i.MX platforms provide large AHB mapped space for QSPI, each
controller has 256MB. However, current driver only maps small
size (AHB buffer size) of AHB space, this implementation
causes i.MX failed to boot M4 with QSPI XIP image.
Add config CONFIG_FSL_QSPI_AHB_FULL_MAP (default enabled for i.MX)
to address above problem.
When the config is set:
1. Full AHB space is divided to each CS.
2. A dedicated LUT entry is used for AHB read only.
3. The MODE instruction in LUT is replaced to standard ADDR instruction
4. The address in spi_mem_op is used to SFAR and AHB read
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Ashish Kumar <Ashish.Kumar@nxp.com>
Reviewed-by: Kuldeep Singh <kuldeep.singh@nxp.com>
Add compatible string and driver data for i.MX7ULP.
Meanwhile, the address set to SFA1AD/SFA2AD/SFB1AD/SFB2AD should
align with 1KB, because the lowest 10 bits are reserved by the
registers definition.
For i.MX7ULP which has only 128Bytes AHB buffer, must align it
when setting the registers and selecting cs.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Ashish Kumar <Ashish.Kumar@nxp.com>
Reviewed-by: Kuldeep Singh <kuldeep.singh@nxp.com>
While the general policy of not taking changes to the project via pull
requests directly on GitHub has not changed, it can be useful to submit
a PR there in order to trigger a CI run on Azure. These are run
automatically and the results are populated back to GitHub. Add a note
to the template to reflect this.
Signed-off-by: Tom Rini <trini@konsulko.com>
Convert pcm058 support to use device trees and the driver model.
Add rudimentary boot scripts to the environment, expand README.
Signed-off-by: Niel Fourie <lusus@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
Add Phytec Mira device tree files, for use with pcm058.
>From Linux 5.6, commit 7111951b8d49 upstream
Signed-off-by: Niel Fourie <lusus@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
As per the SD physical layer specification version 7.10, erase
command (CMD38) and stop transmission command (CMD12) will generate
R1b response.
R1b = R1 + busy signal
A non-zero value after the R1 response indicates card is ready for
next command.
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
Erase block start address (CMD32) and erase block end address (CMD33)
command will generate R1 response for mmc SPI mode.
R1 response is 1 byte long for mmc SPI, so assign 1 byte as a response
for this commands.
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
Send status command (CMD13) will send R1 response under SD mode
but R2 response under SPI mode.
R2 response is 2 bytes long, so read 2 bytes for mmc SPI mode
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
The content of ssr is useful only for erase operations.
This saves erase time.
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
R1 response is 1 byte long for mmc SPI commands as per the updated
physical layer specification version 7.10.
So correct the resp and resp_size for existing commands
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
When variable i will become 0, while(i--) loop breaks but variable i will
again decrement to -1 because of i-- and that's why below condition
"if (!i && (r != resp_match_value)" will never execute, So doing "i--"
inside of while() loop solves this problem.
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
Current codes assume the OPTEE address is at the end of first DRAM bank.
Adjust the process to allow OPTEE in the middle of first bank.
When OPTEE memory is removed from first bank, it may split the first bank
to two banks, adjust the MMU table for the split case,
Since the default CONFIG_NR_DRAM_BANKS is 4, it is enough, just enlarge
i.MX8MP evk to default to avoid issue.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Silvano di Ninno <silvano.dininno@nxp.com>
Tested-by: Silvano di Ninno <silvano.dininno@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
We use non-dm code to configure the clk settings in order to simplify
dm clk driver in future, so remove the duplicated code from clk driver
Signed-off-by: Peng Fan <peng.fan@nxp.com>
To fused part, we need to disable nodes of dtb to let kernel boot.
To mfgtool, USB issue when using super-speed for mfgtool, temporally
work around the problem to use high-speed only.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
To use one defconfig for all boot device, we have to runtime set
env offset and return env medium according to the boot device.
