ((cfg & MACP_FUNC_CFG_FLAGS_ENABLED) &&
(cfg & MACP_FUNC_CFG_FLAGS_FCOE_OFFLOAD))
count++;
}
}
} else { /* SF */
int port, port_cnt = CHIP_MODE_IS_4_PORT(bp) ? 2 : 1;
for (port = 0; port < port_cnt; port++) {
u32 lic = SHMEM_RD(bp,
drv_lic_key[port].max_fcoe_conn) ^
FW_ENCODE_32BIT_PATTERN;
if (lic)
count++;
}
}
return count;
}
static void bnx2x_get_fcoe_info(struct bnx2x *bp)
{
int port = BP_PORT(bp);
int func = BP_ABS_FUNC(bp);
u32 max_fcoe_conn = FW_ENCODE_32BIT_PATTERN ^ SHMEM_RD(bp,
drv_lic_key[port].max_fcoe_conn);
u8 num_fcoe_func = bnx2x_shared_fcoe_funcs(bp);
if (!CNIC_SUPPORT(bp)) {
bp->flags |= NO_FCOE_FLAG;
return;
}
/* Get the number of maximum allowed FCoE connections */
bp->cnic_eth_dev.max_fcoe_conn =
(max_fcoe_conn & BNX2X_MAX_FCOE_INIT_CONN_MASK) >>
BNX2X_MAX_FCOE_INIT_CONN_SHIFT;
/* Calculate the number of maximum allowed FCoE tasks */
bp->cnic_eth_dev.max_fcoe_exchanges = MAX_NUM_FCOE_TASKS_PER_ENGINE;
/* check if FCoE resources must be shared between different functions */
if (num_fcoe_func)
bp->cnic_eth_dev.max_fcoe_exchanges /= num_fcoe_func;
/* Read the WWN: */
if (!IS_MF(bp)) {
/* Port info */
bp->cnic_eth_dev.fcoe_wwn_port_name_hi =
SHMEM_RD(bp,
dev_info.port_hw_config[port].
fcoe_wwn_port_name_upper);
bp->cnic_eth_dev.fcoe_wwn_port_name_lo =
SHMEM_RD(bp,
dev_info.port_hw_config[port].
fcoe_wwn_port_name_lower);
/* Node info */
bp->cnic_eth_dev.fcoe_wwn_node_name_hi =
SHMEM_RD(bp,
dev_info.port_hw_config[port].
fcoe_wwn_node_name_upper);
bp->cnic_eth_dev.fcoe_wwn_node_name_lo =
SHMEM_RD(bp,
dev_info.port_hw_config[port].
fcoe_wwn_node_name_lower);
} else if (!IS_MF_SD(bp)) {
/* Read the WWN info only if the FCoE feature is enabled for
* this function.
*/
if (BNX2X_HAS_MF_EXT_PROTOCOL_FCOE(bp))
bnx2x_get_ext_wwn_info(bp, func);
} else {
if (BNX2X_IS_MF_SD_PROTOCOL_FCOE(bp) && !CHIP_IS_E1x(bp))
bnx2x_get_ext_wwn_info(bp, func);
}
BNX2X_DEV_INFO("max_fcoe_conn 0x%x\n", bp->cnic_eth_dev.max_fcoe_conn);
/*
* If maximum allowed number of connections is zero -
* disable the feature.
*/
if (!bp->cnic_eth_dev.max_fcoe_conn)
bp->flags |= NO_FCOE_FLAG;
}
static void bnx2x_get_cnic_info(struct bnx2x *bp)
{
/*
* iSCSI may be dynamically disabled but reading
* info here we will decrease memory usage by driver
* if the feature is disabled for good
*/
bnx2x_get_iscsi_info(bp);
bnx2x_get_fcoe_info(bp);
}
static void bnx2x_get_cnic_mac_hwinfo(struct bnx2x *bp)
{
u32 val, val2;
int func = BP_ABS_FUNC(bp);
int port = BP_PORT(bp);
u8 *iscsi_mac = bp->cnic_eth_dev.iscsi_mac;
u8 *fip_mac = bp->fip_mac;
if (IS_MF(bp)) {
/* iSCSI and FCoE NPAR MACs: if there is no either iSCSI or
* FCoE MAC then the appropriate feature should be disabled.
* In non SD mode features configuration comes from struct
* func_ext_config.
*/
if (!IS_MF_SD(bp)) {
u32 cfg = MF_CFG_RD(bp, func_ext_config[func].func_cfg);
if (cfg & MACP_FUNC_CFG_FLAGS_ISCSI_OFFLOAD) {
val2 = MF_CFG_RD(bp, func_ext_config[func].
iscsi_mac_addr_upper);
val = MF_CFG_RD(bp, func_ext_config[func].
iscsi_mac_addr_lower);
bnx2x_set_mac_buf(iscsi_mac, val, val2);
BNX2X_DEV_INFO
("Read iSCSI MAC: %pM\n", iscsi_mac);
} else {
bp->flags |= NO_ISCSI_OOO_FLAG | NO_ISCSI_FLAG;
}
if (cfg & MACP_FUNC_CFG_FLAGS_FCOE_OFFLOAD) {
val2 = MF_CFG_RD(bp, func_ext_config[func].
fcoe_mac_addr_upper);
val = MF_CFG_RD(bp, func_ext_config[func].
fcoe_mac_addr_lower);
bnx2x_set_mac_buf(fip_mac, val, val2);
BNX2X_DEV_INFO
("Read FCoE L2 MAC: %pM\n", fip_mac);
} else {
bp->flags |= NO_FCOE_FLAG;
}
bp->mf_ext_config = cfg;
} else { /* SD MODE */
if (BNX2X_IS_MF_SD_PROTOCOL_ISCSI(bp)) {
/* use primary mac as iscsi mac */
memcpy(iscsi_mac, bp->dev->dev_addr, ETH_ALEN);
BNX2X_DEV_INFO("SD ISCSI MODE\n");
BNX2X_DEV_INFO
("Read iSCSI MAC: %pM\n", iscsi_mac);
} else if (BNX2X_IS_MF_SD_PROTOCOL_FCOE(bp)) {
/* use primary mac as fip mac */
memcpy(fip_mac, bp->dev->dev_addr, ETH_ALEN);
BNX2X_DEV_INFO("SD FCoE MODE\n");
BNX2X_DEV_INFO
("Read FIP MAC: %pM\n", fip_mac);
}
}
/* If this is a storage-only interface, use SAN mac as
* primary MAC. Notice that for SD this is already the case,
* as the SAN mac was copied from the primary MAC.
*/
if (IS_MF_FCOE_AFEX(bp))
memcpy(bp->dev->dev_addr, fip_mac, ETH_ALEN);
} else {
val2 = SHMEM_RD(bp, dev_info.port_hw_config[port].
iscsi_mac_upper);
val = SHMEM_RD(bp, dev_info.port_hw_config[port].
iscsi_mac_lower);
bnx2x_set_mac_buf(iscsi_mac, val, val2);
val2 = SHMEM_RD(bp, dev_info.port_hw_config[port].
fcoe_fip_mac_upper);
val = SHMEM_RD(bp, dev_info.port_hw_config[port].
fcoe_fip_mac_lower);
bnx2x_set_mac_buf(fip_mac, val, val2);
}
/* Disable iSCSI OOO if MAC configuration is invalid. */
if (!is_valid_ether_addr(iscsi_mac)) {
bp->flags |= NO_ISCSI_OOO_FLAG | NO_ISCSI_FLAG;
eth_zero_addr(iscsi_mac);
}
/* Disable FCoE if MAC configuration is invalid. */
if (!is_valid_ether_addr(fip_mac)) {
bp->flags |= NO_FCOE_FLAG;
eth_zero_addr(bp->fip_mac);
}
}
static void bnx2x_get_mac_hwinfo(struct bnx2x *bp)
{
u32 val, val2;
int func = BP_ABS_FUNC(bp);
int port = BP_PORT(bp);
/* Zero primary MAC configuration */
eth_zero_addr(bp->dev->dev_addr);
if (BP_NOMCP(bp)) {
BNX2X_ERROR("warning: random MAC workaround active\n");
eth_hw_addr_random(bp->dev);
} else if (IS_MF(bp)) {
val2 = MF_CFG_RD(bp, func_mf_config[func].mac_upper);
val = MF_CFG_RD(bp, func_mf_config[func].mac_lower);
if ((val2 != FUNC_MF_CFG_UPPERMAC_DEFAULT) &&
(val != FUNC_MF_CFG_LOWERMAC_DEFAULT))
bnx2x_set_mac_buf(bp->dev->dev_addr, val, val2);
if (CNIC_SUPPORT(bp))
bnx2x_get_cnic_mac_hwinfo(bp);
} else {
/* in SF read MACs from port configuration */
val2 = SHMEM_RD(bp, dev_info.port_hw_config[port].mac_upper);
val = SHMEM_RD(bp, dev_info.port_hw_config[port].mac_lower);
bnx2x_set_mac_buf(bp->dev->dev_addr, val, val2);
if (CNIC_SUPPORT(bp))
bnx2x_get_cnic_mac_hwinfo(bp);
}
if (!BP_NOMCP(bp)) {
/* Read physical port identifier from shmem */
val2 = SHMEM_RD(bp, dev_info.port_hw_config[port].mac_upper);
val = SHMEM_RD(bp, dev_info.port_hw_config[port].mac_lower);
bnx2x_set_mac_buf(bp->phys_port_id, val, val2);
bp->flags |= HAS_PHYS_PORT_ID;
}
memcpy(bp->link_params.mac_addr, bp->dev->dev_addr, ETH_ALEN);
if (!is_valid_ether_addr(bp->dev->dev_addr))
dev_err(&bp->pdev->dev,
"bad Ethernet MAC address configuration: %pM\n"
"change it manually before bringing up the appropriate network interface\n",
bp->dev->dev_addr);
}
static bool bnx2x_get_dropless_info(struct bnx2x *bp)
{
int tmp;
u32 cfg;
if (IS_VF(bp))
return false;
if (IS_MF(bp) && !CHIP_IS_E1x(bp)) {
/* Take function: tmp = func */
tmp = BP_ABS_FUNC(bp);
cfg = MF_CFG_RD(bp, func_ext_config[tmp].func_cfg);
cfg = !!(cfg & MACP_FUNC_CFG_PAUSE_ON_HOST_RING);
} else {
/* Take port: tmp = port */
tmp = BP_PORT(bp);
cfg = SHMEM_RD(bp,
dev_info.port_hw_config[tmp].generic_features);
cfg = !!(cfg & PORT_HW_CFG_PAUSE_ON_HOST_RING_ENABLED);
}
return cfg;
}
static void validate_set_si_mode(struct bnx2x *bp)
{
u8 func = BP_ABS_FUNC(bp);
u32 val;
val = MF_CFG_RD(bp, func_mf_config[func].mac_upper);
/* check for legal mac (upper bytes) */
if (val != 0xffff) {
bp->mf_mode = MULTI_FUNCTION_SI;
bp->mf_config[BP_VN(bp)] =
MF_CFG_RD(bp, func_mf_config[func].config);
} else
BNX2X_DEV_INFO("illegal MAC address for SI\n");
}
static int bnx2x_get_hwinfo(struct bnx2x *bp)
{
int /*abs*/func = BP_ABS_FUNC(bp);
int vn, mfw_vn;
u32 val = 0, val2 = 0;
int rc = 0;
/* Validate that chip access is feasible */
if (REG_RD(bp, MISC_REG_CHIP_NUM) == 0xffffffff) {
dev_err(&bp->pdev->dev,
"Chip read returns all Fs. Preventing probe from continuing\n");
return -EINVAL;
}
bnx2x_get_common_hwinfo(bp);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/*
* initialize IGU parameters
*/
if (CHIP_IS_E1x(bp)) {
bp->common.int_block = INT_BLOCK_HC;
bp->igu_dsb_id = DEF_SB_IGU_ID;
bp->igu_base_sb = 0;
} else {
bp->common.int_block = INT_BLOCK_IGU;
/* do not allow device reset during IGU info processing */
bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RESET);
val = REG_RD(bp, IGU_REG_BLOCK_CONFIGURATION);
if (val & IGU_BLOCK_CONFIGURATION_REG_BACKWARD_COMP_EN) {
int tout = 5000;
BNX2X_DEV_INFO("FORCING Normal Mode\n");
val &= ~(IGU_BLOCK_CONFIGURATION_REG_BACKWARD_COMP_EN);
REG_WR(bp, IGU_REG_BLOCK_CONFIGURATION, val);
REG_WR(bp, IGU_REG_RESET_MEMORIES, 0x7f);
while (tout && REG_RD(bp, IGU_REG_RESET_MEMORIES)) {
tout--;
usleep_range(1000, 2000);
}
if (REG_RD(bp, IGU_REG_RESET_MEMORIES)) {
dev_err(&bp->pdev->dev,
"FORCING Normal Mode failed!!!\n");
bnx2x_release_hw_lock(bp,
HW_LOCK_RESOURCE_RESET);
return -EPERM;
}
}
if (val & IGU_BLOCK_CONFIGURATION_REG_BACKWARD_COMP_EN) {
BNX2X_DEV_INFO("IGU Backward Compatible Mode\n");
bp->common.int_block |= INT_BLOCK_MODE_BW_COMP;
} else
BNX2X_DEV_INFO("IGU Normal Mode\n");
rc = bnx2x_get_igu_cam_info(bp);
bnx2x_release_hw_lock(bp, HW_LOCK_RESOURCE_RESET);
if (rc)
return rc;
}
/*
* set base FW non-default (fast path) status block id, this value is
* used to initialize the fw_sb_id saved on the fp/queue structure to
* determine the id used by the FW.
*/
if (CHIP_IS_E1x(bp))
bp->base_fw_ndsb = BP_PORT(bp) * FP_SB_MAX_E1x + BP_L_ID(bp);
else /*
* 57712 - we currently use one FW SB per IGU SB (Rx and Tx of
* the same queue are indicated on the same IGU SB). So we prefer
* FW and IGU SBs to be the same value.
*/
bp->base_fw_ndsb = bp->igu_base_sb;
BNX2X_DEV_INFO("igu_dsb_id %d igu_base_sb %d igu_sb_cnt %d\n"
"base_fw_ndsb %d\n", bp->igu_dsb_id, bp->igu_base_sb,
bp->igu_sb_cnt, bp->base_fw_ndsb);
/*
* Initialize MF configuration
*/
bp->mf_ov = 0;
bp->mf_mode = 0;
bp->mf_sub_mode = 0;
vn = BP_VN(bp);
mfw_vn = BP_FW_MB_IDX(bp);
if (!CHIP_IS_E1(bp) && !BP_NOMCP(bp)) {
BNX2X_DEV_INFO("shmem2base 0x%x, size %d, mfcfg offset %d\n",
bp->common.shmem2_base, SHMEM2_RD(bp, size),
(u32)offsetof(struct shmem2_region, mf_cfg_addr));
if (SHMEM2_HAS(bp, mf_cfg_addr))
bp->common.mf_cfg_base = SHMEM2_RD(bp, mf_cfg_addr);
else
bp->common.mf_cfg_base = bp->common.shmem_base +
offsetof(struct shmem_region, func_mb) +
E1H_FUNC_MAX * sizeof(struct drv_func_mb);
/*
* get mf configuration:
* 1. Existence of MF configuration
* 2. MAC address must be legal (check only upper bytes)
* for Switch-Independent mode;
* OVLAN must be legal for Switch-Dependent mode
* 3. SF_MODE configures specific MF mode
*/
if (bp->common.mf_cfg_base != SHMEM_MF_CFG_ADDR_NONE) {
/* get mf configuration */
val = SHMEM_RD(bp,
dev_info.shared_feature_config.config);
val &= SHARED_FEAT_CFG_FORCE_SF_MODE_MASK;
switch (val) {
case SHARED_FEAT_CFG_FORCE_SF_MODE_SWITCH_INDEPT:
validate_set_si_mode(bp);
break;
case SHARED_FEAT_CFG_FORCE_SF_MODE_AFEX_MODE:
if ((!CHIP_IS_E1x(bp)) &&
(MF_CFG_RD(bp, func_mf_config[func].
