Similarly to the previous patch fixing the flow_dissector ctx accesses,
nf_is_valid_access also doesn't check that ctx accesses are aligned.
Contrary to flow_dissector programs, netfilter programs don't have
context conversion. The unaligned ctx accesses are therefore allowed by
the verifier.
Fixes: fd9c663b9ad6 ("bpf: minimal support for programs hooked into netfilter framework") Signed-off-by: Paul Chaignon <paul.chaignon@gmail.com> Acked-by: Yonghong Song <yonghong.song@linux.dev> Acked-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/853ae9ed5edaa5196e8472ff0f1bb1cc24059214.1754039605.git.paul.chaignon@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
flow_dissector_is_valid_access doesn't check that the context access is
aligned. As a consequence, an unaligned access within one of the exposed
field is considered valid and later rejected by
flow_dissector_convert_ctx_access when we try to convert it.
The later rejection is problematic because it's reported as a verifier
bug with a kernel warning and doesn't point to the right instruction in
verifier logs.
Since commit 6e890c5d5021 ("vhost: use vhost_tasks for worker threads"),
the vhost uses vhost_task and operates as a child of the
owner thread. This is required for correct CPU usage accounting,
especially when using containers.
However, this change has caused confusion for some legacy
userspace applications, and we didn't notice until it's too late.
Unfortunately, it's too late to revert - we now have userspace
depending both on old and new behaviour :(
To address the issue, reintroduce kthread mode for vhost workers and
provide a configuration to select between kthread and task worker.
- Add 'fork_owner' parameter to vhost_dev to let users select kthread
or task mode. Default mode is task mode(VHOST_FORK_OWNER_TASK).
- Reintroduce kthread mode support:
* Bring back the original vhost_worker() implementation,
and renamed to vhost_run_work_kthread_list().
* Add cgroup support for the kthread
* Introduce struct vhost_worker_ops:
- Encapsulates create / stop / wake‑up callbacks.
- vhost_worker_create() selects the proper ops according to
inherit_owner.
- Userspace configuration interface:
* New IOCTLs:
- VHOST_SET_FORK_FROM_OWNER lets userspace select task mode
(VHOST_FORK_OWNER_TASK) or kthread mode (VHOST_FORK_OWNER_KTHREAD)
- VHOST_GET_FORK_FROM_OWNER reads the current worker mode
* Expose module parameter 'fork_from_owner_default' to allow system
administrators to configure the default mode for vhost workers
* Kconfig option CONFIG_VHOST_ENABLE_FORK_OWNER_CONTROL controls whether
these IOCTLs and the parameter are available
- The VHOST_NEW_WORKER functionality requires fork_owner to be set
to true, with validation added to ensure proper configuration
This partially reverts or improves upon:
commit 6e890c5d5021 ("vhost: use vhost_tasks for worker threads")
commit 1cdaafa1b8b4 ("vhost: replace single worker pointer with xarray")
Fixes: 6e890c5d5021 ("vhost: use vhost_tasks for worker threads"), Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20250714071333.59794-2-lulu@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Acked-by: Jason Wang <jasowang@redhat.com> Tested-by: Lei Yang <leiyang@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Add missing idr_destroy() call in vduse_exit() to properly free the
vduse_idr radix tree nodes. Without this, module load/unload cycles leak
576-byte radix tree node allocations, detectable by kmemleak as:
The vduse_idr is initialized via DEFINE_IDR() at line 136 and used throughout
the VDUSE (vDPA Device in Userspace) driver for device ID management. The fix
follows the documented pattern in lib/idr.c and matches the cleanup approach
used by other drivers.
This leak was discovered through comprehensive module testing with cumulative
kmemleak detection across 10 load/unload iterations per module.
Fixes: c8a6153b6c59 ("vduse: Introduce VDUSE - vDPA Device in Userspace") Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Message-Id: <20250704125335.1084649-1-anders.roxell@linaro.org> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The commit in the fixes tag made sure that mlx5_vdpa_free()
is the single entrypoint for removing the vdpa device resources
added in mlx5_vdpa_dev_add(), even in the cleanup path of
mlx5_vdpa_dev_add().
This means that all functions from mlx5_vdpa_free() should be able to
handle uninitialized resources. This was not the case though:
mlx5_vdpa_destroy_mr_resources() and mlx5_cmd_cleanup_async_ctx()
were not able to do so. This caused the splat below when adding
a vdpa device without a MAC address.
This patch fixes these remaining issues:
- Makes mlx5_vdpa_destroy_mr_resources() return early if called on
uninitialized resources.
