We'll soon have multiple unit tests, add a runner that will
discover all of them and execute all tests.
It was opted to discover only files that starts with "test",
as this way unittest discover won't try adding libraries or
other stuff that might not contain unittest classes.
docs: kdoc: use tokenizer to handle comments on structs
Better handle comments inside structs. After those changes,
all unittests now pass:
test_private:
TestPublicPrivate:
test balanced_inner_private: OK
test balanced_non_greddy_private: OK
test balanced_private: OK
test no private: OK
test unbalanced_inner_private: OK
test unbalanced_private: OK
test unbalanced_struct_group_tagged_with_private: OK
test unbalanced_two_struct_group_tagged_first_with_private: OK
test unbalanced_without_end_of_line: OK
Ran 9 tests
This also solves a bug when handling STRUCT_GROUP() with a private
comment on it:
@@ -397134,7 +397134,7 @@ basic V4L2 device-level support.
unsigned int max_len;
unsigned int offset;
struct page_pool_params_slow slow;
- STRUCT_GROUP( struct net_device *netdev;
+ struct net_device *netdev;
unsigned int queue_idx;
unsigned int flags;
};
Handling C code purely using regular expressions doesn't work well.
Add a C tokenizer to help doing it the right way.
The tokenizer was written using as basis the Python re documentation
tokenizer example from:
https://docs.python.org/3/library/re.html#writing-a-tokenizer
Parsing a file like drivers/scsi/isci/host.h, which contains
broken kernel-doc markups makes it create a prototype that contains
unmatched end comments.
That causes, for instance, struct sci_power_control to be shown this
this prototype:
struct sci_power_control {
* it is not. */ bool timer_started;
*/ struct sci_timer timer;
* requesters field. */ u8 phys_waiting;
*/ u8 phys_granted_power;
* mapped into requesters via struct sci_phy.phy_index */ struct isci_phy *requesters[SCI_MAX_PHYS];
};
as comments won't start with "/*" anymore.
Fix the logic to detect such cases, and keep adding the comments
inside it.
unittests: add a testbench to check public/private kdoc comments
Add unit tests to check if the public/private and comments strip
is working properly.
Running it shows that, on several cases, public/private is not
doing what it is expected:
test_private:
TestPublicPrivate:
test balanced_inner_private: OK
test balanced_non_greddy_private: OK
test balanced_private: OK
test no private: OK
test unbalanced_inner_private: FAIL
test unbalanced_private: FAIL
test unbalanced_struct_group_tagged_with_private: FAIL
test unbalanced_two_struct_group_tagged_first_with_private: FAIL
test unbalanced_without_end_of_line: FAIL
While python internal libraries have support for unit tests, its
output is not nice. Add a helper module to improve its output.
I wrote this module last year while testing some scripts I used
internally. The initial skeleton was generated with the help of
LLM tools, but it was higly modified to ensure that it will work
as I would expect.
Haoyang LIU [Mon, 9 Mar 2026 07:47:15 +0000 (15:47 +0800)]
tools/docs/checktransupdate.py: fix all issues reported by pylint
This patch fixes all issues reported by pylint, including:
1. Format issue in logging.
2. Variable name style issue.
Fixes: 63e96ce050e5 ("scripts: fix all issues reported by pylint") Signed-off-by: Haoyang LIU <tttturtleruss@gmail.com> Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20260309074716.10739-1-tttturtleruss@gmail.com>
Daniel Tang [Tue, 10 Mar 2026 01:05:21 +0000 (21:05 -0400)]
docs: path-lookup: fix unrenamed WALK_GET
The symbol WALK_GET does not appears in the codebase as of 0031c06807cfa8aa. It was renamed as of 8f64fb1ccef33107. A previous
documentation update, de9414adafe4, renamed one occurrence in
path-lookup.rst, but forgot to change another occurrence later in the
file.
Fixes: de9414adafe4 ("docs: path-lookup: update WALK_GET, WALK_PUT desc") Signed-off-by: Daniel Tang <danielzgtg.opensource@gmail.com> Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <13011949.O9o76ZdvQC@daniel-desktop3>
Julia Lawall [Tue, 10 Mar 2026 12:14:31 +0000 (13:14 +0100)]
coccinelle: update Coccinelle URL
The LIP6 URL no longer functions.
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr> Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20260310121431.362091-1-Julia.Lawall@inria.fr>
docs: reporting-issues: create a proper appendix explaining specialties
Merge "Why some bugs remain unfixed and some reports are ignored" with
the closing words while rewriting and extending the text.
The result spends fewer words on explaining things that are normal in
FLOSS -- while outlining where the kernel is different and how that
makes bug reporting more complicated than in other FLOSS projects.
docs: verify-bugs-… and quickly-build-…: improve feedback section
Mention sending patches in the section about feedback. This syncs them
with a section a earlier patch added to reporting-issues.rst, which
was based on these sections and improved during review.
docs: reporting-issues: tweak the reference section intro
Fine tuning to the intro of the reference section:
* Call the step-by-step guide what it is.
* Reorder the links to the guides on bug reporting to first mention the
most modern one.
* Many small changes to streamline the text and slightly shorten it.
docs: reporting-issues: mention text is best viewed rendered
Add a comment before the step-by-step guide explaining that the document
is best viewed in the rendered form, as there the internal links will
work that later patches will add.
Daniel Castro [Tue, 17 Mar 2026 14:01:34 +0000 (11:01 -0300)]
docs: pt_BR: translate process/1.Intro.rst
Add Brazilian Portuguese translation of the development process
introduction (Documentation/process/1.Intro.rst), covering the
executive summary, importance of mainline code, and licensing.
Assisted-by: Claude:claude-opus-4-6 Signed-off-by: Daniel Castro <arantescastro@gmail.com> Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20260317140136.29256-1-arantescastro@gmail.com>
Daniel Pereira [Thu, 19 Mar 2026 11:54:12 +0000 (08:54 -0300)]
docs/pt_BR: translation of maintainer-soc-clean-dts.rst
Translate Documentation/process/maintainer-soc-clean-dts.rst into Portuguese.