mac_upper) != 0xffff) &&
(SHMEM2_HAS(bp,
afex_driver_support))) {
bp->mf_mode = MULTI_FUNCTION_AFEX;
bp->mf_config[vn] = MF_CFG_RD(bp,
func_mf_config[func].config);
} else {
BNX2X_DEV_INFO("can not configure afex mode\n");
}
break;
case SHARED_FEAT_CFG_FORCE_SF_MODE_MF_ALLOWED:
/* get OV configuration */
val = MF_CFG_RD(bp,
func_mf_config[FUNC_0].e1hov_tag);
val &= FUNC_MF_CFG_E1HOV_TAG_MASK;
if (val != FUNC_MF_CFG_E1HOV_TAG_DEFAULT) {
bp->mf_mode = MULTI_FUNCTION_SD;
bp->mf_config[vn] = MF_CFG_RD(bp,
func_mf_config[func].config);
} else
BNX2X_DEV_INFO("illegal OV for SD\n");
break;
case SHARED_FEAT_CFG_FORCE_SF_MODE_BD_MODE:
bp->mf_mode = MULTI_FUNCTION_SD;
bp->mf_sub_mode = SUB_MF_MODE_BD;
bp->mf_config[vn] =
MF_CFG_RD(bp,
func_mf_config[func].config);
if (SHMEM2_HAS(bp, mtu_size)) {
int mtu_idx = BP_FW_MB_IDX(bp);
u16 mtu_size;
u32 mtu;
mtu = SHMEM2_RD(bp, mtu_size[mtu_idx]);
mtu_size = (u16)mtu;
DP(NETIF_MSG_IFUP, "Read MTU size %04x [%08x]\n",
mtu_size, mtu);
/* if valid: update device mtu */
if (((mtu_size + ETH_HLEN) >=
ETH_MIN_PACKET_SIZE) &&
(mtu_size <=
ETH_MAX_JUMBO_PACKET_SIZE))
bp->dev->mtu = mtu_size;
}
break;
case SHARED_FEAT_CFG_FORCE_SF_MODE_UFP_MODE:
bp->mf_mode = MULTI_FUNCTION_SD;
bp->mf_sub_mode = SUB_MF_MODE_UFP;
bp->mf_config[vn] =
MF_CFG_RD(bp,
func_mf_config[func].config);
break;
case SHARED_FEAT_CFG_FORCE_SF_MODE_FORCED_SF:
bp->mf_config[vn] = 0;
break;
case SHARED_FEAT_CFG_FORCE_SF_MODE_EXTENDED_MODE:
val2 = SHMEM_RD(bp,
dev_info.shared_hw_config.config_3);
val2 &= SHARED_HW_CFG_EXTENDED_MF_MODE_MASK;
switch (val2) {
case SHARED_HW_CFG_EXTENDED_MF_MODE_NPAR1_DOT_5:
validate_set_si_mode(bp);
bp->mf_sub_mode =
SUB_MF_MODE_NPAR1_DOT_5;
break;
default:
/* Unknown configuration */
bp->mf_config[vn] = 0;
BNX2X_DEV_INFO("unknown extended MF mode 0x%x\n",
val);
}
break;
default:
/* Unknown configuration: reset mf_config */
bp->mf_config[vn] = 0;
BNX2X_DEV_INFO("unknown MF mode 0x%x\n", val);
}
}
BNX2X_DEV_INFO("%s function mode\n",
IS_MF(bp) ? "multi" : "single");
switch (bp->mf_mode) {
case MULTI_FUNCTION_SD:
val = MF_CFG_RD(bp, func_mf_config[func].e1hov_tag) &
FUNC_MF_CFG_E1HOV_TAG_MASK;
if (val != FUNC_MF_CFG_E1HOV_TAG_DEFAULT) {
bp->mf_ov = val;
bp->path_has_ovlan = true;
BNX2X_DEV_INFO("MF OV for func %d is %d (0x%04x)\n",
func, bp->mf_ov, bp->mf_ov);
} else if ((bp->mf_sub_mode == SUB_MF_MODE_UFP) ||
(bp->mf_sub_mode == SUB_MF_MODE_BD)) {
dev_err(&bp->pdev->dev,
"Unexpected - no valid MF OV for func %d in UFP/BD mode\n",
func);
bp->path_has_ovlan = true;
} else {
dev_err(&bp->pdev->dev,
"No valid MF OV for func %d, aborting\n",
func);
return -EPERM;
}
break;
case MULTI_FUNCTION_AFEX:
BNX2X_DEV_INFO("func %d is in MF afex mode\n", func);
break;
case MULTI_FUNCTION_SI:
BNX2X_DEV_INFO("func %d is in MF switch-independent mode\n",
func);
break;
default:
if (vn) {
dev_err(&bp->pdev->dev,
"VN %d is in a single function mode, aborting\n",
vn);
return -EPERM;
}
break;
}
/* check if other port on the path needs ovlan:
* Since MF configuration is shared between ports
* Possible mixed modes are only
* {SF, SI} {SF, SD} {SD, SF} {SI, SF}
*/
if (CHIP_MODE_IS_4_PORT(bp) &&
!bp->path_has_ovlan &&
!IS_MF(bp) &&
bp->common.mf_cfg_base != SHMEM_MF_CFG_ADDR_NONE) {
u8 other_port = !BP_PORT(bp);
u8 other_func = BP_PATH(bp) + 2*other_port;
val = MF_CFG_RD(bp,
func_mf_config[other_func].e1hov_tag);
if (val != FUNC_MF_CFG_E1HOV_TAG_DEFAULT)
bp->path_has_ovlan = true;
}
}
/* adjust igu_sb_cnt to MF for E1H */
if (CHIP_IS_E1H(bp) && IS_MF(bp))
bp->igu_sb_cnt = min_t(u8, bp->igu_sb_cnt, E1H_MAX_MF_SB_COUNT);
/* port info */
bnx2x_get_port_hwinfo(bp);
/* Get MAC addresses */
bnx2x_get_mac_hwinfo(bp);
bnx2x_get_cnic_info(bp);
return rc;
}
static void bnx2x_read_fwinfo(struct bnx2x *bp)
{
int cnt, i, block_end, rodi;
char vpd_start[BNX2X_VPD_LEN+1];
char str_id_reg[VENDOR_ID_LEN+1];
char str_id_cap[VENDOR_ID_LEN+1];
char *vpd_data;
char *vpd_extended_data = NULL;
u8 len;
cnt = pci_read_vpd(bp->pdev, 0, BNX2X_VPD_LEN, vpd_start);
memset(bp->fw_ver, 0, sizeof(bp->fw_ver));
if (cnt < BNX2X_VPD_LEN)
goto out_not_found;
/* VPD RO tag should be first tag after identifier string, hence
* we should be able to find it in first BNX2X_VPD_LEN chars
*/
i = pci_vpd_find_tag(vpd_start, 0, BNX2X_VPD_LEN,
PCI_VPD_LRDT_RO_DATA);
if (i < 0)
goto out_not_found;
block_end = i + PCI_VPD_LRDT_TAG_SIZE +
pci_vpd_lrdt_size(&vpd_start[i]);
i += PCI_VPD_LRDT_TAG_SIZE;
if (block_end > BNX2X_VPD_LEN) {
vpd_extended_data = kmalloc(block_end, GFP_KERNEL);
if (vpd_extended_data == NULL)
goto out_not_found;
/* read rest of vpd image into vpd_extended_data */
memcpy(vpd_extended_data, vpd_start, BNX2X_VPD_LEN);
cnt = pci_read_vpd(bp->pdev, BNX2X_VPD_LEN,
block_end - BNX2X_VPD_LEN,
vpd_extended_data + BNX2X_VPD_LEN);
if (cnt < (block_end - BNX2X_VPD_LEN))
goto out_not_found;
vpd_data = vpd_extended_data;
} else
vpd_data = vpd_start;
/* now vpd_data holds full vpd content in both cases */
rodi = pci_vpd_find_info_keyword(vpd_data, i, block_end,
PCI_VPD_RO_KEYWORD_MFR_ID);
if (rodi < 0)
goto out_not_found;
len = pci_vpd_info_field_size(&vpd_data[rodi]);
if (len != VENDOR_ID_LEN)
goto out_not_found;
rodi += PCI_VPD_INFO_FLD_HDR_SIZE;
/* vendor specific info */
snprintf(str_id_reg, VENDOR_ID_LEN + 1, "%04x", PCI_VENDOR_ID_DELL);
snprintf(str_id_cap, VENDOR_ID_LEN + 1, "%04X", PCI_VENDOR_ID_DELL);
if (!strncmp(str_id_reg, &vpd_data[rodi], VENDOR_ID_LEN) ||
!strncmp(str_id_cap, &vpd_data[rodi], VENDOR_ID_LEN)) {
rodi = pci_vpd_find_info_keyword(vpd_data, i, block_end,
PCI_VPD_RO_KEYWORD_VENDOR0);
if (rodi >= 0) {
len = pci_vpd_info_field_size(&vpd_data[rodi]);
rodi += PCI_VPD_INFO_FLD_HDR_SIZE;
if (len < 32 && (len + rodi) <= BNX2X_VPD_LEN) {
memcpy(bp->fw_ver, &vpd_data[rodi], len);
bp->fw_ver[len] = ' ';
}
}
kfree(vpd_extended_data);
return;
}
out_not_found:
kfree(vpd_extended_data);
return;
}
static void bnx2x_set_modes_bitmap(struct bnx2x *bp)
{
u32 flags = 0;
if (CHIP_REV_IS_FPGA(bp))
SET_FLAGS(flags, MODE_FPGA);
else if (CHIP_REV_IS_EMUL(bp))
SET_FLAGS(flags, MODE_EMUL);
else
SET_FLAGS(flags, MODE_ASIC);
if (CHIP_MODE_IS_4_PORT(bp))
SET_FLAGS(flags, MODE_PORT4);
else
SET_FLAGS(flags, MODE_PORT2);
if (CHIP_IS_E2(bp))
SET_FLAGS(flags, MODE_E2);
else if (CHIP_IS_E3(bp)) {
SET_FLAGS(flags, MODE_E3);
if (CHIP_REV(bp) == CHIP_REV_Ax)
SET_FLAGS(flags, MODE_E3_A0);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
else /*if (CHIP_REV(bp) == CHIP_REV_Bx)*/
SET_FLAGS(flags, MODE_E3_B0 | MODE_COS3);
}
if (IS_MF(bp)) {
SET_FLAGS(flags, MODE_MF);
switch (bp->mf_mode) {
case MULTI_FUNCTION_SD:
SET_FLAGS(flags, MODE_MF_SD);
break;
case MULTI_FUNCTION_SI:
SET_FLAGS(flags, MODE_MF_SI);
break;
case MULTI_FUNCTION_AFEX:
SET_FLAGS(flags, MODE_MF_AFEX);
break;
}
} else
SET_FLAGS(flags, MODE_SF);
#if defined(__LITTLE_ENDIAN)
SET_FLAGS(flags, MODE_LITTLE_ENDIAN);
#else /*(__BIG_ENDIAN)*/
SET_FLAGS(flags, MODE_BIG_ENDIAN);
#endif
INIT_MODE_FLAGS(bp) = flags;
}
static int bnx2x_init_bp(struct bnx2x *bp)
{
int func;
int rc;
mutex_init(&bp->port.phy_mutex);
mutex_init(&bp->fw_mb_mutex);
mutex_init(&bp->drv_info_mutex);
sema_init(&bp->stats_lock, 1);
bp->drv_info_mng_owner = false;
INIT_LIST_HEAD(&bp->vlan_reg);
INIT_DELAYED_WORK(&bp->sp_task, bnx2x_sp_task);
INIT_DELAYED_WORK(&bp->sp_rtnl_task, bnx2x_sp_rtnl_task);
INIT_DELAYED_WORK(&bp->period_task, bnx2x_period_task);
INIT_DELAYED_WORK(&bp->iov_task, bnx2x_iov_task);
if (IS_PF(bp)) {
rc = bnx2x_get_hwinfo(bp);
if (rc)
return rc;
} else {
eth_zero_addr(bp->dev->dev_addr);
}
bnx2x_set_modes_bitmap(bp);
rc = bnx2x_alloc_mem_bp(bp);
if (rc)
return rc;
bnx2x_read_fwinfo(bp);
func = BP_FUNC(bp);
/* need to reset chip if undi was active */
if (IS_PF(bp) && !BP_NOMCP(bp)) {
/* init fw_seq */
bp->fw_seq =
SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
DRV_MSG_SEQ_NUMBER_MASK;
BNX2X_DEV_INFO("fw_seq 0x%08x\n", bp->fw_seq);
rc = bnx2x_prev_unload(bp);
if (rc) {
bnx2x_free_mem_bp(bp);
return rc;
}
}
if (CHIP_REV_IS_FPGA(bp))
dev_err(&bp->pdev->dev, "FPGA detected\n");
if (BP_NOMCP(bp) && (func == 0))
dev_err(&bp->pdev->dev, "MCP disabled, must load devices in order!\n");
bp->disable_tpa = disable_tpa;
bp->disable_tpa |= !!IS_MF_STORAGE_ONLY(bp);
/* Reduce memory usage in kdump environment by disabling TPA */
bp->disable_tpa |= is_kdump_kernel();
/* Set TPA flags */
if (bp->disable_tpa) {
bp->dev->hw_features &= ~NETIF_F_LRO;
bp->dev->features &= ~NETIF_F_LRO;
}
if (CHIP_IS_E1(bp))
bp->dropless_fc = 0;
else
bp->dropless_fc = dropless_fc | bnx2x_get_dropless_info(bp);
bp->mrrs = mrrs;
bp->tx_ring_size = IS_MF_STORAGE_ONLY(bp) ? 0 : MAX_TX_AVAIL;
if (IS_VF(bp))
bp->rx_ring_size = MAX_RX_AVAIL;
/* make sure that the numbers are in the right granularity */
bp->tx_ticks = (50 / BNX2X_BTR) * BNX2X_BTR;
bp->rx_ticks = (25 / BNX2X_BTR) * BNX2X_BTR;
bp->current_interval = CHIP_REV_IS_SLOW(bp) ? 5*HZ : HZ;
init_timer(&bp->timer);
bp->timer.expires = jiffies + bp->current_interval;
bp->timer.data = (unsigned long) bp;
bp->timer.function = bnx2x_timer;
if (SHMEM2_HAS(bp, dcbx_lldp_params_offset) &&
SHMEM2_HAS(bp, dcbx_lldp_dcbx_stat_offset) &&
SHMEM2_HAS(bp, dcbx_en) &&
SHMEM2_RD(bp, dcbx_lldp_params_offset) &&
SHMEM2_RD(bp, dcbx_lldp_dcbx_stat_offset) &&
SHMEM2_RD(bp, dcbx_en[BP_PORT(bp)])) {
bnx2x_dcbx_set_state(bp, true, BNX2X_DCBX_ENABLED_ON_NEG_ON);
bnx2x_dcbx_init_params(bp);
} else {
bnx2x_dcbx_set_state(bp, false, BNX2X_DCBX_ENABLED_OFF);
}
if (CHIP_IS_E1x(bp))
bp->cnic_base_cl_id = FP_SB_MAX_E1x;
else
bp->cnic_base_cl_id = FP_SB_MAX_E2;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* multiple tx priority */
if (IS_VF(bp))
bp->max_cos = 1;
else if (CHIP_IS_E1x(bp))
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bp->max_cos = BNX2X_MULTI_TX_COS_E1X;
else if (CHIP_IS_E2(bp) || CHIP_IS_E3A0(bp))
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bp->max_cos = BNX2X_MULTI_TX_COS_E2_E3A0;
else if (CHIP_IS_E3B0(bp))
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bp->max_cos = BNX2X_MULTI_TX_COS_E3B0;
else
BNX2X_ERR("unknown chip %x revision %x\n",
CHIP_NUM(bp), CHIP_REV(bp));
BNX2X_DEV_INFO("set bp->max_cos to %d\n", bp->max_cos);
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* We need at least one default status block for slow-path events,
* second status block for the L2 queue, and a third status block for
* CNIC if supported.
*/
if (IS_VF(bp))
bp->min_msix_vec_cnt = 1;
else if (CNIC_SUPPORT(bp))
bp->min_msix_vec_cnt = 3;
else /* PF w/o cnic */
bp->min_msix_vec_cnt = 2;
BNX2X_DEV_INFO("bp->min_msix_vec_cnt %d", bp->min_msix_vec_cnt);
bp->dump_preset_idx = 1;
if (CHIP_IS_E3B0(bp))
bp->flags |= PTP_SUPPORTED;
return rc;
}
/****************************************************************************
* General service functions
****************************************************************************/
/*
* net_device service functions
*/
/* called with rtnl_lock */
static int bnx2x_open(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
int rc;
bp->stats_init = true;
netif_carrier_off(dev);
bnx2x_set_power_state(bp, PCI_D0);
/* If parity had happen during the unload, then attentions
* and/or RECOVERY_IN_PROGRES may still be set. In this case we
* want the first function loaded on the current engine to
* complete the recovery.
* Parity recovery is only relevant for PF driver.
*/
if (IS_PF(bp)) {
int other_engine = BP_PATH(bp) ? 0 : 1;
bool other_load_status, load_status;
bool global = false;
other_load_status = bnx2x_get_load_status(bp, other_engine);
load_status = bnx2x_get_load_status(bp, BP_PATH(bp));
if (!bnx2x_reset_is_done(bp, BP_PATH(bp)) ||
bnx2x_chk_parity_attn(bp, &global, true)) {
do {
/* If there are attentions and they are in a
* global blocks, set the GLOBAL_RESET bit
* regardless whether it will be this function
* that will complete the recovery or not.
*/
if (global)
bnx2x_set_reset_global(bp);
/* Only the first function on the current
* engine should try to recover in open. In case
* of attentions in global blocks only the first
* in the chip should try to recover.
*/
if ((!load_status &&
(!global || !other_load_status)) &&
bnx2x_trylock_leader_lock(bp) &&
!bnx2x_leader_reset(bp)) {
netdev_info(bp->dev,
"Recovered in open\n");
break;
}
/* recovery has failed... */
bnx2x_set_power_state(bp, PCI_D3hot);
bp->recovery_state = BNX2X_RECOVERY_FAILED;
BNX2X_ERR("Recovery flow hasn't been properly completed yet. Try again later.\n"
"If you still see this message after a few retries then power cycle is required.\n");
return -EAGAIN;
} while (0);
}
}
bp->recovery_state = BNX2X_RECOVERY_DONE;
rc = bnx2x_nic_load(bp, LOAD_OPEN);
if (rc)
return rc;
if (IS_PF(bp))
udp_tunnel_get_rx_info(dev);
return 0;
}
/* called with rtnl_lock */
static int bnx2x_close(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
/* Unload the driver, release IRQs */
bnx2x_nic_unload(bp, UNLOAD_CLOSE, false);
return 0;
}
struct bnx2x_mcast_list_elem_group
{
struct list_head mcast_group_link;
struct bnx2x_mcast_list_elem mcast_elems[];
};
#define MCAST_ELEMS_PER_PG \
((PAGE_SIZE - sizeof(struct bnx2x_mcast_list_elem_group)) / \
sizeof(struct bnx2x_mcast_list_elem))
static void bnx2x_free_mcast_macs_list(struct list_head *mcast_group_list)
{
struct bnx2x_mcast_list_elem_group *current_mcast_group;
while (!list_empty(mcast_group_list)) {
current_mcast_group = list_first_entry(mcast_group_list,
struct bnx2x_mcast_list_elem_group,
mcast_group_link);
list_del(&current_mcast_group->mcast_group_link);
free_page((unsigned long)current_mcast_group);
}
}
static int bnx2x_init_mcast_macs_list(struct bnx2x *bp,
struct bnx2x_mcast_ramrod_params *p,
struct list_head *mcast_group_list)
{
struct bnx2x_mcast_list_elem *mc_mac;
struct netdev_hw_addr *ha;
struct bnx2x_mcast_list_elem_group *current_mcast_group = NULL;
int mc_count = netdev_mc_count(bp->dev);
int offset = 0;
INIT_LIST_HEAD(&p->mcast_list);
netdev_for_each_mc_addr(ha, bp->dev) {
if (!offset) {
current_mcast_group =
(struct bnx2x_mcast_list_elem_group *)
__get_free_page(GFP_ATOMIC);
if (!current_mcast_group) {
bnx2x_free_mcast_macs_list(mcast_group_list);
BNX2X_ERR("Failed to allocate mc MAC list\n");
return -ENOMEM;
}
list_add(&current_mcast_group->mcast_group_link,
mcast_group_list);
}
mc_mac = &current_mcast_group->mcast_elems[offset];
mc_mac->mac = bnx2x_mc_addr(ha);
list_add_tail(&mc_mac->link, &p->mcast_list);
offset++;
if (offset == MCAST_ELEMS_PER_PG)
offset = 0;
}
p->mcast_list_len = mc_count;
return 0;
}
/**
* bnx2x_set_uc_list - configure a new unicast MACs list.
