Miquel Raynal [Sat, 4 Nov 2023 10:50:22 +0000 (11:50 +0100)]
Merge tag 'nand/for-6.7' into mtd/next
The raw NAND subsystem has, as usual, seen a bit of cleanup being done
this cycle, typically return values of platform_get_irq() and
devm_kasprintf(), plus structure annotations for sanitizers. There is
also a better ECC check in the Arasan driver. This comes with smaller
misc changes.
In the SPI-NAND world there is now support for Foresee F35SQA002G,
Winbond W25N and XTX XT26 chips.
Miquel Raynal [Sat, 4 Nov 2023 10:49:52 +0000 (11:49 +0100)]
Merge tag 'spi-nor/for-6.7' into mtd/next
For SPI NOR we cleaned the flash info entries in order to have
them slimmer and self explanatory. In order to make the entries
as slim as possible, we introduced sane default values so that
the actual flash entries don't need to specify them. We now use
a flexible macro to specify the flash ID instead of the previous
INFOx() macros that had hardcoded ID lengths.
We now use:
+ .id = SNOR_ID(0xef, 0x80, 0x20),
+ .name = "w25q512nwm",
+ .otp = SNOR_OTP(256, 3, 0x1000, 0x1000),
We also removed some flash entries: the very old Catalyst
SPI EEPROMs that were introduced once with the SPI-NOR subsystem,
and a Fujitsu MRAM. Both should use the at25 EEPROM driver.
The latter even has device tree bindings for the at25 driver.
We made sure that the conversion didn't introduce any unwanted
changes by comparing the .rodata segment before and after the
conversion. The patches landed in linux-next immediately after
v6.6-rc2, we haven't seen any regressions yet.
Apart of the autumn cleaning we introduced a new flash entry,
at25ff321a, and added block protection support for mt25qu512a.
Linus Walleij [Fri, 20 Oct 2023 20:30:29 +0000 (22:30 +0200)]
mtd: cfi_cmdset_0001: Byte swap OTP info
Currently the offset into the device when looking for OTP
bits can go outside of the address of the MTD NOR devices,
and if that memory isn't readable, bad things happen
on the IXP4xx (added prints that illustrate the problem before
the crash):
cfi_intelext_otp_walk walk OTP on chip 0 start at reg_prot_offset 0x00000100
ixp4xx_copy_from copy from 0x00000100 to 0xc880dd78
cfi_intelext_otp_walk walk OTP on chip 0 start at reg_prot_offset 0x12000000
ixp4xx_copy_from copy from 0x12000000 to 0xc880dd78
8<--- cut here ---
Unable to handle kernel paging request at virtual address db000000
[db000000] *pgd=00000000
(...)
This happens in this case because the IXP4xx is big endian and
the 32- and 16-bit fields in the struct cfi_intelext_otpinfo are not
properly byteswapped. Compare to how the code in read_pri_intelext()
byteswaps the fields in struct cfi_pri_intelext.
Adding a small byte swapping loop for the OTP in read_pri_intelext()
and the crash goes away.
The problem went unnoticed for many years until I enabled
CONFIG_MTD_OTP on the IXP4xx as well, triggering the bug.
Yi Yang [Thu, 19 Oct 2023 06:55:48 +0000 (06:55 +0000)]
mtd: rawnand: meson: check return value of devm_kasprintf()
devm_kasprintf() returns a pointer to dynamically allocated memory
which can be NULL upon failure. Ensure the allocation was successful by
checking the pointer validity.
Yi Yang [Thu, 19 Oct 2023 06:55:37 +0000 (06:55 +0000)]
mtd: rawnand: intel: check return value of devm_kasprintf()
devm_kasprintf() returns a pointer to dynamically allocated memory
which can be NULL upon failure. Ensure the allocation was successful by
checking the pointer validity.
Uwe Kleine-König [Mon, 16 Oct 2023 10:35:41 +0000 (12:35 +0200)]
mtd: rawnand: sh_flctl: Convert to module_platform_driver()
The driver doesn't benefit from the advantages that
module_platform_driver_probe() allows (i.e. putting the probe function
in .init.text and the .remove function into .exit.text).
