Fixed delay 200us is not working in certain platforms. Change to
poll for reset completion status to have more reliable reset process.
Controller will set the rst_comp bit in intr_status register after
controller has completed its reset and initialization process.
Tested-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Radu Bacrau <radu.bacrau@intel.com>
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Always put the controller in reset, then take it out of reset.
This is to make sure controller always in reset state in both SPL and
proper Uboot.
This is preparation for the next patch to poll for reset completion
(rst_comp) bit after reset.
Tested-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Radu Bacrau <radu.bacrau@intel.com>
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The offset at which the image to be loaded from NAND is located is
retrieved from the itb header. The presence of bad blocks in the area
of the NAND where the itb image is located could invalidate the offset
which must therefore be adjusted taking into account the state of the
sectors concerned.
cc: Michael Trimarchi <michael@amarulasolutions.com>
Signed-off-by: Dario Binacchi <dariobin@libero.it>
Reviewed-by: Michael Trimarchi <michael@amarulasolutions.com>
[ Linux commit 9afbe7c0140f663586edb6e823b616bd7076c00a ]
If the write protect signal from this IP is connected to the NAND
device, this IP can handle the WP# pin via the WRITE_PROTECT
register.
The Denali NAND Flash Memory Controller User's Guide describes
this register like follows:
When the controller is in reset, the WP# pin is always asserted
to the device. Once the reset is removed, the WP# is de-asserted.
The software will then have to come and program this bit to
assert/de-assert the same.
1 - Write protect de-assert
0 - Write protect assert
The default value is 1, so the write protect is de-asserted after
the reset is removed. The driver can write to the device unless
someone has explicitly cleared register before booting the kernel.
The boot ROM of some UniPhier SoCs (LD4, Pro4, sLD8, Pro5) is the
case; the boot ROM clears the WRITE_PROTECT register when the system
is booting from the NAND device, so the NAND device becomes read-only.
Set it to 1 in the driver in order to allow the write access to the
device.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
This CONFIG option is only used in denali_spl.c
Move it close to SPL_NAND_DENALI, and make it depend on SPL_NAND_DENALI.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.
Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is bad practice to include common.h in other header files since it can
bring in any number of superfluous definitions. It implies that some C
files don't include it and thus may be missing CONFIG options that are set
up by that file. The C files should include these themselves.
Update some header files in arch/arm to drop this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add more BCH setting mode and remove the unnecessary platform constrain
Signed-off-by: Han Xu <han.xu@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
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>
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>
Since iMX8 has enabled clock uclass, we can parse the clocks from DTB
and enable them in GPMI driver.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
enable the GPMI NAND driver for i.MX8, i.MX8 use similar controller as
i.MX8M
- register definition for i.mx8
- DMA structure must be 32bit address
Signed-off-by: Peng Fan <peng.fan@nxp.com>
imx8mm needs to BCH encode and set NAND page number needed to be
randomized
modify conditional compilation
Use CONFIG_IMX8M, so it apply to imx8mq/mm/mn
Signed-off-by: Alice Guo <alice.guo@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The iMX6SX uses compatible string "fsl,imx6sx-gpmi-nand" for gpmi
node in DTS, so update the driver for the string
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Update the mini driver to add support for getting ecc info from ONFI and
support read image data from page unaligned NAND address.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Update the gpmi/apbh_dma/bch drivers and relevant registers for i.MX8M.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
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.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
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.
"
Signed-off-by: Peng Fan <peng.fan@nxp.com>
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>
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>
Linux Broadcom NAND driver only disabled clock if no childs are initialized.
This section of the code seems to have been accidentally dropped when it was
imported in U-Boot.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
The end_cmd field in the variables cmd_phase_addr and data_phase_addr
contains the value 0xFF when the end_cmd equals NAND_CMD_NONE. This
should be 0x00.
This is caused by comparing NAND_CMD_NONE (int) with end_cmd (u8).
end_cmd will be promoted by the int value -1 and therefore is not equal
to 0xFF. Solved by casting NAND_CMD_NONE to u8 which will avoid int
promotion.
Signed-off-by: Patrick van Gelder <patrick.vangelder@nl.bosch.com>
Reviewed-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Commit 6293b0361d ("mtd: nand: pxa3xx: add raw read support") added the
local data_len variable in handle_data_pio() to track read size, but
forgot to update the condition of drain_fifo() call. That happens to
work when the layout last_chunk_size != 0. But when last_chunk_size ==
0, drain_fifo() is not called to read the last chunk, which leads to
"Wait timeout!!!" error. Fix this.
Fixes: 6293b0361d ("mtd: nand: pxa3xx: add raw read support")
Cc: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
There is no reason to continue when DT status property indicates that NAND
flash is disabled. But that means that NOR flash should be present that's
why try it find it out.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
xnand structure is private data structure and it is handled by core and
probe shouldn't touch it.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Fixed wrong enumeration of nand_config structure. Added chip select
function before reading the nand chip for maf/dev id's, without this
unable to access id's from some of the micron chips. Also added a
print statement to identify if a nand flash is using on-die ecc.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
The implementation of dma_map_single() and dma_unmap_single() is
exactly the same for all the architectures that support them.