*
* @bp: driver handle
*
* We will use zero (0) as a MAC type for these MACs.
*/
static int bnx2x_set_uc_list(struct bnx2x *bp)
{
int rc;
struct net_device *dev = bp->dev;
struct netdev_hw_addr *ha;
struct bnx2x_vlan_mac_obj *mac_obj = &bp->sp_objs->mac_obj;
unsigned long ramrod_flags = 0;
/* First schedule a cleanup up of old configuration */
rc = bnx2x_del_all_macs(bp, mac_obj, BNX2X_UC_LIST_MAC, false);
if (rc < 0) {
BNX2X_ERR("Failed to schedule DELETE operations: %d\n", rc);
return rc;
}
netdev_for_each_uc_addr(ha, dev) {
rc = bnx2x_set_mac_one(bp, bnx2x_uc_addr(ha), mac_obj, true,
BNX2X_UC_LIST_MAC, &ramrod_flags);
if (rc == -EEXIST) {
DP(BNX2X_MSG_SP,
"Failed to schedule ADD operations: %d\n", rc);
/* do not treat adding same MAC as error */
rc = 0;
} else if (rc < 0) {
BNX2X_ERR("Failed to schedule ADD operations: %d\n",
rc);
return rc;
}
}
/* Execute the pending commands */
__set_bit(RAMROD_CONT, &ramrod_flags);
return bnx2x_set_mac_one(bp, NULL, mac_obj, false /* don't care */,
BNX2X_UC_LIST_MAC, &ramrod_flags);
}
static int bnx2x_set_mc_list_e1x(struct bnx2x *bp)
{
LIST_HEAD(mcast_group_list);
struct net_device *dev = bp->dev;
struct bnx2x_mcast_ramrod_params rparam = {NULL};
int rc = 0;
rparam.mcast_obj = &bp->mcast_obj;
/* first, clear all configured multicast MACs */
rc = bnx2x_config_mcast(bp, &rparam, BNX2X_MCAST_CMD_DEL);
if (rc < 0) {
BNX2X_ERR("Failed to clear multicast configuration: %d\n", rc);
return rc;
}
/* then, configure a new MACs list */
if (netdev_mc_count(dev)) {
rc = bnx2x_init_mcast_macs_list(bp, &rparam, &mcast_group_list);
if (rc)
return rc;
/* Now add the new MACs */
rc = bnx2x_config_mcast(bp, &rparam,
BNX2X_MCAST_CMD_ADD);
if (rc < 0)
BNX2X_ERR("Failed to set a new multicast configuration: %d\n",
rc);
bnx2x_free_mcast_macs_list(&mcast_group_list);
}
return rc;
}
static int bnx2x_set_mc_list(struct bnx2x *bp)
{
LIST_HEAD(mcast_group_list);
struct bnx2x_mcast_ramrod_params rparam = {NULL};
struct net_device *dev = bp->dev;
int rc = 0;
/* On older adapters, we need to flush and re-add filters */
if (CHIP_IS_E1x(bp))
return bnx2x_set_mc_list_e1x(bp);
rparam.mcast_obj = &bp->mcast_obj;
if (netdev_mc_count(dev)) {
rc = bnx2x_init_mcast_macs_list(bp, &rparam, &mcast_group_list);
if (rc)
return rc;
/* Override the curently configured set of mc filters */
rc = bnx2x_config_mcast(bp, &rparam,
BNX2X_MCAST_CMD_SET);
if (rc < 0)
BNX2X_ERR("Failed to set a new multicast configuration: %d\n",
rc);
bnx2x_free_mcast_macs_list(&mcast_group_list);
} else {
/* If no mc addresses are required, flush the configuration */
rc = bnx2x_config_mcast(bp, &rparam, BNX2X_MCAST_CMD_DEL);
if (rc)
BNX2X_ERR("Failed to clear multicast configuration %d\n",
rc);
}
return rc;
}
/* If bp->state is OPEN, should be called with netif_addr_lock_bh() */
static void bnx2x_set_rx_mode(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
if (bp->state != BNX2X_STATE_OPEN) {
DP(NETIF_MSG_IFUP, "state is %x, returning\n", bp->state);
return;
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
} else {
/* Schedule an SP task to handle rest of change */
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_RX_MODE,
NETIF_MSG_IFUP);
}
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
}
void bnx2x_set_rx_mode_inner(struct bnx2x *bp)
{
u32 rx_mode = BNX2X_RX_MODE_NORMAL;
DP(NETIF_MSG_IFUP, "dev->flags = %x\n", bp->dev->flags);
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
netif_addr_lock_bh(bp->dev);
if (bp->dev->flags & IFF_PROMISC) {
rx_mode = BNX2X_RX_MODE_PROMISC;
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
} else if ((bp->dev->flags & IFF_ALLMULTI) ||
((netdev_mc_count(bp->dev) > BNX2X_MAX_MULTICAST) &&
CHIP_IS_E1(bp))) {
rx_mode = BNX2X_RX_MODE_ALLMULTI;
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
} else {
if (IS_PF(bp)) {
/* some multicasts */
if (bnx2x_set_mc_list(bp) < 0)
rx_mode = BNX2X_RX_MODE_ALLMULTI;
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
/* release bh lock, as bnx2x_set_uc_list might sleep */
netif_addr_unlock_bh(bp->dev);
if (bnx2x_set_uc_list(bp) < 0)
rx_mode = BNX2X_RX_MODE_PROMISC;
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
netif_addr_lock_bh(bp->dev);
} else {
/* configuring mcast to a vf involves sleeping (when we
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
* wait for the pf's response).
*/
bnx2x_schedule_sp_rtnl(bp,
BNX2X_SP_RTNL_VFPF_MCAST, 0);
}
}
bp->rx_mode = rx_mode;
/* handle ISCSI SD mode */
if (IS_MF_ISCSI_ONLY(bp))
bp->rx_mode = BNX2X_RX_MODE_NONE;
/* Schedule the rx_mode command */
if (test_bit(BNX2X_FILTER_RX_MODE_PENDING, &bp->sp_state)) {
set_bit(BNX2X_FILTER_RX_MODE_SCHED, &bp->sp_state);
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
netif_addr_unlock_bh(bp->dev);
return;
}
if (IS_PF(bp)) {
bnx2x_set_storm_rx_mode(bp);
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
netif_addr_unlock_bh(bp->dev);
} else {
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
/* VF will need to request the PF to make this change, and so
* the VF needs to release the bottom-half lock prior to the
* request (as it will likely require sleep on the VF side)
*/
bnx2x: Revising locking scheme for MAC configuration On very rare occasions, repeated load/unload stress test in the presence of our storage driver (bnx2i/bnx2fc) causes a kernel panic in bnx2x code (NULL pointer dereference). Stack traces indicate the issue happens during MAC configuration; thorough code review showed that indeed several races exist in which one thread can iterate over the list of configured MACs while another deletes entries from the same list. This patch adds a varient on the single-writer/Multiple-reader lock mechanism - It utilizes an already exsiting bottom-half lock, using it so that Whenever a writer is unable to continue due to the existence of another writer/reader, it pends its request for future deliverance. The writer / last readers will check for the existence of such requests and perform them instead of the original initiator. This prevents the writer from having to sleep while waiting for the lock to be accessible, which might cause deadlocks given the locks already held by the writer. Another result of this patch is that setting of Rx Mode is now made in sleepable context - Setting of Rx Mode is made under a bottom-half lock, which was always nontrivial for the bnx2x driver, as the HW/FW configuration requires wait for completions. Since sleep was impossible (due to the sleepless-context), various mechanisms were utilized to prevent the calling thread from sleep, but the truth was that when the caller thread (i.e, the one calling ndo_set_rx_mode()) returned, the Rx mode was still not set in HW/FW. bnx2x_set_rx_mode() will now overtly schedule for the Rx changes to be configured by the sp_rtnl_task which hold the RTNL lock and is sleepable context. Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-08-01 23:30:59 +09:00
netif_addr_unlock_bh(bp->dev);
bnx2x_vfpf_storm_rx_mode(bp);
}
}
/* called with rtnl_lock */
static int bnx2x_mdio_read(struct net_device *netdev, int prtad,
int devad, u16 addr)
{
struct bnx2x *bp = netdev_priv(netdev);
u16 value;
int rc;
DP(NETIF_MSG_LINK, "mdio_read: prtad 0x%x, devad 0x%x, addr 0x%x\n",
prtad, devad, addr);
/* The HW expects different devad if CL22 is used */
devad = (devad == MDIO_DEVAD_NONE) ? DEFAULT_PHY_DEV_ADDR : devad;
bnx2x_acquire_phy_lock(bp);
rc = bnx2x_phy_read(&bp->link_params, prtad, devad, addr, &value);
bnx2x_release_phy_lock(bp);
DP(NETIF_MSG_LINK, "mdio_read_val 0x%x rc = 0x%x\n", value, rc);
if (!rc)
rc = value;
return rc;
}
/* called with rtnl_lock */
static int bnx2x_mdio_write(struct net_device *netdev, int prtad, int devad,
u16 addr, u16 value)
{
struct bnx2x *bp = netdev_priv(netdev);
int rc;
DP(NETIF_MSG_LINK,
"mdio_write: prtad 0x%x, devad 0x%x, addr 0x%x, value 0x%x\n",
prtad, devad, addr, value);
/* The HW expects different devad if CL22 is used */
devad = (devad == MDIO_DEVAD_NONE) ? DEFAULT_PHY_DEV_ADDR : devad;
bnx2x_acquire_phy_lock(bp);
rc = bnx2x_phy_write(&bp->link_params, prtad, devad, addr, value);
bnx2x_release_phy_lock(bp);
return rc;
}
/* called with rtnl_lock */
static int bnx2x_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
{
struct bnx2x *bp = netdev_priv(dev);
struct mii_ioctl_data *mdio = if_mii(ifr);
if (!netif_running(dev))
return -EAGAIN;
switch (cmd) {
case SIOCSHWTSTAMP:
return bnx2x_hwtstamp_ioctl(bp, ifr);
default:
DP(NETIF_MSG_LINK, "ioctl: phy id 0x%x, reg 0x%x, val_in 0x%x\n",
mdio->phy_id, mdio->reg_num, mdio->val_in);
return mdio_mii_ioctl(&bp->mdio, mdio, cmd);
}
}
#ifdef CONFIG_NET_POLL_CONTROLLER
static void poll_bnx2x(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
int i;
for_each_eth_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
napi_schedule(&bnx2x_fp(bp, fp->index, napi));
}
}
#endif
static int bnx2x_validate_addr(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
/* query the bulletin board for mac address configured by the PF */
if (IS_VF(bp))
bnx2x_sample_bulletin(bp);
if (!is_valid_ether_addr(dev->dev_addr)) {
BNX2X_ERR("Non-valid Ethernet address\n");
return -EADDRNOTAVAIL;
}
return 0;
}
static int bnx2x_get_phys_port_id(struct net_device *netdev,
struct netdev_phys_item_id *ppid)
{
struct bnx2x *bp = netdev_priv(netdev);
if (!(bp->flags & HAS_PHYS_PORT_ID))
return -EOPNOTSUPP;
ppid->id_len = sizeof(bp->phys_port_id);
memcpy(ppid->id, bp->phys_port_id, ppid->id_len);
return 0;
}
static netdev_features_t bnx2x_features_check(struct sk_buff *skb,
struct net_device *dev,
netdev_features_t features)
{
features = vlan_features_check(skb, features);
return vxlan_features_check(skb, features);
}
static int __bnx2x_vlan_configure_vid(struct bnx2x *bp, u16 vid, bool add)
{
int rc;
if (IS_PF(bp)) {
unsigned long ramrod_flags = 0;
__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
rc = bnx2x_set_vlan_one(bp, vid, &bp->sp_objs->vlan_obj,
add, &ramrod_flags);
} else {
rc = bnx2x_vfpf_update_vlan(bp, vid, bp->fp->index, add);
}
return rc;
}
static int bnx2x_vlan_configure_vid_list(struct bnx2x *bp)
{
struct bnx2x_vlan_entry *vlan;
int rc = 0;
/* Configure all non-configured entries */
list_for_each_entry(vlan, &bp->vlan_reg, link) {
if (vlan->hw)
continue;
if (bp->vlan_cnt >= bp->vlan_credit)
return -ENOBUFS;
rc = __bnx2x_vlan_configure_vid(bp, vlan->vid, true);
if (rc) {
BNX2X_ERR("Unable to config VLAN %d\n", vlan->vid);
return rc;
}
DP(NETIF_MSG_IFUP, "HW configured for VLAN %d\n", vlan->vid);
vlan->hw = true;
bp->vlan_cnt++;
}
return 0;
}
static void bnx2x_vlan_configure(struct bnx2x *bp, bool set_rx_mode)
{
bool need_accept_any_vlan;
need_accept_any_vlan = !!bnx2x_vlan_configure_vid_list(bp);
if (bp->accept_any_vlan != need_accept_any_vlan) {
bp->accept_any_vlan = need_accept_any_vlan;
DP(NETIF_MSG_IFUP, "Accept all VLAN %s\n",
bp->accept_any_vlan ? "raised" : "cleared");
if (set_rx_mode) {
if (IS_PF(bp))
bnx2x_set_rx_mode_inner(bp);
else
bnx2x_vfpf_storm_rx_mode(bp);
}
}
}
int bnx2x_vlan_reconfigure_vid(struct bnx2x *bp)
{
struct bnx2x_vlan_entry *vlan;
/* The hw forgot all entries after reload */
list_for_each_entry(vlan, &bp->vlan_reg, link)
vlan->hw = false;
bp->vlan_cnt = 0;
/* Don't set rx mode here. Our caller will do it. */
bnx2x_vlan_configure(bp, false);
return 0;
}
static int bnx2x_vlan_rx_add_vid(struct net_device *dev, __be16 proto, u16 vid)
{
struct bnx2x *bp = netdev_priv(dev);
struct bnx2x_vlan_entry *vlan;
DP(NETIF_MSG_IFUP, "Adding VLAN %d\n", vid);
vlan = kmalloc(sizeof(*vlan), GFP_KERNEL);
if (!vlan)
return -ENOMEM;
vlan->vid = vid;
vlan->hw = false;
list_add_tail(&vlan->link, &bp->vlan_reg);
if (netif_running(dev))
bnx2x_vlan_configure(bp, true);
return 0;
}
static int bnx2x_vlan_rx_kill_vid(struct net_device *dev, __be16 proto, u16 vid)
{
struct bnx2x *bp = netdev_priv(dev);
struct bnx2x_vlan_entry *vlan;
bool found = false;
int rc = 0;
DP(NETIF_MSG_IFUP, "Removing VLAN %d\n", vid);
list_for_each_entry(vlan, &bp->vlan_reg, link)
if (vlan->vid == vid) {
found = true;
break;
}
if (!found) {
BNX2X_ERR("Unable to kill VLAN %d - not found\n", vid);
return -EINVAL;
}
if (netif_running(dev) && vlan->hw) {
rc = __bnx2x_vlan_configure_vid(bp, vid, false);
DP(NETIF_MSG_IFUP, "HW deconfigured for VLAN %d\n", vid);
bp->vlan_cnt--;
}
list_del(&vlan->link);
kfree(vlan);
if (netif_running(dev))
bnx2x_vlan_configure(bp, true);
DP(NETIF_MSG_IFUP, "Removing VLAN result %d\n", rc);
return rc;
}
static const struct net_device_ops bnx2x_netdev_ops = {
.ndo_open = bnx2x_open,
.ndo_stop = bnx2x_close,
.ndo_start_xmit = bnx2x_start_xmit,
.ndo_select_queue = bnx2x_select_queue,
.ndo_set_rx_mode = bnx2x_set_rx_mode,
.ndo_set_mac_address = bnx2x_change_mac_addr,
.ndo_validate_addr = bnx2x_validate_addr,
.ndo_do_ioctl = bnx2x_ioctl,
.ndo_change_mtu = bnx2x_change_mtu,
.ndo_fix_features = bnx2x_fix_features,
.ndo_set_features = bnx2x_set_features,
.ndo_tx_timeout = bnx2x_tx_timeout,
.ndo_vlan_rx_add_vid = bnx2x_vlan_rx_add_vid,
.ndo_vlan_rx_kill_vid = bnx2x_vlan_rx_kill_vid,
#ifdef CONFIG_NET_POLL_CONTROLLER
.ndo_poll_controller = poll_bnx2x,
#endif
.ndo_setup_tc = __bnx2x_setup_tc,
#ifdef CONFIG_BNX2X_SRIOV
.ndo_set_vf_mac = bnx2x_set_vf_mac,
.ndo_set_vf_vlan = bnx2x_set_vf_vlan,
.ndo_get_vf_config = bnx2x_get_vf_config,
#endif
#ifdef NETDEV_FCOE_WWNN
.ndo_fcoe_get_wwn = bnx2x_fcoe_get_wwn,
#endif
.ndo_get_phys_port_id = bnx2x_get_phys_port_id,
.ndo_set_vf_link_state = bnx2x_set_vf_link_state,
.ndo_features_check = bnx2x_features_check,
.ndo_udp_tunnel_add = bnx2x_udp_tunnel_add,
.ndo_udp_tunnel_del = bnx2x_udp_tunnel_del,
};
static int bnx2x_set_coherency_mask(struct bnx2x *bp)
{
struct device *dev = &bp->pdev->dev;
Merge branch 'for-linus-dma-masks' of git://git.linaro.org/people/rmk/linux-arm Pull DMA mask updates from Russell King: "This series cleans up the handling of DMA masks in a lot of drivers, fixing some bugs as we go. Some of the more serious errors include: - drivers which only set their coherent DMA mask if the attempt to set the streaming mask fails. - drivers which test for a NULL dma mask pointer, and then set the dma mask pointer to a location in their module .data section - which will cause problems if the module is reloaded. To counter these, I have introduced two helper functions: - dma_set_mask_and_coherent() takes care of setting both the streaming and coherent masks at the same time, with the correct error handling as specified by the API. - dma_coerce_mask_and_coherent() which resolves the problem of drivers forcefully setting DMA masks. This is more a marker for future work to further clean these locations up - the code which creates the devices really should be initialising these, but to fix that in one go along with this change could potentially be very disruptive. The last thing this series does is prise away some of Linux's addition to "DMA addresses are physical addresses and RAM always starts at zero". We have ARM LPAE systems where all system memory is above 4GB physical, hence having DMA masks interpreted by (eg) the block layers as describing physical addresses in the range 0..DMAMASK fails on these platforms. Santosh Shilimkar addresses this in this series; the patches were copied to the appropriate people multiple times but were ignored. Fixing this also gets rid of some ARM weirdness in the setup of the max*pfn variables, and brings ARM into line with every other Linux architecture as far as those go" * 'for-linus-dma-masks' of git://git.linaro.org/people/rmk/linux-arm: (52 commits) ARM: 7805/1: mm: change max*pfn to include the physical offset of memory ARM: 7797/1: mmc: Use dma_max_pfn(dev) helper for bounce_limit calculations ARM: 7796/1: scsi: Use dma_max_pfn(dev) helper for bounce_limit calculations ARM: 7795/1: mm: dma-mapping: Add dma_max_pfn(dev) helper function ARM: 7794/1: block: Rename parameter dma_mask to max_addr for blk_queue_bounce_limit() ARM: DMA-API: better handing of DMA masks for coherent allocations ARM: 7857/1: dma: imx-sdma: setup dma mask DMA-API: firmware/google/gsmi.c: avoid direct access to DMA masks DMA-API: dcdbas: update DMA mask handing DMA-API: dma: edma.c: no need to explicitly initialize DMA masks DMA-API: usb: musb: use platform_device_register_full() to avoid directly messing with dma masks DMA-API: crypto: remove last references to 'static struct device *dev' DMA-API: crypto: fix ixp4xx crypto platform device support DMA-API: others: use dma_set_coherent_mask() DMA-API: staging: use dma_set_coherent_mask() DMA-API: usb: use new dma_coerce_mask_and_coherent() DMA-API: usb: use dma_set_coherent_mask() DMA-API: parport: parport_pc.c: use dma_coerce_mask_and_coherent() DMA-API: net: octeon: use dma_coerce_mask_and_coherent() DMA-API: net: nxp/lpc_eth: use dma_coerce_mask_and_coherent() ...
