So far Linux supported only two levels of MTD devices so we didn't need
a very precise description for this sysfs file. With commit
97519dc5 ("mtd: partitions: add support for subpartitions") there
is support for a tree structure so we should have more precise
description. Using "parent" and "flash device" makes it more accurate.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

From Boris:
"""
This pull request contains the following core changes:

* addition of on-ecc support to Micron driver
* addition of helpers to help drivers choose most appropriate ECC
  settings
* deletion of dead-code (cached programming and ->errstat() hook)
* make sure drivers that do not support the SET/GET FEATURES command
  return ENOTSUPP use a dummy ->set/get_features implementation
  returning -ENOTSUPP (required for Micron on-die ECC)
* change the semantic of ecc->write_page() for drivers setting the
  NAND_ECC_CUSTOM_PAGE_ACCESS flag
* support exiting 'GET STATUS' command in default ->cmdfunc()
  implementations
* change the prototype of ->setup_data_interface()

A bunch of driver related changes:

* various cleanup, fixes and improvements of the MTK driver
* OMAP DT bindings fixes
* support for ->setup_data_interface() in the fsmc driver
* support for imx7 in the gpmi driver
* finalization of the denali driver rework (thanks to Masahiro for the
  work he's done on this driver)
* fix "bitflips in erased pages" handling in the ifc driver
* addition of PM ops and dynamic timing configuration to the atmel
  driver

And as usual we also have a few minor cleanup/fixes/improvements
patches across the subsystem.
"""

From Cyrille:
"""
This pull request contains the following notable changes:
- introduce support to the SPI 1-2-2 and 1-4-4 protocols.
- introduce support to the Double Data Rate (DDR) mode.
- introduce support to the Octo SPI protocols.
- add support to new memory parts for Spansion, Macronix and Winbond.
- add fixes for the Aspeed, STM32 and Cadence QSPI controler drivers.
- clean up the st_spi_fsm driver.
"""

The code checks that ->_point is not NULL, but we should actually check
->_unpoint value which is dereferenced a few lines after the check.
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

We only want to hold the lock on the success path, not this error path.

Fixes: 7ec4a37c ("mtd: nand: mediatek: add support for different MTK NAND FLASH Controller IP")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

The 35h instruction op code has two aliases/macro definitions:
- SPINOR_OP_RDCR from include/linux/mtd/spi-nor.h
- SPINOR_OP_RDSR2 from drivers/mtd/devices/serial_flash_cmds.h

Actually, some manufacturers name the associated internal register Status
Register 2 whereas other manufacturers name it Configuration Register
hence the two different macros for the very same instruction op code.

Since the spi-nor.h file is the reference file for all SPI NOR instruction
op codes, this patch removes the definition of the SPINOR_OP_RDSR2 macro.

Also the SPINOR_OP_RDSR2 macro will be associated to another instruction
op code in a further patch so we need to avoid a conflict defining this
macro twice. Indeed the JESD216 rev B specification, defining the SFDP
tables, also refers to the 3Eh and 3Fh instruction op codes to write/read
the Status Register 2 on some SPI NOR flash memories, the 35h op code
still being used to read the Configuration Register/Status Register 2 on
other memories.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@microchip.com>
Acked-by: Marek Vasut <marek.vasut@gmail.com>

The variable was already marked 'const' before the previous
patch, but the qualifier was in an unusual place, and now the
extra 'const' causes a harmless warning:

drivers/mtd/spi-nor/cadence-quadspi.c:1286:34: error: duplicate 'const' declaration specifier [-Werror=duplicate-decl-specifier]

This removes the other 'const' instead.