Signed-off-by: Daniel Pereira <danielmaraboo@gmail.com> Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20260319115416.495020-3-danielmaraboo@gmail.com>
Daniel Pereira [Thu, 19 Mar 2026 11:54:11 +0000 (08:54 -0300)]
docs/pt_BR: translation of maintainer-soc.rst
Translate Documentation/process/maintainer-soc.rst into Portuguese.
This is part of the effort to localize the kernel documentation.
Signed-off-by: Daniel Pereira <danielmaraboo@gmail.com> Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20260319115416.495020-2-danielmaraboo@gmail.com>
Andrea Righi [Sun, 22 Mar 2026 06:35:46 +0000 (07:35 +0100)]
tools/sched_ext: Add compat handling for sub-scheduler ops
Extend SCX_OPS_OPEN() with compatibility handling for ops.sub_attach()
and ops.sub_detach(), allowing scx C schedulers with sub-scheduler
support to run on kernels both with and without its support.
Cheng-Yang Chou [Sun, 22 Mar 2026 13:48:16 +0000 (21:48 +0800)]
sched_ext: Fix build errors and unused label warning in non-cgroup configs
When building with SCHED_CLASS_EXT=y but CGROUPS=n, clang reports errors
for undeclared cgroup_put() and cgroup_get() calls, and a warning for the
unused err_stop_helper label.
EXT_SUB_SCHED is def_bool y depending only on SCHED_CLASS_EXT, but it
fundamentally requires cgroups (cgroup_path, cgroup_get, cgroup_put,
cgroup_id, etc.). Add the missing CGROUPS dependency to EXT_SUB_SCHED in
init/Kconfig.
Guard cgroup_put() and cgroup_get() in the common paths with:
#if defined(CONFIG_EXT_GROUP_SCHED) || defined(CONFIG_EXT_SUB_SCHED)
Guard the err_stop_helper label with #ifdef CONFIG_EXT_SUB_SCHED since
all gotos targeting it are inside that same ifdef block.
Jens Axboe [Sun, 22 Mar 2026 19:37:45 +0000 (13:37 -0600)]
Merge tag 'md-7.1-20260323' of git://git.kernel.org/pub/scm/linux/kernel/git/mdraid/linux into for-7.1/block
Pull MD changes from Yu Kuia:
"Bug Fixes:
- md: suppress spurious superblock update error message for dm-raid
(Chen Cheng)
- md/raid1: fix the comparing region of interval tree (Xiao Ni)
- md/raid10: fix deadlock with check operation and nowait requests
(Josh Hunt)
- md/raid5: skip 2-failure compute when other disk is R5_LOCKED
(FengWei Shih)
- md/md-llbitmap: raise barrier before state machine transition
(Yu Kuai)
- md/md-llbitmap: skip reading rdevs that are not in_sync (Yu Kuai)
Improvements:
- md/raid5: set chunk_sectors to enable full stripe I/O splitting
(Yu Kuai)
Cleanups:
- md: remove unused mddev argument from export_rdev (Chen Cheng)
- md/raid5: remove stale md_raid5_kick_device() declaration
(Chen Cheng)
- md/raid5: move handle_stripe() comment to correct location
(Chen Cheng)"
* tag 'md-7.1-20260323' of git://git.kernel.org/pub/scm/linux/kernel/git/mdraid/linux:
md: remove unused mddev argument from export_rdev
md/raid5: move handle_stripe() comment to correct location
md/raid5: remove stale md_raid5_kick_device() declaration
md/raid1: fix the comparing region of interval tree
md/raid5: skip 2-failure compute when other disk is R5_LOCKED
md/md-llbitmap: raise barrier before state machine transition
md/md-llbitmap: skip reading rdevs that are not in_sync
md/raid5: set chunk_sectors to enable full stripe I/O splitting
md/raid10: fix deadlock with check operation and nowait requests
md: suppress spurious superblock update error message for dm-raid
Add vhca_id_type bit to alias context which allows indicating the
vhca_id_type to be passed at vhca_id_to_be_accessed, which can be either
HW or SW, note that SW_VHCA_ID must be used to allow alias to work
properly after migration.
Signed-off-by: Patrisious Haddad <phaddad@nvidia.com> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Link: https://patch.msgid.link/20260319122211.27384-3-tariqt@nvidia.com Reviewed-by: Joe Damato <joe@dama.to> Signed-off-by: Leon Romanovsky <leon@kernel.org>
Parav Pandit [Thu, 19 Mar 2026 12:22:10 +0000 (14:22 +0200)]
mlx5: Remove redundant iseg base
iseg_base and base_addr both point to BAR0, making iseg_base redundant.
Remove iseg_base and rely on base_addr instead, reducing the size of
struct mlx5_core_dev.
EDAC/versalnet: Fix memory leak in remove and probe error paths
The mcdi object allocated using kzalloc() in the setup_mcdi() is not freed in
the remove path or in probe's error handling path leading to a memory leak.
Fix it by freeing the allocated memory.
Fixes: d5fe2fec6c40d ("EDAC: Add a driver for the AMD Versal NET DDR controller") Signed-off-by: Prasanna Kumar T S M <ptsm@linux.microsoft.com> Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de> Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260322131139.1684716-1-ptsm@linux.microsoft.com
Mikko Perttunen [Tue, 3 Mar 2026 04:32:11 +0000 (13:32 +0900)]
i2c: tegra: Don't mark devices with pins as IRQ safe
I2C devices with associated pinctrl states (DPAUX I2C controllers)
will change pinctrl state during runtime PM. This requires taking
a mutex, so these devices cannot be marked as IRQ safe.