2013-11-14 07:55:21 +09:00
if (dma_set_mask_and_coherent(dev, DMA_BIT_MASK(64)) != 0 &&
dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32)) != 0) {
dev_err(dev, "System does not support DMA, aborting\n");
return -EIO;
}
return 0;
}
static void bnx2x_disable_pcie_error_reporting(struct bnx2x *bp)
{
if (bp->flags & AER_ENABLED) {
pci_disable_pcie_error_reporting(bp->pdev);
bp->flags &= ~AER_ENABLED;
}
}
static int bnx2x_init_dev(struct bnx2x *bp, struct pci_dev *pdev,
struct net_device *dev, unsigned long board_type)
{
int rc;
u32 pci_cfg_dword;
bool chip_is_e1x = (board_type == BCM57710 ||
board_type == BCM57711 ||
board_type == BCM57711E);
SET_NETDEV_DEV(dev, &pdev->dev);
bp->dev = dev;
bp->pdev = pdev;
rc = pci_enable_device(pdev);
if (rc) {
dev_err(&bp->pdev->dev,
"Cannot enable PCI device, aborting\n");
goto err_out;
}
if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM)) {
dev_err(&bp->pdev->dev,
"Cannot find PCI device base address, aborting\n");
rc = -ENODEV;
goto err_out_disable;
}
if (IS_PF(bp) && !(pci_resource_flags(pdev, 2) & IORESOURCE_MEM)) {
dev_err(&bp->pdev->dev, "Cannot find second PCI device base address, aborting\n");
rc = -ENODEV;
goto err_out_disable;
}
pci_read_config_dword(pdev, PCICFG_REVISION_ID_OFFSET, &pci_cfg_dword);
if ((pci_cfg_dword & PCICFG_REVESION_ID_MASK) ==
PCICFG_REVESION_ID_ERROR_VAL) {
pr_err("PCI device error, probably due to fan failure, aborting\n");
rc = -ENODEV;
goto err_out_disable;
}
if (atomic_read(&pdev->enable_cnt) == 1) {
rc = pci_request_regions(pdev, DRV_MODULE_NAME);
if (rc) {
dev_err(&bp->pdev->dev,
"Cannot obtain PCI resources, aborting\n");
goto err_out_disable;
}
pci_set_master(pdev);
pci_save_state(pdev);
}
if (IS_PF(bp)) {
if (!pdev->pm_cap) {
dev_err(&bp->pdev->dev,
"Cannot find power management capability, aborting\n");
rc = -EIO;
goto err_out_release;
}
}
if (!pci_is_pcie(pdev)) {
dev_err(&bp->pdev->dev, "Not PCI Express, aborting\n");
rc = -EIO;
goto err_out_release;
}
rc = bnx2x_set_coherency_mask(bp);
if (rc)
goto err_out_release;
dev->mem_start = pci_resource_start(pdev, 0);
dev->base_addr = dev->mem_start;
dev->mem_end = pci_resource_end(pdev, 0);
dev->irq = pdev->irq;
bp->regview = pci_ioremap_bar(pdev, 0);
if (!bp->regview) {
dev_err(&bp->pdev->dev,
"Cannot map register space, aborting\n");
rc = -ENOMEM;
goto err_out_release;
}
/* In E1/E1H use pci device function given by kernel.
* In E2/E3 read physical function from ME register since these chips
* support Physical Device Assignment where kernel BDF maybe arbitrary
* (depending on hypervisor).
*/
if (chip_is_e1x) {
bp->pf_num = PCI_FUNC(pdev->devfn);
} else {
/* chip is E2/3*/
pci_read_config_dword(bp->pdev,
PCICFG_ME_REGISTER, &pci_cfg_dword);
bp->pf_num = (u8)((pci_cfg_dword & ME_REG_ABS_PF_NUM) >>
ME_REG_ABS_PF_NUM_SHIFT);
}
BNX2X_DEV_INFO("me reg PF num: %d\n", bp->pf_num);
/* clean indirect addresses */
pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS,
PCICFG_VENDOR_ID_OFFSET);
/* Set PCIe reset type to fundamental for EEH recovery */
pdev->needs_freset = 1;
/* AER (Advanced Error reporting) configuration */
rc = pci_enable_pcie_error_reporting(pdev);
if (!rc)
bp->flags |= AER_ENABLED;
else
BNX2X_DEV_INFO("Failed To configure PCIe AER [%d]\n", rc);
/*
* Clean the following indirect addresses for all functions since it
* is not used by the driver.
*/
if (IS_PF(bp)) {
REG_WR(bp, PXP2_REG_PGL_ADDR_88_F0, 0);
REG_WR(bp, PXP2_REG_PGL_ADDR_8C_F0, 0);
REG_WR(bp, PXP2_REG_PGL_ADDR_90_F0, 0);
REG_WR(bp, PXP2_REG_PGL_ADDR_94_F0, 0);
if (chip_is_e1x) {
REG_WR(bp, PXP2_REG_PGL_ADDR_88_F1, 0);
REG_WR(bp, PXP2_REG_PGL_ADDR_8C_F1, 0);
REG_WR(bp, PXP2_REG_PGL_ADDR_90_F1, 0);
REG_WR(bp, PXP2_REG_PGL_ADDR_94_F1, 0);
}
/* Enable internal target-read (in case we are probed after PF
* FLR). Must be done prior to any BAR read access. Only for
* 57712 and up
*/
if (!chip_is_e1x)
REG_WR(bp,
PGLUE_B_REG_INTERNAL_PFID_ENABLE_TARGET_READ, 1);
}
dev->watchdog_timeo = TX_TIMEOUT;
dev->netdev_ops = &bnx2x_netdev_ops;
bnx2x_set_ethtool_ops(bp, dev);
dev->priv_flags |= IFF_UNICAST_FLT;
dev->hw_features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
use FW 7.2.16 The patch integrates FW 7.2.16 HSI and implements driver part of GRO flow. FW 7.2.16 adds the ability to aggregate packets for GRO (and not just LRO) and also fixes some bugs. 1. Added new aggregation mode: GRO. In this mode packets are aggregated such that the original packets can be reconstructed by the OS. 2. 57712 HW bug workaround - initialized all CAM TM registers to 0x32. 3. Adding the FCoE statistics structures to the BNX2X HSI. 4. Wrong configuration of TX HW input buffer size may cause theoretical performance effect. Performed configuration fix. 5. FCOE - Arrival of packets beyond task IO size can lead to crash. Fix firmware data-in flow. 6. iSCSI - In rare cases of on-chip termination the graceful termination timer hangs, and the termination doesn't complete. Firmware fix to MSL timer tolerance. 7. iSCSI - Chip hangs when target sends FIN out-of-order or with isles open at the initiator side. Firmware implementation corrected to drop FIN received out-of-order or with isles still open. 8. iSCSI - Chip hangs when in case of retransmission not aligned to 4-bytes from the beginning of iSCSI PDU. Firmware implementation corrected to support arbitrary aligned retransmissions. 9. iSCSI - Arrival of target-initiated NOP-IN during intense ISCSI traffic might lead to crash. Firmware fix to relevant flow. Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-02-20 18:59:08 +09:00
NETIF_F_TSO | NETIF_F_TSO_ECN | NETIF_F_TSO6 |
NETIF_F_RXCSUM | NETIF_F_LRO | NETIF_F_GRO |
NETIF_F_RXHASH | NETIF_F_HW_VLAN_CTAG_TX;
if (!chip_is_e1x) {
dev->hw_features |= NETIF_F_GSO_GRE | NETIF_F_GSO_GRE_CSUM |
NETIF_F_GSO_IPXIP4 |
NETIF_F_GSO_UDP_TUNNEL |
NETIF_F_GSO_UDP_TUNNEL_CSUM |
NETIF_F_GSO_PARTIAL;
dev->hw_enc_features =
NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM | NETIF_F_SG |
NETIF_F_TSO | NETIF_F_TSO_ECN | NETIF_F_TSO6 |
NETIF_F_GSO_IPXIP4 |
NETIF_F_GSO_GRE | NETIF_F_GSO_GRE_CSUM |
NETIF_F_GSO_UDP_TUNNEL | NETIF_F_GSO_UDP_TUNNEL_CSUM |
NETIF_F_GSO_PARTIAL;
dev->gso_partial_features = NETIF_F_GSO_GRE_CSUM |
NETIF_F_GSO_UDP_TUNNEL_CSUM;
}
dev->vlan_features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
NETIF_F_TSO | NETIF_F_TSO_ECN | NETIF_F_TSO6 | NETIF_F_HIGHDMA;
if (IS_PF(bp)) {
if (chip_is_e1x)
bp->accept_any_vlan = true;
else
dev->hw_features |= NETIF_F_HW_VLAN_CTAG_FILTER;
}
/* For VF we'll know whether to enable VLAN filtering after
* getting a response to CHANNEL_TLV_ACQUIRE from PF.
*/
dev->features |= dev->hw_features | NETIF_F_HW_VLAN_CTAG_RX;
dev->features |= NETIF_F_HIGHDMA;
/* Add Loopback capability to the device */
dev->hw_features |= NETIF_F_LOOPBACK;
#ifdef BCM_DCBNL
dev->dcbnl_ops = &bnx2x_dcbnl_ops;
#endif
/* get_port_hwinfo() will set prtad and mmds properly */
bp->mdio.prtad = MDIO_PRTAD_NONE;
bp->mdio.mmds = 0;
bp->mdio.mode_support = MDIO_SUPPORTS_C45 | MDIO_EMULATE_C22;
bp->mdio.dev = dev;
bp->mdio.mdio_read = bnx2x_mdio_read;
bp->mdio.mdio_write = bnx2x_mdio_write;
return 0;
err_out_release:
if (atomic_read(&pdev->enable_cnt) == 1)
pci_release_regions(pdev);
err_out_disable:
pci_disable_device(pdev);
err_out:
return rc;
}
static int bnx2x_check_firmware(struct bnx2x *bp)
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
{
const struct firmware *firmware = bp->firmware;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
struct bnx2x_fw_file_hdr *fw_hdr;
struct bnx2x_fw_file_section *sections;
u32 offset, len, num_ops;
__be16 *ops_offsets;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
int i;
const u8 *fw_ver;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
if (firmware->size < sizeof(struct bnx2x_fw_file_hdr)) {
BNX2X_ERR("Wrong FW size\n");
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
return -EINVAL;
}
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
fw_hdr = (struct bnx2x_fw_file_hdr *)firmware->data;
sections = (struct bnx2x_fw_file_section *)fw_hdr;
/* Make sure none of the offsets and sizes make us read beyond
* the end of the firmware data */
for (i = 0; i < sizeof(*fw_hdr) / sizeof(*sections); i++) {
offset = be32_to_cpu(sections[i].offset);
len = be32_to_cpu(sections[i].len);
if (offset + len > firmware->size) {
BNX2X_ERR("Section %d length is out of bounds\n", i);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
return -EINVAL;
}
}
/* Likewise for the init_ops offsets */
offset = be32_to_cpu(fw_hdr->init_ops_offsets.offset);
ops_offsets = (__force __be16 *)(firmware->data + offset);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
num_ops = be32_to_cpu(fw_hdr->init_ops.len) / sizeof(struct raw_op);
for (i = 0; i < be32_to_cpu(fw_hdr->init_ops_offsets.len) / 2; i++) {
if (be16_to_cpu(ops_offsets[i]) > num_ops) {
BNX2X_ERR("Section offset %d is out of bounds\n", i);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
return -EINVAL;
}
}
/* Check FW version */
offset = be32_to_cpu(fw_hdr->fw_version.offset);
fw_ver = firmware->data + offset;
if ((fw_ver[0] != BCM_5710_FW_MAJOR_VERSION) ||
(fw_ver[1] != BCM_5710_FW_MINOR_VERSION) ||
(fw_ver[2] != BCM_5710_FW_REVISION_VERSION) ||
(fw_ver[3] != BCM_5710_FW_ENGINEERING_VERSION)) {
BNX2X_ERR("Bad FW version:%d.%d.%d.%d. Should be %d.%d.%d.%d\n",
fw_ver[0], fw_ver[1], fw_ver[2], fw_ver[3],
BCM_5710_FW_MAJOR_VERSION,
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
BCM_5710_FW_MINOR_VERSION,
BCM_5710_FW_REVISION_VERSION,
BCM_5710_FW_ENGINEERING_VERSION);
return -EINVAL;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
}
return 0;
}
static void be32_to_cpu_n(const u8 *_source, u8 *_target, u32 n)
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
{
const __be32 *source = (const __be32 *)_source;
u32 *target = (u32 *)_target;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
u32 i;
for (i = 0; i < n/4; i++)
target[i] = be32_to_cpu(source[i]);
}
/*
Ops array is stored in the following format:
{op(8bit), offset(24bit, big endian), data(32bit, big endian)}
*/
static void bnx2x_prep_ops(const u8 *_source, u8 *_target, u32 n)
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
{
const __be32 *source = (const __be32 *)_source;
struct raw_op *target = (struct raw_op *)_target;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
u32 i, j, tmp;
for (i = 0, j = 0; i < n/8; i++, j += 2) {
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
tmp = be32_to_cpu(source[j]);
target[i].op = (tmp >> 24) & 0xff;
target[i].offset = tmp & 0xffffff;
target[i].raw_data = be32_to_cpu(source[j + 1]);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
}
}
/* IRO array is stored in the following format:
* {base(24bit), m1(16bit), m2(16bit), m3(16bit), size(16bit) }
*/
static void bnx2x_prep_iro(const u8 *_source, u8 *_target, u32 n)
{
const __be32 *source = (const __be32 *)_source;
struct iro *target = (struct iro *)_target;
u32 i, j, tmp;
for (i = 0, j = 0; i < n/sizeof(struct iro); i++) {
target[i].base = be32_to_cpu(source[j]);
j++;
tmp = be32_to_cpu(source[j]);
target[i].m1 = (tmp >> 16) & 0xffff;
target[i].m2 = tmp & 0xffff;
j++;
tmp = be32_to_cpu(source[j]);
target[i].m3 = (tmp >> 16) & 0xffff;
target[i].size = tmp & 0xffff;
j++;
}
}
static void be16_to_cpu_n(const u8 *_source, u8 *_target, u32 n)
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
{
const __be16 *source = (const __be16 *)_source;
u16 *target = (u16 *)_target;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
u32 i;
for (i = 0; i < n/2; i++)
target[i] = be16_to_cpu(source[i]);
}
#define BNX2X_ALLOC_AND_SET(arr, lbl, func) \
do { \
u32 len = be32_to_cpu(fw_hdr->arr.len); \
bp->arr = kmalloc(len, GFP_KERNEL); \
if (!bp->arr) \
goto lbl; \
func(bp->firmware->data + be32_to_cpu(fw_hdr->arr.offset), \
(u8 *)bp->arr, len); \
} while (0)
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
static int bnx2x_init_firmware(struct bnx2x *bp)
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
{
const char *fw_file_name;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
struct bnx2x_fw_file_hdr *fw_hdr;
int rc;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
if (bp->firmware)
return 0;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
if (CHIP_IS_E1(bp))
fw_file_name = FW_FILE_NAME_E1;
else if (CHIP_IS_E1H(bp))
fw_file_name = FW_FILE_NAME_E1H;
else if (!CHIP_IS_E1x(bp))
fw_file_name = FW_FILE_NAME_E2;
else {
BNX2X_ERR("Unsupported chip revision\n");
return -EINVAL;
}
BNX2X_DEV_INFO("Loading %s\n", fw_file_name);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
rc = request_firmware(&bp->firmware, fw_file_name, &bp->pdev->dev);
if (rc) {
BNX2X_ERR("Can't load firmware file %s\n",
fw_file_name);
goto request_firmware_exit;
}
rc = bnx2x_check_firmware(bp);
if (rc) {
BNX2X_ERR("Corrupt firmware file %s\n", fw_file_name);
goto request_firmware_exit;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
}
fw_hdr = (struct bnx2x_fw_file_hdr *)bp->firmware->data;
/* Initialize the pointers to the init arrays */
/* Blob */
rc = -ENOMEM;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
BNX2X_ALLOC_AND_SET(init_data, request_firmware_exit, be32_to_cpu_n);
/* Opcodes */
BNX2X_ALLOC_AND_SET(init_ops, init_ops_alloc_err, bnx2x_prep_ops);
/* Offsets */
BNX2X_ALLOC_AND_SET(init_ops_offsets, init_offsets_alloc_err,
be16_to_cpu_n);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
/* STORMs firmware */
INIT_TSEM_INT_TABLE_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->tsem_int_table_data.offset);
INIT_TSEM_PRAM_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->tsem_pram_data.offset);
INIT_USEM_INT_TABLE_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->usem_int_table_data.offset);
INIT_USEM_PRAM_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->usem_pram_data.offset);
INIT_XSEM_INT_TABLE_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->xsem_int_table_data.offset);