Fixes: f993c123 ("mtd: spi-nor: cqspi: make of_device_ids const")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

It's an 8 MiB flash with 4 KiB erase sectors.
Signed-off-by: Harry Chou <HarryYC.Chou@gmail.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

This chip supports stateless 4-byte opcodes, dual and quad read and
uniform 4K-byte erase.
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

Currently, we use the fixed ACC timing 0x10804211. This is not the best
setting for each case. Actually, MTK NAND controller can adapt ACC timings
dynamically according to nfi clock frequence.
Implement the ->setup_data_interface() hook to optimize driver performance.
Signed-off-by: Xiaolei Li <xiaolei.li@mediatek.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

There is no need to add mtk_ecc_hw_init during ecc resume, because there
always takes mtk_ecc_wait_idle in the function mtk_ecc_enable.
Signed-off-by: Xiaolei Li <xiaolei.li@mediatek.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

chip->select_chip will do nfc runtime configuration. There is no need to
do mtk_nfc_hw_init before it.
Signed-off-by: Xiaolei Li <xiaolei.li@mediatek.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Currently, ecc encode irq is enabled when writing page with hwecc, but
we actually do not wait for this irq done. Because NFI and ECC work in
parallel, nfi irq and ecc irq almost come together.

Now, there are two steps to check whether page data are totally written.
First, wait for nfi irq INTR_AHB_DONE. This is to ensure all data
in RAM are received by NFI.
Second, polling the register NFI_ADDRCNTR till all data include ecc
parity data runtime generated by ECC are sent to NAND device.

So, it is redunant to enable ecc irq without waiting for it.
Signed-off-by: Xiaolei Li <xiaolei.li@mediatek.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Currently, we trigger ECC HW before setting ecc irq. It is incorrect.
Because ECC starts working once the register ECC_CTL_REG is set as
ECC_OP_ENABLE. And this may lead an abnormal behavior of ecc irq.
So, should enable ecc irq at first, then trigger ECC.

Fixes: 1d6b1e46 ("mtd: mediatek: driver for MTK Smart Device")
Signed-off-by: Xiaolei Li <xiaolei.li@mediatek.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Some recent patches caused churn around this area, and checkpatch
noticed the existing issues.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

This fixes following warning:
include/linux/kern_levels.h:4:18: warning: format '%X' expects argument of type 'unsigned int', but argument 2 has type 'size_t {aka long unsigned int}' [-Wformat=]
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

SPI NOR branches are now hosted on MTD repos, spi-nor/next is on l2-mtd
and spi-nor/fixes is on linux-mtd.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>
Acked-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

This makes TRX parsing code reusable with other platforms and parsers.

Please note this patch doesn't really change anything in the existing
code, just moves it. There is still some place for improvement (e.g.
working on non-hacky method of checking rootfs format) but it's not
really a subject of this change.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Some devices have partitions that are kind of containers with extra
subpartitions / volumes instead of e.g. a simple filesystem data. To
support such cases we need to first create normal flash device
partitions and then take care of these special ones.

It's very common case for home routers. Depending on the vendor there
are formats like TRX, Seama, TP-Link, WRGG & more. All of them are used
to embed few partitions into a single one / single firmware file.

Ideally all vendors would use some well documented / standardized format
like UBI (and some probably start doing so), but there are still
countless devices on the market using these poor vendor specific
formats.

This patch extends MTD subsystem by allowing to specify list of parsers
that should be tried for a given partition. Supporting such poor formats
is highly unlikely to be the top priority so these changes try to
minimize maintenance cost to the minimum. It reuses existing code for
these new parsers and just adds a one property and one new function.

This implementation requires setting partition parsers in a flash
parser. A proper change of bcm47xxpart will follow and in the future we
will hopefully also find a solution for doing it with ofpart
("fixed-partitions").
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

Some flash device partitions can be containers with extra subpartitions
(volumes). All callbacks are already capable of this additional level of
indirection.

This patch makes sure we always display subpartitions using a tree
structure and takes care of deleting subpartitions when parent gets
removed.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

This prepares mtd subsystem for the new feature: subpartitions. In some
cases flash device partition can be a container with extra subpartitions
(volumes).

So far there was a flat structure implemented. One master (flash device)
could be partitioned into few partitions. Every partition got its master
and it was enough to get things running.