- Moves mlx5_cmd_init_async_ctx() early on during device addition
because it can't fail. This means that mlx5_cmd_cleanup_async_ctx()
also can't fail. To mirror this, move the call site of
mlx5_cmd_cleanup_async_ctx() in mlx5_vdpa_free().
An additional comment was added in mlx5_vdpa_free() to document
the expectations of functions called from this context.
As part of the normal initiator side scanning the guest's scsi layer
will loop over all possible targets and send an inquiry. Since the
max number of targets for virtio-scsi is 256, this can result in 255
error messages about targets not existing if you only have a single
target. When there's more than 1 vhost-scsi device each with a single
target, then you get N * 255 log messages.
It looks like the log message was added by accident in:
commit 3f8ca2e115e5 ("vhost/scsi: Extract common handling code from
control queue handler")
when we added common helpers. Then in:
commit 09d7583294aa ("vhost/scsi: Use common handling code in request
queue handler")
we converted the scsi command processing path to use the new
helpers so we started to see the extra log messages during scanning.
The patches were just making some code common but added the vq_err
call and I'm guessing the patch author forgot to enable the vq_err
call (vq_err is implemented by pr_debug which defaults to off). So
this patch removes the call since it's expected to hit this path
during device discovery.
Fixes: 09d7583294aa ("vhost/scsi: Use common handling code in request queue handler") Signed-off-by: Mike Christie <michael.christie@oracle.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20250611210113.10912-1-michael.christie@oracle.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
needs_teardown is a device flag that indicates when virtual queues need
to be recreated. This happens for certain configuration changes: queue
size and some specific features.
Currently, the needs_teardown state can be incorrectly reset by
subsequent .set_vq_num() calls. For example, for 1 rx VQ with size 512
and 1 tx VQ with size 256:
.set_vq_num(0, 512) -> sets needs_teardown to true (rx queue has a
non-default size)
.set_vq_num(1, 256) -> sets needs_teardown to false (tx queue has a
default size)
This change takes into account the previous value of the needs_teardown
flag when re-calculating it during VQ size configuration.
Fixes: 0fe963d6fc16 ("vdpa/mlx5: Re-create HW VQs under certain conditions") Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com> Reviewed-by: Shahar Shitrit <shshitrit@nvidia.com> Reviewed-by: Si-Wei Liu <si-wei.liu@oracle.com> Tested-by: Si-Wei Liu<si-wei.liu@oracle.com>
Message-Id: <20250604184802.2625300-1-dtatulea@nvidia.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Acked-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
It post-processes samples to find which DSO has samples. Based on that
info, it can save used DSOs in the build-ID cache directory. But for
some reason, it saves all DSOs without checking the hit mark. Skipping
unused DSOs can give some speedup especially with --buildid-mmap being
default.
On my idle machine, `time perf record -a sleep 1` goes down from 3 sec
to 1.5 sec with this change.
Fixes: e29386c8f7d71fa5 ("perf record: Add --buildid-mmap option to enable PERF_RECORD_MMAP2's build id") Reviewed-by: Arnaldo Carvalho de Melo <acme@redhat.com> Link: https://lore.kernel.org/r/20250731070330.57116-1-namhyung@kernel.org Signed-off-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In dbAllocCtl(), read_metapage() increases the reference count of the
metapage. However, when dp->tree.budmin < 0, the function returns -EIO
without calling release_metapage() to decrease the reference count,
leading to a memory leak.
Add release_metapage(mp) before the error return to properly manage
the metapage reference count and prevent the leak.
Fixes: a5f5e4698f8abbb25fe4959814093fb5bfa1aa9d ("jfs: fix shift-out-of-bounds in dbSplit") Signed-off-by: Zheng Yu <zheng.yu@northwestern.edu> Signed-off-by: Dave Kleikamp <dave.kleikamp@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
fb_add_videomode() can fail with -ENOMEM when its internal kmalloc() cannot
allocate a struct fb_modelist. If that happens, the modelist stays empty but
the driver continues to register. Add a check for its return value to prevent
poteintial null-ptr-deref, which is similar to the commit 17186f1f90d3 ("fbdev:
Fix do_register_framebuffer to prevent null-ptr-deref in fb_videomode_to_var").
The `adf_ring_next()` function in the QAT debug transport interface
fails to correctly update the position index when reaching the end of
the ring elements. This triggers the following kernel warning when
reading ring files, such as
/sys/kernel/debug/qat_c6xx_<D:B:D:F>/transport/bank_00/ring_00:
[27725.022965] seq_file: buggy .next function adf_ring_next [intel_qat] did not update position index
Ensure that the `*pos` index is incremented before returning NULL when
after the last element in the ring is found, satisfying the seq_file API
requirements and preventing the warning.
