| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Pull MTD fixes from Miquel Raynal:
"The most important set of fixes are around the handling of the QE bit
in SPI NAND.
There are also a couple of behavioral fixes (mutex issue in SPI-NOR,
spurious bitflips on vf610_nfc, OOB bytes count in SPI NAND and
cfi_cmdset stack usage).
The rest is mostly AI fuzzing results"
* tag 'mtd/fixes-for-7.3-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux:
mtd: spinand: Do not update the QE bit on devices without one
mtd: spi-nor: core: Fix mutex leak in spi_nor_rww_start_exclusive()
mtd: rawnand: cadence: Initialize IRQ state before requesting IRQ
mtd: rawnand: vf610_nfc: fix false bitflips on reads of erased pages
mtd: rawnand: vf610_nfc: fix reads on chips with more than 64 bytes of OOB
mtd: spinand: fix zero oobavail when no ECC engine is used
mtd: spinand: fix NULL pointer dereference with no ECC engine
mtd: mtd_intel_dg: reset poll counter for each erase
mtd: cfi_cmdset_0001: shrink do_write_buffer() stack frame
mtd: core: call _get_device() with the master MTD
mtd: core: avoid double-free of OTP NVMEM device
mtd: spinand: Enable QE on all dies
mtd: block2mtd: Fix divide error when erase_size is zero
|
|
Commit be0b86c648bf ("mtd: spinand: Gather all the bus interface
steps in one single function") moved quad-enable setup into
spinand_configure_chip(). The new code only determines whether quad
mode is needed when SPINAND_HAS_QE_BIT is set, but calls
spinand_init_quad_enable() unconditionally. This clears configuration
register bit 0 on devices without a QE bit.
That bit is not universally a QE bit. On the Winbond W25N02KV it is
H-DIS, which disables the active-low HOLD function. Clearing H-DIS
enables HOLD during single and dual I/O operations. If IO3 is not kept
high, the flash can pause a command and ignore clock and data.
H-DIS is not restored by the FFh reset command, allowing the incorrect
state to survive an SoC warm reboot while the flash remains powered.
Before the refactoring, spinand_init_quad_enable() returned without
touching the configuration register on devices without
SPINAND_HAS_QE_BIT. Restore that behavior by only calling the helper
when the flag is set.
Return zero explicitly once SSDR configuration completes, as all errors
are returned immediately. This avoids returning an uninitialized value
when neither optional configuration step runs.
The regression was reproduced on a JioRouter JIDU6401 with an MT7986
SoC and a W25N02KV. With Linux 6.18.44, sysupgrade failed and the
following warm reboot hung in BL2. With this change applied, both
sysupgrade and warm reboot completed successfully.
Fixes: be0b86c648bf ("mtd: spinand: Gather all the bus interface steps in one single function")
Cc: stable@vger.kernel.org
Suggested-by: Miquel Raynal <miquel.raynal@bootlin.com>
Assisted-by: LLM
Signed-off-by: Sagnik Sasmal <sagnik@sagnik.me>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Use the same guard(mutex) pattern as the other RWW helpers so nor->lock
is released on both the busy and successful return paths.
Fixes: 03e7bb864d9a ("mtd: spi-nor: use scope-based mutex cleanup helpers")
Cc: stable@vger.kernel.org
Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@linaro.org>
[mw: rephrased commit message]
Signed-off-by: Michael Walle <mwalle@kernel.org>
|
|
This is another run of the Coccinelle script for converting kmalloc()
family of allocations to kmalloc_obj() via the existing rules in
scripts/coccinelle/api/kmalloc_objs.cocci
This catches both the set of kmalloc() uses added since the first
kmalloc_obj() conversions in v7.0 and adds a large group missed in the
first pass due to Coccinelle not interacting well with the cleanup.h
scoped_...() family of macros[1]. I worked around this with spatch's
"--macro-file" argument to a file with all the scoped_...() macros mapped
to Coccinelle's YACFE_ITERATOR[2] as that was the closest viable control
flow indicator I could find.
Build tested allmodconfig on x86, arm64, arm, loongarch, mips, powerpc,
riscv, and s390 with no new warnings.
Link: https://lore.kernel.org/lkml/202609021314.8A9C0B8@keescook/ [1]
Link: https://github.com/coccinelle/coccinelle/blob/master/standard.h [2]
Signed-off-by: Kees Cook <kees+treewide@kernel.org>
|
|
The Cadence NAND interrupt handler uses both the IRQ lock and completion
object. Registering the IRQ before initializing them leaves a window in
which a pending interrupt can access uninitialized synchronization state.
Initialize them before registering the handler.
Fixes: ec4ba01e894d ("mtd: rawnand: Add new Cadence NAND driver to MTD subsystem")
Cc: stable@vger.kernel.org
Assisted-by: Codex:GPT-5
Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
When the ECC engine fails to decode a page, the driver re-reads the OOB
area with the engine bypassed, but runs the erased-page check for the
data area on the buffer left in the controller SRAM by the failed
transfer.
That buffer does not hold what is on the flash: the failing engine
writes a bogus single-bit "correction" into it. In the 60-byte ECC mode
the all-0xff content of an erased page always decodes to the same error
location, so every erased page shows one stale zero bit at data offset
0x5FD, which the erased-page check then reports as a corrected bitflip.