So use module_platform_driver() instead which allows to bind the driver
also after booting (or module loading) and unbinding via sysfs.
Mamta Shukla [Tue, 17 Oct 2023 07:47:11 +0000 (10:47 +0300)]
mtd: spi-nor: micron-st: use SFDP table for mt25qu512a
Parse SFDP table to get size and functions of mt25qu512a. BFPT wrongly
advertises 16bit SR support and made the locking fail. Add a post BFPT
fixup hook to clear the 16bit SR support.
Mamta Shukla [Tue, 17 Oct 2023 07:47:10 +0000 (10:47 +0300)]
mtd: spi-nor: micron-st: enable lock/unlock for mt25qu512a
mt25qu512a supports locking/unlocking through the SR BP bits. Enable
locking support. Tested with mtd-utils- flash_lock/flash_unlock on
MT25QU512ABB8E12.
Kees Cook [Fri, 6 Oct 2023 20:17:34 +0000 (13:17 -0700)]
mtd: rawnand: cadence: Annotate struct cdns_nand_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct cdns_nand_chip.
Cc: Miquel Raynal <miquel.raynal@bootlin.com> Cc: Richard Weinberger <richard@nod.at> Cc: Vignesh Raghavendra <vigneshr@ti.com> Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org> Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Cc: Nicolas Ferre <nicolas.ferre@microchip.com> Cc: Roger Quadros <rogerq@kernel.org> Cc: Thierry Reding <treding@nvidia.com> Cc: Yang Yingliang <yangyingliang@huawei.com> Cc: "Uwe Kleine-König" <u.kleine-koenig@pengutronix.de> Cc: Valentin Korenblit <vkorenblit@sequans.com> Cc: ye xingchen <ye.xingchen@zte.com.cn> Cc: linux-mtd@lists.infradead.org Cc: linux-hardening@vger.kernel.org Link: https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci Signed-off-by: Kees Cook <keescook@chromium.org> Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/linux-mtd/20231006201734.work.060-kees@kernel.org
Kees Cook [Fri, 6 Oct 2023 20:17:28 +0000 (13:17 -0700)]
mtd: rawnand: Annotate struct mtk_nfc_nand_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct
mtk_nfc_nand_chip.
Cc: Miquel Raynal <miquel.raynal@bootlin.com> Cc: Richard Weinberger <richard@nod.at> Cc: Vignesh Raghavendra <vigneshr@ti.com> Cc: Matthias Brugger <matthias.bgg@gmail.com> Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org> Cc: Heiko Stuebner <heiko@sntech.de> Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Cc: Thierry Reding <treding@nvidia.com> Cc: Roger Quadros <rogerq@kernel.org> Cc: "Uwe Kleine-König" <u.kleine-koenig@pengutronix.de> Cc: Cai Huoqing <cai.huoqing@linux.dev> Cc: Chuanhong Guo <gch981213@gmail.com> Cc: Rob Herring <robh@kernel.org> Cc: Li Zetao <lizetao1@huawei.com> Cc: linux-mtd@lists.infradead.org Cc: linux-arm-kernel@lists.infradead.org Cc: linux-mediatek@lists.infradead.org Cc: linux-hardening@vger.kernel.org Link: https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci Signed-off-by: Kees Cook <keescook@chromium.org> Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/linux-mtd/20231006201728.work.750-kees@kernel.org
Christophe JAILLET [Sun, 1 Oct 2023 07:44:04 +0000 (09:44 +0200)]
mtd: rawnand: rockchip: Use struct_size()
Use struct_size() instead of hand writing it.
This is less verbose and more robust.
While at it, prepare for the coming implementation by GCC and Clang of the
__counted_by attribute. Flexible array members annotated with __counted_by
can have their accesses bounds-checked at run-time checking via
CONFIG_UBSAN_BOUNDS (for array indexing) and CONFIG_FORTIFY_SOURCE (for
strcpy/memcpy-family functions).
