Patrick Delaunay [Wed, 18 Mar 2020 08:24:56 +0000 (09:24 +0100)]
stm32mp: stm32prog: add support of ssbl copy
For reliability of boot from NAND/SPI-NAND (with read-disturb issue)
the SSBL can be present several time, when it is indicated in the
flashlayout with "Binary(X)".
The received binary is copied X times by U-Boot on the target.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:24:55 +0000 (09:24 +0100)]
stm32mp: stm32prog: adapt the MTD partitions
Dynamically adapt the MTD partitions in NOR/NAND/SPI-NAND when stm32prog
command detects in the parsed flash layout files:
- a fsbl partition in NOR.
- a tee partition in NOR/NAND/SPI-NAND
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:24:52 +0000 (09:24 +0100)]
stm32mp: stm32prog: add support of boot partition for eMMC device
Add support of eMMC device boot partition with
part_id = -1 for offset="boot1"
or = -2 for offset="boot2"
The stm32prog command configures the MMC DFU backend with "mmcpart"
and configure the eMMC (command "mmc bootbus" and "mmc partconf")
when the update is done.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:24:50 +0000 (09:24 +0100)]
stm32mp: stm32prog: add flash layout parsing
Build the list of device and of partition with
a tab separated value file with a stm32 header: the FlashLayout.tsv
(https://wiki.st.com/stm32mpu/wiki/STM32CubeProgrammer_flashlayout)
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:24:49 +0000 (09:24 +0100)]
stm32mp: add the command stm32prog
Add a specific command stm32prog for STM32MP soc family
witch allows to program the boot devices with the tool
STM32CubeProgrammer (http://www.st.com/STM32CubeProg).
This command uses the same UART STM32 protocol than MCU
STM32 with or USB with DFU protocol v1.1 (ithe MCU ST
extension are no supported).
The executed actions are based on a tab separated value file
with a stm32 header, the FlashLayout file
(https://wiki.st.com/stm32mpu/wiki/STM32CubeProgrammer_flashlayout).
This file is parsed by the U-Boot command to:
- initialize the devices
- create the partition table on each device
- initialize the DFU backend to access to not volatile memory
(NOR/NAND/SD/eMMC) or to virtual device (OTP/PMIC)
Up to STM32PROG_MAX_DEV (5) devices can be updated with a FlashLayout.
The communication between U-Boot and STM32CubeProgrammer is done with
the specific alternate configuration (see "AN5275: USB DFU/USART protocols
used in STM32MP1 Series bootloaders" for details).
The command stm32prog is executed when a boot from USB is detected
(selected with bootpins) and we can program the boot devices with
a simple command (on Windows or Linux):
PC $> STM32_Programmer_CLI -c port=usb1 -w flaslayout.tsv
1/ the ROM code loads TF-A in embedded RAM (DFU or uart)
2/ TF-A loads flashlayout file and U-Boot in DDR (DFU or uart)
3/ U-Boot executes the stm32prog command (DFU or uart)
Patrick Delaunay [Wed, 18 Mar 2020 08:22:53 +0000 (09:22 +0100)]
board: stm32mp1: adapt MTD partition for BOOT from NOR or NAND
Dynamically adapt the MTD partitions in NAND and SPI-NAND when boot from
NOR or NAND/SPI-NAND is detected.
This patch avoids to define the save MTD partition name for NOR and NAND
devices and issue with latest kernel: only the needed MTD partitions
are defined.
For boot from NOR
1/ bootloader (TF-A, U-Boot and OP-TE) in NOR
2/ one large UBI partition in NAND
For boot from NAND
1/ bootloader (TF-A, U-Boot and OP-TE) in MTD raw partition
2/ one large UBI partition
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:22:48 +0000 (09:22 +0100)]
board: stm32mp1: reserve memory for OP-TEE in device tree
Add reserve memory for OP-TEE in U-Boot and in kernel device tree:
- no more reduce the DDR size in "memory" node:
CONFIG_SYS_MEM_TOP_HIDE is no more used
- U-Boot device-tree defines the needed "reserved-memory" for OP-TEE
and U-Boot should not use this reserved memory: board_get_usable_ram_top
use lmb lib to found the first free region, the not reserved
memory, enough to relocate U-Boot: the needed size of U-Boot
is estimated with gd->mon_len + CONFIG_SYS_MALLOC_LEN.
