Pull mtd updates from Miquel Raynal:
"MTD:
- Apart from preventing the mtdblk to run on top of ftl or ubiblk
(which may cause security issues and has no meaning anyway), there
are a few misc fixes.
Raw NAND:
- Two meaningful changes this time. The conversion of the brcmnand
driver to the ->exec_op() API, this series brought additional
changes to the core in order to help controller drivers to handle
themselves the WP pin during destructive operations when relevant.
- There is also a series bringing important fixes to the sequential
read feature.
- As always, there is as well a whole bunch of miscellaneous W=1
fixes, together with a few runtime fixes (double free, timeout
value, OOB layout, missing register initialization) and the usual
load of remove callbacks turned into void (which led to switch the
txx9ndfmc driver to use module_platform_driver()).
SPI NOR:
- SPI NOR comes with die erase support for multi die flashes, with
new octal protocols (1-1-8 and 1-8-8) parsed from SFDP and with an
updated documentation about what the contributors shall consider
when proposing flash additions or updates.
- Michael Walle stepped out from the reviewer role to maintainer"
* tag 'mtd/for-6.8' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux: (39 commits)
mtd: rawnand: Clarify conditions to enable continuous reads
mtd: rawnand: Prevent sequential reads with on-die ECC engines
mtd: rawnand: Fix core interference with sequential reads
mtd: rawnand: Prevent crossing LUN boundaries during sequential reads
mtd: Fix gluebi NULL pointer dereference caused by ftl notifier
dt-bindings: mtd: partitions: u-boot: Fix typo
mtd: rawnand: s3c2410: fix Excess struct member description kernel-doc warnings
MAINTAINERS: change my mail to the kernel.org one
mtd: spi-nor: sfdp: get the 1-1-8 and 1-8-8 protocol from SFDP
mtd: spi-nor: drop superfluous debug prints
mtd: spi-nor: sysfs: hide the flash name if not set
mtd: spi-nor: mark the flash name as obsolete
mtd: spi-nor: print flash ID instead of name
mtd: maps: vmu-flash: Fix the (mtd core) switch to ref counters
mtd: ssfdc: Remove an unused variable
mtd: rawnand: diskonchip: fix a potential double free in doc_probe
mtd: rawnand: rockchip: Add missing title to a kernel doc comment
mtd: rawnand: rockchip: Rename a structure
mtd: rawnand: pl353: Fix kernel doc
mtd: spi-nor: micron-st: Add support for mt25qu01g
...
* Raw NAND
The most meaningful change being the conversion of the brcmnand driver
to the ->exec_op() API, this series brought additional changes to the
core in order to help controller drivers to handle themselves the WP pin
during destructive operations when relevant.
As always, there is as well a whole bunch of miscellaneous W=1 fixes,
together with a few runtime fixes (double free, timeout value, OOB
layout, missing register initialization) and the usual load of remove
callbacks turned into void (which led to switch the txx9ndfmc driver to
use module_platform_driver()).
SPI NOR comes with die erase support for multi die flashes, with new
octal protocols (1-1-8 and 1-8-8) parsed from SFDP and with an updated
documentation about what the contributors shall consider when proposing
flash additions or updates. Michael Walle stepped out from the reviewer
role to maintainer.
The current logic is probably fine but is a bit convoluted. Plus, we
don't want partial pages to be part of the sequential operation just in
case the core would optimize the page read with a subpage read (which
would break the sequence). This may happen on the first and last page
only, so if the start offset or the end offset is not aligned with a
page boundary, better avoid them to prevent any risk.
Cc: stable@vger.kernel.org
Fixes: 003fe4b954 ("mtd: rawnand: Support for sequential cache reads")
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Tested-by: Martin Hundebøll <martin@geanix.com>
Link: https://lore.kernel.org/linux-mtd/20231215123208.516590-5-miquel.raynal@bootlin.com
The ONFI specification states that devices do not need to support
sequential reads across LUN boundaries. In order to prevent such event
from happening and possibly failing, let's introduce the concept of
"pause" in the sequential read to handle these cases. The first/last
pages remain the same but any time we cross a LUN boundary we will end
and restart (if relevant) the sequential read operation.