Edward Karpicz discovered this behaviour and identified the offset on a
Colibri VF61; the analysis and the fix build on his finding. Measured
with an instrumented driver on a Colibri VF50 (MX30LF1G18AC, 32-bit
ECC): reading a 126 MiB partition with nanddump increased the corrected
counter by 18035, exactly one per erased page, while raw reads of the
same pages return clean 0xff. A v4.4 kernel on the VF61 (MX30LF4G28AC)
accumulates the same false counts, so the behaviour follows the
controller rather than the chip or the driver generation. Neither the
Vybrid reference manual nor the published mask set errata (VFXXX_2N02G)
document it. The 45-byte ECC mode is not affected.
Restoring the known byte is not enough: on pages that fail to decode
with content other than all-0xff the engine writes its correction
wherever the syndrome points (measured at a different offset on such a
page), so the check has to run on what the flash holds. Re-read the data
area with the ECC engine bypassed, exactly as already done for the OOB
area. The corrected counter then stays at zero on both boards.
Reported-by: Edward Karpicz <webmaster@toradex.com>
Link: https://community.toradex.com/t/colibri-vf50-vf61-on-the-current-bsp-mainline-u-boot-v2026-07-and-linux-6-18-lts/30735
Signed-off-by: Mehmet Fide <mehmet.fide@screeningeagle.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The controller transfers 64 spare bytes per page and the driver only
implements the matching 64-byte ECC layout, so attach_chip() shrinks
mtd->oobsize when the chip provides more. That clamp does not survive:
nand_scan_tail() runs nanddev_init() after ->attach_chip(), and it
restores mtd->oobsize from the memory organization, which still holds
the value detected from the chip. The driver then transfers writesize
plus the chip's full OOB size, the hardware ECC parity ends up at a
different offset than the layout the controller was set up for, and
every ECC-protected read fails with -EBADMSG.
Measured on a Colibri VF61 (MX30LF4G28AC, 2048-byte pages, 112 bytes of
OOB): with the clamp lost, UBI cannot read the erase counter headers of
the pages U-Boot has just written, and the on-flash bad block table
written by an older kernel reads back with ECC errors, so the board
does not boot. Kernels before commit a7ab085d7c16 ("mtd: rawnand:
Initialize the nand_device object") are not affected because nothing
overwrote the clamp there, which is why the same chip works with a v4.4
kernel and with U-Boot, whose copy of this driver has no memory
organization to restore the value from. Edward Karpicz reported that
the clamp no longer takes effect on this chip; see the link below.
Instead of modifying the memory organization, keep the detected OOB
size and give the driver its own mtd_ooblayout_ops: the same layout the
NAND core uses for large pages, but computed on the first 64 OOB bytes
instead of the whole OOB, so the ECC bytes stay where U-Boot and the
old kernels put them. The data paths transfer writesize plus those 64
bytes, as the controller always has.
Since mtd->oobsize now reports the chip's real spare size, fill the
tail of oob_poi with 0xff after the 64 transferred bytes on ECC page
reads: the core may copy the full mtd->oobsize from it, which would
otherwise expose whatever the buffer held before. 0xff also matches
what a raw read returns from flash, since the write path only ever
programs the first 64 spare bytes.
Reported-by: Edward Karpicz <webmaster@toradex.com>
Link: https://community.toradex.com/t/colibri-vf50-vf61-on-the-current-bsp-mainline-u-boot-v2026-07-and-linux-6-18-lts/30735
Suggested-by: Miquel Raynal <miquel.raynal@bootlin.com>
Fixes: a7ab085d7c16 ("mtd: rawnand: Initialize the nand_device object")
Cc: stable@vger.kernel.org
Signed-off-by: Mehmet Fide <mehmet.fide@screeningeagle.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Commit 00c15b78b4b4 ("mtd: spinand: Allow the case where there is no
ECC engine") made the OOB free bytes count conditional on having an
ECC engine so that probing would not fail when none is requested.
However mtd->oobavail is still assigned from ret just after that
block, and ret is 0 there, so the device ends up advertising no
available OOB bytes at all.
mtd_oobavail() returns mtd->oobavail for MTD_OPS_AUTO_OOB, so a zero
value makes every automatic OOB access fail with -EINVAL. JFFS2 fares
worse: it keeps its cleanmarker in the OOB area on NAND and refuses to
mount outright, reporting "inconsistent device description".
No ECC engine also means no ooblayout was ever installed, which would
make the count return -ENOTSUPP, so install the same fallback layout
the on-die path already uses before counting unconditionally.
Fixes: 00c15b78b4b4 ("mtd: spinand: Allow the case where there is no ECC engine")
Cc: stable@vger.kernel.org
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
When "nand-no-ecc-engine" is set in DT, nanddev_get_ecc_engine() takes
the NAND_ECC_ENGINE_TYPE_NONE path and returns success while leaving
nand->ecc.engine NULL. The SPI-NAND code nevertheless dereferences it
unconditionally to test for a pipelined engine, so probing such a
device oopses immediately.
Rather than open-coding the test three times, add a
nand_ecc_is_pipelined() helper to the NAND core that folds the NULL
check into the integration comparison, and use it everywhere. Future
callers then cannot reintroduce the problem.
Fixes: f9d7c7265bcf ("mtd: spinand: Create direct mapping descriptors for ECC operations")
Cc: stable@vger.kernel.org
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The non-posted erase polling counter is initialized only once per MTD
erase request. Large requests therefore share the polling budget across
all 4K erase commands and can fail with -ETIME even though no individual
command timed out.