- the optee node ("optee@...": firmware with compatible "linaro,optee-tz")
and the associated "reserved-memory" are deactivated in kernel device
tree when OP-TEE is not detected by U-Boot to prevent kernel issue
(memory is reserved but not used, optee driver probe failed).
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:22:46 +0000 (09:22 +0100)]
stm32mp1: dynamically build DFU_ALT_INFO
This patch reduces the stm32mp1 environment size and
builds dynamically the DFU board configuration with gpt
and mtd partitions and information from defconfig
(CONFIG_DFU_ALT_RAM0).
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:22:45 +0000 (09:22 +0100)]
board: stm32mp1: move set_dfu_alt_info in st common directory
Move the stm32mp1 common code set_dfu_alt_info() in common directory,
this patch reduce the maintenance effort on this generic part (not board
dependent).
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Patrick Delaunay [Wed, 18 Mar 2020 08:22:44 +0000 (09:22 +0100)]
board: stm32mp1: move board_get_mtdparts in st common directory
Move the stm32mp1 common code board_get_mtdparts() in common directory,
this patch reduce the maintenance effort on this generic part (not board
dependent).
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Marek Vasut [Wed, 22 Apr 2020 11:18:14 +0000 (13:18 +0200)]
ARM: stm32: Implement DDR3 coding on DHCOR SoM
The DHCOR board does exist in multiple variants with different DDR3
DRAM sizes. To cater for all of them, implement DDR3 code handling.
There are two GPIOs which code the DRAM size populated on the SoM,
read them out and use the value to pick the correct DDR3 config.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com> Signed-off-by: Marek Vasut <marex@denx.de> Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Cc: Patrick Delaunay <patrick.delaunay@st.com> Cc: Patrice Chotard <patrice.chotard@st.com>
Marek Vasut [Wed, 22 Apr 2020 11:18:13 +0000 (13:18 +0200)]
ARM: dts: stm32: Rework DDR DT inclusion
Adjust the DDR configuration dtsi such that they only generate the
DRAM configuration node, the DDR controller node is moved into the
stm32mp157-u-boot.dtsi itself. This permits including multiple DDR
configuration dtsi files in board DT.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com> Signed-off-by: Marek Vasut <marex@denx.de> Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Cc: Patrick Delaunay <patrick.delaunay@st.com> Cc: Patrice Chotard <patrice.chotard@st.com>
Marek Vasut [Wed, 22 Apr 2020 11:18:12 +0000 (13:18 +0200)]
ram: stm32mp1: Add support for multiple configs
Add support for multiple DRAM configuration subnodes, while retaining
the support for a single flat DRAM configuration node. This is useful
on systems which can be manufactured in multiple configurations and
where the DRAM configuration can be determined at runtime.
The code is augmented by a function which can be overridden on board
level, allowing a match on the configuration node name, very much like
the fitImage configuration node name matching works. The default match
is on the single top-level DRAM configuration, if matching on subnodes
is required, then this board_stm32mp1_ddr_config_name_match() must be
overridden.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Cc: Patrick Delaunay <patrick.delaunay@st.com> Cc: Patrice Chotard <patrice.chotard@st.com>
Marek Vasut [Wed, 22 Apr 2020 11:18:11 +0000 (13:18 +0200)]
ARM: stm32: Implement board coding on AV96
The AV96 board does exist in multiple variants. To cater for all of
them, implement board code handling. There are two GPIOs which code
the type of the board, read them out and use the value to pick the
correct device tree from an fitImage.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com> Signed-off-by: Marek Vasut <marex@denx.de> Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Cc: Patrick Delaunay <patrick.delaunay@st.com> Cc: Patrice Chotard <patrice.chotard@st.com>
Change-Id: Iddb330b9a66500495885457cbe17edc0eacaaf43
Marek Vasut [Wed, 22 Apr 2020 11:18:10 +0000 (13:18 +0200)]
ARM: stm32: Add board_early_init_f() to SPL
Add weak implementation of board_early_init_f() hook into the
STM32MP1 SPL. This can be used to read out e.g. configuration
straps before initializing the DRAM.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com> Signed-off-by: Marek Vasut <marex@denx.de> Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Cc: Patrick Delaunay <patrick.delaunay@st.com> Cc: Patrice Chotard <patrice.chotard@st.com>
Jagan Teki [Wed, 6 May 2020 07:50:25 +0000 (13:20 +0530)]
usb: dwc3: add dis_del_phy_power_chg_quirk
Add a quirk to clear the GUSB3PIPECTL.DELAYP1TRANS bit,
which specifies whether disable delay PHY power change
from P0 to P1/P2/P3 when link state changing from U0
to U1/U2/U3 respectively.