This patch overrides the env_get_offset and env_get_location to
implement the feature.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
iMX8MP has 6 fused parts in each qualification tier, with core, VPU,
ISP, NPU or DSP fused respectively.
The configuration tables for enabled modules:
MIMX8ML8DVNLZAA Quad Core, VPU, NPU, ISP, DSP
MIMX8ML7DVNLZAA Quad Core, NPU, ISP
MIMX8ML6DVNLZAA Quad Core, VPU, ISP
MIMX8ML5DVNLZAA Quad Core, VPU
MIMX8ML4DVNLZAA Quad Lite
MIMX8ML3DVNLZAA Dual Core, VPU, NPU, ISP, DSP
Add the support in U-Boot
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
ROM SError happens on two cases:
1. ERR050342, on iMX8MQ HDCP enabled parts ROM writes to GPV1 register, but
when ROM patch lock is fused, this write will cause SError.
2. ERR050350, on iMX8MQ/MM/MN, when the field return fuse is burned, HAB
is field return mode, but the last 4K of ROM is still protected and cause
SError.
Since ROM mask SError until ATF unmask it, so then ATF always meets the
exception. This patch works around the issue in SPL by enabling SPL
Exception vectors table and the SError exception, take the exception
to eret immediately to clear the SError.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add imx_eqos_txclk_set_rate/imx_get_eqos_csr_clk to override the
weak function in driver
Add set_clk_eqos to configure eQoS clk
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Current DM CLK is a bit complicated, for simplity, let DM clk only
support enable/disable/get_rate. For the expected rate settings,
we use non-DM clk to do that. Then we could have simple DM clk for
i.MX and could also share between SPL/U-Boot proper.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
A53 CCM root max support 1GHz, to support high freq, we need
to switch ARM clk sources from ARM PLL directly.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add clocks for FEC and flexspi, and add set parent clock callback,
so DTS can assign clocks
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add USB relevant clocks to support usb clock settings for both
DM USB host and gadget drivers
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add set clock parent support.
Add ENET and flexspi related clocks to support assigned clocks
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The soc info without initialization value should be put into
data section. The driver could be used before relocation,
with it in BSS section could cause issue, since BSS section
is not initializated and it might overwrite other areas that
used by others, such as dtb.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The soc info without initialization value should be put into
data section. The driver could be used before relocation,
with it in BSS section could cause issue, since BSS section
is not initializated and it might overwrite other areas that
used by others, such as dtb.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The soc info without initialization value should be put into
data section. The driver could be used before relocation,
with it in BSS section could cause issue, since BSS section
is not initializated and it might overwrite other areas that
used by others, such as dtb.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add logic to automatically update umctl2's setting based
on phy training CDD value for rank to rank space issue
Acked-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Oliver Chen <Oliver.Chen@nxp.com>
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
In uMCTL2 Databook, for LPDDR4, it is recommended to set
this register to 1. This can avoid ddr bandwidth is lower
after booting with running for a while.
Reviewed-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Jian Li <jian.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
1. set SCHED.rdwr_idle_gap=0
2. set SCHED.pageclose=1
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Signed-off-by: Jian Li <jian.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
the DRAM Controller in i.MX8MP will support a feature called "Inline ECC".
This is supported for all 3 supported DRAM technologies (LPDDR4, DDR4 and
DDR3L). When this feature is enabled by software, the DRAM Controller
reserves 12.5% of DRAM capacity for ECC information, and presents only
the non-ECC portion (lower 87.5% of the installed capacity of DRAM) to
the rest of the SoC.
The DRAM memory can be divided into 8 regions so that if a use case only
requires ECC protection on a subset of memory, then only that subset of
memory need support inline ECC. If this occurs, then there is no
performance penalty accessing the non-ECC-protected memory (no need to
access ECC for this portion of the memory map). This is all configured
with the DRAM Controller.
Signed-off-by: Sherry Sun <sherry.sun@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>