Reset the counter for each 4K erase command so every operation gets the
intended completion timeout.
Cc: stable@vger.kernel.org
Fixes: a1c940cbf505 ("drm/xe/nvm: add support for non-posted erase")
Signed-off-by: Menachem Adin <menachem.adin@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
arm32 allmodconfig fails to build with gcc:
drivers/mtd/chips/cfi_cmdset_0001.c:1883:1: error: the frame size of 1296
bytes is larger than 1280 bytes [-Werror=frame-larger-than=]
With MTD_MAP_BANK_WIDTH_32 a map_word is 32 bytes, and with KASAN_STACK
every temporary gets its own redzoned slot. do_write_buffer() builds a
map_word for each of the twelve commands it issues, and those temporaries
dominate its frame.
Issue the commands through a small noinline helper that takes the
command as a plain value and builds the map_word in its own frame. The
sequence of reads and writes to the chip is unchanged.
The helper must not be inlined, and reusing one map_word local is not
enough: clang cannot assume the callee does not observe the local, so it
still returns each CMD() into a per-call-site temporary, and the frame
grows (1760 -> 1824 bytes in cfi_intelext_writev(), into which clang
inlines do_write_buffer()).
Frame sizes with MTD_MAP_BANK_WIDTH_32 and KASAN_STACK:
gcc 15.2 do_write_buffer() 1296 -> 520
clang 21 cfi_intelext_writev() 1760 -> 896
The helper itself takes 168 (gcc) / 160 (clang) bytes.
Reported-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Closes: https://lore.kernel.org/all/20260408211407.2295175-1-andriy.shevchenko@linux.intel.com/
Suggested-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/all/87ik9cfm6g.fsf@bootlin.com/
Assisted-by: LLM
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
__get_mtd_device() calls the master's _get_device() callback with mtd,
which may be a partition. __put_mtd_device() passes the master.
This breaks gluebi partitions because gluebi_get_device() uses
container_of() and therefore requires the master MTD. Pass the master
to _get_device(), matching _put_device().
Fixes: 46b5889cc2c5 ("mtd: implement proper partition handling")
Assisted-by: Codex:gpt-5.6-sol
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
If factory OTP setup fails after the user OTP NVMEM device has been
registered, mtd_otp_nvmem_add() unregisters the user device but leaves
mtd->otp_user_nvmem set. On an error, the caller unregisters it again.
For -EOPNOTSUPP, registration continues and normal teardown unregisters
it again.
Clear mtd->otp_user_nvmem after unregistering it.
Fixes: e0489f6e221f ("mtd: core: fix error path for nvmem provider")
Fixes: fe0b8213c012 ("mtd: core: Don't fail mtd_otp_nvmem_add() if OTP is unsupported")
Assisted-by: Codex:gpt-5.6-sol
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The QUAD ENABLE (QE) bit is stored in a per-die configuration
register on some SPI-NAND devices. When a device contains multiple
dies, updating the QE bit only on the currently selected die can
leave the remaining dies operating in non-quad mode.
Iterate over all targets and update the QE setting on each die
during initialization to ensure consistent quad I/O operation
across the entire device.
Tested on ISSI IS38SMW04G8B.
Fixes: 7529df465248 ("mtd: nand: Add core infrastructure to support SPI NANDs")
Cc: stable@vger.kernel.org
Signed-off-by: Han Xu <han.xu@nxp.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The erase size is parsed from the "block2mtd" module parameter and can
be set to zero. add_device() then evaluates
if (size % erase_size)
with a zero divisor, which triggers a divide error:
divide error: 0000 [#1] PREEMPT SMP PTI
RIP: 0010:add_device drivers/mtd/devices/block2mtd.c:296 [inline]
RIP: 0010:block2mtd_setup2+0x592/0xda0 drivers/mtd/devices/block2mtd.c:459
Call Trace:
block2mtd_setup+0x27/0xe0 drivers/mtd/devices/block2mtd.c:476
param_attr_store+0x214/0x310 kernel/params.c:589
module_attr_store+0x65/0x90 kernel/params.c:904
kernfs_fop_write_iter+0x3a4/0x540 fs/kernfs/file.c:345
...
Reject a zero erase size before performing the modulo operation so the
existing "erasesize must be a divisor of device size" error path
reports the invalid argument and frees the device.