Neil Armstrong [Tue, 5 May 2020 08:43:17 +0000 (10:43 +0200)]
gpio: emulate open drain & open source in dm_gpio_set_value()
Handle the GPIOD_OPEN_DRAIN & GPIOD_OPEN_SOURCE flags to emulate open drain
and open source by setting the GPIO line as input depending on the
requested value.
The behaviour is taken from the Linux gpiolib.
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Pali Rohár [Tue, 31 Mar 2020 22:35:15 +0000 (00:35 +0200)]
Nokia RX-51: Enable CONFIG_CONSOLE_MUX
After this change both device display and serial console would contain
U-Boto output automatically without any future configuration. This would
allow easier debugging on real device as access to serial console is hard
and also in qemu emulator where it is easier to copy+paste from serial
console as from SDL framebuffer.
Pali Rohár [Tue, 31 Mar 2020 22:35:13 +0000 (00:35 +0200)]
Nokia RX-51: Remove PART* macros
Now when code for defining partitions is duplicated at two locations
(option CONFIG_MTDPARTS_DEFAULT in nokia_rx51_defconfig file and macro
OMAP_TAG_PARTITION_CONFIG in rx51.c file) there is no need to have common
macros. Lets inline PART* macros to rx51.c file.
Pali Rohár [Tue, 31 Mar 2020 22:35:11 +0000 (00:35 +0200)]
Nokia RX-51: Move code from defconfig back to C header file
In commit commit 37304aaf60bf ("Convert CONFIG_USE_PREBOOT and
CONFIG_PREBOOT to Kconfig") was moved complicated multiline script code
from C header to oneliner in defconfig. After this change multiline to wide
oneliner it is hard to read this code and even harder to debug. Moreover
this script code should be at place where are other scripts, so move it
back to C header file.
Define new env variable preboot which stores this script and in option
CONFIG_PREBOOT calls this preboot variable.
Pali Rohár [Tue, 31 Mar 2020 22:35:10 +0000 (00:35 +0200)]
Nokia RX-51: Move comment about CONFIG_SYS_TEXT_BASE to correct place
In commit commit 278b90ce786f ("configs: Migrate CONFIG_SYS_TEXT_BASE") was
moved definition for CONFIG_SYS_TEXT_BASE option but author probably forgot
to move also comment for lines which are moving. So do it now!
The memory allocated to store the FIT image containing SYSFW and board
configuration data is statically defined to the largest size expected.
Some additions to the board configuration data has pushed us slightly
over the current defined size on some HS devices, expand to 278000.
spi: nxp-fspi: Use new readl_poll_sleep_timeout API
Board gets reset when performing burst read/write operations. On the
other hand, no such behaviour is observed on small size operations.
In Linux, readl_poll_timeout API already adds delay of 1us which is
further skipped in U-boot. Hence, use new "readl_poll_sleep_timeout" API
which adds delay alongwith timeout functionality.
Jagan Teki [Sat, 2 May 2020 07:15:01 +0000 (12:45 +0530)]
iopoll: Add read_poll_timeout common API
Add read_poll_timeout common API similar to Linux iopoll.
readx_poll_timeout will trigger read_poll_timeout with
proper op. This will help to extend the functionalities
like sleep_us to poll timeout in future.
Jagan Teki [Fri, 1 May 2020 18:15:08 +0000 (23:45 +0530)]
clk: Fix clk func names in comments
clk function names in comments should be prefix with
clk instead of clock.
Fix it.
Cc: Simon Glass <sjg@chromium.org> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Jagan Teki <jagan@amarulasolutions.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Jagan Teki [Fri, 1 May 2020 16:34:50 +0000 (22:04 +0530)]
spi: Zap lpc32xx_ssp driver-related code
lpc32xx_ssp driver is deprecated, no active updates
and no board user, hence dropped the same.
Cc: Vladimir Zapolskiy <vz@mleia.com> Cc: Albert ARIBAUD <albert.aribaud@3adev.fr> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Ye Li [Tue, 5 May 2020 12:28:42 +0000 (20:28 +0800)]
imx8: Select boot device dynamically
For fspi build, we will enable both SPL NOR support and SPL SPI
support. SPL will dynamically check the resource owner then
select corresponding boot device.
Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Peng Fan [Tue, 5 May 2020 12:28:41 +0000 (20:28 +0800)]
imx: imx8qm/qxp: Recover SPL data section for partition reboot
When doing partition reboot, the boot image won't be reloaded by ROM,
it is just CPU reset to boot entry. The SW has to keep the boot image
inside the RAM unchanged. It includes both the TEXT section and DATA
section.
For SPL, the problem is DATA section will be updated at runtime, so in
next partition reboot the data is not same as the initial value from
cold boot. If any code depends on the initial value, then it will have
problem.
This patch introduces a mechanism to recover the data section
for partition reboot. It adds a new section in image for saving
data section. When from cold boot, the data section will be saved
to that new section at SPL early phase. When from partition reboot,
the data section will be restored from the new section.
Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Peng Fan [Tue, 5 May 2020 12:28:40 +0000 (20:28 +0800)]
imx: imx8qm/imx8qxp: Power down the resources before SPL jump to u-boot
Make sure that all devices that are powered up by SPL are powered down
before entering into the u-boot. Otherwise the subsystem/device will
never be powered down by SCFW, due to SPL and u-boot are in different
partitions.
Benefiting from power domain driver, this patch implements the function
"imx8_power_off_pd_devices" to power off all active devices.
Ye Li [Tue, 5 May 2020 12:28:38 +0000 (20:28 +0800)]
imx: imx8qm/qxp: Fix issue in get_effective_memsize
When Trusty OS allocates the mem region from 0xfe0000000-0xffffffff,
the get_effective_memsize does not return correct memory size.
There is a check in get_effective_memsize to find the memreg where
the u-boot is running, and return the size of that memreg as the result
of get_effective_memsize. When using aligned start, the value is
0x80200000 since it is 2MB aligned. Thus the finding of memreg will
fail and return the PHYS_SDRAM_1_SIZE because u-boot text base is
0x80020000. This cause u-boot is relocated to the high memory where has
been occupied by Trusty OS.
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>
Peng Fan [Tue, 5 May 2020 12:28:37 +0000 (20:28 +0800)]
imx: imx8qm/qxp: reserving DDR memory for M4
The DDR memory from 0x88000000 to 0x8FFFFFFF is assigned to M4 on
QM and QXP. The M4 can allocate this memory by two ways,
in SCD or u-boot.
In this patch, u-boot addes the memory reserve node to DTB to pass
the info to kernel, no matter the M4 memory is reserved in SCD
or u-boot. So kernel won't access M4 reserved memory.
Han Xu [Tue, 5 May 2020 14:04:03 +0000 (22:04 +0800)]
cmd: nandbcb: Reconstruct the nandbcb tool for all platforms
The original nandbcb tool was designed for imx6 only, when trying to
leverage it to replace the kobs-ng tool, we found the design is not
friendly for supporting all platforms. To support all iMX6/7/8 platforms
and for easy further maintain, I reconstruct the structure of the tool.
The main changes including:
1. Use platform_data to determine the logic branches rather than simply
use SOC name.
2. More data structures as parameter for functions.
3. Global variables to define the FCB/DBBT/FW locations.
4. Implement the kobs-ng default 4 FCB/4 DBBT/2 FW layout.
5. Support Hamming coding/ 40bit BCH/ 62bit BCH coding FCB.
6. Dump and compare all written FCB/DBBT to verify data integrity.
The tool has been verified on iMX6Q/DL, 6SX, 7D, 6ULL, iMX8QX, iMX8MM.
Signed-off-by: Han Xu <han.xu@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Alice Guo [Tue, 5 May 2020 14:04:00 +0000 (22:04 +0800)]
nandbcb: support i.MX8M
Tested on i.MX8MM EVK, imx8mm evk uses BCH
encoding and randomizer
modify macro and print size_t with %zx
use CONFIG_IMX8M because it should apply to imx8mq/mm/mn
Signed-off-by: Alice Guo <alice.guo@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Leonard Crestez [Mon, 4 May 2020 13:16:54 +0000 (21:16 +0800)]
imx8: Replace SC_R_LAST with SC_R_NONE in DTB
We are currently using SC_R_LAST as a marker for imx8 power domain tree
nodes without a resource attached. This value is compiled into dtb as
part of the linux build and used by uboot.