Also remove a useless comment about the position of a flex-array in a
structure.
mtd: rawnand: arasan: Include ECC syndrome along with in-band data while checking for ECC failure
Following an ECC failure condition upon page reads, we shall distinguish
between a real ECC failure and an empty page. This is handled with a call
to nand_check_erased_ecc_chunk() which looks at the data and counts the
number of bits which are not 'ones'. If we get less zeros than the ECC
strength, we assume the page was erased and we are in the presence of
natural bitflips. Otherwise, if we are above, we assume some data was
written and the ECC engine could not recover it all, so we report an ECC
failure.
In order for this logic to be as close as the reality as we can (this is
already a simplified condition but we can hardly be more precise), we
should check all the data that is covered by the ECC step not only the
in-band data, so we should also include the ECC syndrome in the check.
Rob Herring [Mon, 9 Oct 2023 17:28:56 +0000 (12:28 -0500)]
mtd: Use device_get_match_data()
Use preferred device_get_match_data() instead of of_match_device() to
get the driver match data. With this, adjust the includes to explicitly
include the correct headers.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:43 +0000 (22:01 +0200)]
mtd: spi-nor: nxp-spifi: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:42 +0000 (22:01 +0200)]
mtd: spi-nor: hisi-sfc: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:40 +0000 (22:01 +0200)]
mtd: maps: sun_uflash: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:39 +0000 (22:01 +0200)]
mtd: maps: sa1100-flash: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:38 +0000 (22:01 +0200)]
mtd: maps: pxa2xx-flash: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:37 +0000 (22:01 +0200)]
mtd: maps: plat-ram: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:36 +0000 (22:01 +0200)]
mtd: maps: physmap-core: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:35 +0000 (22:01 +0200)]
mtd: maps: lantiq-flash: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:34 +0000 (22:01 +0200)]
mtd: lpddr2_nvm: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:33 +0000 (22:01 +0200)]
mtd: hyperbus: rpc-if: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:32 +0000 (22:01 +0200)]
mtd: hyperbus: hbmc-am654: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:31 +0000 (22:01 +0200)]
mtd: st_spi_fsm: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:30 +0000 (22:01 +0200)]
mtd: spear_smi: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:29 +0000 (22:01 +0200)]
mtd: powernv_flash: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:28 +0000 (22:01 +0200)]
mtd: phram: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:27 +0000 (22:01 +0200)]
mtd: docg3: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Uwe Kleine-König [Sun, 8 Oct 2023 20:01:26 +0000 (22:01 +0200)]
mtd: bcm47xxsflash: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
mtd: map_ram: prevent use of point and unpoint when NO_XIP is set
When the DT property no-unaligned-direct-access is set, map->phys is set
to NO_XIP. With this property set, the flash should not be exposed
directly to MTD users, since it cannot be mapped.
map_ram() exposes the flash direct access unconditionally which leads to
access errors (when the bus width does not match the RAM width).
Therefore do not set point and unpoint when NO_XIP is set.
mtd: rawnand: sunxi: Annotate struct sunxi_nand_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct sunxi_nand_chip.
mtd: rawnand: renesas: Annotate struct rnand_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct rnand_chip.
mtd: rawnand: meson: Annotate struct meson_nfc_nand_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct meson_nfc_nand_chip.
mtd: rawnand: marvell: Annotate struct marvell_nand_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct marvell_nand_chip.
mtd: rawnand: ingenic: Annotate struct ingenic_nfc with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct ingenic_nfc.
mtd: rawnand: denali: Annotate struct denali_chip with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct denali_chip.
mtd: rawnand: atmel: Annotate struct atmel_nand with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct atmel_nand.
mtd: cfi: Annotate struct cfi_private with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct cfi_private.
mtd: Annotate struct lpddr_private with __counted_by
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct lpddr_private.
Michael Walle [Fri, 8 Sep 2023 10:16:58 +0000 (12:16 +0200)]
mtd: spi-nor: atmel: drop duplicate entry
The Atmel AT26DF321 and AT25DF321 have the same ID. Both were just
discovered by reading their IDs, that is, there is no probing by name.
Thus only the first one (the AT25DF321) in the list was ever probed.
Luckily, the AT25DF is also the newer series. Drop the AT26DF321.