Linus Torvalds [Sun, 22 Mar 2026 18:16:06 +0000 (11:16 -0700)]
Merge tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Pull bpf fixes from Alexei Starovoitov:
- Fix how linked registers track zero extension of subregisters (Daniel
Borkmann)
- Fix unsound scalar fork for OR instructions (Daniel Wade)
- Fix exception exit lock check for subprogs (Ihor Solodrai)
- Fix undefined behavior in interpreter for SDIV/SMOD instructions
(Jenny Guanni Qu)
- Release module's BTF when module is unloaded (Kumar Kartikeya
Dwivedi)
- Fix constant blinding for PROBE_MEM32 instructions (Sachin Kumar)
- Reset register ID for END instructions to prevent incorrect value
tracking (Yazhou Tang)
* tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
selftests/bpf: Add a test cases for sync_linked_regs regarding zext propagation
bpf: Fix sync_linked_regs regarding BPF_ADD_CONST32 zext propagation
selftests/bpf: Add tests for maybe_fork_scalars() OR vs AND handling
bpf: Fix unsound scalar forking in maybe_fork_scalars() for BPF_OR
selftests/bpf: Add tests for sdiv32/smod32 with INT_MIN dividend
bpf: Fix undefined behavior in interpreter sdiv/smod for INT_MIN
selftests/bpf: Add tests for bpf_throw lock leak from subprogs
bpf: Fix exception exit lock checking for subprogs
bpf: Release module BTF IDR before module unload
selftests/bpf: Fix pkg-config call on static builds
bpf: Fix constant blinding for PROBE_MEM32 stores
selftests/bpf: Add test for BPF_END register ID reset
bpf: Reset register ID for BPF_END value tracking
Xiao Ni [Thu, 5 Mar 2026 01:18:33 +0000 (09:18 +0800)]
md/raid1: fix the comparing region of interval tree
Interval tree uses [start, end] as a region which stores in the tree.
In raid1, it uses the wrong end value. For example:
bio(A,B) is too big and needs to be split to bio1(A,C-1), bio2(C,B).
The region of bio1 is [A,C] and the region of bio2 is [C,B]. So bio1 and
bio2 overlap which is not right.
Fix this problem by using right end value of the region.
Linus Torvalds [Sun, 22 Mar 2026 18:10:31 +0000 (11:10 -0700)]
Merge tag 'trace-v7.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull tracing fixes from Steven Rostedt:
- Revert "tracing: Remove pid in task_rename tracing output"
A change was made to remove the pid field from the task_rename event
because it was thought that it was always done for the current task
and recording the pid would be redundant. This turned out to be
incorrect and there are a few corner case where this is not true and
caused some regressions in tooling.
- Fix the reading from user space for migration
The reading of user space uses a seq lock type of logic where it uses
a per-cpu temporary buffer and disables migration, then enables
preemption, does the copy from user space, disables preemption,
enables migration and checks if there was any schedule switches while
preemption was enabled. If there was a context switch, then it is
considered that the per-cpu buffer could be corrupted and it tries
again. There's a protection check that tests if it takes a hundred
tries, it issues a warning and exits out to prevent a live lock.
This was triggered because the task was selected by the load balancer
to be migrated to another CPU, every time preemption is enabled the
migration task would schedule in try to migrate the task but can't
because migration is disabled and let it run again. This caused the
scheduler to schedule out the task every time it enabled preemption
and made the loop never exit (until the 100 iteration test
triggered).
Fix this by enabling and disabling preemption and keeping migration
enabled if the reading from user space needs to be done again. This
will let the migration thread migrate the task and the copy from user
space will likely pass on the next iteration.
- Fix trace_marker copy option freeing
The "copy_trace_marker" option allows a tracing instance to get a
copy of a write to the trace_marker file of the top level instance.
This is managed by a link list protected by RCU. When an instance is
removed, a check is made if the option is set, and if so
synchronized_rcu() is called.
The problem is that an iteration is made to reset all the flags to
what they were when the instance was created (to perform clean ups)
was done before the check of the copy_trace_marker option and that
option was cleared, so the synchronize_rcu() was never called.
Move the clearing of all the flags after the check of
copy_trace_marker to do synchronize_rcu() so that the option is still
set if it was before and the synchronization is performed.
- Fix entries setting when validating the persistent ring buffer
When validating the persistent ring buffer on boot up, the number of
events per sub-buffer is added to the sub-buffer meta page. The
validator was updating cpu_buffer->head_page (the first sub-buffer of
the per-cpu buffer) and not the "head_page" variable that was
iterating the sub-buffers. This was causing the first sub-buffer to
be assigned the entries for each sub-buffer and not the sub-buffer
that was supposed to be updated.
- Use "hash" value to update the direct callers
When updating the ftrace direct callers, it assigned a temporary
callback to all the callback functions of the ftrace ops and not just
the functions represented by the passed in hash. This causes an
unnecessary slow down of the functions of the ftrace_ops that is not
being modified. Only update the functions that are going to be
modified to call the ftrace loop function so that the update can be
made on those functions.
* tag 'trace-v7.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
ftrace: Use hash argument for tmp_ops in update_ftrace_direct_mod
ring-buffer: Fix to update per-subbuf entries of persistent ring buffer
tracing: Fix trace_marker copy link list updates
tracing: Fix failure to read user space from system call trace events
tracing: Revert "tracing: Remove pid in task_rename tracing output"
Linus Torvalds [Sun, 22 Mar 2026 18:05:34 +0000 (11:05 -0700)]
Merge tag 'i2c-for-7.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux
Pull i2c fixes from Wolfram Sang:
- fix broken I2C communication on Armada 3700 with recovery
- fix device_node reference leak in probe (fsi)
- fix NULL-deref when serial string is missing (cp2615)
* tag 'i2c-for-7.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
i2c: pxa: defer reset on Armada 3700 when recovery is used
i2c: fsi: Fix a potential leak in fsi_i2c_probe()
i2c: cp2615: fix serial string NULL-deref at probe
Linus Torvalds [Sun, 22 Mar 2026 17:54:12 +0000 (10:54 -0700)]
Merge tag 'x86-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar:
- Improve Qemu MCE-injection behavior by only using AMD SMCA MSRs if
the feature bit is set
- Fix the relative path of gettimeofday.c inclusion in vclock_gettime.c
- Fix a boot crash on UV clusters when a socket is marked as
'deconfigured' which are mapped to the SOCK_EMPTY node ID by
the UV firmware, while Linux APIs expect NUMA_NO_NODE.