Fixes: ea6d833a3fdd ("mtd: block2mtd: check device size")
Reported-by: syzbot+b320a4d5f65a61dbbf89@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/lkml/6a7b58ba.ac361c09.22ff0a.0045.GAE@google.com/
Suggested-by: Jörn Engel <joern@barelysecure.org>
Cc: stable@vger.kernel.org
Signed-off-by: Pei Xiao <xiaopei01@kylinos.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Pull UBI and UBIFS updates from Richard Weinberger:
"UBI:
- Support for a per-device wear-leveling threshold
- Various fixes and cleanups of error paths
- Correctly preserve torture flag up wear-leveling
UBIFS:
- Various fixes and cleanups of error paths and kernel-doc"
* tag 'ubifs-for-linus-7.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
UBI: support per-device wear-leveling threshold
UBI: fix two issues in the ubi.mtd MODULE_PARM_DESC
mtd: ubi: Release device reference on busy detach
ubi: Fix rollback for explicit UBI device numbers
ubifs: fix out-of-bounds read in signature length check
UBI: fastmap: Pass to_be_tortured when reusing old fastmap PEBs
UBI: Preserve torture flag when rescheduling failed erasures
ubifs: ubifs.h: clean up kernel-doc comments
ubifs: key.h: use correct function parameter name
ubifs: debug.h: fix kernel-doc struct prototypes
|
|
Pull mtd updates from Miquel Raynal:
"Raw NAND changes:
- Sunxi: Support added for the H616 compatible
- Qcom: Support added for the MDM9607 compatible
- Support for the Toshiba TC58NVG1S3H part
- GPMI: New debugfs entry to expose the chip geometry
- PL353: Timing updates and software ECC support have been fixed
SPI NAND changes:
- fmsh: Support added for FM25G{01,02}B chips
- HeYangTek: Support added for HYF1GQ4UDACAE
Aside from these main changes, there is a high load of misc fixes and
hardening changes, and exceptionally no SPI NOR change"
* tag 'mtd/for-7.3' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux: (30 commits)
mtd: rawnand: sunxi: fix H6/H616 controller timings
mtd: rawnand: sunxi: describe tADL and tWHR delays
mtd: rawnand: sunxi: group controller delay tables
mtd: maps: remove dead select of MTD_CFI_BE_BYTE_SWAP
mtd: rawnand: gpmi: add debugfs entry for BCH geometry
mtd: rawnand: validate ONFI extended parameter page sections
mtd: rawnand: sunxi: add H616 MBUS DMA support
mtd: spinand: fmsh: fix FM25G01B/FM25G02B Quad I/O read dummy cycles
mtd: part: reject MTDPART_OFS_RETAIN in mtd_add_partition()
mtd: mpc5121_nfc: use platform for irq and ioremap
mtd: mtdoops: free page bitmap when the backing MTD is removed
mtd: mtdswap: Avoid freeing registered blktrans device twice
mtd: afs: validate v2 image info bounds
mtd: intel-dg: Fix runtime PM error path in probe
mtd: nand-omap2: Move omap_nand_ids[] to raw nand driver
mtd: rawnand: add Toshiba TC58NVG1S3H
mtd: nand: realtek-ecc: add missing MODULE_DEVICE_TABLE()
mtd: rawnand: qcom: Add MDM9607 compatible
mtd: rawnand: qcom: Make has_onfi_read_op separate from qpic_version2
mtd: rawnand: qcom: Make "aon" clock optional
...
|
|
* Raw NAND changes
- Sunxi: Support added for the H616 compatible
- Qcom: Support added for the MDM9607 compatible
- Support for the Toshiba TC58NVG1S3H part
- GPMI: New debugfs entry to expose the chip geometry
- PL353: Timing updates and software ECC support have been fixed
* SPI NAND changes
- fmsh: Support added for FM25G{01,02}B chips
- HeYangTek: Support added for HYF1GQ4UDACAE
Aside from these main changes, there is the usual load of misc fixes and
hardening changes.
|
|
The UBI wear-leveling threshold (CONFIG_MTD_UBI_WL_THRESHOLD) is
currently a compile-time constant shared by all UBI devices. When a
single kernel image must support multiple NAND flashes with different
erase lifetimes, one global threshold cannot suit all devices.
Add a per-device configurable wl_threshold parameter:
- UAPI: add __s32 wl_threshold to struct ubi_attach_req, carved
from the existing padding.
- Module parameter: ubi.mtd gains a new optional token
"wl_threshold":
ubi.mtd=0,0,0,0,0,0,256 ubi.mtd=1,0,0,0,0,0,4096
0 means "use the kernel default", the accepted range is
2-65536.
Signed-off-by: Ran Hongyun <ranhongyun1@huawei.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
|
|
Fix two issues introduced before:
- The parameter format string was missing the enable_fm and
need_resv_pool tokens introduced in an earlier commit.
- The bad-block reservation note was misplaced after Example 5
instead of after Example 3. It was misplaced due to an earlier patch.
Fixes: 83ff59a06663 ("UBI: support ubi_num on mtd.ubi command line")
Signed-off-by: Ran Hongyun <ranhongyun1@huawei.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
|
|
ubi_detach_mtd_dev() obtains a device reference through ubi_get_device()
before checking whether the UBI device is busy. The busy return path drops
ubi->ref_count but leaves the device reference held, so the device object
cannot be released after a later detach.
Drop the device reference before returning -EBUSY.
Fixes: 7e84c961b2eb ("mtd: ubi: introduce pre-removal notification for UBI volumes")
Signed-off-by: Yuho Choi <dbgh9129@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
|
|
ubi_init_attach() rolls back module initialization failures by scanning
ubi_devices[0..i-1], where i is the mtd= parameter index. That assumes
the parameter index matches the UBI device number.
That assumption is not true when mtd= specifies an explicit ubi_num. A
successfully attached device can be stored at a higher ubi_devices[]
slot, and a later failure can miss it during rollback.
Scan the full ubi_devices[] array and detach by the actual array index,
matching the way UBI devices are stored.
Fixes: 83ff59a06663 ("UBI: support ubi_num on mtd.ubi command line")
Signed-off-by: Yuho Choi <dbgh9129@gmail.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
|
|
In ubi_update_fastmap(), when no fresh PEB is available to replace an
old fastmap PEB, the old PEB is reused by erasing it synchronously
via ubi_sync_erase() with the torture flag hardcoded to zero.