The SC_R_LAST constant changes frequently as SCFW resources are added
(by design) and every time we need to update linux and uboot headers
together or boot can fail.
Fix this by replacing SC_R_LAST usage with a new constant SC_R_NONE
defined to be 0xFFF0.
Reviewed-by: Peng Fan <peng.fan@nxp.com> Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Ye Li [Mon, 4 May 2020 13:16:53 +0000 (21:16 +0800)]
power: imx8: Check owned resource in power on
When fspi is assigned to M4, we have to let the fspi probe failed when
its power domain is failed to power up. Because not all devices have power
domain (for example, external devices on the board). Current checking
resource owner in power domain probe is not good, change to check it in
power on.
Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Peng Fan [Mon, 4 May 2020 13:16:52 +0000 (21:16 +0800)]
power: imx8: remove the resource owned check before power off
For all the devices used and set ACTIVE in U-Boot, U-Boot needs
to power off all of them without the check of resource owner.
When we create software partition before booting Linux, the resource
own checkw will return false, and cause the power domain not powered
off. If without the check of resource owner, the power domain in
the other software partition could be powered off with parent
partition could access child partition resources.
Ye Li [Mon, 4 May 2020 13:16:51 +0000 (21:16 +0800)]
power: imx8-power-domain: Set DM_FLAG_DEFAULT_PD_CTRL_OFF flag
If without this flag, calling dev_power_domain_ctrl will iteratively remove
the power domain device will causes iteratively power off parent PD. This is
not expected by imx8-power-domain-legacy driver. Power off parent PD is
controlled by the driver internally.
So set DM_FLAG_DEFAULT_PD_CTRL_OFF to avoid such issue
Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Peng Fan [Mon, 4 May 2020 13:16:50 +0000 (21:16 +0800)]
imx8: power: Add PD device lookup interface to power domain uclass
Add power_domain_lookup_name interface to power domain uclass to find
a power domain device by its DTB node name, not using its associated
client device.
Through this interface, we can operate the power domain devices directly.
This is needed for non-DM drivers.
Modified from Ye's NXP downstream patch
only for legacy imx8 power domain driver, since we have not migrated
to use new power domain driver.
Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Alice Guo [Mon, 4 May 2020 14:09:03 +0000 (22:09 +0800)]
nand: enable the Randomizer module for i.mx7 and i.mx8
To enable the Randomizer module, set GPMI_ECCCTRL[RANDOMIZER_ENABLE] to
1, then set GPMI_ECCCOUNT[RANDOMIZER_PAGE] to select randomizer page
number needed to be randomized.
Signed-off-by: Alice Guo <alice.guo@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Han Xu [Mon, 4 May 2020 14:09:02 +0000 (22:09 +0800)]
mxs_nand: don't check zero count when ECC reading with randomizer
When enabled randomizer during ECC reading, the controller reported it's
erased page. Checking zero count will cause data get modified to all
0xFF. Stop checking during randomizer to workaround this issue.
Signed-off-by: Han Xu <han.xu@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Peng Fan [Mon, 4 May 2020 14:08:53 +0000 (22:08 +0800)]
nand: mxs: correct bitflip for erased NAND page
This patch is a porting of
http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/
commit/?h=imx_4.1.15_1.0.0_ga&id=e4dacc44d22e9474ec456cb330df525cd805ea38
"
i.MX6QP and i.MX7D BCH module integrated a new feature to detect the
bitflip number for erased NAND page. So for these two platform, set the
erase threshold to gf/2 and if bitflip detected, GPMI driver will
correct the data to all 0xFF.
Also updated the imx6qp dts file to ditinguish the GPMI module for i.MX6Q
with the one for i.MX6QP.
"
In this patch, i.MX6UL is added and threshold changed to use ecc_strength.
Peng Fan [Mon, 4 May 2020 14:08:52 +0000 (22:08 +0800)]
nand: mxs: fix the bitflips for erased page when uncorrectable error
This patch is porting from linux:
http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/commit/
?h=imx_4.1.15_1.0.0_ga&id=3d42fcece496224fde59f9343763fb2dfc5b0768
"
We may meet the bitflips in reading an erased page(contains all 0xFF),
this may causes the UBIFS corrupt, please see the log from Elie:
-----------------------------------------------------------------
[ 3.831323] UBI warning: ubi_io_read: error -74 (ECC error) while reading 16384 bytes from PEB 443:245760, read only 16384 bytes, retry
[ 3.845026] UBI warning: ubi_io_read: error -74 (ECC error) while reading 16384 bytes from PEB 443:245760, read only 16384 bytes, retry
[ 3.858710] UBI warning: ubi_io_read: error -74 (ECC error) while reading 16384 bytes from PEB 443:245760, read only 16384 bytes, retry
[ 3.872408] UBI error: ubi_io_read: error -74 (ECC error) while reading 16384 bytes from PEB 443:245760, read 16384 bytes
...