INIT_XSEM_PRAM_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->xsem_pram_data.offset);
INIT_CSEM_INT_TABLE_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->csem_int_table_data.offset);
INIT_CSEM_PRAM_DATA(bp) = bp->firmware->data +
be32_to_cpu(fw_hdr->csem_pram_data.offset);
/* IRO */
BNX2X_ALLOC_AND_SET(iro_arr, iro_alloc_err, bnx2x_prep_iro);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
return 0;
iro_alloc_err:
kfree(bp->init_ops_offsets);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
init_offsets_alloc_err:
kfree(bp->init_ops);
init_ops_alloc_err:
kfree(bp->init_data);
request_firmware_exit:
release_firmware(bp->firmware);
bp->firmware = NULL;
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
return rc;
}
static void bnx2x_release_firmware(struct bnx2x *bp)
{
kfree(bp->init_ops_offsets);
kfree(bp->init_ops);
kfree(bp->init_data);
release_firmware(bp->firmware);
bp->firmware = NULL;
}
static struct bnx2x_func_sp_drv_ops bnx2x_func_sp_drv = {
.init_hw_cmn_chip = bnx2x_init_hw_common_chip,
.init_hw_cmn = bnx2x_init_hw_common,
.init_hw_port = bnx2x_init_hw_port,
.init_hw_func = bnx2x_init_hw_func,
.reset_hw_cmn = bnx2x_reset_common,
.reset_hw_port = bnx2x_reset_port,
.reset_hw_func = bnx2x_reset_func,
.gunzip_init = bnx2x_gunzip_init,
.gunzip_end = bnx2x_gunzip_end,
.init_fw = bnx2x_init_firmware,
.release_fw = bnx2x_release_firmware,
};
void bnx2x__init_func_obj(struct bnx2x *bp)
{
/* Prepare DMAE related driver resources */
bnx2x_setup_dmae(bp);
bnx2x_init_func_obj(bp, &bp->func_obj,
bnx2x_sp(bp, func_rdata),
bnx2x_sp_mapping(bp, func_rdata),
bnx2x_sp(bp, func_afex_rdata),
bnx2x_sp_mapping(bp, func_afex_rdata),
&bnx2x_func_sp_drv);
}
/* must be called after sriov-enable */
static int bnx2x_set_qm_cid_count(struct bnx2x *bp)
{
int cid_count = BNX2X_L2_MAX_CID(bp);
bnx2x: Separated FW from the source. >From now on FW will be downloaded from the binary file using request_firmware. There will be different files for every supported chip. Currently 57710 (e1) and 57711 (e1h). File names have the following format: bnx2x-<chip version>-<FW version>.fw. ihex versions of current FW files are submitted in the next patch. Each binary file has a header in the following format: struct bnx2x_fw_file_section { __be32 len; __be32 offset; } struct bnx2x_fw_file_hdr { struct bnx2x_fw_file_section init_ops; struct bnx2x_fw_file_section init_ops_offsets; struct bnx2x_fw_file_section init_data; struct bnx2x_fw_file_section tsem_int_table_data; struct bnx2x_fw_file_section tsem_pram_data; struct bnx2x_fw_file_section usem_int_table_data; struct bnx2x_fw_file_section usem_pram_data; struct bnx2x_fw_file_section csem_int_table_data; struct bnx2x_fw_file_section csem_pram_data; struct bnx2x_fw_file_section xsem_int_table_data; struct bnx2x_fw_file_section xsem_pram_data; struct bnx2x_fw_file_section fw_version; } Each bnx2x_fw_file_section contains the length and the offset of the appropriate section in the binary file. Values are stored in the big endian format. Data types of arrays: init_data __be32 init_ops_offsets __be16 XXsem_pram_data u8 XXsem_int_table_data u8 init_ops struct raw_op { u8 op; __be24 offset; __be32 data; } fw_version u8 >From now boundaries of a specific initialization stage are stored in init_ops_offsets array instead of being defined by separate macroes. The index in init_ops_offsets is calculated by BLOCK_OPS_IDX macro: #define BLOCK_OPS_IDX(block, stage, end) \ (2*(((block)*STAGE_IDX_MAX) + (stage)) + (end)) Security: In addition to sanity check of array boundaries bnx2x will check a FW version. Additional checks might be added in the future. Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-27 19:27:43 +09:00
if (IS_SRIOV(bp))
cid_count += BNX2X_VF_CIDS;
if (CNIC_SUPPORT(bp))
cid_count += CNIC_CID_MAX;
return roundup(cid_count, QM_CID_ROUND);
}
/**
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
* bnx2x_get_num_none_def_sbs - return the number of none default SBs
*
* @dev: pci device
*
*/
static int bnx2x_get_num_non_def_sbs(struct pci_dev *pdev, int cnic_cnt)
{
int index;
u16 control = 0;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/*
* If MSI-X is not supported - return number of SBs needed to support
* one fast path queue: one FP queue + SB for CNIC
*/
if (!pdev->msix_cap) {
dev_info(&pdev->dev, "no msix capability found\n");
return 1 + cnic_cnt;
}
dev_info(&pdev->dev, "msix capability found\n");
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/*
* The value in the PCI configuration space is the index of the last
* entry, namely one less than the actual size of the table, which is
* exactly what we want to return from this function: number of all SBs
* without the default SB.
* For VFs there is no default SB, then we return (index+1).
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
*/
pci_read_config_word(pdev, pdev->msix_cap + PCI_MSIX_FLAGS, &control);
index = control & PCI_MSIX_FLAGS_QSIZE;
return index;
}
static int set_max_cos_est(int chip_id)
{
switch (chip_id) {
case BCM57710:
case BCM57711:
case BCM57711E:
return BNX2X_MULTI_TX_COS_E1X;
case BCM57712:
case BCM57712_MF:
return BNX2X_MULTI_TX_COS_E2_E3A0;
case BCM57800:
case BCM57800_MF:
case BCM57810:
case BCM57810_MF:
case BCM57840_4_10:
case BCM57840_2_20:
case BCM57840_O:
case BCM57840_MFO:
case BCM57840_MF:
case BCM57811:
case BCM57811_MF:
return BNX2X_MULTI_TX_COS_E3B0;
case BCM57712_VF:
case BCM57800_VF:
case BCM57810_VF:
case BCM57840_VF:
case BCM57811_VF:
return 1;
default:
pr_err("Unknown board_type (%d), aborting\n", chip_id);
return -ENODEV;
}
}
static int set_is_vf(int chip_id)
{
switch (chip_id) {
case BCM57712_VF:
case BCM57800_VF:
case BCM57810_VF:
case BCM57840_VF:
case BCM57811_VF:
return true;
default:
return false;
}
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* nig_tsgen registers relative address */
#define tsgen_ctrl 0x0
#define tsgen_freecount 0x10
#define tsgen_synctime_t0 0x20
#define tsgen_offset_t0 0x28
#define tsgen_drift_t0 0x30
#define tsgen_synctime_t1 0x58
#define tsgen_offset_t1 0x60
#define tsgen_drift_t1 0x68
/* FW workaround for setting drift */
static int bnx2x_send_update_drift_ramrod(struct bnx2x *bp, int drift_dir,
int best_val, int best_period)
{
struct bnx2x_func_state_params func_params = {NULL};
struct bnx2x_func_set_timesync_params *set_timesync_params =
&func_params.params.set_timesync;
/* Prepare parameters for function state transitions */
__set_bit(RAMROD_COMP_WAIT, &func_params.ramrod_flags);
__set_bit(RAMROD_RETRY, &func_params.ramrod_flags);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_SET_TIMESYNC;
/* Function parameters */
set_timesync_params->drift_adjust_cmd = TS_DRIFT_ADJUST_SET;
set_timesync_params->offset_cmd = TS_OFFSET_KEEP;
set_timesync_params->add_sub_drift_adjust_value =
drift_dir ? TS_ADD_VALUE : TS_SUB_VALUE;
set_timesync_params->drift_adjust_value = best_val;
set_timesync_params->drift_adjust_period = best_period;
return bnx2x_func_state_change(bp, &func_params);
}
static int bnx2x_ptp_adjfreq(struct ptp_clock_info *ptp, s32 ppb)
{
struct bnx2x *bp = container_of(ptp, struct bnx2x, ptp_clock_info);
int rc;
int drift_dir = 1;
int val, period, period1, period2, dif, dif1, dif2;
int best_dif = BNX2X_MAX_PHC_DRIFT, best_period = 0, best_val = 0;
DP(BNX2X_MSG_PTP, "PTP adjfreq called, ppb = %d\n", ppb);
if (!netif_running(bp->dev)) {
DP(BNX2X_MSG_PTP,
"PTP adjfreq called while the interface is down\n");
return -ENETDOWN;
}
if (ppb < 0) {
ppb = -ppb;
drift_dir = 0;
}
if (ppb == 0) {
best_val = 1;
best_period = 0x1FFFFFF;
} else if (ppb >= BNX2X_MAX_PHC_DRIFT) {
best_val = 31;
best_period = 1;
} else {
/* Changed not to allow val = 8, 16, 24 as these values
* are not supported in workaround.
*/
for (val = 0; val <= 31; val++) {
if ((val & 0x7) == 0)
continue;
period1 = val * 1000000 / ppb;
period2 = period1 + 1;
if (period1 != 0)
dif1 = ppb - (val * 1000000 / period1);
else
dif1 = BNX2X_MAX_PHC_DRIFT;
if (dif1 < 0)
dif1 = -dif1;
dif2 = ppb - (val * 1000000 / period2);
if (dif2 < 0)
dif2 = -dif2;
dif = (dif1 < dif2) ? dif1 : dif2;
period = (dif1 < dif2) ? period1 : period2;
if (dif < best_dif) {
best_dif = dif;
best_val = val;
best_period = period;
}
}
}
rc = bnx2x_send_update_drift_ramrod(bp, drift_dir, best_val,
best_period);
if (rc) {
BNX2X_ERR("Failed to set drift\n");
return -EFAULT;
}
DP(BNX2X_MSG_PTP, "Configured val = %d, period = %d\n", best_val,
best_period);
return 0;
}
static int bnx2x_ptp_adjtime(struct ptp_clock_info *ptp, s64 delta)
{
struct bnx2x *bp = container_of(ptp, struct bnx2x, ptp_clock_info);
if (!netif_running(bp->dev)) {
DP(BNX2X_MSG_PTP,
"PTP adjtime called while the interface is down\n");
return -ENETDOWN;
}
DP(BNX2X_MSG_PTP, "PTP adjtime called, delta = %llx\n", delta);
timecounter_adjtime(&bp->timecounter, delta);
return 0;
}
static int bnx2x_ptp_gettime(struct ptp_clock_info *ptp, struct timespec64 *ts)
{
struct bnx2x *bp = container_of(ptp, struct bnx2x, ptp_clock_info);
u64 ns;
if (!netif_running(bp->dev)) {
DP(BNX2X_MSG_PTP,
"PTP gettime called while the interface is down\n");
return -ENETDOWN;
}
ns = timecounter_read(&bp->timecounter);
DP(BNX2X_MSG_PTP, "PTP gettime called, ns = %llu\n", ns);
*ts = ns_to_timespec64(ns);
return 0;
}
static int bnx2x_ptp_settime(struct ptp_clock_info *ptp,
const struct timespec64 *ts)
{
struct bnx2x *bp = container_of(ptp, struct bnx2x, ptp_clock_info);
u64 ns;
if (!netif_running(bp->dev)) {
DP(BNX2X_MSG_PTP,
"PTP settime called while the interface is down\n");
return -ENETDOWN;
}
ns = timespec64_to_ns(ts);
DP(BNX2X_MSG_PTP, "PTP settime called, ns = %llu\n", ns);
/* Re-init the timecounter */
timecounter_init(&bp->timecounter, &bp->cyclecounter, ns);
return 0;
}
/* Enable (or disable) ancillary features of the phc subsystem */
static int bnx2x_ptp_enable(struct ptp_clock_info *ptp,
struct ptp_clock_request *rq, int on)
{
struct bnx2x *bp = container_of(ptp, struct bnx2x, ptp_clock_info);
BNX2X_ERR("PHC ancillary features are not supported\n");
return -ENOTSUPP;
}
static void bnx2x_register_phc(struct bnx2x *bp)
{
/* Fill the ptp_clock_info struct and register PTP clock*/
bp->ptp_clock_info.owner = THIS_MODULE;
snprintf(bp->ptp_clock_info.name, 16, "%s", bp->dev->name);
bp->ptp_clock_info.max_adj = BNX2X_MAX_PHC_DRIFT; /* In PPB */
bp->ptp_clock_info.n_alarm = 0;
bp->ptp_clock_info.n_ext_ts = 0;
bp->ptp_clock_info.n_per_out = 0;
bp->ptp_clock_info.pps = 0;
bp->ptp_clock_info.adjfreq = bnx2x_ptp_adjfreq;
bp->ptp_clock_info.adjtime = bnx2x_ptp_adjtime;
bp->ptp_clock_info.gettime64 = bnx2x_ptp_gettime;
bp->ptp_clock_info.settime64 = bnx2x_ptp_settime;
bp->ptp_clock_info.enable = bnx2x_ptp_enable;
bp->ptp_clock = ptp_clock_register(&bp->ptp_clock_info, &bp->pdev->dev);
if (IS_ERR(bp->ptp_clock)) {
bp->ptp_clock = NULL;
BNX2X_ERR("PTP clock registeration failed\n");
}
}
static int bnx2x_init_one(struct pci_dev *pdev,
const struct pci_device_id *ent)
{
struct net_device *dev = NULL;
struct bnx2x *bp;
enum pcie_link_width pcie_width;
enum pci_bus_speed pcie_speed;
int rc, max_non_def_sbs;
int rx_count, tx_count, rss_count, doorbell_size;
int max_cos_est;
bool is_vf;
int cnic_cnt;
/* Management FW 'remembers' living interfaces. Allow it some time
* to forget previously living interfaces, allowing a proper re-load.
*/
if (is_kdump_kernel()) {
ktime_t now = ktime_get_boottime();
ktime_t fw_ready_time = ktime_set(5, 0);
if (ktime_before(now, fw_ready_time))
msleep(ktime_ms_delta(fw_ready_time, now));
}
/* An estimated maximum supported CoS number according to the chip
* version.
* We will try to roughly estimate the maximum number of CoSes this chip
* may support in order to minimize the memory allocated for Tx
* netdev_queue's. This number will be accurately calculated during the
* initialization of bp->max_cos based on the chip versions AND chip
* revision in the bnx2x_init_bp().
*/
max_cos_est = set_max_cos_est(ent->driver_data);
if (max_cos_est < 0)
return max_cos_est;
is_vf = set_is_vf(ent->driver_data);
cnic_cnt = is_vf ? 0 : 1;
max_non_def_sbs = bnx2x_get_num_non_def_sbs(pdev, cnic_cnt);
/* add another SB for VF as it has no default SB */
max_non_def_sbs += is_vf ? 1 : 0;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* Maximum number of RSS queues: one IGU SB goes to CNIC */
rss_count = max_non_def_sbs - cnic_cnt;
if (rss_count < 1)
return -EINVAL;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* Maximum number of netdev Rx queues: RSS + FCoE L2 */
rx_count = rss_count + cnic_cnt;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
/* Maximum number of netdev Tx queues:
* Maximum TSS queues * Maximum supported number of CoS + FCoE L2
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
*/
tx_count = rss_count * max_cos_est + cnic_cnt;
/* dev zeroed in init_etherdev */
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
dev = alloc_etherdev_mqs(sizeof(*bp), tx_count, rx_count);
if (!dev)
return -ENOMEM;
bp = netdev_priv(dev);
bp->flags = 0;
if (is_vf)
bp->flags |= IS_VF_FLAG;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bp->igu_sb_cnt = max_non_def_sbs;
bp->igu_base_addr = IS_VF(bp) ? PXP_VF_ADDR_IGU_START : BAR_IGU_INTMEM;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bp->msg_enable = debug;
bp->cnic_support = cnic_cnt;
bp->cnic_probe = bnx2x_cnic_probe;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
pci_set_drvdata(pdev, dev);
rc = bnx2x_init_dev(bp, pdev, dev, ent->driver_data);
if (rc < 0) {
free_netdev(dev);
return rc;
}
BNX2X_DEV_INFO("This is a %s function\n",
IS_PF(bp) ? "physical" : "virtual");
BNX2X_DEV_INFO("Cnic support is %s\n", CNIC_SUPPORT(bp) ? "on" : "off");
BNX2X_DEV_INFO("Max num of status blocks %d\n", max_non_def_sbs);
BNX2X_DEV_INFO("Allocated netdev with %d tx and %d rx queues\n",
tx_count, rx_count);
rc = bnx2x_init_bp(bp);
if (rc)
goto init_one_exit;
/* Map doorbells here as we need the real value of bp->max_cos which
* is initialized in bnx2x_init_bp() to determine the number of
* l2 connections.