Cc: stable@vger.kernel.org
Fixes: 003fe4b954 ("mtd: rawnand: Support for sequential cache reads")
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Tested-by: Martin Hundebøll <martin@geanix.com>
Link: https://lore.kernel.org/linux-mtd/20231215123208.516590-2-miquel.raynal@bootlin.com
If both ftl.ko and gluebi.ko are loaded, the notifier of ftl
triggers NULL pointer dereference when trying to access
‘gluebi->desc’ in gluebi_read().
ubi_gluebi_init
ubi_register_volume_notifier
ubi_enumerate_volumes
ubi_notify_all
gluebi_notify nb->notifier_call()
gluebi_create
mtd_device_register
mtd_device_parse_register
add_mtd_device
blktrans_notify_add not->add()
ftl_add_mtd tr->add_mtd()
scan_header
mtd_read
mtd_read_oob
mtd_read_oob_std
gluebi_read mtd->read()
gluebi->desc - NULL
Detailed reproduction information available at the Link [1],
In the normal case, obtain gluebi->desc in the gluebi_get_device(),
and access gluebi->desc in the gluebi_read(). However,
gluebi_get_device() is not executed in advance in the
ftl_add_mtd() process, which leads to NULL pointer dereference.
The solution for the gluebi module is to run jffs2 on the UBI
volume without considering working with ftl or mtdblock [2].
Therefore, this problem can be avoided by preventing gluebi from
creating the mtdblock device after creating mtd partition of the
type MTD_UBIVOLUME.
Fixes: 2ba3d76a1e ("UBI: make gluebi a separate module")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217992 [1]
Link: https://lore.kernel.org/lkml/441107100.23734.1697904580252.JavaMail.zimbra@nod.at/ [2]
Signed-off-by: ZhaoLong Wang <wangzhaolong1@huawei.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Acked-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231220024619.2138625-1-wangzhaolong1@huawei.com
BFPT 17th DWORD contains the information about 1-1-8 and 1-8-8.
Parse BFPT DWORD[17] instruction to determine whether flash
supports 1-1-8 and 1-8-8, and set its dummy cycles accordingly.
Validated only the 1-1-8 read using a macronix flash with
Xilinx board zynq-picozed.
Signed-off-by: JaimeLiao <jaimeliao@mxic.com.tw>
Reviewed-by: Michael Walle <mwalle@kernel.org>
Link: https://lore.kernel.org/r/20231219102103.92738-2-jaimeliao.tw@gmail.com
[ta: update commit message, get rid of extra dereference]
Signed-off-by: Tudor Ambarus <tudor.ambarus@linaro.org>
When nand_scan() fails, it has cleaned up related resources
in its error paths. Therefore, the following nand_cleanup()
may lead to a double-free. One possible trace is:
doc_probe
|-> nand_scan
| |-> nand_scan_with_ids
| |-> nand_scan_tail
| |-> kfree(chip->data_buf) [First free]
|
|-> nand_cleanup
|-> kfree(chip->data_buf) [Double free here]
Fix this by removing nand_cleanup() on failure of
nand_scan().
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231214072946.10285-1-dinghao.liu@zju.edu.cn
JESD216 mentions die erase, but does not provide an opcode for it.
Check BFPT dword 11, bits 30:24, "Chip Erase, Typical time", it says:
"Typical time to erase one chip (die). User must poll device busy to
determine if the operation has completed. For a device consisting of
multiple dies, that are individually accessed, the time is for each die
to which a chip erase command is applied."
So when a flash consists of a single die, this is the erase time for the
full chip (die) erase, and when it consists of multiple dies, it's the
die erase time. Chip and die are the same thing.
Add support for die erase. For now, benefit of the die erase when addr
and len are aligned with die size. This could be improved however for
the uniform and non-uniform erases cases to use the die erase when
possible. For example if one requests that an erase of a 2 die device
starting from the last 64KB of the first die to the end of the flash
size, we could use just 2 commands, a 64KB erase and a die erase.
This improvement is left as an exercise for the reader.
Tested-by: Fabio Estevam <festevam@denx.de>
Link: https://lore.kernel.org/r/20231125123529.55686-2-tudor.ambarus@linaro.org
Signed-off-by: Tudor Ambarus <tudor.ambarus@linaro.org>
This commit updates the SPI subsystem, particularly affecting "SPI MEM"
drivers and core parts, by replacing the -ENOTSUPP error code with
-EOPNOTSUPP.
