When using the Quad Read opcode, SPI masters still use Single SPI
transfers, as spi_transfer.rx_nbits defaults to SPI_NBITS_SINGLE.
Use SPI_NBITS_QUAD to fix this.

While an earlier version of commit 3487a639


("drivers: mtd: m25p80: add quad read support") did this correctly, it was
forgotten in the version that got merged.

Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Spansion s25fl512s supports Quad SPI transfers, hence set the
M25P80_QUAD_READ flag.

Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

plat/regs-nand.h is used only by S3C2410 nand driver. Since there
are no other users, merge this file into the driver code to remove
platform dependency. While at it also remove unused macros.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

mtdram_init_device() wasn't updated along with mtd_partition.name.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>

In the following commit (in -next):

    commit 8552b439


    drivers: mtd: m25p80: convert "bool" read check into an enum

We converted the boolean 'fast_read' property to become an enum
'flash_read', but at the same time, we changed the conditional path so
that it doesn't choose a default value in some cases (technically, we
choose the correct default simply by virtue of devm_kzalloc(), which
zeroes this out to be a NORMAL read operation, but still...).

Fix this by setting a default for the 'else' clause.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Cc: Sourav Poddar <sourav.poddar@ti.com>
Acked-by: Marek Vasut <marex@denx.de>

Remove unneeded error handling on the result of a call
to platform_get_resource() when the value is passed to
devm_ioremap_resource().

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Reviewed-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Remove unneeded error handling on the result of a call
to platform_get_resource() when the value is passed to
devm_ioremap_resource(). And move those two call together
to make the connection between them more clear.

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Reviewed-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Add the Intel manufacturer Id.
Tested with Intel JS29F32G08ACMD1(4096 + 224) which is ONFI 2.0 compliant
nand.

Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Add the manufactor ID for SanDisk.
Make preparation for SanDisk SDTNRGAMA-008G.

Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Assume that:
          tmp = ((extid >> 2) & 0x04) | (extid & 0x03));

From the K9LCG08U0B's datasheet, we know that:
  the oob size is 640 when tmp is 6;
  the oob size is 1024 when tmp is 7;

Signed-off-by: Huang Shijie <b32955@freescale.com>
[Brian: fixed compile issue]
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

This commit adds support for devices with 2048B page sizes and
4-bit ECC strength requirements. This is achieved by enabling the BCH
ECC engine, which provides a higher strength: 16-bit over 2048 bytes.

Additionally, add a proper ECC layout to model the controller's view
of the device (where 'U' means unused and 'B' is the bad block marker):

 ----------------------------------------------------
 | 2048B data | B | B | 30B spare | 30B ECC | U | U |
 ----------------------------------------------------

Signed-off-by: Rodolfo Giometti <giometti@linux.it>
[Brian: updated with Ezequiel's patch description]
Acked-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

commit 3487a639

 ("drivers: mtd: m25p80: add
quad read support") in -next added both the 3-byte OPCODE_QUAD_READ and the
4-byte OPCODE_QUAD_READ_4B, but incorrectly uses OPCODE_QUAD_READ for both
3-byte and 4-byte addressing.

Use OPCODE_QUAD_READ_4B in the 4-byte case to fix this.

Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Acked-by: Marek Vasut <marex@denx.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

According to the Open NAND Flash Interface Specification (ONFI) Revision
3.1 "Parameters are always transferred on the lower 8-bits of the data
bus." for the Get Features and Set Features commands.

So using read_buf and write_buf is wrong for 16-bit wide nand chips as
they use I/O[15:0]. The Get Features command is easily fixed using 4
times the read_byte callback. For Set Features implement a new
overwritable callback "write_byte". Still I expect the default to work
just fine for all controllers and making it overwriteable was just done
for symmetry.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
[Brian: fixed warning]
Tested-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

To be consistent with the rest of include/linux/mtd/nand.h, we should
use the __packed shorthand instead of __attribute__((packed)).

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>

Micron provides READ RETRY support via the ONFI vendor-specific
parameter block (to indicate how many read-retry modes are available)
and the ONFI {GET,SET}_FEATURES commands with a vendor-specific feature
address (to support reading/switching the current read-retry mode).

The recommended sequence is as follows:

  1. Perform PAGE_READ operation
  2. If no ECC error, we are done
  3. Run SET_FEATURES with feature address 89h, mode 1
  4. Retry PAGE_READ operation
  5. If ECC error and there are remaining supported modes, increment the
     mode and return to step 3. Otherwise, this is a true ECC error.
  6. Run SET_FEATURES with feature address 89h, mode 0, to return to the
     default state.

This patch implements the chip->setup_read_retry() callback for
Micron and fills in the chip->read_retries.

Tested on Micron MT29F32G08CBADA, which supports 8 read-retry modes.

The Micron vendor-specific table was checked against the datasheets for
the following Micron NAND:

Needs retry   Cell-type    Part number          Vendor revision    Byte 180
-----------   ---------    ----------------     ---------------    ------------
No            SLC          MT29F16G08ABABA      1                  Reserved (0)
No            MLC          MT29F32G08CBABA      1                  Reserved (0)
No            SLC          MT29F1G08AACWP       1                  0
Yes           MLC          MT29F32G08CBADA      1                  08h
Yes           MLC          MT29F64G08CBABA      2                  08h

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>

Modern MLC (and even SLC?) NAND can experience a large number of
bitflips (beyond the recommended correctability capacity) due to drifts
in the voltage threshold (Vt). These bitflips can cause ECC errors to
occur well within the expected lifetime of the flash. To account for
this, some manufacturers provide a mechanism for shifting the Vt
threshold after a corrupted read.