Michael Walle [Fri, 8 Sep 2023 10:16:57 +0000 (12:16 +0200)]
mtd: spi-nor: winbond: sort flash_info entries
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:56 +0000 (12:16 +0200)]
mtd: spi-nor: sst: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:55 +0000 (12:16 +0200)]
mtd: spi-nor: spansion: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:54 +0000 (12:16 +0200)]
mtd: spi-nor: micron-st: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:53 +0000 (12:16 +0200)]
mtd: spi-nor: macronix: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:52 +0000 (12:16 +0200)]
mtd: spi-nor: issi: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:50 +0000 (12:16 +0200)]
mtd: spi-nor: eon: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:49 +0000 (12:16 +0200)]
mtd: spi-nor: atmel: sort flash_info database
The flash ID is the new primary key into our database. Sort the entry by
it. Keep the most specific ones first, because there might be ID
collisions between shorter and longer ones.
Michael Walle [Fri, 8 Sep 2023 10:16:47 +0000 (12:16 +0200)]
mtd: spi-nor: xilinx: use new macros in S3AN_INFO()
There won't be any new entries, nor are the entries that much different
and the very odd page and sector sizes make the new format hard to read.
Therefore, convert the old S3AN_INFO() macro.
Michael Walle [Fri, 8 Sep 2023 10:16:33 +0000 (12:16 +0200)]
mtd: spi-nor: add SNOR_ID() and SNOR_OTP()
After all the preparation, it is now time to introduce the new macros to
specify flashes in our database: SNOR_ID() and SNOR_OTP(). An flash_info
entry might now look like:
{
.id = SNOR_ID(0xef, 0x60, 0x16),
.otp = SNOR_OTP(256, 3, 0x1000, 0x1000),
.flags = SPI_NOR_HAS_LOCK | SPI_NOR_HAS_TB,
}
Michael Walle [Fri, 8 Sep 2023 10:16:32 +0000 (12:16 +0200)]
mtd: spi-nor: rename .otp_org to .otp and make it a pointer
Move the OTP ops out of the flash_info structure. Besides of saving some
space, there will be a new macro SNOR_OTP() which can be used to set the
ops:
.otp = SNOR_OTP(...),
Michael Walle [Fri, 8 Sep 2023 10:16:31 +0000 (12:16 +0200)]
mtd: spi-nor: move the .id and .id_len into an own structure
Create a new structure to hold a flash ID and its length. The goal is to
have a new macro SNOR_ID() which can have a flexible id length. This way
we can get rid of all the individual INFOx() macros.
Michael Walle [Fri, 8 Sep 2023 10:16:30 +0000 (12:16 +0200)]
mtd: spi-nor: introduce (temporary) INFO0()
The id will be converted to an own structure. To differentiate between
flashes with and without IDs, introduce a temporary macro INFO0() and
convert all flashes with no ID to use it. The difference between INFO0()
and INFOx() is that the former, doesn't have a pointer to the id
structure. Something which isn't possible to do within the INFOx()
macro.
After the flash_info conversion, that macro will be removed along with
all the other INFOx() macros.
Michael Walle [Fri, 8 Sep 2023 10:16:29 +0000 (12:16 +0200)]
mtd: spi-nor: drop .parse_sfdp
Drop the size parameter to indicate we need to do SFDP, we can do that
because it is guaranteed that the size will be set by SFDP and because
PARSE_SFDP forced the SFDP parsing it must be overwritten.
There is a (very tiny) chance that this might break block protection
support: we now rely on the SFDP reported size of the flash for the
BP calculation. OTOH, if the flash reports its size wrong, we are
in bigger trouble than just having the BP calculation wrong.
Michael Walle [Fri, 8 Sep 2023 10:16:28 +0000 (12:16 +0200)]
mtd: spi-nor: make sector_size optional
Most of the (old, non-SFDP) flashes use a sector size of 64k. Make that
a default value so it can be optional in the flash_info database.
As a preparation for conversion to the new database format, set the
sector size to zero if the default value is used. This way, the actual
change is happening with this patch ant not with a later conversion
patch.