The key motivations for this change are as follows:
1. The spi-nor driver currently uses EOPNOTSUPP, whereas calls to spi-mem
might return ENOTSUPP. This update aims to unify the error reporting
within the SPI subsystem for clarity and consistency.
2. The use of ENOTSUPP has been flagged by checkpatch as inappropriate,
mainly being reserved for NFS-related errors. To align with kernel coding
standards and recommendations, this change is being made.
3. By using EOPNOTSUPP, we provide more specific context to the error,
indicating that a particular operation is not supported. This helps
differentiate from the more generic ENOTSUPP error, allowing drivers to
better handle and respond to different error scenarios.
Risks and Considerations:
While this change is primarily intended as a code cleanup and error code
unification, there is a minor risk of breaking user-space applications
that rely on specific return codes for unsupported operations. However,
this risk is considered low, as such use-cases are unlikely to be common
or critical. Nevertheless, developers and users should be aware of this
change, especially if they have scripts or tools that specifically handle
SPI error codes.
This commit does not introduce any functional changes to the SPI subsystem
or the affected drivers.
Signed-off-by: "Chia-Lin Kao (AceLan)" <acelan.kao@canonical.com>
Acked-by: Tudor Ambarus <tudor.ambarus@linaro.org>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Acked-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20231129064311.272422-1-acelan.kao@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Under heavy load it is likely that the controller is done
with its own task but the thread unlocking the wait is not
scheduled in time. Increasing IFC_TIMEOUT_MSECS allows the
controller to respond within allowable timeslice of 1 sec.
fsl,ifc-nand 7e800000.nand: Controller is not responding
[<804b2047>] (nand_get_device) from [<804b5335>] (nand_write_oob+0x1b/0x4a)
[<804b5335>] (nand_write_oob) from [<804a3585>] (mtd_write+0x41/0x5c)
[<804a3585>] (mtd_write) from [<804c1d47>] (ubi_io_write+0x17f/0x22c)
[<804c1d47>] (ubi_io_write) from [<804c047b>] (ubi_eba_write_leb+0x5b/0x1d0)
Fixes: 82771882d9 ("NAND Machine support for Integrated Flash Controller")
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Ronald Monthero <debug.penguin32@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231118083156.776887-1-debug.penguin32@gmail.com
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.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231102220246.3336154-10-u.kleine-koenig@pengutronix.de
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().
By changing the function brcmnand_remove() to return void several
drivers that use this function as remove callback can be converted to
.remove_new().
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20231102220246.3336154-7-u.kleine-koenig@pengutronix.de
Pull MIPS updates from Thomas Bogendoerfer:
- removed AR7 platform support
- cleanups and fixes
* tag 'mips_6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
MIPS: AR7: remove platform
watchdog: ar7_wdt: remove driver to prepare for platform removal
vlynq: remove bus driver
mtd: parsers: ar7: remove support
serial: 8250: remove AR7 support
arch: mips: remove ReiserFS from defconfig
MIPS: lantiq: Remove unnecessary include of <linux/of_irq.h>
MIPS: lantiq: Fix pcibios_plat_dev_init() "no previous prototype" warning
MIPS: KVM: Fix a build warning about variable set but not used
MIPS: Remove dead code in relocate_new_kernel
mips: dts: ralink: mt7621: rename to GnuBee GB-PC1 and GnuBee GB-PC2
mips: dts: ralink: mt7621: define each reset as an item
mips: dts: ingenic: Remove unneeded probe-type properties
MIPS: loongson32: Remove dma.h and nand.h
Pull UBI and UBIFS updates from Richard Weinberger:
- UBI Fastmap improvements
- Minor issues found by static analysis bots in both UBI and UBIFS
- Fix for wrong dentry length UBIFS in fscrypt mode
* tag 'ubifs-for-linus-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
ubifs: ubifs_link: Fix wrong name len calculating when UBIFS is encrypted
ubi: block: Fix use-after-free in ubiblock_cleanup
ubifs: fix possible dereference after free
ubi: fastmap: Add control in 'UBI_IOCATT' ioctl to reserve PEBs for filling pools
ubi: fastmap: Add module parameter to control reserving filling pool PEBs
ubi: fastmap: Fix lapsed wear leveling for first 64 PEBs
ubi: fastmap: Get wl PEB even ec beyonds the 'max' if free PEBs are run out
ubi: fastmap: may_reserve_for_fm: Don't reserve PEB if fm_anchor exists
ubi: fastmap: Remove unneeded break condition while filling pools
ubi: fastmap: Wait until there are enough free PEBs before filling pools
ubi: fastmap: Use free pebs reserved for bad block handling
ubi: Replace erase_block() with sync_erase()
ubi: fastmap: Allocate memory with GFP_NOFS in ubi_update_fastmap
ubi: fastmap: erase_block: Get erase counter from wl_entry rather than flash
ubi: fastmap: Fix missed ec updating after erasing old fastmap data block
ubifs: Fix missing error code err
ubifs: Fix memory leak of bud->log_hash
ubifs: Fix some kernel-doc comments
Pull mtd updates from Miquel Raynal:
"The main set of changes is related to Uwe's work converting platform
remove callbacks to return void. Comes next (in number of changes)
Kees' additional structures annotations to improve the sanitizers. The
usual amount of cleanups apply.