The generic pattern seems to be that a particular flash has N read retry
modes (where N = 0, traditionally), and after an ECC failure, the host
should reconfigure the flash to use the next available mode, then retry
the read operation. This process repeats until all bitfips can be
corrected or until the host has tried all available retry modes.

This patch adds the infrastructure support for a
vendor-specific/flash-specific callback, used for setting the read-retry
mode (i.e., voltage threshold).

For now, this patch always returns the flash to mode 0 (the default
mode) after a successful read-retry, according to the flowchart found in
Micron's datasheets. This may need to change in the future if it is
determined that eventually, mode 0 is insufficient for the majority of
the flash cells (and so for performance reasons, we should leave the
flash in mode 1, 2, etc.).

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>

ECC failures can be tracked at the page level, not the do_read_ops level
(i.e., a potentially multi-page transaction).

This helps prepare for READ RETRY support.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>

Now that the driver can support the Armada 370/XP SoC NAND controller,
add the devicetree compatible string, enabling its use.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Acked-by: Jason Cooper <jason@lakedaemon.net>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

This patch kills the NAND_MAX_PAGESIZE/NAND_MAX_OOBSIZE by the following
way:
 1.) change the @buf field of nand_buf{} from an array to a pointer.
     also remove the DENALI_BUF_SIZE macro.

 2.) Before we call the nand_scan_ident, we allocate a temporary buffer
     whose size is PAGE_SIZE.

 3.) After we finish the nand_scan_ident, we have already getten the
     page size and oob size. We will allocate the right buffer size
     again.

Signed-off-by: Huang Shijie <shijie8@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

We kill the NAND_MAX_PAGESIZE/NAND_MAX_OOBSIZE by the following way:
 1.) Before we call the nand_scan_ident, we allocate a temporary buffer
     whose size is PAGE_SIZE.
 2.) After we finish the nand_scan_ident, we have already getten the
     page size and oob size. We will allocate the right buffer size
     again.

Signed-off-by: Huang Shijie <shijie8@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

mtd.name is assigned to IFC NAND physical address. Assignment type is u32.
It is not providing correct physical address of IFC NAND.

Update assignment type to u64.

Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

This patch assigned the type->name to mtd->name when mtd->name is
NULL in function "find_full_id_nand".
mtd->name is NULL may cause some problem.

Signed-off-by: Cai Zhiyong <caizhiyong@huawei.com>
Acked-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

commit 93115b7f

 ("mtd: onenand/samsung: make regs-onenand.h file local")
moved the file to the current location but forgot to remove the pointer to
its previous location. Clean it up.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Commit c02cecb9

 ("ARM: orion: move platform_data definitions")
moved the file to the current location but forgot to remove the pointer
to its previous location. Clean it up. While at it also change the header
file protection macros appropriately.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Commit 2203747c

 ("ARM: omap: move platform_data definitions")
moved the files to the current location but forgot to remove the pointer
to its previous location. Clean it up.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Pekon Gupta <pekon@ti.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

The kernel already has this information, and individual drivers
shouldn't duplicate that.  This also eliminates the use of __DATE__ and
__TIME__, which make the build non-deterministic.

Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Some devices (like WNDR3700v3) have board_data without MPFR magic, some
extra header or extra NVRAM around 0x100. In such case we have to look
for another magic which is BD 0B 0D BD (BD probably stands for Board
Data). It's located "far far away", so instead of extending buffer add
another mtd_read.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Some devices have even nicer-to-recognize CFE thanks to the magic.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Make this option a hidden one and get a cleaner configuration.
This option just selects a common infrastructure for MTD-based devices
to expose a block interface. There is no point in allowing a separate
enable/disable.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
[Brian: keep symbol as tristate]
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Fixes this sparse warning:

    CHECK   drivers/mtd/onenand/generic.c
  drivers/mtd/onenand/generic.c:61:62: warning: Using plain integer as NULL pointer

Signed-off-by: Brian Norris <computersforpeace@gmail.com>

The omap_{read,write}_buf{8,16}() functions are identical to the default
nand_base versions. Just let nand_base assign them in the
NAND_OMAP_POLLED case.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Tested-by: Pekon Gupta <pekon@ti.com>

We shouldn't try to allocate a resource until we're sure the
of_property_read_u64() call didn't fail. This is especially important if
we use this code for both CONFIG_OF and !CONFIG_OF builds, since
of_property_read_u64() will always return -ENOSYS for !CONFIG_OF.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>

The site-specific OOM messages are unnecessary, because they
duplicate the MM subsystem generic OOM message.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Use devm_*() functions to make cleanup paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Use devm_*() functions to make cleanup paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Use dev_err() instead of printk() to provide a better message
to userspace.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Use dev_err() instead of printk() to provide a better message
to userspace.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Use dev_warn() instead of printk() to provide a better message
to userspace.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Use dev_err() instead of printk() to provide a better message
to userspace.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Josh Wu <josh.wu@atmel.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>