However, old_fm->to_be_tortured[] may be non-zero: it is set during
fastmap scanning when ubi_io_read_data()/ubi_io_read_ec_hdr() reports
UBI_IO_BITFLIPS, meaning the PEB showed signs of bit decay and should
be tortured to verify it is still reliable.
When the old PEB is instead returned to the WL sub-system via
ubi_wl_put_fm_peb(), the to_be_tortured flag is honored. The
synchronous reuse path is inconsistent: it silently skips the torture
test for PEBs that exhibited bit-flips.
Fix it by passing &old_fm->to_be_tortured[i] to ubi_sync_erase().
Fixes: dbb7d2a88d2a ("UBI: Add fastmap core")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
|
|
In __erase_worker(), when ubi_sync_erase() returns -EINTR, -ENOMEM,
-EAGAIN, or -EBUSY, the physical eraseblock is rescheduled for erasure
via schedule_erase() with the torture flag hardcoded to 0. This
unconditionally drops the torture request.
If the error occurred before torture_peb() completed (e.g., -ENOMEM
from the ec_hdr allocation in ubi_sync_erase(), errors from
self_check_not_bad()/nor_erase_prepare() in ubi_io_sync_erase(), or
torture_peb() failing mid-test because its internal do_sync_erase()/
ubi_io_read()/ubi_io_write() returned one of the above error codes),
the torture test was never finished. Such a PEB may reach the free
pool without being tortured, defeating the purpose of the torture
test for detecting marginally-bad eraseblocks.
If the error occurred after torture_peb() succeeded (the final
do_sync_erase() in ubi_io_sync_erase() or ubi_io_write_ec_hdr() in
ubi_sync_erase() failed), the torture has already been done and need
not be repeated.
Pass the torture flag by pointer to ubi_sync_erase() and
ubi_io_sync_erase(), and clear it to zero once torture_peb() has
completed successfully. __erase_worker() then forwards the updated
wl_wrk->torture to schedule_erase(): torture already done is dropped,
torture not completed is preserved across the reschedule.
Fixes: 784c145444e7 ("UBI: fix error handling in erase worker")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
|
|
The NAND timing calculation assumes that command and address setup and
hold intervals T1-T4, T7 and T11 each take one controller clock. It also
uses the original A10 delay encodings for tWB, tADL, tWHR and tRHW.
The H6/H616 NDFC defines the setup and hold intervals as two internal
clock cycles and uses different delay encodings. Add the H616 timing
characteristics and select them through the controller capability data so
the clock solver and timing fields match the hardware.
Fixes: 88fd4e4deae8 ("mtd: rawnand: sunxi: Add support for H616 nand controller")
Cc: stable@vger.kernel.org
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The tADL and tWHR timing fields use four encoded delays, but the driver
currently derives their values with a shift. This hides the actual
controller timing characteristics and lets the clock solver select a
32-cycle delay that the fields cannot encode.
Describe the legacy 7, 15, 23 and 31 cycle thresholds explicitly and use
the tables for both clock selection and field lookup. This prepares the
driver for controllers with different encodings.
Fixes: 88fd4e4deae8 ("mtd: rawnand: sunxi: Add support for H616 nand controller")
Cc: stable@vger.kernel.org
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The tWB and tRHW timing field encodings are controller properties, but
they currently live in standalone lookup tables.
Group them in a timing descriptor selected through the controller
capability data. Point every existing controller at the legacy values so
this is a pure preparation change.
Fixes: 88fd4e4deae8 ("mtd: rawnand: sunxi: Add support for H616 nand controller")
Cc: stable@vger.kernel.org
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
'select' does not work on config options in a 'choice', so currently it is
possible to enable MTD_PHYSMAP_IXP4XX without MTD_CFI_BE_BYTE_SWAP.
From a previous conversation, it was decided to remove the select entirely:
https://lore.kernel.org/all/c6268184-4904-49d9-b14f-0e11dce8bdad@app.fastmail.com/
This dead select was found by kconfirm, a static analysis tool for Kconfig.
Signed-off-by: Julian Braha <julianbraha@gmail.com>
Link: $URL [1]"
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Export the BCH geometry parameters via debugfs to aid debugging and
provide the necessary information for legacy kobs-ng tool. The debugfs
directory "gpmi-nand" is created under the root debugfs tree, exposing:
bch_geometry - blob containing the struct bch_geometry fields
raw_mode - flag indicating raw mode status
The implementation is guarded with #ifdef CONFIG_DEBUG_FS to avoid
build failures when debugfs is disabled.
Signed-off-by: Han Xu <han.xu@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
nand_flash_detect_ext_param_page() allocates the length declared by the
ONFI parameter page, then treats the data as a fixed header followed by
variable-length sections. It reads that header and advances over sections
without first proving that the fixed page and each current section fit in
the allocation.
Reject pages shorter than the fixed header, track the remaining variable
area while walking sections, and require the ECC section to contain every
field read from struct onfi_ext_ecc_info. Use device-scoped diagnostics
that identify the malformed ONFI section.