To support subpartitions we need to store pointer to the parent for each
partition. This is required to implement more natural tree structure and
handle all recursion and offsets calculation.

To make code consistent this patch renamed "master" to the "parent" in
places where we can be dealing with subpartitions.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

When support for sysfs "offset" file was added it missed to update the
del_mtd_partitions function. It deletes partitions just like
mtd_del_partition does so both should also take care of removing sysfs
files.

This change moves sysfs_remove_files call to the shared function to fix
this issue.

Fixes: a62c24d7 ("mtd: part: Add sysfs variable for offset of partition")
Cc: Dan Ehrenberg <dehrenberg@chromium.org>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

There are two similar functions handling deletion. One handles single
partition and another the whole MTD flash device. They share (duplicate)
some code so it makes sense to add a small helper for that part.

Function del_mtd_partitions has been moved a bit to keep all deleting
stuff together.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>

The Denali NAND controller driver (drivers/mtd/nand/denali*) has been
largely reworked by me.  Add myself as its maintainer now.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

The binding bus/ti-gpmc.txt has been moved to
memory-controllers/omap-gpmc.txt.  Update all references to this in
order to make reading and understanding a given binding easier.

Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Brian Norris <computersforpeace@gmail.com>
Cc:Boris Brezillon <boris.brezillon@free-electrons.com>
Cc: Marek Vasut <marek.vasut@gmail.com>
Cc: Richard Weinberger <richard@nod.at>
Cc: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

The binding says that the compatible string must be "ti,am33xx-elm"
but the code checks only for, and all functioning users set, this as
"ti,am3352-elm" so correct the binding.

Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Brian Norris <computersforpeace@gmail.com>
Cc: Boris Brezillon <boris.brezillon@free-electrons.com>
Cc: Marek Vasut <marek.vasut@gmail.com>
Cc: Richard Weinberger <richard@nod.at>
Cc: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

The segment registers of the SMC controller provide a way to configure
the mapping windows of the chips on the AHB bus. The settings are
required to be correct when the controller operates in Command mode,
which is the case for DMAs and the LPC mapping.

This tries to set the segment registers of each chip depending on the
size of the flash device and depending on the previous segment
settings, in order to have a contiguous window across multiple chips.

Unfortunately, the AST2500 SPI controller has a bug and it is not
possible to configure a full 128MB window for a chip of the same
size. The window size needs to be restricted to 120MB. This issue only
applies to CE0.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

There is no need to keep the dummy bytes in the control register if
the command mode is not kept also. This could lead to an inconsistent
setting : normal read mode (command 0x3) and dummy bytes. It is to be
noted that the HW allows such a configuration.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

These devices are used on OpenPOWER systems. The SPI_NOR_DUAL_READ
flags is added for the Aspeed SoCs which do not support QUAD reads.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

These modules are used on the OpenPOWER Witherspoon systems to hold
the POWER9 host firmware image. The SPI_NOR_DUAL_READ flags is added
for the Aspeed SoCs which do not support QUAD reads.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

of_device_ids are not supposed to change at runtime. All functions
working with of_device_ids provided by <linux/of.h> work with const
of_device_ids. So mark the non-const structs as const.
Signed-off-by: Arvind Yadav <arvind.yadav.cs@gmail.com>
Reviewed-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

Similar to the other ones, different size. The "JV" suffix is in
the datasheet, I haven't seen mentions of a different one.

The datasheet indicates DUAL and QUAD are supported.

 http://www.winbond.com/resource-files/w25m512jv%20revc%2001062017.pdfSigned-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>

This code is going to be reused for parsers matched using OF so let's
factor it out to make this easier.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: Brian Norris <computersforpeace@gmail.com>

Currently the only documented partitioning is "fixed-partitions" but
there are more methods in use that we may want to support in the future.
Mention them and make it clear Fixed Partitions are just a single case.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: Rob Herring <robh@kernel.org>

All IFC version >= 1.0 use 28nm technology for SRAM. Here SRAM has
a requirement to initialize before any read operation performed for
avoiding ECC Error.