About the more substancial contribution, one main function of the
partitions core could return an error which was not checked, this is
now fixed. On the bindings side, fixed partitions can now have a
compression property. Finally, an erroneous situation is now always
avoided in the MAP RAM driver.
CFI:
- A several years old byte swap has been fixed.
NAND:
- The subsystem has, as usual, seen a bit of cleanup being done this
cycle, typically return values of platform_get_irq() and
devm_kasprintf(). 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.
SPI NOR:
- 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.
Instead of:
{ "w25q512nwm", INFO(0xef8020, 0, 64 * 1024, 0)
OTP_INFO(256, 3, 0x1000, 0x1000) },
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"
* tag 'mtd/for-6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux: (91 commits)
mtd: cfi_cmdset_0001: Byte swap OTP info
mtd: rawnand: meson: check return value of devm_kasprintf()
mtd: rawnand: intel: check return value of devm_kasprintf()
mtd: rawnand: sh_flctl: Convert to module_platform_driver()
mtd: spi-nor: micron-st: use SFDP table for mt25qu512a
mtd: spi-nor: micron-st: enable lock/unlock for mt25qu512a
mtd: rawnand: Remove unused of_gpio.h inclusion
mtd: spinand: Add support for XTX XT26xxxDxxxxx
mtd: spinand: winbond: add support for serial NAND flash
mtd: rawnand: cadence: Annotate struct cdns_nand_chip with __counted_by
mtd: rawnand: Annotate struct mtk_nfc_nand_chip with __counted_by
mtd: spinand: add support for FORESEE F35SQA002G
mtd: rawnand: rockchip: Use struct_size()
mtd: rawnand: arasan: Include ECC syndrome along with in-band data while checking for ECC failure
mtd: Use device_get_match_data()
mtd: spi-nor: nxp-spifi: Convert to platform remove callback returning void
mtd: spi-nor: hisi-sfc: Convert to platform remove callback returning void
mtd: maps: sun_uflash: Convert to platform remove callback returning void
mtd: maps: sa1100-flash: Convert to platform remove callback returning void
mtd: maps: pxa2xx-flash: Convert to platform remove callback returning void
...
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.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
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.
Instead of:
- { "w25q512nwm", INFO(0xef8020, 0, 64 * 1024, 0)
- OTP_INFO(256, 3, 0x1000, 0x1000) },
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.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Pull char/misc updates from Greg KH:
"Here is the big set of char/misc and other small driver subsystem
changes for 6.7-rc1. Included in here are:
- IIO subsystem driver updates and additions (largest part of this
pull request)
- FPGA subsystem driver updates
- Counter subsystem driver updates
- ICC subsystem driver updates
- extcon subsystem driver updates
- mei driver updates and additions
- nvmem subsystem driver updates and additions
- comedi subsystem dependency fixes
- parport driver fixups
- cdx subsystem driver and core updates
- splice support for /dev/zero and /dev/full
- other smaller driver cleanups
All of these have been in linux-next for a while with no reported
issues"
* tag 'char-misc-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (326 commits)
cdx: add sysfs for subsystem, class and revision
cdx: add sysfs for bus reset
cdx: add support for bus enable and disable
cdx: Register cdx bus as a device on cdx subsystem
cdx: Create symbol namespaces for cdx subsystem
cdx: Introduce lock to protect controller ops
cdx: Remove cdx controller list from cdx bus system
dts: ti: k3-am625-beagleplay: Add beaglecc1352
greybus: Add BeaglePlay Linux Driver
dt-bindings: net: Add ti,cc1352p7
dt-bindings: eeprom: at24: allow NVMEM cells based on old syntax
dt-bindings: nvmem: SID: allow NVMEM cells based on old syntax
Revert "nvmem: add new config option"
MAINTAINERS: coresight: Add missing Coresight files
misc: pci_endpoint_test: Add deviceID for J721S2 PCIe EP device support
firmware: xilinx: Move EXPORT_SYMBOL_GPL next to zynqmp_pm_feature definition
uacce: make uacce_class constant
ocxl: make ocxl_class constant
cxl: make cxl_class constant
misc: phantom: make phantom_class constant
...