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
*/
if (IS_VF(bp)) {
bp->doorbells = bnx2x_vf_doorbells(bp);
rc = bnx2x_vf_pci_alloc(bp);
if (rc)
goto init_one_freemem;
} else {
doorbell_size = BNX2X_L2_MAX_CID(bp) * (1 << BNX2X_DB_SHIFT);
if (doorbell_size > pci_resource_len(pdev, 2)) {
dev_err(&bp->pdev->dev,
"Cannot map doorbells, bar size too small, aborting\n");
rc = -ENOMEM;
goto init_one_freemem;
}
bp->doorbells = ioremap_nocache(pci_resource_start(pdev, 2),
doorbell_size);
}
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
if (!bp->doorbells) {
dev_err(&bp->pdev->dev,
"Cannot map doorbell space, aborting\n");
rc = -ENOMEM;
goto init_one_freemem;
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
}
if (IS_VF(bp)) {
rc = bnx2x_vfpf_acquire(bp, tx_count, rx_count);
if (rc)
goto init_one_freemem;
#ifdef CONFIG_BNX2X_SRIOV
/* VF with OLD Hypervisor or old PF do not support filtering */
if (bp->acquire_resp.pfdev_info.pf_cap & PFVF_CAP_VLAN_FILTER) {
dev->hw_features |= NETIF_F_HW_VLAN_CTAG_FILTER;
dev->features |= NETIF_F_HW_VLAN_CTAG_FILTER;
}
#endif
}
/* Enable SRIOV if capability found in configuration space */
rc = bnx2x_iov_init_one(bp, int_mode, BNX2X_MAX_NUM_OF_VFS);
if (rc)
goto init_one_freemem;
/* calc qm_cid_count */
bnx2x: Multiple concurrent l2 traffic classes Overview: Support mapping of priorities to traffic classes and traffic classes to transmission queues ranges in the net device. The queue ranges are (count, offset) pairs relating to the txq array. This can be done via DCBX negotiation or by kernel. As a result Enhanced Transmission Selection (ETS) and Priority Flow Control (PFC) are supported between L2 network traffic classes. Mapping: This patch uses the netdev_set_num_tc, netdev_set_prio_tc_map and netdev_set_tc_queue functions to map priorities to traffic classes and traffic classes to transmission queue ranges. This mapping is performed by bnx2x_setup_tc function which is connected to the ndo_setup_tc. This function is always called at nic load where by default it maps all priorities to tc 0, and it may also be called by the kernel or by the bnx2x upon DCBX negotiation to modify the mapping. rtnl lock: When the ndo_setup_tc is called at nic load or by kernel the rtnl lock is already taken. However, when DCBX negotiation takes place the lock is not taken. The work is therefore scheduled to be handled by the sp_rtnl task. Fastpath: The fastpath structure of the bnx2x which was previously used to hold the information of one tx queue and one rx queue was redesigned to represent multiple tx queues, one for each traffic class. The transmission queue supplied in the skb by the kernel can no longer be interpreted as a straightforward index into the fastpath structure array, but it must rather be decoded to the appropriate fastpath index and the tc within that fastpath. Slowpath: The bnx2x's queue object was redesigned to accommodate multiple transmission queues. The queue object's state machine was enhanced to allow opening multiple transmission-only connections on top of the regular tx-rx connection. Firmware: This feature relies on the tx-only queue feature introduced in the bnx2x 7.0.23 firmware and the FW likewise must have the bnx2x multi cos support. Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-07-14 17:31:57 +09:00
bp->qm_cid_count = bnx2x_set_qm_cid_count(bp);
BNX2X_DEV_INFO("qm_cid_count %d\n", bp->qm_cid_count);
/* disable FCOE L2 queue for E1x*/
if (CHIP_IS_E1x(bp))
bp->flags |= NO_FCOE_FLAG;
/* Set bp->num_queues for MSI-X mode*/
bnx2x_set_num_queues(bp);
/* Configure interrupt mode: try to enable MSI-X/MSI if
* needed.
*/
rc = bnx2x_set_int_mode(bp);
if (rc) {
dev_err(&pdev->dev, "Cannot set interrupts\n");
goto init_one_freemem;
}
BNX2X_DEV_INFO("set interrupts successfully\n");
/* register the net device */
rc = register_netdev(dev);
if (rc) {
dev_err(&pdev->dev, "Cannot register net device\n");
goto init_one_freemem;
}
BNX2X_DEV_INFO("device name after netdev register %s\n", dev->name);
if (!NO_FCOE(bp)) {
/* Add storage MAC address */
rtnl_lock();
dev_addr_add(bp->dev, bp->fip_mac, NETDEV_HW_ADDR_T_SAN);
rtnl_unlock();
}
if (pcie_get_minimum_link(bp->pdev, &pcie_speed, &pcie_width) ||
pcie_speed == PCI_SPEED_UNKNOWN ||
pcie_width == PCIE_LNK_WIDTH_UNKNOWN)
BNX2X_DEV_INFO("Failed to determine PCI Express Bandwidth\n");
else
BNX2X_DEV_INFO(
"%s (%c%d) PCI-E x%d %s found at mem %lx, IRQ %d, node addr %pM\n",
board_info[ent->driver_data].name,
(CHIP_REV(bp) >> 12) + 'A', (CHIP_METAL(bp) >> 4),
pcie_width,
pcie_speed == PCIE_SPEED_2_5GT ? "2.5GHz" :
pcie_speed == PCIE_SPEED_5_0GT ? "5.0GHz" :
pcie_speed == PCIE_SPEED_8_0GT ? "8.0GHz" :
"Unknown",
dev->base_addr, bp->pdev->irq, dev->dev_addr);
bnx2x_register_phc(bp);
if (!IS_MF_SD_STORAGE_PERSONALITY_ONLY(bp))
bnx2x_set_os_driver_state(bp, OS_DRIVER_STATE_DISABLED);
return 0;
init_one_freemem:
bnx2x_free_mem_bp(bp);
init_one_exit:
bnx2x_disable_pcie_error_reporting(bp);
if (bp->regview)
iounmap(bp->regview);
if (IS_PF(bp) && bp->doorbells)
iounmap(bp->doorbells);
free_netdev(dev);
if (atomic_read(&pdev->enable_cnt) == 1)
pci_release_regions(pdev);
pci_disable_device(pdev);
return rc;
}
static void __bnx2x_remove(struct pci_dev *pdev,
struct net_device *dev,
struct bnx2x *bp,
bool remove_netdev)
{
if (bp->ptp_clock) {
ptp_clock_unregister(bp->ptp_clock);
bp->ptp_clock = NULL;
}
/* Delete storage MAC address */
if (!NO_FCOE(bp)) {
rtnl_lock();
dev_addr_del(bp->dev, bp->fip_mac, NETDEV_HW_ADDR_T_SAN);
rtnl_unlock();
}
#ifdef BCM_DCBNL
/* Delete app tlvs from dcbnl */
bnx2x_dcbnl_update_applist(bp, true);
#endif
if (IS_PF(bp) &&
!BP_NOMCP(bp) &&
(bp->flags & BC_SUPPORTS_RMMOD_CMD))
bnx2x_fw_command(bp, DRV_MSG_CODE_RMMOD, 0);
/* Close the interface - either directly or implicitly */
if (remove_netdev) {
unregister_netdev(dev);
} else {
rtnl_lock();
dev_close(dev);
rtnl_unlock();
}
bnx2x_iov_remove_one(bp);
/* Power on: we can't let PCI layer write to us while we are in D3 */
if (IS_PF(bp)) {
bnx2x_set_power_state(bp, PCI_D0);
bnx2x_set_os_driver_state(bp, OS_DRIVER_STATE_NOT_LOADED);
/* Set endianity registers to reset values in case next driver
* boots in different endianty environment.
*/
bnx2x_reset_endianity(bp);
}
/* Disable MSI/MSI-X */
bnx2x_disable_msi(bp);
/* Power off */
if (IS_PF(bp))
bnx2x_set_power_state(bp, PCI_D3hot);
/* Make sure RESET task is not scheduled before continuing */
cancel_delayed_work_sync(&bp->sp_rtnl_task);
/* send message via vfpf channel to release the resources of this vf */
if (IS_VF(bp))
bnx2x_vfpf_release(bp);
/* Assumes no further PCIe PM changes will occur */
if (system_state == SYSTEM_POWER_OFF) {
pci_wake_from_d3(pdev, bp->wol);
pci_set_power_state(pdev, PCI_D3hot);
}
bnx2x_disable_pcie_error_reporting(bp);
if (remove_netdev) {
if (bp->regview)
iounmap(bp->regview);
/* For vfs, doorbells are part of the regview and were unmapped
* along with it. FW is only loaded by PF.
*/
if (IS_PF(bp)) {
if (bp->doorbells)
iounmap(bp->doorbells);
bnx2x_release_firmware(bp);
} else {
bnx2x_vf_pci_dealloc(bp);
}
bnx2x_free_mem_bp(bp);
free_netdev(dev);
if (atomic_read(&pdev->enable_cnt) == 1)
pci_release_regions(pdev);
pci_disable_device(pdev);
}
}
static void bnx2x_remove_one(struct pci_dev *pdev)
{
struct net_device *dev = pci_get_drvdata(pdev);
struct bnx2x *bp;
if (!dev) {
dev_err(&pdev->dev, "BAD net device from bnx2x_init_one\n");
return;
}
bp = netdev_priv(dev);
__bnx2x_remove(pdev, dev, bp, true);
}
static int bnx2x_eeh_nic_unload(struct bnx2x *bp)
{
bp->state = BNX2X_STATE_CLOSING_WAIT4_HALT;
bp->rx_mode = BNX2X_RX_MODE_NONE;
if (CNIC_LOADED(bp))
bnx2x_cnic_notify(bp, CNIC_CTL_STOP_CMD);
/* Stop Tx */
bnx2x_tx_disable(bp);
/* Delete all NAPI objects */
bnx2x_del_all_napi(bp);
if (CNIC_LOADED(bp))
bnx2x_del_all_napi_cnic(bp);
netdev_reset_tc(bp->dev);
del_timer_sync(&bp->timer);
cancel_delayed_work_sync(&bp->sp_task);
cancel_delayed_work_sync(&bp->period_task);
if (!down_timeout(&bp->stats_lock, HZ / 10)) {
bp->stats_state = STATS_STATE_DISABLED;
up(&bp->stats_lock);
}
bnx2x_save_statistics(bp);
netif_carrier_off(bp->dev);
return 0;
}
/**
* bnx2x_io_error_detected - called when PCI error is detected
* @pdev: Pointer to PCI device
* @state: The current pci connection state
*
* This function is called after a PCI bus error affecting
* this device has been detected.
*/
static pci_ers_result_t bnx2x_io_error_detected(struct pci_dev *pdev,
pci_channel_state_t state)
{
struct net_device *dev = pci_get_drvdata(pdev);
struct bnx2x *bp = netdev_priv(dev);
rtnl_lock();
BNX2X_ERR("IO error detected\n");
netif_device_detach(dev);
if (state == pci_channel_io_perm_failure) {
rtnl_unlock();
return PCI_ERS_RESULT_DISCONNECT;
}
if (netif_running(dev))
bnx2x_eeh_nic_unload(bp);
bnx2x_prev_path_mark_eeh(bp);
pci_disable_device(pdev);
rtnl_unlock();
/* Request a slot reset */
return PCI_ERS_RESULT_NEED_RESET;
}
/**
* bnx2x_io_slot_reset - called after the PCI bus has been reset
* @pdev: Pointer to PCI device
*
* Restart the card from scratch, as if from a cold-boot.
*/
static pci_ers_result_t bnx2x_io_slot_reset(struct pci_dev *pdev)
{
struct net_device *dev = pci_get_drvdata(pdev);
struct bnx2x *bp = netdev_priv(dev);
int i;
rtnl_lock();
BNX2X_ERR("IO slot reset initializing...\n");
if (pci_enable_device(pdev)) {
dev_err(&pdev->dev,
"Cannot re-enable PCI device after reset\n");
rtnl_unlock();
return PCI_ERS_RESULT_DISCONNECT;
}
pci_set_master(pdev);
pci_restore_state(pdev);
pci_save_state(pdev);
if (netif_running(dev))
bnx2x_set_power_state(bp, PCI_D0);
if (netif_running(dev)) {
BNX2X_ERR("IO slot reset --> driver unload\n");
/* MCP should have been reset; Need to wait for validity */
if (bnx2x_init_shmem(bp)) {
rtnl_unlock();
return PCI_ERS_RESULT_DISCONNECT;
}
if (IS_PF(bp) && SHMEM2_HAS(bp, drv_capabilities_flag)) {
u32 v;
v = SHMEM2_RD(bp,
drv_capabilities_flag[BP_FW_MB_IDX(bp)]);
SHMEM2_WR(bp, drv_capabilities_flag[BP_FW_MB_IDX(bp)],
v & ~DRV_FLAGS_CAPABILITIES_LOADED_L2);
}
bnx2x_drain_tx_queues(bp);
bnx2x_send_unload_req(bp, UNLOAD_RECOVERY);
bnx2x_netif_stop(bp, 1);
bnx2x_free_irq(bp);
/* Report UNLOAD_DONE to MCP */
bnx2x_send_unload_done(bp, true);
bp->sp_state = 0;
bp->port.pmf = 0;
bnx2x_prev_unload(bp);
/* We should have reseted the engine, so It's fair to
* assume the FW will no longer write to the bnx2x driver.
*/
bnx2x_squeeze_objects(bp);
bnx2x_free_skbs(bp);
for_each_rx_queue(bp, i)
bnx2x_free_rx_sge_range(bp, bp->fp + i, NUM_RX_SGE);
bnx2x_free_fp_mem(bp);
bnx2x_free_mem(bp);
bp->state = BNX2X_STATE_CLOSED;
}
rtnl_unlock();
/* If AER, perform cleanup of the PCIe registers */
if (bp->flags & AER_ENABLED) {
if (pci_cleanup_aer_uncorrect_error_status(pdev))
BNX2X_ERR("pci_cleanup_aer_uncorrect_error_status failed\n");
else
DP(NETIF_MSG_HW, "pci_cleanup_aer_uncorrect_error_status succeeded\n");
}
return PCI_ERS_RESULT_RECOVERED;
}
/**
* bnx2x_io_resume - called when traffic can start flowing again
* @pdev: Pointer to PCI device
*
* This callback is called when the error recovery driver tells us that
* its OK to resume normal operation.
*/
static void bnx2x_io_resume(struct pci_dev *pdev)
{
struct net_device *dev = pci_get_drvdata(pdev);
struct bnx2x *bp = netdev_priv(dev);
if (bp->recovery_state != BNX2X_RECOVERY_DONE) {
netdev_err(bp->dev, "Handling parity error recovery. Try again later\n");
return;
}
rtnl_lock();
bp->fw_seq = SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
DRV_MSG_SEQ_NUMBER_MASK;
if (netif_running(dev))
bnx2x_nic_load(bp, LOAD_NORMAL);
netif_device_attach(dev);
rtnl_unlock();
}
static const struct pci_error_handlers bnx2x_err_handler = {
.error_detected = bnx2x_io_error_detected,
.slot_reset = bnx2x_io_slot_reset,
.resume = bnx2x_io_resume,
};
static void bnx2x_shutdown(struct pci_dev *pdev)
{
struct net_device *dev = pci_get_drvdata(pdev);
struct bnx2x *bp;
if (!dev)
return;
bp = netdev_priv(dev);
if (!bp)
return;
rtnl_lock();
netif_device_detach(dev);
rtnl_unlock();
/* Don't remove the netdevice, as there are scenarios which will cause
* the kernel to hang, e.g., when trying to remove bnx2i while the
* rootfs is mounted from SAN.
*/
__bnx2x_remove(pdev, dev, bp, false);
}
static struct pci_driver bnx2x_pci_driver = {
.name = DRV_MODULE_NAME,
.id_table = bnx2x_pci_tbl,
.probe = bnx2x_init_one,
.remove = bnx2x_remove_one,
.suspend = bnx2x_suspend,
.resume = bnx2x_resume,
.err_handler = &bnx2x_err_handler,
#ifdef CONFIG_BNX2X_SRIOV
.sriov_configure = bnx2x_sriov_configure,
#endif
.shutdown = bnx2x_shutdown,
};
static int __init bnx2x_init(void)
{
int ret;
pr_info("%s", version);
bnx2x_wq = create_singlethread_workqueue("bnx2x");
if (bnx2x_wq == NULL) {
pr_err("Cannot create workqueue\n");
return -ENOMEM;
}
bnx2x_iov_wq = create_singlethread_workqueue("bnx2x_iov");
if (!bnx2x_iov_wq) {
pr_err("Cannot create iov workqueue\n");
destroy_workqueue(bnx2x_wq);
return -ENOMEM;
}
ret = pci_register_driver(&bnx2x_pci_driver);
if (ret) {
pr_err("Cannot register driver\n");
destroy_workqueue(bnx2x_wq);
destroy_workqueue(bnx2x_iov_wq);
}
return ret;
}
static void __exit bnx2x_cleanup(void)
{
struct list_head *pos, *q;
pci_unregister_driver(&bnx2x_pci_driver);
destroy_workqueue(bnx2x_wq);
destroy_workqueue(bnx2x_iov_wq);
/* Free globally allocated resources */
list_for_each_safe(pos, q, &bnx2x_prev_list) {
struct bnx2x_prev_path_list *tmp =
list_entry(pos, struct bnx2x_prev_path_list, list);
list_del(pos);
kfree(tmp);
}
}
void bnx2x_notify_link_changed(struct bnx2x *bp)
{
REG_WR(bp, MISC_REG_AEU_GENERAL_ATTN_12 + BP_FUNC(bp)*sizeof(u32), 1);
}
module_init(bnx2x_init);
module_exit(bnx2x_cleanup);
/**
* bnx2x_set_iscsi_eth_mac_addr - set iSCSI MAC(s).
*
* @bp: driver handle
* @set: set or clear the CAM entry
*
* This function will wait until the ramrod completion returns.
* Return 0 if success, -ENODEV if ramrod doesn't return.
*/
static int bnx2x_set_iscsi_eth_mac_addr(struct bnx2x *bp)
{
unsigned long ramrod_flags = 0;
__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
return bnx2x_set_mac_one(bp, bp->cnic_eth_dev.iscsi_mac,
&bp->iscsi_l2_mac_obj, true,
BNX2X_ISCSI_ETH_MAC, &ramrod_flags);
}
/* count denotes the number of new completions we have seen */
static void bnx2x_cnic_sp_post(struct bnx2x *bp, int count)
{
struct eth_spe *spe;
int cxt_index, cxt_offset;
#ifdef BNX2X_STOP_ON_ERROR
if (unlikely(bp->panic))
return;
#endif
spin_lock_bh(&bp->spq_lock);
BUG_ON(bp->cnic_spq_pending < count);
bp->cnic_spq_pending -= count;
for (; bp->cnic_kwq_pending; bp->cnic_kwq_pending--) {
u16 type = (le16_to_cpu(bp->cnic_kwq_cons->hdr.type)
& SPE_HDR_CONN_TYPE) >>
SPE_HDR_CONN_TYPE_SHIFT;
u8 cmd = (le32_to_cpu(bp->cnic_kwq_cons->hdr.conn_and_cmd_data)
>> SPE_HDR_CMD_ID_SHIFT) & 0xff;
/* Set validation for iSCSI L2 client before sending SETUP
* ramrod
*/
if (type == ETH_CONNECTION_TYPE) {
if (cmd == RAMROD_CMD_ID_ETH_CLIENT_SETUP) {
cxt_index = BNX2X_ISCSI_ETH_CID(bp) /
ILT_PAGE_CIDS;
cxt_offset = BNX2X_ISCSI_ETH_CID(bp) -
(cxt_index * ILT_PAGE_CIDS);
bnx2x_set_ctx_validation(bp,
&bp->context[cxt_index].
vcxt[cxt_offset].eth,
BNX2X_ISCSI_ETH_CID(bp));
}
}
/*
* There may be not more than 8 L2, not more than 8 L5 SPEs
* and in the air. We also check that number of outstanding
* COMMON ramrods is not more than the EQ and SPQ can
* accommodate.