So update condition check to initialize SRAM for all IFC version >= 1.0.0
Signed-off-by: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Introduce some macros and helpers to avoid magic numbers and
rename macros/functions for clarification.

- We see '| 2' in several places.  This means Data Cycle in MAP11 mode.
  The Denali User's Guide says bit[1:0] of MAP11 is like follows:

  b'00 = Command Cycle
  b'01 = Address Cycle
  b'10 = Data Cycle

  So, this commit added DENALI_MAP11_{CMD,ADDR,DATA} macros.

- We see 'denali->flash_mem + 0x10' in several places, but 0x10 is a
  magic number.  Actually, this accesses the data port of the Host
  Data/Command Interface.  So, this commit added DENALI_HOST_DATA.
  On the other hand, 'denali->flash_mem' gets access to the address
  port, so DENALI_HOST_ADDR was also added.

- We see 'index_addr(denali, cmd, 0x1)' in denali_erase(), but 0x1
  is a magic number.  0x1 means the erase operation.  Replace 0x1
  with DENALI_ERASE.

- Rename index_addr() to denali_host_write() for clarification

- Denali User's Guide says MAP{00,01,10,11} for access mode.  Match
  the macros with terminology in the IP document.

- Rename struct members as follows:
  flash_bank   -> active_bank    (currently selected bank)
  flash_reg    -> reg            (base address of registers)
  flash_mem    -> host           (base address of host interface)
  devnum       -> devs_per_cs    (devices connected in parallel)
  bbtskipbytes -> oob_skip_bytes (number of bytes to skip in OOB)
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Now this driver is ready to remove NAND_SKIP_BBTSCAN.

The BBT descriptors in denali.c are equivalent to the ones in
nand_bbt.c.  There is no need to duplicate the equivalent structures.
The with-oob decriptors do not work for this driver anyway.

The bbt_pattern (offs = 8) and the version (veroffs = 12) area
overlaps the ECC area.  Set NAND_BBT_NO_OOB flag to use the no_oob
variant of the BBT descriptors.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

As Russell and Lars stated in the discussion [1], using
devm_k*alloc() with DMA is not a good idea.

Let's use kmalloc (not kzalloc because no need for zero-out).
Also, allocate the buffer as late as possible because it must be
freed for any error that follows.

[1] https://lkml.org/lkml/2017/3/8/693Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Russell King <rmk+kernel@armlinux.org.uk>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

For ecc->read_page() and ecc->write_page(), it is possible to call
dma_map_single() against the given buffer.  This bypasses the driver
internal bounce buffer and save the memcpy().
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Recent versions of this IP support automatic erased page detection.
If an erased page is detected on reads, the controller does not set
INTR__ECC_UNCOR_ERR, but INTR__ERASED_PAGE.

The detection of erased pages is based on the number of zeros in a
page; if the number of zeros is less than the value in the field
ERASED_THRESHOLD, the page is assumed as erased.

Please note ERASED_THRESHOLD specifies the number of zeros in a _page_
instead of an ECC chunk.  Moreover, the controller does not provide a
way to know the actual number of bitflips.

Actually, an erased page (all 0xff) is not an ECC correctable pattern
on the Denali ECC engine.  In other words, there may be overlap between
the following two:

[1] a bit pattern reachable from a valid payload + ECC pattern within
    ecc.strength bitflips
[2] a bit pattern reachable from an erased state (all 0xff) within
    ecc.strength bitflips

So, this feature may intercept ECC correctable patterns, then replace
[1] with [2].

After all, this feature can work safely only when ECC_THRESHOLD == 1,
i.e. detect erased pages without any bitflips.  This should be the
case most of the time.  If there is a bitflip or more, the driver will
fallback to the software method by using nand_check_erased_ecc_chunk().

Strangely enough, the driver still has to fill the buffer with 0xff
in case of INTR__ERASED_PAGE because the ECC correction engine has
already manipulated the data in the buffer before it judges erased
pages.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>