linux-brain/drivers/ata/pata_cs5530.c

358 lines
8.6 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* pata-cs5530.c - CS5530 PATA for new ATA layer
* (C) 2005 Red Hat Inc
*
* based upon cs5530.c by Mark Lord.
*
* Loosely based on the piix & svwks drivers.
*
* Documentation:
* Available from AMD web site.
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/blkdev.h>
#include <linux/delay.h>
#include <scsi/scsi_host.h>
#include <linux/libata.h>
#include <linux/dmi.h>
#define DRV_NAME "pata_cs5530"
#define DRV_VERSION "0.7.4"
static void __iomem *cs5530_port_base(struct ata_port *ap)
{
unsigned long bmdma = (unsigned long)ap->ioaddr.bmdma_addr;
return (void __iomem *)((bmdma & ~0x0F) + 0x20 + 0x10 * ap->port_no);
}
/**
* cs5530_set_piomode - PIO setup
* @ap: ATA interface
* @adev: device on the interface
*
* Set our PIO requirements. This is fairly simple on the CS5530
* chips.
*/
static void cs5530_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
static const unsigned int cs5530_pio_timings[2][5] = {
{0x00009172, 0x00012171, 0x00020080, 0x00032010, 0x00040010},
{0xd1329172, 0x71212171, 0x30200080, 0x20102010, 0x00100010}
};
void __iomem *base = cs5530_port_base(ap);
u32 tuning;
int format;
/* Find out which table to use */
tuning = ioread32(base + 0x04);
format = (tuning & 0x80000000UL) ? 1 : 0;
/* Now load the right timing register */
if (adev->devno)
base += 0x08;
iowrite32(cs5530_pio_timings[format][adev->pio_mode - XFER_PIO_0], base);
}
/**
* cs5530_set_dmamode - DMA timing setup
* @ap: ATA interface
* @adev: Device being configured
*
* We cannot mix MWDMA and UDMA without reloading timings each switch
* master to slave. We track the last DMA setup in order to minimise
* reloads.
*/
static void cs5530_set_dmamode(struct ata_port *ap, struct ata_device *adev)
{
void __iomem *base = cs5530_port_base(ap);
u32 tuning, timing = 0;
u8 reg;
/* Find out which table to use */
tuning = ioread32(base + 0x04);
switch(adev->dma_mode) {
case XFER_UDMA_0:
timing = 0x00921250;break;
case XFER_UDMA_1:
timing = 0x00911140;break;
case XFER_UDMA_2:
timing = 0x00911030;break;
case XFER_MW_DMA_0:
timing = 0x00077771;break;
case XFER_MW_DMA_1:
timing = 0x00012121;break;
case XFER_MW_DMA_2:
timing = 0x00002020;break;
default:
BUG();
}
/* Merge in the PIO format bit */
timing |= (tuning & 0x80000000UL);
if (adev->devno == 0) /* Master */
iowrite32(timing, base + 0x04);
else {
if (timing & 0x00100000)
tuning |= 0x00100000; /* UDMA for both */
else
tuning &= ~0x00100000; /* MWDMA for both */
iowrite32(tuning, base + 0x04);
iowrite32(timing, base + 0x0C);
}
/* Set the DMA capable bit in the BMDMA area */
reg = ioread8(ap->ioaddr.bmdma_addr + ATA_DMA_STATUS);
reg |= (1 << (5 + adev->devno));
iowrite8(reg, ap->ioaddr.bmdma_addr + ATA_DMA_STATUS);
/* Remember the last DMA setup we did */
ap->private_data = adev;
}
/**
* cs5530_qc_issue - command issue
* @qc: command pending
*
* Called when the libata layer is about to issue a command. We wrap
* this interface so that we can load the correct ATA timings if
* necessary. Specifically we have a problem that there is only
* one MWDMA/UDMA bit.