*/
if (type == ETH_CONNECTION_TYPE) {
if (!atomic_read(&bp->cq_spq_left))
break;
else
atomic_dec(&bp->cq_spq_left);
} else if (type == NONE_CONNECTION_TYPE) {
if (!atomic_read(&bp->eq_spq_left))
break;
else
atomic_dec(&bp->eq_spq_left);
} else if ((type == ISCSI_CONNECTION_TYPE) ||
(type == FCOE_CONNECTION_TYPE)) {
if (bp->cnic_spq_pending >=
bp->cnic_eth_dev.max_kwqe_pending)
break;
else
bp->cnic_spq_pending++;
} else {
BNX2X_ERR("Unknown SPE type: %d\n", type);
bnx2x_panic();
break;
}
spe = bnx2x_sp_get_next(bp);
*spe = *bp->cnic_kwq_cons;
DP(BNX2X_MSG_SP, "pending on SPQ %d, on KWQ %d count %d\n",
bp->cnic_spq_pending, bp->cnic_kwq_pending, count);
if (bp->cnic_kwq_cons == bp->cnic_kwq_last)
bp->cnic_kwq_cons = bp->cnic_kwq;
else
bp->cnic_kwq_cons++;
}
bnx2x_sp_prod_update(bp);
spin_unlock_bh(&bp->spq_lock);
}
static int bnx2x_cnic_sp_queue(struct net_device *dev,
struct kwqe_16 *kwqes[], u32 count)
{
struct bnx2x *bp = netdev_priv(dev);
int i;
#ifdef BNX2X_STOP_ON_ERROR
if (unlikely(bp->panic)) {
BNX2X_ERR("Can't post to SP queue while panic\n");
return -EIO;
}
#endif
if ((bp->recovery_state != BNX2X_RECOVERY_DONE) &&
(bp->recovery_state != BNX2X_RECOVERY_NIC_LOADING)) {
BNX2X_ERR("Handling parity error recovery. Try again later\n");
return -EAGAIN;
}
spin_lock_bh(&bp->spq_lock);
for (i = 0; i < count; i++) {
struct eth_spe *spe = (struct eth_spe *)kwqes[i];
if (bp->cnic_kwq_pending == MAX_SP_DESC_CNT)
break;
*bp->cnic_kwq_prod = *spe;
bp->cnic_kwq_pending++;
DP(BNX2X_MSG_SP, "L5 SPQE %x %x %x:%x pos %d\n",
spe->hdr.conn_and_cmd_data, spe->hdr.type,
spe->data.update_data_addr.hi,
spe->data.update_data_addr.lo,
bp->cnic_kwq_pending);
if (bp->cnic_kwq_prod == bp->cnic_kwq_last)
bp->cnic_kwq_prod = bp->cnic_kwq;
else
bp->cnic_kwq_prod++;
}
spin_unlock_bh(&bp->spq_lock);
if (bp->cnic_spq_pending < bp->cnic_eth_dev.max_kwqe_pending)
bnx2x_cnic_sp_post(bp, 0);
return i;
}
static int bnx2x_cnic_ctl_send(struct bnx2x *bp, struct cnic_ctl_info *ctl)
{
struct cnic_ops *c_ops;
int rc = 0;
mutex_lock(&bp->cnic_mutex);
c_ops = rcu_dereference_protected(bp->cnic_ops,
lockdep_is_held(&bp->cnic_mutex));
if (c_ops)
rc = c_ops->cnic_ctl(bp->cnic_data, ctl);
mutex_unlock(&bp->cnic_mutex);
return rc;
}
static int bnx2x_cnic_ctl_send_bh(struct bnx2x *bp, struct cnic_ctl_info *ctl)
{
struct cnic_ops *c_ops;
int rc = 0;
rcu_read_lock();
c_ops = rcu_dereference(bp->cnic_ops);
if (c_ops)
rc = c_ops->cnic_ctl(bp->cnic_data, ctl);
rcu_read_unlock();
return rc;
}
/*
* for commands that have no data
*/
int bnx2x_cnic_notify(struct bnx2x *bp, int cmd)
{
struct cnic_ctl_info ctl = {0};
ctl.cmd = cmd;
return bnx2x_cnic_ctl_send(bp, &ctl);
}
static void bnx2x_cnic_cfc_comp(struct bnx2x *bp, int cid, u8 err)
{
struct cnic_ctl_info ctl = {0};
/* first we tell CNIC and only then we count this as a completion */
ctl.cmd = CNIC_CTL_COMPLETION_CMD;
ctl.data.comp.cid = cid;
ctl.data.comp.error = err;
bnx2x_cnic_ctl_send_bh(bp, &ctl);
bnx2x_cnic_sp_post(bp, 0);
}
/* Called with netif_addr_lock_bh() taken.
* Sets an rx_mode config for an iSCSI ETH client.
* Doesn't block.
* Completion should be checked outside.
*/
static void bnx2x_set_iscsi_eth_rx_mode(struct bnx2x *bp, bool start)
{
unsigned long accept_flags = 0, ramrod_flags = 0;
u8 cl_id = bnx2x_cnic_eth_cl_id(bp, BNX2X_ISCSI_ETH_CL_ID_IDX);
int sched_state = BNX2X_FILTER_ISCSI_ETH_STOP_SCHED;
if (start) {
/* Start accepting on iSCSI L2 ring. Accept all multicasts
* because it's the only way for UIO Queue to accept
* multicasts (in non-promiscuous mode only one Queue per
* function will receive multicast packets (leading in our
* case).
*/
__set_bit(BNX2X_ACCEPT_UNICAST, &accept_flags);
__set_bit(BNX2X_ACCEPT_ALL_MULTICAST, &accept_flags);
__set_bit(BNX2X_ACCEPT_BROADCAST, &accept_flags);
__set_bit(BNX2X_ACCEPT_ANY_VLAN, &accept_flags);
/* Clear STOP_PENDING bit if START is requested */
clear_bit(BNX2X_FILTER_ISCSI_ETH_STOP_SCHED, &bp->sp_state);
sched_state = BNX2X_FILTER_ISCSI_ETH_START_SCHED;
} else
/* Clear START_PENDING bit if STOP is requested */
clear_bit(BNX2X_FILTER_ISCSI_ETH_START_SCHED, &bp->sp_state);
if (test_bit(BNX2X_FILTER_RX_MODE_PENDING, &bp->sp_state))
set_bit(sched_state, &bp->sp_state);
else {
__set_bit(RAMROD_RX, &ramrod_flags);
bnx2x_set_q_rx_mode(bp, cl_id, 0, accept_flags, 0,
ramrod_flags);
}
}
static int bnx2x_drv_ctl(struct net_device *dev, struct drv_ctl_info *ctl)
{
struct bnx2x *bp = netdev_priv(dev);
int rc = 0;
switch (ctl->cmd) {
case DRV_CTL_CTXTBL_WR_CMD: {
u32 index = ctl->data.io.offset;
dma_addr_t addr = ctl->data.io.dma_addr;
bnx2x_ilt_wr(bp, index, addr);
break;
}
case DRV_CTL_RET_L5_SPQ_CREDIT_CMD: {
int count = ctl->data.credit.credit_count;
bnx2x_cnic_sp_post(bp, count);
break;
}
/* rtnl_lock is held. */
case DRV_CTL_START_L2_CMD: {
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
unsigned long sp_bits = 0;
/* Configure the iSCSI classification object */
bnx2x_init_mac_obj(bp, &bp->iscsi_l2_mac_obj,
cp->iscsi_l2_client_id,
cp->iscsi_l2_cid, BP_FUNC(bp),
bnx2x_sp(bp, mac_rdata),
bnx2x_sp_mapping(bp, mac_rdata),
BNX2X_FILTER_MAC_PENDING,
&bp->sp_state, BNX2X_OBJ_TYPE_RX,
&bp->macs_pool);
/* Set iSCSI MAC address */
rc = bnx2x_set_iscsi_eth_mac_addr(bp);
if (rc)
break;
mmiowb();
barrier();
/* Start accepting on iSCSI L2 ring */
netif_addr_lock_bh(dev);
bnx2x_set_iscsi_eth_rx_mode(bp, true);
netif_addr_unlock_bh(dev);
/* bits to wait on */
__set_bit(BNX2X_FILTER_RX_MODE_PENDING, &sp_bits);
__set_bit(BNX2X_FILTER_ISCSI_ETH_START_SCHED, &sp_bits);
if (!bnx2x_wait_sp_comp(bp, sp_bits))
BNX2X_ERR("rx_mode completion timed out!\n");
break;
}
/* rtnl_lock is held. */
case DRV_CTL_STOP_L2_CMD: {
unsigned long sp_bits = 0;
/* Stop accepting on iSCSI L2 ring */
netif_addr_lock_bh(dev);
bnx2x_set_iscsi_eth_rx_mode(bp, false);
netif_addr_unlock_bh(dev);
/* bits to wait on */
__set_bit(BNX2X_FILTER_RX_MODE_PENDING, &sp_bits);
__set_bit(BNX2X_FILTER_ISCSI_ETH_STOP_SCHED, &sp_bits);
if (!bnx2x_wait_sp_comp(bp, sp_bits))
BNX2X_ERR("rx_mode completion timed out!\n");
mmiowb();
barrier();
/* Unset iSCSI L2 MAC */
rc = bnx2x_del_all_macs(bp, &bp->iscsi_l2_mac_obj,
BNX2X_ISCSI_ETH_MAC, true);
break;
}
case DRV_CTL_RET_L2_SPQ_CREDIT_CMD: {
int count = ctl->data.credit.credit_count;
smp_mb__before_atomic();
atomic_add(count, &bp->cq_spq_left);
smp_mb__after_atomic();
break;
}
case DRV_CTL_ULP_REGISTER_CMD: {
int ulp_type = ctl->data.register_data.ulp_type;
if (CHIP_IS_E3(bp)) {
int idx = BP_FW_MB_IDX(bp);
u32 cap = SHMEM2_RD(bp, drv_capabilities_flag[idx]);
int path = BP_PATH(bp);
int port = BP_PORT(bp);
int i;
u32 scratch_offset;
u32 *host_addr;
/* first write capability to shmem2 */
if (ulp_type == CNIC_ULP_ISCSI)
cap |= DRV_FLAGS_CAPABILITIES_LOADED_ISCSI;
else if (ulp_type == CNIC_ULP_FCOE)
cap |= DRV_FLAGS_CAPABILITIES_LOADED_FCOE;
SHMEM2_WR(bp, drv_capabilities_flag[idx], cap);
if ((ulp_type != CNIC_ULP_FCOE) ||
(!SHMEM2_HAS(bp, ncsi_oem_data_addr)) ||
(!(bp->flags & BC_SUPPORTS_FCOE_FEATURES)))
break;
/* if reached here - should write fcoe capabilities */
scratch_offset = SHMEM2_RD(bp, ncsi_oem_data_addr);
if (!scratch_offset)
break;
scratch_offset += offsetof(struct glob_ncsi_oem_data,
fcoe_features[path][port]);
host_addr = (u32 *) &(ctl->data.register_data.
fcoe_features);
for (i = 0; i < sizeof(struct fcoe_capabilities);
i += 4)
REG_WR(bp, scratch_offset + i,
*(host_addr + i/4));
}
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_GET_DRV_VERSION, 0);
break;
}
case DRV_CTL_ULP_UNREGISTER_CMD: {
int ulp_type = ctl->data.ulp_type;
if (CHIP_IS_E3(bp)) {
int idx = BP_FW_MB_IDX(bp);
u32 cap;
cap = SHMEM2_RD(bp, drv_capabilities_flag[idx]);
if (ulp_type == CNIC_ULP_ISCSI)
cap &= ~DRV_FLAGS_CAPABILITIES_LOADED_ISCSI;
else if (ulp_type == CNIC_ULP_FCOE)
cap &= ~DRV_FLAGS_CAPABILITIES_LOADED_FCOE;
SHMEM2_WR(bp, drv_capabilities_flag[idx], cap);
}
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_GET_DRV_VERSION, 0);
break;
}
default:
BNX2X_ERR("unknown command %x\n", ctl->cmd);
rc = -EINVAL;
}
/* For storage-only interfaces, change driver state */
if (IS_MF_SD_STORAGE_PERSONALITY_ONLY(bp)) {
switch (ctl->drv_state) {
case DRV_NOP:
break;
case DRV_ACTIVE:
bnx2x_set_os_driver_state(bp,
OS_DRIVER_STATE_ACTIVE);
break;
case DRV_INACTIVE:
bnx2x_set_os_driver_state(bp,
OS_DRIVER_STATE_DISABLED);
break;
case DRV_UNLOADED:
bnx2x_set_os_driver_state(bp,
OS_DRIVER_STATE_NOT_LOADED);
break;
default:
BNX2X_ERR("Unknown cnic driver state: %d\n", ctl->drv_state);
}
}
return rc;
}
static int bnx2x_get_fc_npiv(struct net_device *dev,
struct cnic_fc_npiv_tbl *cnic_tbl)
{
struct bnx2x *bp = netdev_priv(dev);
struct bdn_fc_npiv_tbl *tbl = NULL;
u32 offset, entries;
int rc = -EINVAL;
int i;
if (!SHMEM2_HAS(bp, fc_npiv_nvram_tbl_addr[0]))
goto out;
DP(BNX2X_MSG_MCP, "About to read the FC-NPIV table\n");
tbl = kmalloc(sizeof(*tbl), GFP_KERNEL);
if (!tbl) {
BNX2X_ERR("Failed to allocate fc_npiv table\n");
goto out;
}
offset = SHMEM2_RD(bp, fc_npiv_nvram_tbl_addr[BP_PORT(bp)]);
if (!offset) {
DP(BNX2X_MSG_MCP, "No FC-NPIV in NVRAM\n");
goto out;
}
DP(BNX2X_MSG_MCP, "Offset of FC-NPIV in NVRAM: %08x\n", offset);
/* Read the table contents from nvram */
if (bnx2x_nvram_read(bp, offset, (u8 *)tbl, sizeof(*tbl))) {
BNX2X_ERR("Failed to read FC-NPIV table\n");
goto out;
}
/* Since bnx2x_nvram_read() returns data in be32, we need to convert
* the number of entries back to cpu endianness.
*/
entries = tbl->fc_npiv_cfg.num_of_npiv;
entries = (__force u32)be32_to_cpu((__force __be32)entries);
tbl->fc_npiv_cfg.num_of_npiv = entries;
if (!tbl->fc_npiv_cfg.num_of_npiv) {
DP(BNX2X_MSG_MCP,
"No FC-NPIV table [valid, simply not present]\n");
goto out;
} else if (tbl->fc_npiv_cfg.num_of_npiv > MAX_NUMBER_NPIV) {
BNX2X_ERR("FC-NPIV table with bad length 0x%08x\n",
tbl->fc_npiv_cfg.num_of_npiv);
goto out;
} else {
DP(BNX2X_MSG_MCP, "Read 0x%08x entries from NVRAM\n",
tbl->fc_npiv_cfg.num_of_npiv);
}
/* Copy the data into cnic-provided struct */
cnic_tbl->count = tbl->fc_npiv_cfg.num_of_npiv;
for (i = 0; i < cnic_tbl->count; i++) {
memcpy(cnic_tbl->wwpn[i], tbl->settings[i].npiv_wwpn, 8);
memcpy(cnic_tbl->wwnn[i], tbl->settings[i].npiv_wwnn, 8);
}
rc = 0;
out:
kfree(tbl);
return rc;
}
void bnx2x_setup_cnic_irq_info(struct bnx2x *bp)
{
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
if (bp->flags & USING_MSIX_FLAG) {
cp->drv_state |= CNIC_DRV_STATE_USING_MSIX;
cp->irq_arr[0].irq_flags |= CNIC_IRQ_FL_MSIX;
cp->irq_arr[0].vector = bp->msix_table[1].vector;
} else {
cp->drv_state &= ~CNIC_DRV_STATE_USING_MSIX;
cp->irq_arr[0].irq_flags &= ~CNIC_IRQ_FL_MSIX;
}
if (!CHIP_IS_E1x(bp))
cp->irq_arr[0].status_blk = (void *)bp->cnic_sb.e2_sb;
else
cp->irq_arr[0].status_blk = (void *)bp->cnic_sb.e1x_sb;
cp->irq_arr[0].status_blk_num = bnx2x_cnic_fw_sb_id(bp);
cp->irq_arr[0].status_blk_num2 = bnx2x_cnic_igu_sb_id(bp);
cp->irq_arr[1].status_blk = bp->def_status_blk;
cp->irq_arr[1].status_blk_num = DEF_SB_ID;
cp->irq_arr[1].status_blk_num2 = DEF_SB_IGU_ID;
cp->num_irq = 2;
}
void bnx2x_setup_cnic_info(struct bnx2x *bp)
{
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
cp->ctx_tbl_offset = FUNC_ILT_BASE(BP_FUNC(bp)) +
bnx2x_cid_ilt_lines(bp);
cp->starting_cid = bnx2x_cid_ilt_lines(bp) * ILT_PAGE_CIDS;
cp->fcoe_init_cid = BNX2X_FCOE_ETH_CID(bp);
cp->iscsi_l2_cid = BNX2X_ISCSI_ETH_CID(bp);
DP(NETIF_MSG_IFUP, "BNX2X_1st_NON_L2_ETH_CID(bp) %x, cp->starting_cid %x, cp->fcoe_init_cid %x, cp->iscsi_l2_cid %x\n",
BNX2X_1st_NON_L2_ETH_CID(bp), cp->starting_cid, cp->fcoe_init_cid,
cp->iscsi_l2_cid);
if (NO_ISCSI_OOO(bp))
cp->drv_state |= CNIC_DRV_STATE_NO_ISCSI_OOO;
}
static int bnx2x_register_cnic(struct net_device *dev, struct cnic_ops *ops,
void *data)
{
struct bnx2x *bp = netdev_priv(dev);
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
int rc;
DP(NETIF_MSG_IFUP, "Register_cnic called\n");
if (ops == NULL) {
BNX2X_ERR("NULL ops received\n");
return -EINVAL;
}
if (!CNIC_SUPPORT(bp)) {
BNX2X_ERR("Can't register CNIC when not supported\n");
return -EOPNOTSUPP;
}
if (!CNIC_LOADED(bp)) {
rc = bnx2x_load_cnic(bp);
if (rc) {
BNX2X_ERR("CNIC-related load failed\n");
return rc;
}
}
bp->cnic_enabled = true;
bp->cnic_kwq = kzalloc(PAGE_SIZE, GFP_KERNEL);
if (!bp->cnic_kwq)
return -ENOMEM;
bp->cnic_kwq_cons = bp->cnic_kwq;
bp->cnic_kwq_prod = bp->cnic_kwq;
bp->cnic_kwq_last = bp->cnic_kwq + MAX_SP_DESC_CNT;
bp->cnic_spq_pending = 0;
bp->cnic_kwq_pending = 0;
bp->cnic_data = data;
cp->num_irq = 0;
cp->drv_state |= CNIC_DRV_STATE_REGD;
cp->iro_arr = bp->iro_arr;
bnx2x_setup_cnic_irq_info(bp);
rcu_assign_pointer(bp->cnic_ops, ops);
/* Schedule driver to read CNIC driver versions */
bnx2x_schedule_sp_rtnl(bp, BNX2X_SP_RTNL_GET_DRV_VERSION, 0);
return 0;
}
static int bnx2x_unregister_cnic(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
mutex_lock(&bp->cnic_mutex);
cp->drv_state = 0;
RCU_INIT_POINTER(bp->cnic_ops, NULL);
mutex_unlock(&bp->cnic_mutex);
synchronize_rcu();
bp->cnic_enabled = false;
kfree(bp->cnic_kwq);
bp->cnic_kwq = NULL;
return 0;
}
static struct cnic_eth_dev *bnx2x_cnic_probe(struct net_device *dev)
{
struct bnx2x *bp = netdev_priv(dev);
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
/* If both iSCSI and FCoE are disabled - return NULL in
* order to indicate CNIC that it should not try to work
* with this device.