Pull gpio updates from Bartosz Golaszewski:
"We don't have any new drivers. The loongson driver is getting extended
with support for new models. There's a big refactor of gpio-pca953x
and many small improvements to others.
The GPIO code in the kernel has acquired a lot of cruft over the years
as well as many abusers of the API across the kernel tree. This
release cycle we have started a major cleanup and improvement effort
that will most likely span several releases. We have started by
converting external users of struct gpio_chip to accessing the wrapper
around it - struct gpio_device. This is because the latter is
reference counted while the former is removed when the provider is
unbound. We also removed several instances of drivers accessing
private GPIOLIB structures and including the private header from
drivers/gpio/.
To that end you'll see several commits aimed at different subsystems
(acked by relevant maintainers) as well as two merges from the
x86/platform tree.
We'll then rework the locking in GPIOLIB which currently uses a big
spinlock for many different things and could use becoming more
fine-grained, especially as it doesn't even get the locking right.
We'll also use SRCU for protecting the gpio_chip pointer against
in-kernel hot-unplug crashes similar to what we saw triggered from
user-space and fixed with semaphores in gpiolib-cdev. The core GPIOLIB
is still vulnerable to these use-cases. I'm just mentioning the plans
here, this is not part of this PR.
You'll see some new instances of using __free(). We've added a
gpio_device_put cleanup helper similar to the put_device one
introduced by Peter Zijlstra and used it according to the preferred
pattern except where it didn't make sense.
GPIOLIB core:
- provide interfaces allowing users to retrieve, manage and query the
reference counted GPIO device instead of accessing the private
gpio_chip structure
- replace gpiochip_find() with gpio_device_find()
- remove unused acpi_get_and_request_gpiod()
- improve the ignore_interrupt functionality in GPIO ACPI
- correct notifier return codes in gpiolib-of
- unexport gpiod_set_transitory() as it's unused outside of core GPIO
code
- while there are still external users accessing struct gpio_chip,
let's make gpiochip_get_desc() public so that they at least use the
preferred helper
- improve locking for lookup tables
- annotate struct linereq with __counted_by
- improve GPIOLIB docs
- add an OF quirk for LED trigger sources
Driver improvements:
- convert all GPIO drivers with .remove() callbacks to using the new
variant returning void instead of int
- stop accessing the GPIOLIB private structures in gpio-mockup,
i2c-mux-gpio, hte-tegra194, gpio-sim
- use the recommended pattern for autofree variables in gpio-sim
- add support for more models to gpio-loongson
- use a notifier chain to notify other blocks about interrupts in
gpio-eic-sprd instead of looking up GPIO devices on every interrupt
- convert gpio-pca953x and gpio-fx6408 to using the maple tree regmap
cache
- don't include GPIOLIB internal headers in drivers which don't need
them
- move the ingenic NAND quirk into gpiolib-of
- add an ignore interrupt quirk for Peaq C1010
- drop static GPIO base from gpio-omap, gpio-f7188x
- use the preferred device_get_match_data() function in drivers that
still don't
- refactor gpio-pca953x: switch to using DEFINE_SIMPLE_DEV_PM_OPS(),
use cleanup helpers, use dev_err_probe() where it makes sense,
fully convert to using devres and some other minor tweaks
DT bindings:
- add support for a new model to gpio-vf610 and update existing
properties
- add support for more loongson models
- add missing support for imx models that are used but undocumented
- convert bindings for Intel IXP4xx to schema
Minor stuff:
- deprecate gpio-mockup in favor of gpio-sim
- include missing headers here and there
- stop using gpiochip_find() in OMAP1 board files
- minor tweaks in gpio-vf610, gpio-hisi