*/
static unsigned int cs5530_qc_issue(struct ata_queued_cmd *qc)
{
struct ata_port *ap = qc->ap;
struct ata_device *adev = qc->dev;
struct ata_device *prev = ap->private_data;
/* See if the DMA settings could be wrong */
if (ata_dma_enabled(adev) && adev != prev && prev != NULL) {
/* Maybe, but do the channels match MWDMA/UDMA ? */
if ((ata_using_udma(adev) && !ata_using_udma(prev)) ||
(ata_using_udma(prev) && !ata_using_udma(adev)))
/* Switch the mode bits */
cs5530_set_dmamode(ap, adev);
}
return ata_bmdma_qc_issue(qc);
}
static struct scsi_host_template cs5530_sht = {
ATA_BMDMA_SHT(DRV_NAME),
.sg_tablesize = LIBATA_DUMB_MAX_PRD,
};
static struct ata_port_operations cs5530_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 12:22:49 +09:00
.inherits = &ata_bmdma_port_ops,
.qc_prep = ata_bmdma_dumb_qc_prep,
.qc_issue = cs5530_qc_issue,
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 12:22:49 +09:00
.cable_detect = ata_cable_40wire,
.set_piomode = cs5530_set_piomode,
.set_dmamode = cs5530_set_dmamode,
};
static const struct dmi_system_id palmax_dmi_table[] = {
{
.ident = "Palmax PD1100",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Cyrix"),
DMI_MATCH(DMI_PRODUCT_NAME, "Caddis"),
},
},
{ }
};
static int cs5530_is_palmax(void)
{
if (dmi_check_system(palmax_dmi_table)) {
printk(KERN_INFO "Palmax PD1100: Disabling DMA on docking port.\n");
return 1;
}
return 0;
}
/**
* cs5530_init_chip - Chipset init
*
* Perform the chip initialisation work that is shared between both
* setup and resume paths
*/
static int cs5530_init_chip(void)
{
struct pci_dev *master_0 = NULL, *cs5530_0 = NULL, *dev = NULL;
while ((dev = pci_get_device(PCI_VENDOR_ID_CYRIX, PCI_ANY_ID, dev)) != NULL) {
switch (dev->device) {
case PCI_DEVICE_ID_CYRIX_PCI_MASTER:
master_0 = pci_dev_get(dev);
break;
case PCI_DEVICE_ID_CYRIX_5530_LEGACY:
cs5530_0 = pci_dev_get(dev);
break;
}
}
if (!master_0) {
printk(KERN_ERR DRV_NAME ": unable to locate PCI MASTER function\n");
goto fail_put;
}
if (!cs5530_0) {
printk(KERN_ERR DRV_NAME ": unable to locate CS5530 LEGACY function\n");
goto fail_put;
}
pci_set_master(cs5530_0);
pci_try_set_mwi(cs5530_0);
/*
* Set PCI CacheLineSize to 16-bytes:
* --> Write 0x04 into 8-bit PCI CACHELINESIZE reg of function 0 of the cs5530
*
* Note: This value is constant because the 5530 is only a Geode companion
*/
pci_write_config_byte(cs5530_0, PCI_CACHE_LINE_SIZE, 0x04);
/*
* Disable trapping of UDMA register accesses (Win98 hack):
* --> Write 0x5006 into 16-bit reg at offset 0xd0 of function 0 of the cs5530
*/
pci_write_config_word(cs5530_0, 0xd0, 0x5006);
/*
* Bit-1 at 0x40 enables MemoryWriteAndInvalidate on internal X-bus:
* The other settings are what is necessary to get the register
* into a sane state for IDE DMA operation.
*/
pci_write_config_byte(master_0, 0x40, 0x1e);
/*
* Set max PCI burst size (16-bytes seems to work best):
* 16bytes: set bit-1 at 0x41 (reg value of 0x16)
* all others: clear bit-1 at 0x41, and do:
* 128bytes: OR 0x00 at 0x41
* 256bytes: OR 0x04 at 0x41
* 512bytes: OR 0x08 at 0x41
* 1024bytes: OR 0x0c at 0x41
*/
pci_write_config_byte(master_0, 0x41, 0x14);
/*
* These settings are necessary to get the chip
* into a sane state for IDE DMA operation.
*/
pci_write_config_byte(master_0, 0x42, 0x00);
pci_write_config_byte(master_0, 0x43, 0xc1);
pci_dev_put(master_0);
pci_dev_put(cs5530_0);
return 0;
fail_put:
pci_dev_put(master_0);
pci_dev_put(cs5530_0);
return -ENODEV;
}
/**
* cs5530_init_one - Initialise a CS5530
* @dev: PCI device
* @id: Entry in match table
*
* Install a driver for the newly found CS5530 companion chip. Most of
* this is just housekeeping. We have to set the chip up correctly and
* turn off various bits of emulation magic.