Fixes: 6dcbe0cdd83f ("mtd: get the ECC info from the Extended Parameter Page")
Cc: stable@vger.kernel.org
Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The H616 NAND controller uses a descriptor-based internal MBUS DMA
engine instead of the direct address and count registers used by the
A23/A33 controller. Since the driver does not support these descriptors,
it currently attempts to request an external rxtx DMA channel and falls
back to PIO when none is provided.
Add a single-descriptor backend to the existing ECC page DMA paths.
Allocate the descriptor coherently, constrain data mappings to the
controller's 32-bit address range, program the H6-style data block mask,
and request an interrupt for both command and DMA completion. Keep the
existing external DMA and legacy MBUS DMA paths unchanged, and fall back
to PIO if the descriptor cannot be allocated.
With identical kernels except for this patch, running
flash_speed -d -b 1906 -c 100 /dev/mtd6 on an H616 board with 2 KiB-page
SLC NAND reported:
PIO descriptor DMA
eraseblock write 3365 KiB/s 4192 KiB/s
eraseblock read 6454 KiB/s 16040 KiB/s
page write 3254 KiB/s 4021 KiB/s
page read 6419 KiB/s 15686 KiB/s
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The FM25G01B/FM25G02B datasheets specify a single dummy byte for the
0xEB Quad I/O read-from-cache operation, but the generic
read_cache_variants set uses two dummy bytes for the 1S-4S-4S variant.
The extra dummy byte shifts the data phase and returns corrupted data
with no ECC error, breaking boot on boards using these chips.
Use a dedicated read-from-cache variant set with ndummy=1 for the
1S-4S-4S (0xEB) operation.
FM25G01B datasheet: https://www.fmsh.com/nvm/FM25G01B_ds_eng.pdf
FM25G02B datasheet: https://www.fmsh.com/nvm/FM25G02B_ds_eng.pdf
Fixes: d5a5c9eb2ee9 ("mtd: spinand: fmsh: add support for FM25G{01,02}B")
Cc: stable@vger.kernel.org
Signed-off-by: Aleksandr Mineev <sanderrrs@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Pull mtd fixes from Miquel Raynal:
"Among the most important fixes that we have here, there are:
- the revert of the uclinux map driver which was presumed to
be no longer used but in fact was
- the use of SPI match data to get chip capabilities in the
mchp23k256 driver
- several fixes addressing the newly introduced virt-concat
support
- a missing build dependency on ndfc
as well as the usual load (if not actually bigger than usual) of
uninitialized variables, leaks, double free, and AI fuzzed issues
being fixed"
* tag 'mtd/fixes-for-7.2-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux:
Revert "mtd: maps: remove uclinux map driver"
mtd: onenand: samsung: report DMA completion timeouts
mtd: rawnand: fsl_ifc: return errors for failed page reads
mtd: mchp23k256: use SPI match data for chip caps
mtd: rawnand: lpc32xx_slc: fail DMA transfer on completion timeout
mtd: rawnand: lpc32xx_mlc: fail DMA transfers on timeout
mtd: fix double free and WARN_ON in add_mtd_device() error paths
mtd: virt-concat: free duplicate generated name
mtd: nand: mtk-ecc: stop on ECC idle timeouts
mtd: mtdswap: remove debugfs stats file on teardown
mtd: mtdpart: validate partition bounds in mtd_add_partition()
mtd: mtdpart: fix uninitialized erasesize on MTDPART_OFS_RETAIN error path
mtd: rawnand: ndfc: add CONFIG_OF dependency
mtd: spinand: initialize ret in regular page reads
mtd: virt_concat: fix use-after-free in mtd_virt_concat_destroy()
mtd: rawnand: ingenic: handle ECC clock enable failures
mtd: nand: ecc-mtk: handle ECC clock enable failures
mtd: virt_concat: fix use-after-free in mtd_virt_concat_destroy_joins()
mtd: rawnand: ndfc: fix gcc uninitialized var
|
|
mtd_add_partition() does not reject the special offset value
MTDPART_OFS_RETAIN (-3), which leads to a WARN_ON in
add_mtd_device() when called through the BLKPG ioctl on NAND
devices. The RETAIN value depends on cur_offset being the end of
the previous partition, but in the dynamic partition path
cur_offset equals the offset argument itself, causing undefined
behavior.
Commit 5daa7b21496a ("mtd: prepare partition add and del functions
for ioctl requests") introduced mtd_add_partition() and correctly
rejected MTDPART_OFS_APPEND (-1) and MTDPART_OFS_NXTBLK (-2),
since those special offsets rely on cur_offset tracking the
previous partition's end. However, commit 1a31368bf92e ("mtd: add a flags
for partitions which should just leave smth. after them")
later added MTDPART_OFS_RETAIN (-3) for the static
partition table path without updating mtd_add_partition() to
also reject this value.
With offset=-3 passed via BLKPG, the RETAIN size calculation in
allocate_partition() underflows (parent_size - 0xFFFFFFFFFFFFFFFD
= parent_size + 3). If the underflow result does not appear to
leave enough space, allocate_partition() jumps to out_register via
goto, skipping erasesize initialization. This results in
erasesize=0, which triggers:
WARN_ON((!mtd->erasesize || !master->_erase) &&
!(mtd->flags & MTD_NO_ERASE))
in add_mtd_device(). If the underflow result appears to leave
enough space, a bogus partition size is calculated, but the
"out of reach" sanity check catches the invalid offset and
creates a disabled empty partition (offset=0, size=0) instead
of returning an error.