Michael Walle [Fri, 8 Sep 2023 10:16:27 +0000 (12:16 +0200)]
mtd: spi-nor: push 4k SE handling into spi_nor_select_uniform_erase()
4k sector erase sizes are only a thing with uniform erase types. Push
the "we want 4k erase sizes" handling into spi_nor_select_uniform_erase().
One might wonder why the former sector_size isn't used anymore. It is
because we either search for the largest erase size or if selected
through kconfig, the 4k erase size. Now, why is that correct? For this,
we have to differentiate between (1) flashes with SFDP and (2) without
SFDP. For (1), we just set one (or two if SECT_4K is set) erase types
and wanted_size is exactly one of these.
For (2) things are a bit more complicated. For flashes which we don't
have in our flash_info database, the generic driver is used and
sector_size was already 0, which in turn selected the largest erase
size. For flashes which had SFDP and an entry in flash_info, sector_size
was always the largest sector and thus the largest erase type.
Michael Walle [Fri, 8 Sep 2023 10:16:26 +0000 (12:16 +0200)]
mtd: spi-nor: default .n_banks to 1
If .n_banks is not set in the flash_info database, the default value
should be 1. This way, we don't have to always set the .n_banks
parameter in flash_info.
Michael Walle [Fri, 8 Sep 2023 10:16:25 +0000 (12:16 +0200)]
mtd: spi-nor: store .n_banks in struct spi_nor_flash_parameter
First, fixups might want to replace the n_banks parameter, thus we need
it in the (writable) parameter struct. Secondly, this way we can have a
default in the core and just skip setting the n_banks in the flash_info
database. Most of the flashes doesn't have more than one bank.
Michael Walle [Fri, 8 Sep 2023 10:16:24 +0000 (12:16 +0200)]
mtd: spi-nor: default page_size to 256 bytes
The INFO() macro always set the page_size to 256 bytes. Make that an
optional parameter. This default is a sane one for all older flashes,
newer ones will set the page size by its SFDP tables anyway.
Michael Walle [Fri, 8 Sep 2023 10:16:23 +0000 (12:16 +0200)]
mtd: spi-nor: convert .n_sectors to .size
.n_sectors is rarely used. In fact it is only used in swp.c and to
calculate the flash size in the core. The use in swp.c might be
converted to use the (largest) flash erase size. For now, we just
locally calculate the sector size.
Simplify the flash_info database and set the size of the flash directly.
This also let us use the SZ_x macros.
Verified that there's no flash that specifies BP and sector size of zero
to make sure we avoid a division by zero in
spi_nor_get_min_prot_length_sr(). We'll protect from a possible division
by zero in a further patch by introducing a default value for
sector_size.
Michael Walle [Fri, 8 Sep 2023 10:16:21 +0000 (12:16 +0200)]
mtd: spi-nor: xilinx: use SPI_NOR_ID() in S3AN_INFO()
In commit 59273180299a ("mtd: spi-nor: Create macros to define chip IDs
and geometries") SPI_NOR_ID() were introduced, but it did only update
the INFO() macro in core.h. Also use it in S3AN_INFO().
Michael Walle [Fri, 8 Sep 2023 10:16:20 +0000 (12:16 +0200)]
mtd: spi-nor: remove Fujitsu MB85RS1MT support
This part is not a flash but an EEPROM like FRAM. It is even has a DT
binding for the (correct) driver (at25), see
Documentation/devicetree/bindings/eeprom/at25.yaml. Just remove it.
Michael Walle [Fri, 8 Sep 2023 10:16:19 +0000 (12:16 +0200)]
mtd: spi-nor: remove catalyst 'flashes'
CAT25xx are actually EEPROMs manufactured by Catalyst. The devices are
ancient (DS are from 1998), there are not in-tree users, nor are there
any device tree bindings. Remove it. The correct driver is the at25.