Factor them out to <linux/dma-mapping.h>, and make all drivers to
include <linux/dma-mapping.h> instead of <asm/dma-mapping.h>.
If we need to differentiate them for some architectures, we can
move the generic definitions to <asm-generic/dma-mapping.h>.
Add some comments to the helpers. The concept is quite similar to
the DMA-API of Linux kernel. Drivers are agnostic about what is
going on behind the scene. Just call dma_map_single() before the
DMA, and dma_unmap_single() after it.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
At present dm/device.h includes the linux-compatible features. This
requires including linux/compat.h which in turn includes a lot of headers.
One of these is malloc.h which we thus end up including in every file in
U-Boot. Apart from the inefficiency of this, it is problematic for sandbox
which needs to use the system malloc() in some files.
Move the compatibility features into a separate header file.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present devres.h is included in all files that include dm.h but few
make use of it. Also this pulls in linux/compat which adds several more
headers. Drop the automatic inclusion and require files to include devres
themselves. This provides a good indication of which files use devres.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Anatolij Gustschin <agust@denx.de>
This function name conflicts with our desire to #define free() to
something else on sandbox. Since it deals with resources, rename it to
rfree().
Signed-off-by: Simon Glass <sjg@chromium.org>
UCLASS_MTD is a better fit for NAND drivers.
Make NAND_DENALI_DT depend on DM_MTD, which is needed to compile
drivers/mtd/mtd-uclass.c
Also, make ARCH_UNIPHIER select DM_MTD because all the defconfig
of this platform enables NAND_DENALI_DT.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Currently, the denali NAND driver in U-Boot configures the
SPARE_AREA_SKIP_BYTES based on the CONFIG option.
Recently, Linux kernel merged a patch that associates the proper
value for this register with the DT compatible string.
Do likewise in U-Boot too.
The denali_spl.c still uses CONFIG_NAND_DENALI_SPARE_AREA_SKIP_BYTES.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
When the reset signal is de-asserted, the HW-controlled bootstrap
starts running unless it is disabled in the SoC integration.
It issues some commands to detect a NAND chip, and sets up registers
automatically. Until this process finishes, software should avoid
any register access.
Without this delay function, some of UniPhier boards hangs up while
executing nand_scan_ident(). (denali_read_byte() is blocked)
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The Denali NAND driver in mainline Linux currently cannot deassert the
reset. The upcoming Linux 5.6 will support the reset controlling, and
also set up SPARE_AREA_SKIP_BYTES correctly. So, the Denali driver in
the future kernel will work without relying on any bootloader or firmware.
However, we still need to take care of stable kernel versions for a while.
U-boot should not assert the reset of this controller.
Fixes: ed784ac382 ("mtd: rawnand: denali: add reset handling")
Signed-off-by: Marek Vasut <marex@denx.de>
[yamada.masahiro: reword the commit description]
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The "nand_x" and "ecc" clocks are currently optional. Make the core
clock optional in the same way. This will allow platforms with no clock
driver support to use this driver.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Tested-by: Marek Vasut <marex@denx.de> # On SoCFPGA Arria V
On Altera SoCFPGA, upon either cold-boot or power-on reset, the
Denali NAND IP is initialized by the BootROM ; upon warm-reset,
the Denali NAND IP is NOT initialized by BootROM. In fact, upon
warm-reset, the SoCFPGA BootROM checks whether the SPL image in
on-chip RAM is valid and if so, completely skips re-loading the
SPL from the boot media.
This does sometimes lead to problems where the software left
the boot media in inconsistent state before warm-reset, and
because the BootROM does not reset the boot media, the boot
media is left in this inconsistent state, often until another
component attempts to access the boot media and fails with an
difficult to debug failure. To mitigate this problem, the SPL
on Altera SoCFPGA always resets all the IPs on the SoC early
on boot.
This results in a couple of register values, pre-programmed by
the BootROM, to be lost during this reset. To restore correct
operation of the IP on SoCFPGA, these values must be programmed
back into the controller by the driver. Note that on other SoCs
which do not use the HW-controlled bootstrap, more registers
may have to be programmed.
This also aligns the SPL behavior with the full Denali NAND
driver, which sets these values in denali_hw_init().
Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Drop local dma_map_single() and dma_unmap_single() and use arch specific
common implementation
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Acked-by: Masahiro Yamada <yamada.masahiro@socionext.com>
At present panic() is in the vsprintf.h header file. That does not seem
like an obvious choice for hang(), even though it relates to panic(). So
let's put hang() in its own header.
Signed-off-by: Simon Glass <sjg@chromium.org>
[trini: Migrate a few more files]
Signed-off-by: Tom Rini <trini@konsulko.com>
Remove hardcoded base addresses of smc controller and nand controller.
Get those addresses from dt and replace wherever they are used.
Remove smc and nand base address from header file too.
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