The difference being (0xffff [unsigned short ~0]) vs [int -1]
* tag 'x86-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/platform/uv: Handle deconfigured sockets
x86/entry/vdso: Fix path of included gettimeofday.c
x86/mce/amd: Check SMCA feature bit before accessing SMCA MSRs
Linus Torvalds [Sun, 22 Mar 2026 17:31:51 +0000 (10:31 -0700)]
Merge tag 'perf-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fixes from Ingo Molnar:
- Fix a PMU driver crash on AMD EPYC systems, caused by
a race condition in x86_pmu_enable()
- Fix a possible counter-initialization bug in x86_pmu_enable()
- Fix a counter inheritance bug in inherit_event() and
__perf_event_read()
- Fix an Intel PMU driver branch constraints handling bug
found by UBSAN
- Fix the Intel PMU driver's new Off-Module Response (OMR)
support code for Diamond Rapids / Nova lake, to fix a snoop
information parsing bug
* tag 'perf-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel: Fix OMR snoop information parsing issues
perf/x86/intel: Add missing branch counters constraint apply
perf: Make sure to use pmu_ctx->pmu for groups
x86/perf: Make sure to program the counter value for stopped events on migration
perf/x86: Move event pointer setup earlier in x86_pmu_enable()
Linus Torvalds [Sun, 22 Mar 2026 17:17:50 +0000 (10:17 -0700)]
Merge tag 'objtool-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull objtool fixes from Ingo Molnar:
"Fix three more livepatching related build environment bugs, and a
false positive warning with Clang jump tables"
* tag 'objtool-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
objtool: Fix Clang jump table detection
livepatch/klp-build: Fix inconsistent kernel version
objtool/klp: fix mkstemp() failure with long paths
objtool/klp: fix data alignment in __clone_symbol()
Linus Torvalds [Sun, 22 Mar 2026 16:57:20 +0000 (09:57 -0700)]
Merge tag 'locking-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull locking fix from Ingo Molnar:
"Fix a sparse build error regression in <linux/local_lock_internal.h>
caused by the locking context-analysis changes"
* tag 'locking-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
include/linux/local_lock_internal.h: Make this header file again compatible with sparse
Linus Torvalds [Sun, 22 Mar 2026 16:55:58 +0000 (09:55 -0700)]
Merge tag 'irq-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq fix from Ingo Molnar:
"Fix a mailbox channel leak in the riscv-rpmi-sysmsi irqchip driver"
* tag 'irq-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/riscv-rpmi-sysmsi: Fix mailbox channel leak in rpmi_sysmsi_probe()
drm/mediatek: dsi: Store driver data before invoking mipi_dsi_host_register
The call to mipi_dsi_host_register triggers a callback to mtk_dsi_bind,
which uses dev_get_drvdata to retrieve the mtk_dsi struct, so this
structure needs to be stored inside the driver data before invoking it.
As drvdata is currently uninitialized it leads to a crash when
registering the DSI DRM encoder right after acquiring
the mode_config.idr_mutex, blocking all subsequent DRM operations.
Fixes the following crash during mediatek-drm probe (tested on Xiaomi
Smart Clock x04g):
Fixes: e4732b590a77 ("drm/mediatek: dsi: Register DSI host after acquiring clocks and PHY") Signed-off-by: Luca Leonardo Scorcia <l.scorcia@gmail.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Reviewed-by: CK Hu <ck.hu@mediatek.com> Link: https://patchwork.kernel.org/project/dri-devel/patch/20260225094047.76780-1-l.scorcia@gmail.com/ Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org>
drm/mediatek: Remove all conflicting aperture devices during probe
If a device has a framebuffer available it might be already used as
display by simple-framebuffer or simpledrm when mediatek-drm is probed.
This is actually helpful when porting to a new device as
framebuffers are simple to setup in device trees and fbcon can be
used to monitor the kernel boot process.
When drm-mediatek loads a new fb device is initialized, however
fbcon remains attached to the initial framebuffer which is no longer
connected to the actual display - the early fb is never removed.
We can gracefully transition from framebuffer handling to drm-managed
display by calling aperture_remove_all_conflicting_devices before
registering mediatek-drm. This takes care of unloading other fb
devices/drivers and disconnects fbcon which then automatically
reconnects to mediatekdrmfb as soon as it's available.
The function is invoked just before drm_dev_register() to kick out
the existing framebuffer as late as possible to reduce the time the
screen is unresponsive.
Oleksij Rempel [Tue, 10 Feb 2026 13:51:10 +0000 (14:51 +0100)]
iio: dac: ds4424: add Rfs-based scale and per-variant limits
Parse optional maxim,rfs-ohms values to derive the per-channel output
current scale (mA per step) for the IIO current ABI.
Behavior changes:
- If maxim,rfs-ohms is present, IIO_CHAN_INFO_SCALE becomes available
and reports mA/step derived from Rfs.
- If maxim,rfs-ohms is missing, SCALE is not exposed to keep older DTs
working without requiring updates.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The Maxim DS4422/DS4424 and DS4402/DS4404 current DACs determine their
full-scale output current via external resistors (Rfs) connected to the
FSx pins. Without knowing these values, the full-scale range of the
hardware is undefined.
Add the 'maxim,rfs-ohms' property to describe these physical components.
This property is required to provide a complete description of the
hardware configuration.