*/
static int cs5530_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
{
libata: clean up SFF init mess The intention of using port_mask in SFF init helpers was to eventually support exoctic configurations such as combination of legacy and native port on the same controller. This never became actually necessary and the related code always has been subtly broken one way or the other. Now that new init model is in place, there is no reason to make common helpers capable of handling all corner cases. Exotic cases can simply dealt within LLDs as necessary. This patch removes port_mask handling in SFF init helpers. SFF init helpers don't take n_ports argument and interpret it into port_mask anymore. All information is carried via port_info. n_ports argument is dropped and always two ports are allocated. LLD can tell SFF to skip certain port by marking it dummy. Note that SFF code has been treating unuvailable ports this way for a long time until recent breakage fix from Linus and is consistent with how other drivers handle with unavailable ports. This fixes 1-port legacy host handling still broken after the recent native mode fix and simplifies SFF init logic. The following changes are made... * ata_pci_init_native_host() and ata_init_legacy_host() both now try to initialized whatever they can and mark failed ports dummy. They return 0 if any port is successfully initialized. * ata_pci_prepare_native_host() and ata_pci_init_one() now doesn't take n_ports argument. All info should be specified via port_info array. Always two ports are allocated. * ata_pci_init_bmdma() exported to be used by LLDs in exotic cases. * port_info handling in all LLDs are standardized - all port_info arrays are const stack variable named ppi. Unless the second port is different from the first, its port_info is specified as NULL (tells libata that it's identical to the last non-NULL port_info). * pata_hpt37x/hpt3x2n: don't modify static variable directly. Make an on-stack copy instead as ata_piix does. * pata_uli: It has 4 ports instead of 2. Don't use ata_pci_prepare_native_host(). Allocate the host explicitly and use init helpers. It's simple enough. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-05-04 19:43:58 +09:00
static const struct ata_port_info info = {
.flags = ATA_FLAG_SLAVE_POSS,
.pio_mask = ATA_PIO4,
.mwdma_mask = ATA_MWDMA2,
.udma_mask = ATA_UDMA2,
.port_ops = &cs5530_port_ops
};
/* The docking connector doesn't do UDMA, and it seems not MWDMA */
libata: clean up SFF init mess The intention of using port_mask in SFF init helpers was to eventually support exoctic configurations such as combination of legacy and native port on the same controller. This never became actually necessary and the related code always has been subtly broken one way or the other. Now that new init model is in place, there is no reason to make common helpers capable of handling all corner cases. Exotic cases can simply dealt within LLDs as necessary. This patch removes port_mask handling in SFF init helpers. SFF init helpers don't take n_ports argument and interpret it into port_mask anymore. All information is carried via port_info. n_ports argument is dropped and always two ports are allocated. LLD can tell SFF to skip certain port by marking it dummy. Note that SFF code has been treating unuvailable ports this way for a long time until recent breakage fix from Linus and is consistent with how other drivers handle with unavailable ports. This fixes 1-port legacy host handling still broken after the recent native mode fix and simplifies SFF init logic. The following changes are made... * ata_pci_init_native_host() and ata_init_legacy_host() both now try to initialized whatever they can and mark failed ports dummy. They return 0 if any port is successfully initialized. * ata_pci_prepare_native_host() and ata_pci_init_one() now doesn't take n_ports argument. All info should be specified via port_info array. Always two ports are allocated. * ata_pci_init_bmdma() exported to be used by LLDs in exotic cases. * port_info handling in all LLDs are standardized - all port_info arrays are const stack variable named ppi. Unless the second port is different from the first, its port_info is specified as NULL (tells libata that it's identical to the last non-NULL port_info). * pata_hpt37x/hpt3x2n: don't modify static variable directly. Make an on-stack copy instead as ata_piix does. * pata_uli: It has 4 ports instead of 2. Don't use ata_pci_prepare_native_host(). Allocate the host explicitly and use init helpers. It's simple enough. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-05-04 19:43:58 +09:00
static const struct ata_port_info info_palmax_secondary = {
.flags = ATA_FLAG_SLAVE_POSS,
.pio_mask = ATA_PIO4,
.port_ops = &cs5530_port_ops
};