- remove unneeded 'extern' specifiers from headers"
* tag 'gpio-updates-for-v6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux: (108 commits)
hte: tegra194: add GPIOLIB dependency
hte: tegra194: don't access struct gpio_chip
gpiolib: provide gpio_device_get_base()
i2c: mux: gpio: don't fiddle with GPIOLIB internals
gpiolib: provide gpiod_to_gpio_device()
gpiolib: provide gpio_device_to_device()
gpio: hisi: Fix format specifier
gpiolib: provide gpio_device_find_by_fwnode()
gpio: acpi: remove acpi_get_and_request_gpiod()
gpio: Use device_get_match_data()
gpio: vf610: update comment for i.MX8ULP and i.MX93 legacy compatibles
platform/x86: int3472: Switch to devm_get_gpiod()
platform/x86: int3472: Stop using gpiod_toggle_active_low()
platform/x86: int3472: Add new skl_int3472_gpiod_get_from_temp_lookup() helper
platform/x86: int3472: Add new skl_int3472_fill_gpiod_lookup() helper
gpio: vf610: simplify code by dropping data check
gpio: vf610: add i.MX8ULP of_device_id entry
dt-bindings: gpio: vf610: add i.MX95 compatible
dt-bindings: gpio: vf610: correct i.MX8ULP and i.MX93
dt-bindings: gpio: vf610: update gpio-ranges
...
The following BUG is reported when a ubiblock is removed:
==================================================================
BUG: KASAN: slab-use-after-free in ubiblock_cleanup+0x88/0xa0 [ubi]
Read of size 4 at addr ffff88810c8f3804 by task ubiblock/1716
CPU: 5 PID: 1716 Comm: ubiblock Not tainted 6.6.0-rc2+ #135
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014
Call Trace:
<TASK>
dump_stack_lvl+0x37/0x50
print_report+0xd0/0x620
kasan_report+0xb6/0xf0
ubiblock_cleanup+0x88/0xa0 [ubi]
ubiblock_remove+0x121/0x190 [ubi]
vol_cdev_ioctl+0x355/0x630 [ubi]
__x64_sys_ioctl+0xc7/0x100
do_syscall_64+0x3f/0x90
entry_SYSCALL_64_after_hwframe+0x6e/0xd8
RIP: 0033:0x7f08d7445577
Code: b3 66 90 48 8b 05 11 89 2c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d e1 8
RSP: 002b:00007ffde05a3018 EFLAGS: 00000206 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 00000000ffffffff RCX: 00007f08d7445577
RDX: 0000000000000000 RSI: 0000000000004f08 RDI: 0000000000000003
RBP: 0000000000816010 R08: 00000000008163a7 R09: 0000000000000000
R10: 0000000000000003 R11: 0000000000000206 R12: 0000000000000003
R13: 00007ffde05a3130 R14: 0000000000000000 R15: 0000000000000000
</TASK>
Allocated by task 1715:
kasan_save_stack+0x22/0x50
kasan_set_track+0x25/0x30
__kasan_kmalloc+0x7f/0x90
__alloc_disk_node+0x40/0x2b0
__blk_mq_alloc_disk+0x3e/0xb0
ubiblock_create+0x2ba/0x620 [ubi]
vol_cdev_ioctl+0x581/0x630 [ubi]
__x64_sys_ioctl+0xc7/0x100
do_syscall_64+0x3f/0x90
entry_SYSCALL_64_after_hwframe+0x6e/0xd8
Freed by task 0:
kasan_save_stack+0x22/0x50
kasan_set_track+0x25/0x30
kasan_save_free_info+0x2b/0x50
__kasan_slab_free+0x10e/0x190
__kmem_cache_free+0x96/0x220
bdev_free_inode+0xa4/0xf0
rcu_core+0x496/0xec0
__do_softirq+0xeb/0x384
The buggy address belongs to the object at ffff88810c8f3800
which belongs to the cache kmalloc-1k of size 1024
The buggy address is located 4 bytes inside of
freed 1024-byte region [ffff88810c8f3800, ffff88810c8f3c00)
The buggy address belongs to the physical page:
page:00000000d03de848 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x10c8f0
head:00000000d03de848 order:3 entire_mapcount:0 nr_pages_mapped:0 pincount:0
flags: 0x200000000000840(slab|head|node=0|zone=2)
page_type: 0xffffffff()
raw: 0200000000000840 ffff888100042dc0 ffffea0004244400 dead000000000002
raw: 0000000000000000 0000000080100010 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
ffff88810c8f3700: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff88810c8f3780: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff88810c8f3800: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
^
ffff88810c8f3880: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff88810c8f3900: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Fix it by using a local variable to record the gendisk ID.