[ 4.011529] UBIFS error (pid 36): ubifs_recover_leb: corrupt empty space LEB 27:237568, corruption starts at 9815
[ 4.021897] UBIFS error (pid 36): ubifs_scanned_corruption: corruption at LEB 27:247383
[ 4.030000] UBIFS error (pid 36): ubifs_scanned_corruption: first 6569 bytes from LEB 27:247383
-----------------------------------------------------------------
This patch does a check for the uncorrectable failure in the following steps:
[0] set the threshold.
The threshold is set based on the truth:
"A single 0 bit will lead to gf_len(13 or 14) bits 0 after the BCH
do the ECC."
For the sake of safe, we will set the threshold with half the gf_len, and
do not make it bigger the ECC strength.
[1] count the bitflips of the current ECC chunk, assume it is N.
[2] if the (N <= threshold) is true, we continue to read out the page with
ECC disabled. and we count the bitflips again, assume it is N2.
(We read out the whole page, not just a chunk, this makes the check
more strictly, and make the code more simple.)
[3] if the (N2 <= threshold) is true again, we can regard this is a erased
page. This is because a real erased page is full of 0xFF(maybe also has
several bitflips), while a page contains the 0xFF data will definitely
has many bitflips in the ECC parity areas.
[4] if the [3] fails, we can regard this is a page filled with the '0xFF'
data.
"
Ye Li [Mon, 4 May 2020 14:08:51 +0000 (22:08 +0800)]
mtd: gpmi: provide the option to use legacy bch geometry
Provide an option in DT to use legacy bch geometry, which compatible
with the 3.10 kernel bch setting. To enable the feature, adding
"fsl,legacy-bch-geometry" under gpmi-nand node.
NOTICE: The feature must be enabled/disabled in both u-boot and kernel.
Signed-off-by: Han Xu <han.xu@nxp.com> Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
Ye Li [Mon, 4 May 2020 14:08:50 +0000 (22:08 +0800)]
mtd: gpmi: change the BCH layout setting for large oob NAND
The code change updated the NAND driver BCH ECC layout algorithm to
support large oob size NAND chips(oob > 1024 bytes) and proposed a new
way to set ECC layout.
Current implementation requires each chunk size larger than oob size so
the bad block marker (BBM) can be guaranteed located in data chunk. The
ECC layout always using the unbalanced layout(Ecc for both meta and
Data0 chunk), but for the NAND chips with oob larger than 1k, the driver
cannot support because BCH doesn’t support GF 15 for 2K chunk.
The change keeps the data chunk no larger than 1k and adjust the ECC
strength or ECC layout to locate the BBM in data chunk. General idea for
large oob NAND chips is
1.Try all ECC strength from the minimum value required by NAND spec to
the maximum one that works, any ECC makes the BBM locate in data chunk
can be chosen.
2.If none of them works, using separate ECC for meta, which will add one
extra ecc with the same ECC strength as other data chunks. This extra
ECC can guarantee BBM located in data chunk, of course, we need to check
if oob can afford it.
Previous code has two methods for ECC layout setting, the
legacy_calc_ecc_layout and calc_ecc_layout_by_info, the difference
between these two methods is, legacy_calc_ecc_layout set the chunk size
larger chan oob size and then set the maximum ECC strength that oob can
afford. While the calc_ecc_layout_by_info set chunk size and ECC
strength according to NAND spec. It has been proved that the first
method cannot provide safe ECC strength for some modern NAND chips, so
in current code,
1. Driver read NAND parameters first and then chose the proper ECC
layout setting method.
2. If the oob is large or NAND required data chunk larger than oob size,
chose calc_ecc_for_large_oob, otherwise use calc_ecc_layout_by_info
3. legacy_calc_ecc_layout only used for some NAND chips does not contains
necessary information. So this is only a backup plan, it is NOT
recommended to use these NAND chips.
Signed-off-by: Han Xu <b45815@freescale.com> Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com>
projects / u-boot.git / log