*/
if (NO_ISCSI(bp) && NO_FCOE(bp))
return NULL;
cp->drv_owner = THIS_MODULE;
cp->chip_id = CHIP_ID(bp);
cp->pdev = bp->pdev;
cp->io_base = bp->regview;
cp->io_base2 = bp->doorbells;
cp->max_kwqe_pending = 8;
cp->ctx_blk_size = CDU_ILT_PAGE_SZ;
cp->ctx_tbl_offset = FUNC_ILT_BASE(BP_FUNC(bp)) +
bnx2x_cid_ilt_lines(bp);
cp->ctx_tbl_len = CNIC_ILT_LINES;
cp->starting_cid = bnx2x_cid_ilt_lines(bp) * ILT_PAGE_CIDS;
cp->drv_submit_kwqes_16 = bnx2x_cnic_sp_queue;
cp->drv_ctl = bnx2x_drv_ctl;
cp->drv_get_fc_npiv_tbl = bnx2x_get_fc_npiv;
cp->drv_register_cnic = bnx2x_register_cnic;
cp->drv_unregister_cnic = bnx2x_unregister_cnic;
cp->fcoe_init_cid = BNX2X_FCOE_ETH_CID(bp);
cp->iscsi_l2_client_id =
bnx2x_cnic_eth_cl_id(bp, BNX2X_ISCSI_ETH_CL_ID_IDX);
cp->iscsi_l2_cid = BNX2X_ISCSI_ETH_CID(bp);
if (NO_ISCSI_OOO(bp))
cp->drv_state |= CNIC_DRV_STATE_NO_ISCSI_OOO;
if (NO_ISCSI(bp))
cp->drv_state |= CNIC_DRV_STATE_NO_ISCSI;
if (NO_FCOE(bp))
cp->drv_state |= CNIC_DRV_STATE_NO_FCOE;
BNX2X_DEV_INFO(
"page_size %d, tbl_offset %d, tbl_lines %d, starting cid %d\n",
cp->ctx_blk_size,
cp->ctx_tbl_offset,
cp->ctx_tbl_len,
cp->starting_cid);
return cp;
}
static u32 bnx2x_rx_ustorm_prods_offset(struct bnx2x_fastpath *fp)
{
struct bnx2x *bp = fp->bp;
u32 offset = BAR_USTRORM_INTMEM;
if (IS_VF(bp))
return bnx2x_vf_ustorm_prods_offset(bp, fp);
else if (!CHIP_IS_E1x(bp))
offset += USTORM_RX_PRODS_E2_OFFSET(fp->cl_qzone_id);
else
offset += USTORM_RX_PRODS_E1X_OFFSET(BP_PORT(bp), fp->cl_id);
return offset;
}
/* called only on E1H or E2.
* When pretending to be PF, the pretend value is the function number 0...7
* When pretending to be VF, the pretend val is the PF-num:VF-valid:ABS-VFID
* combination
*/
int bnx2x_pretend_func(struct bnx2x *bp, u16 pretend_func_val)
{
u32 pretend_reg;
if (CHIP_IS_E1H(bp) && pretend_func_val >= E1H_FUNC_MAX)
return -1;
/* get my own pretend register */
pretend_reg = bnx2x_get_pretend_reg(bp);
REG_WR(bp, pretend_reg, pretend_func_val);
REG_RD(bp, pretend_reg);
return 0;
}
static void bnx2x_ptp_task(struct work_struct *work)
{
struct bnx2x *bp = container_of(work, struct bnx2x, ptp_task);
int port = BP_PORT(bp);
u32 val_seq;
u64 timestamp, ns;
struct skb_shared_hwtstamps shhwtstamps;
/* Read Tx timestamp registers */
val_seq = REG_RD(bp, port ? NIG_REG_P1_TLLH_PTP_BUF_SEQID :
NIG_REG_P0_TLLH_PTP_BUF_SEQID);
if (val_seq & 0x10000) {
/* There is a valid timestamp value */
timestamp = REG_RD(bp, port ? NIG_REG_P1_TLLH_PTP_BUF_TS_MSB :
NIG_REG_P0_TLLH_PTP_BUF_TS_MSB);
timestamp <<= 32;
timestamp |= REG_RD(bp, port ? NIG_REG_P1_TLLH_PTP_BUF_TS_LSB :
NIG_REG_P0_TLLH_PTP_BUF_TS_LSB);
/* Reset timestamp register to allow new timestamp */
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_BUF_SEQID :
NIG_REG_P0_TLLH_PTP_BUF_SEQID, 0x10000);
ns = timecounter_cyc2time(&bp->timecounter, timestamp);
memset(&shhwtstamps, 0, sizeof(shhwtstamps));
shhwtstamps.hwtstamp = ns_to_ktime(ns);
skb_tstamp_tx(bp->ptp_tx_skb, &shhwtstamps);
dev_kfree_skb_any(bp->ptp_tx_skb);
bp->ptp_tx_skb = NULL;
DP(BNX2X_MSG_PTP, "Tx timestamp, timestamp cycles = %llu, ns = %llu\n",
timestamp, ns);
} else {
DP(BNX2X_MSG_PTP, "There is no valid Tx timestamp yet\n");
/* Reschedule to keep checking for a valid timestamp value */
schedule_work(&bp->ptp_task);
}
}
void bnx2x_set_rx_ts(struct bnx2x *bp, struct sk_buff *skb)
{
int port = BP_PORT(bp);
u64 timestamp, ns;
timestamp = REG_RD(bp, port ? NIG_REG_P1_LLH_PTP_HOST_BUF_TS_MSB :
NIG_REG_P0_LLH_PTP_HOST_BUF_TS_MSB);
timestamp <<= 32;
timestamp |= REG_RD(bp, port ? NIG_REG_P1_LLH_PTP_HOST_BUF_TS_LSB :
NIG_REG_P0_LLH_PTP_HOST_BUF_TS_LSB);
/* Reset timestamp register to allow new timestamp */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_HOST_BUF_SEQID :
NIG_REG_P0_LLH_PTP_HOST_BUF_SEQID, 0x10000);
ns = timecounter_cyc2time(&bp->timecounter, timestamp);
skb_hwtstamps(skb)->hwtstamp = ns_to_ktime(ns);
DP(BNX2X_MSG_PTP, "Rx timestamp, timestamp cycles = %llu, ns = %llu\n",
timestamp, ns);
}
/* Read the PHC */
static cycle_t bnx2x_cyclecounter_read(const struct cyclecounter *cc)
{
struct bnx2x *bp = container_of(cc, struct bnx2x, cyclecounter);
int port = BP_PORT(bp);
u32 wb_data[2];
u64 phc_cycles;
REG_RD_DMAE(bp, port ? NIG_REG_TIMESYNC_GEN_REG + tsgen_synctime_t1 :
NIG_REG_TIMESYNC_GEN_REG + tsgen_synctime_t0, wb_data, 2);
phc_cycles = wb_data[1];
phc_cycles = (phc_cycles << 32) + wb_data[0];
DP(BNX2X_MSG_PTP, "PHC read cycles = %llu\n", phc_cycles);
return phc_cycles;
}
static void bnx2x_init_cyclecounter(struct bnx2x *bp)
{
memset(&bp->cyclecounter, 0, sizeof(bp->cyclecounter));
bp->cyclecounter.read = bnx2x_cyclecounter_read;
bp->cyclecounter.mask = CYCLECOUNTER_MASK(64);
bp->cyclecounter.shift = 0;
bp->cyclecounter.mult = 1;
}
static int bnx2x_send_reset_timesync_ramrod(struct bnx2x *bp)
{
struct bnx2x_func_state_params func_params = {NULL};
struct bnx2x_func_set_timesync_params *set_timesync_params =
&func_params.params.set_timesync;
/* Prepare parameters for function state transitions */
__set_bit(RAMROD_COMP_WAIT, &func_params.ramrod_flags);
__set_bit(RAMROD_RETRY, &func_params.ramrod_flags);
func_params.f_obj = &bp->func_obj;
func_params.cmd = BNX2X_F_CMD_SET_TIMESYNC;
/* Function parameters */
set_timesync_params->drift_adjust_cmd = TS_DRIFT_ADJUST_RESET;
set_timesync_params->offset_cmd = TS_OFFSET_KEEP;
return bnx2x_func_state_change(bp, &func_params);
}
static int bnx2x_enable_ptp_packets(struct bnx2x *bp)
{
struct bnx2x_queue_state_params q_params;
int rc, i;
/* send queue update ramrod to enable PTP packets */
memset(&q_params, 0, sizeof(q_params));
__set_bit(RAMROD_COMP_WAIT, &q_params.ramrod_flags);
q_params.cmd = BNX2X_Q_CMD_UPDATE;
__set_bit(BNX2X_Q_UPDATE_PTP_PKTS_CHNG,
&q_params.params.update.update_flags);
__set_bit(BNX2X_Q_UPDATE_PTP_PKTS,
&q_params.params.update.update_flags);
/* send the ramrod on all the queues of the PF */
for_each_eth_queue(bp, i) {
struct bnx2x_fastpath *fp = &bp->fp[i];
/* Set the appropriate Queue object */
q_params.q_obj = &bnx2x_sp_obj(bp, fp).q_obj;
/* Update the Queue state */
rc = bnx2x_queue_state_change(bp, &q_params);
if (rc) {
BNX2X_ERR("Failed to enable PTP packets\n");
return rc;
}
}
return 0;
}
int bnx2x_configure_ptp_filters(struct bnx2x *bp)
{
int port = BP_PORT(bp);
int rc;
if (!bp->hwtstamp_ioctl_called)
return 0;
switch (bp->tx_type) {
case HWTSTAMP_TX_ON:
bp->flags |= TX_TIMESTAMPING_EN;
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_PARAM_MASK :
NIG_REG_P0_TLLH_PTP_PARAM_MASK, 0x6AA);
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_RULE_MASK :
NIG_REG_P0_TLLH_PTP_RULE_MASK, 0x3EEE);
break;
case HWTSTAMP_TX_ONESTEP_SYNC:
BNX2X_ERR("One-step timestamping is not supported\n");
return -ERANGE;
}
switch (bp->rx_filter) {
case HWTSTAMP_FILTER_NONE:
break;
case HWTSTAMP_FILTER_ALL:
case HWTSTAMP_FILTER_SOME:
bp->rx_filter = HWTSTAMP_FILTER_NONE;
break;
case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
bp->rx_filter = HWTSTAMP_FILTER_PTP_V1_L4_EVENT;
/* Initialize PTP detection for UDP/IPv4 events */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_PARAM_MASK :
NIG_REG_P0_LLH_PTP_PARAM_MASK, 0x7EE);
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_RULE_MASK :
NIG_REG_P0_LLH_PTP_RULE_MASK, 0x3FFE);
break;
case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
bp->rx_filter = HWTSTAMP_FILTER_PTP_V2_L4_EVENT;
/* Initialize PTP detection for UDP/IPv4 or UDP/IPv6 events */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_PARAM_MASK :
NIG_REG_P0_LLH_PTP_PARAM_MASK, 0x7EA);
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_RULE_MASK :
NIG_REG_P0_LLH_PTP_RULE_MASK, 0x3FEE);
break;
case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
bp->rx_filter = HWTSTAMP_FILTER_PTP_V2_L2_EVENT;
/* Initialize PTP detection L2 events */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_PARAM_MASK :
NIG_REG_P0_LLH_PTP_PARAM_MASK, 0x6BF);
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_RULE_MASK :
NIG_REG_P0_LLH_PTP_RULE_MASK, 0x3EFF);
break;
case HWTSTAMP_FILTER_PTP_V2_EVENT:
case HWTSTAMP_FILTER_PTP_V2_SYNC:
case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
bp->rx_filter = HWTSTAMP_FILTER_PTP_V2_EVENT;
/* Initialize PTP detection L2, UDP/IPv4 or UDP/IPv6 events */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_PARAM_MASK :
NIG_REG_P0_LLH_PTP_PARAM_MASK, 0x6AA);
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_RULE_MASK :
NIG_REG_P0_LLH_PTP_RULE_MASK, 0x3EEE);
break;
}
/* Indicate to FW that this PF expects recorded PTP packets */
rc = bnx2x_enable_ptp_packets(bp);
if (rc)
return rc;
/* Enable sending PTP packets to host */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_TO_HOST :
NIG_REG_P0_LLH_PTP_TO_HOST, 0x1);
return 0;
}
static int bnx2x_hwtstamp_ioctl(struct bnx2x *bp, struct ifreq *ifr)
{
struct hwtstamp_config config;
int rc;
DP(BNX2X_MSG_PTP, "HWTSTAMP IOCTL called\n");
if (copy_from_user(&config, ifr->ifr_data, sizeof(config)))
return -EFAULT;
DP(BNX2X_MSG_PTP, "Requested tx_type: %d, requested rx_filters = %d\n",
config.tx_type, config.rx_filter);
if (config.flags) {
BNX2X_ERR("config.flags is reserved for future use\n");
return -EINVAL;
}
bp->hwtstamp_ioctl_called = 1;
bp->tx_type = config.tx_type;
bp->rx_filter = config.rx_filter;
rc = bnx2x_configure_ptp_filters(bp);
if (rc)
return rc;
config.rx_filter = bp->rx_filter;
return copy_to_user(ifr->ifr_data, &config, sizeof(config)) ?
-EFAULT : 0;
}
/* Configures HW for PTP */
static int bnx2x_configure_ptp(struct bnx2x *bp)
{
int rc, port = BP_PORT(bp);
u32 wb_data[2];
/* Reset PTP event detection rules - will be configured in the IOCTL */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_PARAM_MASK :
NIG_REG_P0_LLH_PTP_PARAM_MASK, 0x7FF);
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_RULE_MASK :
NIG_REG_P0_LLH_PTP_RULE_MASK, 0x3FFF);
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_PARAM_MASK :
NIG_REG_P0_TLLH_PTP_PARAM_MASK, 0x7FF);
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_RULE_MASK :
NIG_REG_P0_TLLH_PTP_RULE_MASK, 0x3FFF);
/* Disable PTP packets to host - will be configured in the IOCTL*/
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_TO_HOST :
NIG_REG_P0_LLH_PTP_TO_HOST, 0x0);
/* Enable the PTP feature */
REG_WR(bp, port ? NIG_REG_P1_PTP_EN :
NIG_REG_P0_PTP_EN, 0x3F);
/* Enable the free-running counter */
wb_data[0] = 0;
wb_data[1] = 0;
REG_WR_DMAE(bp, NIG_REG_TIMESYNC_GEN_REG + tsgen_ctrl, wb_data, 2);
/* Reset drift register (offset register is not reset) */
rc = bnx2x_send_reset_timesync_ramrod(bp);
if (rc) {
BNX2X_ERR("Failed to reset PHC drift register\n");
return -EFAULT;
}
/* Reset possibly old timestamps */
REG_WR(bp, port ? NIG_REG_P1_LLH_PTP_HOST_BUF_SEQID :
NIG_REG_P0_LLH_PTP_HOST_BUF_SEQID, 0x10000);
REG_WR(bp, port ? NIG_REG_P1_TLLH_PTP_BUF_SEQID :
NIG_REG_P0_TLLH_PTP_BUF_SEQID, 0x10000);
return 0;
}
/* Called during load, to initialize PTP-related stuff */
void bnx2x_init_ptp(struct bnx2x *bp)
{
int rc;
/* Configure PTP in HW */
rc = bnx2x_configure_ptp(bp);
if (rc) {
BNX2X_ERR("Stopping PTP initialization\n");
return;
}
/* Init work queue for Tx timestamping */
INIT_WORK(&bp->ptp_task, bnx2x_ptp_task);
/* Init cyclecounter and timecounter. This is done only in the first
* load. If done in every load, PTP application will fail when doing
* unload / load (e.g. MTU change) while it is running.
*/
if (!bp->timecounter_init_done) {
bnx2x_init_cyclecounter(bp);
timecounter_init(&bp->timecounter, &bp->cyclecounter,
ktime_to_ns(ktime_get_real()));
bp->timecounter_init_done = 1;
}
DP(BNX2X_MSG_PTP, "PTP initialization ended successfully\n");
}