Oleksij Rempel [Tue, 10 Feb 2026 13:51:07 +0000 (14:51 +0100)]
iio: dac: ds4424: support per-variant output range limits
The DS4402/DS4404 variants operate with a 5-bit resolution (31 steps),
whereas the DS4422/DS4424 support 7-bit (127 steps).
Previously, the driver enforced a hardcoded 7-bit mask (DS4424_DAC_MASK)
for all variants. This allowed users to write values exceeding the 5-bit
range to DS4402/DS4404 devices, resulting in silent truncation or
undefined behavior.
Add a `result_mask` field to the chip_info structure to define the valid
data range for each variant. Use this mask to:
1. Correctly mask register values in read_raw().
2. Return -EINVAL in write_raw() if the input value exceeds the
variant's capabilities.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Oleksij Rempel [Tue, 10 Feb 2026 13:51:04 +0000 (14:51 +0100)]
iio: dac: ds4424: use fsleep() instead of usleep_range()
The DS4422/DS4424 and DS4402/DS4404 datasheets do not specify a minimum
delay between power-up (POR) and the availability of the I2C interface.
The driver previously used `usleep_range(1000, 1200)` to enforce a ~1ms
delay. Replace this with `fsleep(1000)` to allow the kernel to select
the most efficient sleep mechanism while retaining the existing
conservative delay to ensure device readiness.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Gary Bisson [Tue, 20 Jan 2026 11:36:59 +0000 (12:36 +0100)]
drm/mediatek: mtk_dsi: enable hs clock during pre-enable
Some bridges, such as the TI SN65DSI83, require the HS clock to be
running in order to lock its PLL during its own pre-enable function.
Without this change, the bridge gives the following error:
sn65dsi83 14-002c: failed to lock PLL, ret=-110
sn65dsi83 14-002c: Unexpected link status 0x01
sn65dsi83 14-002c: reset the pipe
Move the necessary functions from enable to pre-enable.
Signed-off-by: Gary Bisson <bisson.gary@gmail.com> Reviewed-by: CK Hu <ck.hu@mediatek.com> Tested-by: Chen-Yu Tsai <wenst@chromium.org> # Chromebooks Tested-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Link: https://patchwork.kernel.org/project/dri-devel/patch/20260120-mtkdsi-v1-1-b0f4094f3ac3@gmail.com/ Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org>
Oleksij Rempel [Tue, 10 Feb 2026 13:51:03 +0000 (14:51 +0100)]
iio: dac: ds4424: use device match data for chip info
Refactor the driver to use device match data instead of checking ID enums
in a switch statement.
Define a `ds4424_chip_info` structure to hold variant-specific attributes
(currently just the channel count) and attach it directly to the I2C and
OF device ID tables.
This simplifies the probe function and makes it easier to add support for
new variants like DS4402/DS4404.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Oleksij Rempel [Tue, 10 Feb 2026 13:51:02 +0000 (14:51 +0100)]
iio: dac: ds4424: rename iio_info struct to avoid ambiguity
Rename the static `ds4424_info` structure to `ds4424_iio_info`.
The previous name was generic and could be confused with chip-specific
data structures (like the upcoming `ds4424_chip_info`). The new name
explicitly indicates that this structure holds the IIO framework
callbacks.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Oleksij Rempel [Tue, 10 Feb 2026 13:50:59 +0000 (14:50 +0100)]
iio: dac: ds4424: refactor raw access to use bitwise operations
Refactor the raw access logic to use standard GENMASK() and BIT()
macros. Use abs() for magnitude calculation to simplify the logic and
make the data flow clearer.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Felix Gu [Sun, 22 Feb 2026 09:45:39 +0000 (17:45 +0800)]
iio: adc: nxp-sar-adc: Fix DMA channel leak in trigger mode
The DMA channel was requested in nxp_sar_adc_buffer_postenable() but
was only released in nxp_sar_adc_buffer_software_do_predisable().
This caused a DMA channel resource leak when operating in trigger mode.
Fix this by moving dma_request_chan() from
nxp_sar_adc_buffer_postenable() into
nxp_sar_adc_buffer_software_do_postenable(), ensuring the DMA channel
is only requested in software mode.
Fixes: 4434072a893e ("iio: adc: Add the NXP SAR ADC support for the s32g2/3 platforms") Signed-off-by: Felix Gu <ustc.gu@gmail.com> Cc: <Stable@vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
arm64: dts: ti: am62-verdin: Enable pullup for eMMC data pins
Verdin AM62 board does not have external pullups on eMMC DAT1-DAT7 pins.
Enable internal pullups on DAT1-DAT7 considering:
- without a host-side pullup, these lines rely solely on the eMMC
device's internal pullup (R_int, 10kohm-150kohm per JEDEC), which may
exceed the recommended 50kohm max for 1.8V VCCQ
- JEDEC JESD84-B51 Table 200 requires host-side pullups (R_DAT,
10kohm-100kohm) on all data lines to prevent bus floating
Randy Dunlap [Thu, 12 Mar 2026 05:13:45 +0000 (22:13 -0700)]
iio: st_sensors: correct kernel-doc issues
Use the proper kernel-doc format and struct member names to avoid
kernel-doc warnings:
Warning: include/linux/iio/common/st_sensors.h:184 struct member 'int1'
not described in 'st_sensor_data_ready_irq'
Warning: ../include/linux/iio/common/st_sensors.h:184 struct member 'int2'
not described in 'st_sensor_data_ready_irq'
Warning: ../include/linux/iio/common/st_sensors.h:184 struct member
'stat_drdy' not described in 'st_sensor_data_ready_irq'
Warning: ../include/linux/iio/common/st_sensors.h:184 struct member 'ig1'
not described in 'st_sensor_data_ready_irq'
Warning: ../include/linux/iio/common/st_sensors.h:219 struct member
'num_ch' not described in 'st_sensor_settings'
Warning: ../include/linux/iio/common/st_sensors.h:263 struct member
'num_data_channels' not described in 'st_sensor_data'
Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Valek Andrej [Fri, 13 Mar 2026 09:24:13 +0000 (10:24 +0100)]
iio: accel: fix ADXL355 temperature signature value
Temperature was wrongly represented as 12-bit signed, confirmed by checking
the datasheet. Even if the temperature is negative, the value in the
register stays unsigned.