Fix this by adding MTDPART_OFS_RETAIN to the rejection list in
mtd_add_partition(), consistent with the existing handling of
APPEND and NXTBLK.
Fixes: 1a31368bf92e ("mtd: add a flags for partitions which should just leave smth. after them")
Signed-off-by: zhouminqiang <zhouminqiang2@huawei.com>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Replace the open-coded of_address_to_resource() plus devm_request_mem_region()
and devm_ioremap() sequence with a single devm_platform_ioremap_resource()
call, which folds the resource lookup, region reservation and mapping into
one step and returns an ERR_PTR on failure, checked with IS_ERR() and
propagated via PTR_ERR().
Switch IRQ acquisition from irq_of_parse_and_map() to platform_get_irq(),
which only retrieves the interrupt the OF/platform core has already set up
rather than transferring mapping ownership to the driver. Drop the now
unneeded of_irq.h include.
This is behaviorally equivalent: the driver already reserved the region
with devm_request_mem_region(), so the non-overlapping reg requirement of
devm_platform_ioremap_resource() was already satisfied.
Drop the now-unused regs_paddr / regs_size locals, which previously only
fed the open-coded request/ioremap calls. Keep the linux/of_address.h
include, as of_iomap() is still used elsewhere in the driver.
Built for PowerPC (mpc512x_defconfig + CONFIG_MTD_NAND_MPC5121_NFC) with
LLVM=1; drivers/mtd/nand/raw/mpc5121_nfc.o compiles cleanly.
Assisted-by: opencode:hy3-free
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
mtdoops_notify_add() allocates oops_page_used when the configured MTD
device is registered. mtdoops_notify_remove() detaches from that device
but leaves the bitmap allocated. If the same MTD device is later
registered again, the add path allocates a new bitmap and overwrites the
old pointer, leaking one vmalloc allocation per remove/add cycle.
This is only visible when the backing MTD device can disappear and be
registered again while mtdoops remains loaded, so the usual static MTD
case does not expose it.
Free the bitmap after unregistering the dumper and flushing the pending
workers, then clear the pointer and page count before a later attach can
allocate fresh state. Clearing the pointer also keeps the module exit
path from freeing the same bitmap a second time after a remove event.
Fixes: be95745f0167 ("mtd: mtdoops: keep track of used/unused pages in an array")
Cc: stable@vger.kernel.org
Signed-off-by: Xu Rao <raoxu@uniontech.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
In mtdswap_add_mtd(), debugfs setup failure after successful blktrans
registration can free mbd_dev twice.
add_mtd_blktrans_dev() initializes the blktrans device reference and
publishes the disk. Once that succeeds, del_mtd_blktrans_dev() tears the
disk down and drops the blktrans reference; when that reference reaches
zero, blktrans_dev_release() frees the mtd_blktrans_dev.
The debugfs failure path called del_mtd_blktrans_dev(mbd_dev), then fell
through the common cleanup label and called kfree(mbd_dev) again. Clear
the local pointer after deregistration so the common cleanup can still
release the mtdswap state without freeing the blktrans object twice.
This issue was found by a static analysis checker and confirmed by
manual source review.
Fixes: e8e3edb95ce6 ("mtd: create per-device and module-scope debugfs entries")
Signed-off-by: Ruoyu Wang <ruoyuw560@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The AFS v2 parser uses footer[8] to locate the image information block
inside the current erase block, then uses the image information
region_count to walk entries from a fixed local array. The footer offset
and region count come from flash contents and are not checked against the
erase block or the local image-info array before use.
Reject v2 entries whose image information offset would underflow the
erase block calculation, and reject region counts that cannot fit in the
local image-info array before walking region entries.
Fixes: b7cf5e2830bb ("mtd: afs: add v2 partition parsing")
Cc: stable@vger.kernel.org
Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
Acked-by: Linus Walleij <linusw@kernel.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
intel_dg_mtd_probe() allocates region names before enabling runtime PM
and before calling pm_runtime_resume_and_get().
If kasprintf() fails while building a region name, the error path jumps
to err, which calls pm_runtime_put(). At that point there has not been a
successful pm_runtime_resume_and_get() call to balance, so the runtime PM
usage count can underflow.
Jump to err_norpm from the kasprintf() failure path, as the runtime PM
reference has not been acquired yet.
Fixes: 779c59274d03 ("mtd: intel-dg: Fix accessing regions before setting nregions")
Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
Reviewed-by: Raag Jadav <raag.jadav@intel.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Defining a static array in a header results in each .c file that
includes the header (here: drivers/memory/omap-gpmc.c and
drivers/mtd/nand/raw/omap2.c) to contain a copy of that array when
compiled to an object file.
With sizeof(struct of_device_id[3]) ≥ 588 having omap_nand_ids[] twice
just to do two string comparisons is quite some bloat. So move
omap_nand_ids[] to the nand driver which actually needs that array for
its module meta data and do the compatible check by hand.
bloat-o-meter reports for drivers/memory/omap-gpmc.o (ARCH=arm):
add/remove: 1/2 grow/shrink: 1/0 up/down: 28/-588 (-560)
Function old new delta
gpmc_probe_generic_child 2108 2136 +28
omap_nand_ids 588 - -588
Total: Before=18114, After=17554, chg -3.09%
(drivers/mtd/nand/raw/omap2.o doesn't change).