libata: clean up SFF init mess The intention of using port_mask in SFF init helpers was to eventually support exoctic configurations such as combination of legacy and native port on the same controller. This never became actually necessary and the related code always has been subtly broken one way or the other. Now that new init model is in place, there is no reason to make common helpers capable of handling all corner cases. Exotic cases can simply dealt within LLDs as necessary. This patch removes port_mask handling in SFF init helpers. SFF init helpers don't take n_ports argument and interpret it into port_mask anymore. All information is carried via port_info. n_ports argument is dropped and always two ports are allocated. LLD can tell SFF to skip certain port by marking it dummy. Note that SFF code has been treating unuvailable ports this way for a long time until recent breakage fix from Linus and is consistent with how other drivers handle with unavailable ports. This fixes 1-port legacy host handling still broken after the recent native mode fix and simplifies SFF init logic. The following changes are made... * ata_pci_init_native_host() and ata_init_legacy_host() both now try to initialized whatever they can and mark failed ports dummy. They return 0 if any port is successfully initialized. * ata_pci_prepare_native_host() and ata_pci_init_one() now doesn't take n_ports argument. All info should be specified via port_info array. Always two ports are allocated. * ata_pci_init_bmdma() exported to be used by LLDs in exotic cases. * port_info handling in all LLDs are standardized - all port_info arrays are const stack variable named ppi. Unless the second port is different from the first, its port_info is specified as NULL (tells libata that it's identical to the last non-NULL port_info). * pata_hpt37x/hpt3x2n: don't modify static variable directly. Make an on-stack copy instead as ata_piix does. * pata_uli: It has 4 ports instead of 2. Don't use ata_pci_prepare_native_host(). Allocate the host explicitly and use init helpers. It's simple enough. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-05-04 19:43:58 +09:00
const struct ata_port_info *ppi[] = { &info, NULL };
int rc;
rc = pcim_enable_device(pdev);
if (rc)
return rc;
/* Chip initialisation */
if (cs5530_init_chip())
return -ENODEV;
if (cs5530_is_palmax())
libata: clean up SFF init mess The intention of using port_mask in SFF init helpers was to eventually support exoctic configurations such as combination of legacy and native port on the same controller. This never became actually necessary and the related code always has been subtly broken one way or the other. Now that new init model is in place, there is no reason to make common helpers capable of handling all corner cases. Exotic cases can simply dealt within LLDs as necessary. This patch removes port_mask handling in SFF init helpers. SFF init helpers don't take n_ports argument and interpret it into port_mask anymore. All information is carried via port_info. n_ports argument is dropped and always two ports are allocated. LLD can tell SFF to skip certain port by marking it dummy. Note that SFF code has been treating unuvailable ports this way for a long time until recent breakage fix from Linus and is consistent with how other drivers handle with unavailable ports. This fixes 1-port legacy host handling still broken after the recent native mode fix and simplifies SFF init logic. The following changes are made... * ata_pci_init_native_host() and ata_init_legacy_host() both now try to initialized whatever they can and mark failed ports dummy. They return 0 if any port is successfully initialized. * ata_pci_prepare_native_host() and ata_pci_init_one() now doesn't take n_ports argument. All info should be specified via port_info array. Always two ports are allocated. * ata_pci_init_bmdma() exported to be used by LLDs in exotic cases. * port_info handling in all LLDs are standardized - all port_info arrays are const stack variable named ppi. Unless the second port is different from the first, its port_info is specified as NULL (tells libata that it's identical to the last non-NULL port_info). * pata_hpt37x/hpt3x2n: don't modify static variable directly. Make an on-stack copy instead as ata_piix does. * pata_uli: It has 4 ports instead of 2. Don't use ata_pci_prepare_native_host(). Allocate the host explicitly and use init helpers. It's simple enough. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-05-04 19:43:58 +09:00
ppi[1] = &info_palmax_secondary;
/* Now kick off ATA set up */
return ata_pci_bmdma_init_one(pdev, ppi, &cs5530_sht, NULL, 0);
}
#ifdef CONFIG_PM_SLEEP
static int cs5530_reinit_one(struct pci_dev *pdev)
{
struct ata_host *host = pci_get_drvdata(pdev);
int rc;
rc = ata_pci_device_do_resume(pdev);
if (rc)
return rc;
/* If we fail on resume we are doomed */
if (cs5530_init_chip())
return -EIO;
ata_host_resume(host);
return 0;
}
#endif /* CONFIG_PM_SLEEP */
static const struct pci_device_id cs5530[] = {
{ PCI_VDEVICE(CYRIX, PCI_DEVICE_ID_CYRIX_5530_IDE), },
{ },
};
static struct pci_driver cs5530_pci_driver = {
.name = DRV_NAME,
.id_table = cs5530,
.probe = cs5530_init_one,
.remove = ata_pci_remove_one,
#ifdef CONFIG_PM_SLEEP
.suspend = ata_pci_device_suspend,
.resume = cs5530_reinit_one,
#endif
};
module_pci_driver(cs5530_pci_driver);
MODULE_AUTHOR("Alan Cox");
MODULE_DESCRIPTION("low-level driver for the Cyrix/NS/AMD 5530");
MODULE_LICENSE("GPL");
MODULE_DEVICE_TABLE(pci, cs5530);
MODULE_VERSION(DRV_VERSION);