Fixes: 12ed27863ea3 iio: accel: Add driver support for ADXL355 Signed-off-by: Valek Andrej <andrej.v@skyrain.eu> Cc: <Stable@vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Thomas Weißschuh [Thu, 19 Mar 2026 16:20:17 +0000 (17:20 +0100)]
tools/nolibc: rename sys_foo() functions to _sys_foo()
The sys_foo() naming scheme used by the syscall wrappers may collide
with application symbols. Especially as 'sys_' is an obvious naming
scheme an application may choose for its own custom systemcall wrappers.
Avoid these conflicts by using an leading underscore which moves the
names into the implementation's namespace. This naming scheme was chosen
over a '__nolibc_' prefix, as these functions are not an implementation
detail but a documented interface meant to be used by applications.
While this may break some existing users, adapting them should be
straightforward. Given that nolibc is most-likely vendored, no
unexpected breakage should happen. No in-tree users are affected.
These conflicts happen when compiling some of the kernel selftests
with nolibc.
Thomas Weißschuh [Wed, 18 Mar 2026 16:12:42 +0000 (17:12 +0100)]
tools/nolibc: avoid -Wundef warning for __STDC_VERSION__
With -std=c89 the macro __STDC_VERSION__ is not defined.
While undefined identifiers in '#if' directives are assumed to be '0',
with -Wundef a warning is emitted.
Avoid the warning by explicitly falling back to '0' if __STDC_VERSION__
is not provided by the preprocessor.
Thomas Weißschuh [Wed, 18 Mar 2026 15:50:16 +0000 (16:50 +0100)]
tools/nolibc: add support for program_invocation_{,short_}name
Add support for the GNU extensions 'program_invocation_name' and
'program_invocation_short_name'. These are useful to print error
messages, which by convention include the program name.
As these are global variables which take up memory even if not used,
similar to 'errno', gate them behind NOLIBC_IGNORE_ERRNO.
Juergen Gross [Mon, 19 Jan 2026 18:26:29 +0000 (19:26 +0100)]
block/floppy: Don't use REALLY_SLOW_IO for delays
Instead of defining REALLY_SLOW_IO before including io.h, add the
required additional calls of native_io_delay() to the related functions
in arch/x86/include/asm/floppy.h.
Drop REALLY_SLOW_IO now too as it has no users.
[ bp: Merge the REALLY_SLOW_IO removal into this patch. ]
Juergen Gross [Mon, 19 Jan 2026 18:26:28 +0000 (19:26 +0100)]
x86/paravirt: Replace io_delay() hook with a bool
The io_delay() paravirt hook is in no way performance critical and all users
setting it to a different function than native_io_delay() are using an empty
function as replacement.
Allow replacing the hook with a bool indicating whether native_io_delay()
should be called.
Juergen Gross [Mon, 19 Jan 2026 18:26:27 +0000 (19:26 +0100)]
x86/irqflags: Preemptively move include paravirt.h directive where it belongs
Commit
22cc5ca5de52 ("x86/paravirt: Move halt paravirt calls under CONFIG_PARAVIRT")
moved some paravirt hooks from the CONFIG_PARAVIRT_XXL umbrella to
CONFIG_PARAVIRT, but missed to move the associated "#include <asm/paravirt.h>"
in irqflags.h from CONFIG_PARAVIRT_XXL to CONFIG_PARAVIRT.
This hasn't resulted in build failures yet, as all use cases of irqflags.h had
paravirt.h included via other header files, even without CONFIG_PARAVIRT_XXL
being set.
In order to allow changing those other header files, e.g. by no longer
including paravirt.h, fix irqflags.h by moving inclusion of paravirt.h under
the CONFIG_PARAVIRT umbrella.
Cosmin Tanislav [Fri, 30 Jan 2026 12:23:53 +0000 (14:23 +0200)]
counter: rz-mtu3-cnt: do not use struct rz_mtu3_channel's dev member
The counter driver can use HW channels 1 and 2, while the PWM driver can
use HW channels 0, 1, 2, 3, 4, 6, 7.
The dev member is assigned both by the counter driver and the PWM driver
for channels 1 and 2, to their own struct device instance, overwriting
the previous value.
The sub-drivers race to assign their own struct device pointer to the
same struct rz_mtu3_channel's dev member.
The dev member of struct rz_mtu3_channel is used by the counter
sub-driver for runtime PM.
Depending on the probe order of the counter and PWM sub-drivers, the
dev member may point to the wrong struct device instance, causing the
counter sub-driver to do runtime PM actions on the wrong device.
To fix this, use the parent pointer of the counter, which is assigned
during probe to the correct struct device, not the struct device pointer
inside the shared struct rz_mtu3_channel.
At the same time, hardware registers end up being accessed with clocks
off in rz_mtu3_terminate_counter() to disable an already disabled
channel.
If user writes 1 to the sysfs enable file multiple times, runtime PM
usage count will be incremented each time, requiring the same number of
0 writes to get it back to 0.
If user writes 0 to the sysfs enable file while PWM is in progress, PWM
is stopped without counter being the owner of the underlying MTU3
channel.
Check against the cached count_is_enabled value and exit if the user
is trying to set the same enable value.
Song Liu [Sun, 22 Mar 2026 03:30:45 +0000 (20:30 -0700)]
workqueue: Fix false positive stall reports
On weakly ordered architectures (e.g., arm64), the lockless check in
wq_watchdog_timer_fn() can observe a reordering between the worklist
insertion and the last_progress_ts update. Specifically, the watchdog
can see a non-empty worklist (from a list_add) while reading a stale
last_progress_ts value, causing a false positive stall report.