This allows to drop <linux/mod_devicetable.h> from
include/linux/platform_data/mtd-nand-omap2.h (which is my original
motivation for this change). Note that this header isn't needed in the
two drivers because omap-gpmc.c doesn't use any device id struct and for
the nand driver omap2.c of_device_id is already provided via
<linux/platform_device.h>.
Signed-off-by: Uwe Kleine-König (The Capable Hub) <u.kleine-koenig@baylibre.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Without a full-ID entry, the non-ONFI TC58NVG1S3H falls back to the
generic 0xda extended-ID decoding. It is therefore identified only as a
generic 256 MiB Toshiba NAND. Without a model-specific interface
configuration, the core leaves it in the conservative reset timing mode.
Add its full five-byte ID with the datasheet geometry, OOB size and ECC
requirements. Its interface timings match the mode-4-derived profile
already used by the related TH58NVG2S3HBAI4 and TH58NVG3S0HBAI4, so
share that profile and select it for the new model.
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The Realtek external ECC engine driver has an OF match table wired into
its platform driver, but the table is not exported with
MODULE_DEVICE_TABLE().
When the driver is built as a module, the missing OF module alias
prevents automatic module loading from the compatible string.
Add the missing MODULE_DEVICE_TABLE() entry.
Fixes: 3148d0e5b1c5 ("mtd: nand: realtek-ecc: Add Realtek external ECC engine support")
Cc: stable@vger.kernel.org
Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
MDM9607 has QPIC v1.5 that supports the OP_PAGE_READ_ONFI_READ command, but
is missing the rest of the hardware changes in QPIC v2. Add the new
qcom,mdm9607-nand compatible and set it to use has_onfi_read_op without
also setting qpic_version2.
Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
QPIC v1.5 requires using the OP_PAGE_READ_ONFI_READ command, but is missing
the rest of the hardware changes that are currently covered by the QPIC v2
(qpic_version2) check in the driver. Split that into an extra
has_onfi_read_op feature flag so it can be separately enabled.
No functional change.
Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
Some SoCs (e.g. MDM9607, SDX55) have only a single separately controllable
clock for the NAND controller. The actual clocks in the hardware are
managed by the firmware and turned on all together when needed. In this
case, there is no separate "aon" clock that can be described in the device
tree from the OS point of view.
Make the second "aon" clock optional to avoid an error when it is missing.
For platforms that really need it, the dt-bindings are responsible for
validating that.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
RedBoot FIS partition names are stored in a fixed 16-byte field that is
expected to be NUL-terminated. parse_redboot_partitions() used strlen()
to size the names area and later copied the same field with strcpy(), so
a malformed table entry without a terminator could make both operations
read beyond the descriptor.
Validate each accepted FIS name with strnlen() before adding it to the
partition list.
Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The comment above makes it clear that this is a single element for
legacy handling. Clarify that with struct_size and avoid manual pointer
math.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The uclinux MTD map driver is still actively used on ColdFire (m68k)
platforms. The ColdFire startup code in arch/m68k/coldfire/head.S copies
an attached filesystem blob to _bss_stop, and the uclinux driver picks
it up from there. Most ColdFire defconfigs enable CONFIG_MTD_UCLINUX, as
the driver seems to serve as a replacement for CONFIG_BLK_DEV_INITRD on
these non-MMU targets.
The driver was removed as part of a series cleaning up AMD Élan specific
drivers, but unlike the other three drivers in that series, this driver
seemed to not be as Élan specific (nor unused) as initially thought.
This reverts commit c584b8a7ad01a334a89732f3c5791ba14e58642b.
Reported-by: Greg Ungerer <gregungerer@westnet.com.au>
Closes: https://lore.kernel.org/all/b55ac37f-4523-46d0-8e77-3a17d7235ffe@westnet.com.au/
Acked-by: Arnd Bergmann <arnd@kernel.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
The S5PC110 OneNAND DMA helpers have bounded waits for transfer
completion. The polling helper falls out of its timeout loop and returns
success, and the IRQ helper ignores wait_for_completion_timeout().
Return -ETIMEDOUT when the DMA transfer-done bit or completion does not
arrive before the timeout so callers can treat the buffer transfer as
failed.
Fixes: e23abf4b7743 ("mtd: OneNAND: S5PC110: Implement DMA interrupt method")
Cc: stable@vger.kernel.org
Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
fsl_ifc_run_command() logs controller timeout and other non-OPC
completion states in ctrl->nand_stat. fsl_ifc_read_page() then only
increments the ECC failure counter for non-OPC status and still returns
max_bitflips, which can be zero.
Return -ETIMEDOUT when the command did not complete at all and -EIO for
other non-OPC read completions so the NAND core does not treat a failed
page read as a clean page.
Fixes: 82771882d960 ("NAND Machine support for Integrated Flash Controller")
Cc: stable@vger.kernel.org
Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
Linus Torvalds
Sagnik Sasmal
Runyu Xiao
Kees Cook
Mehmet Fide
Nuno Sá
Menachem Adin
Karl Mehltretter
Han Xu
Pei Xiao
Miquel Raynal
Ran Hongyun
Yuho Choi
Zhihao Cheng
James Hilliard
Julian Braha
Pengpeng Hou
Aleksandr Mineev
zhouminqiang
Rosen Penev
Xu Rao
Ruoyu Wang
Guangshuo Li
Uwe Kleine-König (The Capable Hub)
Stephan Gerhold