Fixes: 77567b25ab ("ubi: use blk_mq_alloc_disk and blk_cleanup_disk")
Signed-off-by: ZhaoLong Wang <wangzhaolong1@huawei.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Adding 6th module parameter in 'mtd=xxx' to control whether or not
reserving PEBs for filling pool/wl_pool.
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
The anchor PEB must be picked from first 64 PEBs, these PEBs could have
large erase counter greater than other PEBs especially when free space
is nearly running out.
The ubi_update_fastmap will be called as long as pool/wl_pool is empty,
old anchor PEB is erased when updating fastmap. Given an UBI device with
N PEBs, free PEBs is nearly running out and pool will be filled with 1
PEB every time ubi_update_fastmap invoked. So t=N/POOL_SIZE[1]/64 means
that in worst case the erase counter of first 64 PEBs is t times greater
than other PEBs in theory.
After running fsstress for 24h, the erase counter statistics for two UBI
devices shown as follow(CONFIG_MTD_UBI_WL_THRESHOLD=128):
Device A(1024 PEBs, pool=50, wl_pool=25):
=========================================================
from to count min avg max
---------------------------------------------------------
0 .. 9: 0 0 0 0
10 .. 99: 0 0 0 0
100 .. 999: 0 0 0 0
1000 .. 9999: 0 0 0 0
10000 .. 99999: 960 29224 29282 29362
100000 .. inf: 64 117897 117934 117940
---------------------------------------------------------
Total : 1024 29224 34822 117940
Device B(8192 PEBs, pool=256, wl_pool=128):
=========================================================
from to count min avg max
---------------------------------------------------------
0 .. 9: 0 0 0 0
10 .. 99: 0 0 0 0
100 .. 999: 0 0 0 0
1000 .. 9999: 8128 2253 2321 2387
10000 .. 99999: 64 35387 35387 35388
100000 .. inf: 0 0 0 0
---------------------------------------------------------
Total : 8192 2253 2579 35388
The key point is reducing fastmap updating frequency by enlarging
POOL_SIZE, so let UBI reserve ubi->fm_pool.max_size PEBs during
attaching. Then POOL_SIZE will become ubi->fm_pool.max_size/2 even
in free space running out case.
Given an UBI device with 8192 PEBs(16384\8192\4096 is common
large-capacity flash), t=8192/128/64=1. The fastmap updating will
happen in either wl_pool or pool is empty, so setting fm_pool_rsv_cnt
as ubi->fm_pool.max_size can fill wl_pool in full state.
After pool reservation, running fsstress for 24h:
Device A(1024 PEBs, pool=50, wl_pool=25):
=========================================================
from to count min avg max
---------------------------------------------------------
0 .. 9: 0 0 0 0
10 .. 99: 0 0 0 0
100 .. 999: 0 0 0 0
1000 .. 9999: 0 0 0 0
10000 .. 99999: 1024 33801 33997 34056
100000 .. inf: 0 0 0 0
---------------------------------------------------------
Total : 1024 33801 33997 34056
Device B(8192 PEBs, pool=256, wl_pool=128):
=========================================================
from to count min avg max
---------------------------------------------------------
0 .. 9: 0 0 0 0
10 .. 99: 0 0 0 0
100 .. 999: 0 0 0 0
1000 .. 9999: 8192 2205 2397 2460
10000 .. 99999: 0 0 0 0
100000 .. inf: 0 0 0 0
---------------------------------------------------------
Total : 8192 2205 2397 2460
The difference of erase counter between first 64 PEBs and others is
under WL_FREE_MAX_DIFF(2*UBI_WL_THRESHOLD=2*128=256).
Device A: 34056 - 33801 = 255
Device B: 2460 - 2205 = 255
Next patch will add a switch to control whether UBI needs to reserve
PEBs for filling pool.
Fixes: dbb7d2a88d ("UBI: Add fastmap core")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217787
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