Merge tag 'x86-urgent-2023-09-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar:
"Misc fixes:
- Fix an UV boot crash
- Skip spurious ENDBR generation on _THIS_IP_
- Fix ENDBR use in putuser() asm methods
- Fix corner case boot crashes on 5-level paging
- and fix a false positive WARNING on LTO kernels"
* tag 'x86-urgent-2023-09-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/purgatory: Remove LTO flags
x86/boot/compressed: Reserve more memory for page tables
x86/ibt: Avoid duplicate ENDBR in __put_user_nocheck*()
x86/ibt: Suppress spurious ENDBR
x86/platform/uv: Use alternate source for socket to node data
Merge tag 'sched-urgent-2023-09-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fixes from Ingo Molnar:
"Fix a performance regression on large SMT systems, an Intel SMT4
balancing bug, and a topology setup bug on (Intel) hybrid processors"
* tag 'sched-urgent-2023-09-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/sched: Restore the SD_ASYM_PACKING flag in the DIE domain
sched/fair: Fix SMT4 group_smt_balance handling
sched/fair: Optimize should_we_balance() for large SMT systems
The choose_32_64() macros were added to deal with an odd inconsistency
between the 32-bit and 64-bit layout of 'struct stat' way back when in
commit a52dd971f947 ("vfs: de-crapify "cp_new_stat()" function").
Then a decade later Mikulas noticed that said inconsistency had been a
mistake in the early x86-64 port, and shouldn't have existed in the
first place. So commit 932aba1e1690 ("stat: fix inconsistency between
struct stat and struct compat_stat") removed the uses of the helpers.
Merge tag '6.6-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fixes from Steve French:
"Three small SMB3 client fixes, one to improve a null check and two
minor cleanups"
* tag '6.6-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb3: fix some minor typos and repeated words
smb3: correct places where ENOTSUPP is used instead of preferred EOPNOTSUPP
smb3: move server check earlier when setting channel sequence number
Merge tag 'ext4_for_linus-6.6-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 fixes from Ted Ts'o:
"Regression and bug fixes for ext4"
* tag 'ext4_for_linus-6.6-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: fix rec_len verify error
ext4: do not let fstrim block system suspend
ext4: move setting of trimmed bit into ext4_try_to_trim_range()
jbd2: Fix memory leak in journal_init_common()
jbd2: Remove page size assumptions
buffer: Make bh_offset() work for compound pages
This causes WARNING from kexec_purgatory_setup_sechdrs():
WARNING: CPU: 26 PID: 110894 at kernel/kexec_file.c:919
kexec_load_purgatory+0x37f/0x390
Fix this by disabling LTO for purgatory.
[ AFAICT, x86 is the only arch that supports LTO and purgatory. ]
We could also fix this with an explicit linker script to rejoin .text.*
sections back into .text. However, given the benefit of LTOing purgatory
is small, simply disable the production of more .text.* sections for now.
Fixes: b33fff07e3e3 ("x86, build: allow LTO to be selected") Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Ingo Molnar <mingo@kernel.org> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com> Reviewed-by: Sami Tolvanen <samitolvanen@google.com> Link: https://lore.kernel.org/r/20230914170138.995606-1-song@kernel.org
x86/boot/compressed: Reserve more memory for page tables
The decompressor has a hard limit on the number of page tables it can
allocate. This limit is defined at compile-time and will cause boot
failure if it is reached.
The kernel is very strict and calculates the limit precisely for the
worst-case scenario based on the current configuration. However, it is
easy to forget to adjust the limit when a new use-case arises. The
worst-case scenario is rarely encountered during sanity checks.
In the case of enabling 5-level paging, a use-case was overlooked. The
limit needs to be increased by one to accommodate the additional level.
This oversight went unnoticed until Aaron attempted to run the kernel
via kexec with 5-level paging and unaccepted memory enabled.
Update wost-case calculations to include 5-level paging.
To address this issue, let's allocate some extra space for page tables.
128K should be sufficient for any use-case. The logic can be simplified
by using a single value for all kernel configurations.
[ Also add a warning, should this memory run low - by Dave Hansen. ]
Fixes: 34bbb0009f3b ("x86/boot/compressed: Enable 5-level paging during decompression stage") Reported-by: Aaron Lu <aaron.lu@intel.com> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@kernel.org> Link: https://lore.kernel.org/r/20230915070221.10266-1-kirill.shutemov@linux.intel.com
projects / users / hch / block.git / log