This was confirmed by reading pool->last_progress_ts again after holding
pool->lock in wq_watchdog_timer_fn():
To avoid slowing down the hot path (queue_work, etc.), recheck
last_progress_ts with pool->lock held. This will eliminate the false
positive with minimal overhead.
Remove two extra empty lines in wq_watchdog_timer_fn() as we are on it.
Fixes: 82607adcf9cd ("workqueue: implement lockup detector") Cc: stable@vger.kernel.org # v4.5+ Assisted-by: claude-code:claude-opus-4-6 Signed-off-by: Song Liu <song@kernel.org> Signed-off-by: Tejun Heo <tj@kernel.org>
Kit Dallege [Sun, 15 Mar 2026 14:57:22 +0000 (15:57 +0100)]
crypto: add missing kernel-doc for anonymous union members
Document the anonymous SKCIPHER_ALG_COMMON and COMP_ALG_COMMON struct
members in skcipher_alg, scomp_alg, and acomp_alg, following the
existing pattern used by HASH_ALG_COMMON in shash_alg.
This fixes the following kernel-doc warnings:
include/crypto/skcipher.h:166: struct member 'SKCIPHER_ALG_COMMON' not described in 'skcipher_alg'
include/crypto/internal/scompress.h:39: struct member 'COMP_ALG_COMMON' not described in 'scomp_alg'
include/crypto/internal/acompress.h:55: struct member 'COMP_ALG_COMMON' not described in 'acomp_alg'
Signed-off-by: Kit Dallege <xaum.io@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Eric Biggers [Sat, 14 Mar 2026 21:37:20 +0000 (14:37 -0700)]
crypto: simd - Remove unused skcipher support
Remove the skcipher algorithm support from crypto/simd.c. It is no
longer used, and it is unlikely to gain any new user in the future,
given the performance issues with this code.
Signed-off-by: Eric Biggers <ebiggers@kernel.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Thorsten Blum [Sat, 14 Mar 2026 19:36:29 +0000 (20:36 +0100)]
crypto: atmel-sha204a - Fix potential UAF and memory leak in remove path
Unregister the hwrng to prevent new ->read() calls and flush the Atmel
I2C workqueue before teardown to prevent a potential UAF if a queued
callback runs while the device is being removed.
Drop the early return to ensure sysfs entries are removed and
->hwrng.priv is freed, preventing a memory leak.
Fixes: da001fb651b0 ("crypto: atmel-i2c - add support for SHA204A random number generator") Cc: stable@vger.kernel.org Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Daniel Jordan [Fri, 13 Mar 2026 15:24:33 +0000 (11:24 -0400)]
padata: Put CPU offline callback in ONLINE section to allow failure
syzbot reported the following warning:
DEAD callback error for CPU1
WARNING: kernel/cpu.c:1463 at _cpu_down+0x759/0x1020 kernel/cpu.c:1463, CPU#0: syz.0.1960/14614
at commit 4ae12d8bd9a8 ("Merge tag 'kbuild-fixes-7.0-2' of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux")
which tglx traced to padata_cpu_dead() given it's the only
sub-CPUHP_TEARDOWN_CPU callback that returns an error.
Failure isn't allowed in hotplug states before CPUHP_TEARDOWN_CPU
so move the CPU offline callback to the ONLINE section where failure is
possible.
Fixes: 894c9ef9780c ("padata: validate cpumask without removed CPU during offline") Reported-by: syzbot+123e1b70473ce213f3af@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/69af0a05.050a0220.310d8.002f.GAE@google.com/ Debugged-by: Thomas Gleixner <tglx@kernel.org> Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Zongyu Wu [Fri, 13 Mar 2026 09:40:39 +0000 (17:40 +0800)]
crypto: hisilicon - add device load query functionality to debugfs
The accelerator device supports usage statistics. This patch enables
obtaining the accelerator's usage through the "dev_usage" file.
The returned number expressed as a percentage as a percentage.
Signed-off-by: Zongyu Wu <wuzongyu1@huawei.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Randy Dunlap [Thu, 12 Mar 2026 05:13:23 +0000 (22:13 -0700)]
hwrng: core - avoid kernel-doc warnings
Mark internal fields as "private:" so that kernel-doc comments
are not needed for them, eliminating kernel-doc warnings:
Warning: include/linux/hw_random.h:54 struct member 'list' not described
in 'hwrng'
Warning: include/linux/hw_random.h:54 struct member 'ref' not described
in 'hwrng'
Warning: include/linux/hw_random.h:54 struct member 'cleanup_work' not
described in 'hwrng'
Warning: include/linux/hw_random.h:54 struct member 'cleanup_done' not
described in 'hwrng'
Warning: include/linux/hw_random.h:54 struct member 'dying' not described
in 'hwrng'
Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Thorsten Blum [Wed, 11 Mar 2026 15:56:49 +0000 (16:56 +0100)]
crypto: nx - fix context leak in nx842_crypto_free_ctx
Since the scomp conversion, nx842_crypto_alloc_ctx() allocates the
context separately, but nx842_crypto_free_ctx() never releases it. Add
the missing kfree(ctx) to nx842_crypto_free_ctx(), and reuse
nx842_crypto_free_ctx() in the allocation error path.
Thorsten Blum [Wed, 11 Mar 2026 15:56:47 +0000 (16:56 +0100)]
crypto: nx - fix bounce buffer leaks in nx842_crypto_{alloc,free}_ctx
The bounce buffers are allocated with __get_free_pages() using
BOUNCE_BUFFER_ORDER (order 2 = 4 pages), but both the allocation error
path and nx842_crypto_free_ctx() release the buffers with free_page().
Use free_pages() with the matching order instead.
Thorsten Blum [Wed, 11 Mar 2026 11:39:28 +0000 (12:39 +0100)]
crypto: atmel-aes - guard unregister on error in atmel_aes_register_algs
Ensure the device supports XTS and GCM with 'has_xts' and 'has_gcm'
before unregistering algorithms when XTS or authenc registration fails,
which would trigger a WARN in crypto_unregister_alg().
Currently, with the capabilities defined in atmel_aes_get_cap(), this
bug cannot happen because all devices that support XTS and authenc also
support GCM, but the error handling should still be correct regardless
of hardware capabilities.
Fixes: d52db5188a87 ("crypto: atmel-aes - add support to the XTS mode") Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
George Abraham P [Wed, 11 Mar 2026 08:22:45 +0000 (13:52 +0530)]
crypto: qat - add wireless mode support for QAT GEN6
Add wireless mode support for QAT GEN6 devices.
When the WCP_WAT fuse bit is clear, the device operates in wireless
cipher mode (wcy_mode). In this mode all accelerator engines load the
wireless firmware and service configuration via 'cfg_services' sysfs
attribute is restricted to 'sym' only.
The get_accel_cap() function is extended to report wireless-specific
capabilities (ZUC, ZUC-256, 5G, extended algorithm chaining) gated by
their respective slice-disable fuse bits. The set_ssm_wdtimer() function
is updated to configure WCP (wireless cipher) and WAT (wireless
authentication) watchdog timers. The adf_gen6_cfg_dev_init() function is
updated to use adf_6xxx_is_wcy() to enforce sym-only service selection
for WCY devices during initialization.
Co-developed-by: Aviraj Cj <aviraj.cj@intel.com> Signed-off-by: Aviraj Cj <aviraj.cj@intel.com> Signed-off-by: George Abraham P <george.abraham.p@intel.com> Reviewed-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Thorsten Blum [Wed, 11 Mar 2026 02:07:35 +0000 (03:07 +0100)]
crypto: atmel-aes - Fix 3-page memory leak in atmel_aes_buff_cleanup
atmel_aes_buff_init() allocates 4 pages using __get_free_pages() with
ATMEL_AES_BUFFER_ORDER, but atmel_aes_buff_cleanup() frees only the
first page using free_page(), leaking the remaining 3 pages. Use
free_pages() with ATMEL_AES_BUFFER_ORDER to fix the memory leak.
Fixes: bbe628ed897d ("crypto: atmel-aes - improve performances of data transfer") Cc: stable@vger.kernel.org Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
FengWei Shih [Thu, 19 Mar 2026 05:33:51 +0000 (13:33 +0800)]
md/raid5: skip 2-failure compute when other disk is R5_LOCKED
When skip_copy is enabled on a doubly-degraded RAID6, a device that is
being written to will be in R5_LOCKED state with R5_UPTODATE cleared.
If a new read triggers fetch_block() while the write is still in
flight, the 2-failure compute path may select this locked device as a
compute target because it is not R5_UPTODATE.
Because skip_copy makes the device page point directly to the bio page,
reconstructing data into it might be risky. Also, since the compute
marks the device R5_UPTODATE, it triggers WARN_ON in ops_run_io()
which checks that R5_SkipCopy and R5_UPTODATE are not both set.
This can be reproduced by running small-range concurrent read/write on
a doubly-degraded RAID6 with skip_copy enabled, for example:
slot_free() basically completely resets the slots by clearing all of
its flags and attributes. While zram_writeback_complete() restores
some of flags back (those that are necessary for async read
decompression) we still lose a lot of slot's metadata. For example,
slot's ac-time, or ZRAM_INCOMPRESSIBLE.
More importantly, restoring flags/attrs requires extra attention as
some of the flags are directly affecting zram device stats. And the
original code did not pay that attention. Namely ZRAM_HUGE slots
handling in zram_writeback_complete(). The call to slot_free() would
decrement ->huge_pages, however when zram_writeback_complete() restored
the slot's ZRAM_HUGE flag, it would not get reflected in an incremented
->huge_pages. So when the slot would finally get freed, slot_free()
would decrement ->huge_pages again, leading to underflow.
Fix this by open-coding the required memory free and stats updates in
zram_writeback_complete(), rather than calling the destructive
slot_free(). Since we now preserve the ZRAM_HUGE flag on written-back
slots (for the deferred decompression path), we also update slot_free()
to skip decrementing ->huge_pages if ZRAM_WB is set.
SeongJae Park [Thu, 19 Mar 2026 14:52:17 +0000 (07:52 -0700)]
mm/damon/core: avoid use of half-online-committed context
One major usage of damon_call() is online DAMON parameters update. It is
done by calling damon_commit_ctx() inside the damon_call() callback
function. damon_commit_ctx() can fail for two reasons: 1) invalid
parameters and 2) internal memory allocation failures. In case of
failures, the damon_ctx that attempted to be updated (commit destination)
can be partially updated (or, corrupted from a perspective), and therefore
shouldn't be used anymore. The function only ensures the damon_ctx object
can safely deallocated using damon_destroy_ctx().
The API callers are, however, calling damon_commit_ctx() only after
asserting the parameters are valid, to avoid damon_commit_ctx() fails due
to invalid input parameters. But it can still theoretically fail if the
internal memory allocation fails. In the case, DAMON may run with the
partially updated damon_ctx. This can result in unexpected behaviors
including even NULL pointer dereference in case of damos_commit_dests()
failure [1]. Such allocation failure is arguably too small to fail, so
the real world impact would be rare. But, given the bad consequence, this
needs to be fixed.
Avoid such partially-committed (maybe-corrupted) damon_ctx use by saving
the damon_commit_ctx() failure on the damon_ctx object. For this,
introduce damon_ctx->maybe_corrupted field. damon_commit_ctx() sets it
when it is failed. kdamond_call() checks if the field is set after each
damon_call_control->fn() is executed. If it is set, ignore remaining
callback requests and return. All kdamond_call() callers including
kdamond_fn() also check the maybe_corrupted field right after
kdamond_call() invocations. If the field is set, break the kdamond_fn()
main loop so that DAMON sill doesn't use the context that might be
corrupted.
projects / thirdparty / linux.git / log