Fixes: a672a9dc872e ("crypto: qat - Intel(R) QAT transport code") Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Reviewed-by: Ahsan Atta <ahsan.atta@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
QAT devices perform an additional integrity check during compression by
decompressing the output. Starting from QAT GEN4, this verification is
done in-line by the hardware. However, on GEN2 devices, the hardware
reads back the compressed output from the destination buffer and performs
a decompression operation using it as the source.
In the current QAT driver, destination buffers are always marked as
write-only. This is incorrect for QAT GEN2 compression, where the buffer
is also read during verification. Since commit 6f5dc7658094
("iommu/vt-d: Restore WO permissions on second-level paging entries"),
merged in v6.16-rc1, write-only permissions are strictly enforced, leading
to DMAR errors when using QAT GEN2 devices for compression, if VT-d is
enabled.
Mark the destination buffers as DMA_BIDIRECTIONAL. This ensures
compatibility with GEN2 devices, even though it is not required for
QAT GEN4 and later.
Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Fixes: cf5bb835b7c8 ("crypto: qat - fix DMA transfer direction") Reviewed-by: Ahsan Atta <ahsan.atta@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
arch/sh/Makefile defines and exports ld-bfd to be used by
arch/sh/boot/compressed/Makefile and arch/sh/boot/romimage/Makefile.
However some shells, including dash, will not pass through environment
variables whose name includes a hyphen. Usually GNU make does not use
a shell to recurse, but if e.g. $(srctree) contains '~' it will use a
shell here.
Other instances of this problem were previously fixed by commits 2bfbe7881ee0 "kbuild: Do not use hyphen in exported variable name"
and 82977af93a0d "sh: rename suffix-y to suffix_y".
Rename the variable to ld_bfd.
References: https://buildd.debian.org/status/fetch.php?pkg=linux&arch=sh4&ver=4.13%7Erc5-1%7Eexp1&stamp=1502943967&raw=0 Fixes: 7b022d07a0fd ("sh: Tidy up the ldscript output format specifier.") Signed-off-by: Ben Hutchings <benh@debian.org> Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
There is no PHY for the XCVR module on i.MX93, the channel status needs
to be obtained from FSL_XCVR_RX_CS_DATA_* registers. And channel status
acknowledge (CSA) bit should be set once channel status is processed.
Fixes: e240b9329a30 ("ASoC: fsl_xcvr: Add support for i.MX93 platform") Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com> Link: https://patch.msgid.link/20250710030405.3370671-2-shengjiu.wang@nxp.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The do_k_string() and do_c_string() functions do essentially the same
thing which is they add a string and a comma onto the end of an existing
string. At the end, the caller will overwrite the last comma with a
newline. Later, in orangefs_kernel_debug_init(), we add a newline to
the string.
The change to do_k_string() is just cosmetic. I moved the "- 1" to
the other side of the comparison and made it "+ 1". This has no
effect on runtime, I just wanted the functions to match each other
and the rest of the file.
However in do_c_string(), I removed the "- 2" which allows us to print
two extra characters. I noticed this issue while reviewing the code
and I doubt affects anything in real life. My guess is that this was
double counting the comma and the newline. The "+ 1" accounts for
the newline, and the caller will delete the final comma which ensures
there is enough space for the newline.
Removing the "- 2" lets us print 2 more characters, but mainly it makes
the code more consistent and understandable for reviewers.
Fixes: 44f4641073f1 ("orangefs: clean up debugfs globals") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Signed-off-by: Mike Marshall <hubcap@omnibond.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When enabling runtime PM for clock suppliers that also belong to a power
domain, the following crash is thrown:
error: synchronous external abort: 0000000096000010 [#1] PREEMPT SMP
Workqueue: events_unbound deferred_probe_work_func
pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : clk_mux_get_parent+0x60/0x90
lr : clk_core_reparent_orphans_nolock+0x58/0xd8
Call trace:
clk_mux_get_parent+0x60/0x90
clk_core_reparent_orphans_nolock+0x58/0xd8
of_clk_add_hw_provider.part.0+0x90/0x100
of_clk_add_hw_provider+0x1c/0x38
imx95_bc_probe+0x2e0/0x3f0
platform_probe+0x70/0xd8
Enabling runtime PM without explicitly resuming the device caused
the power domain cut off after clk_register() is called. As a result,
a crash happens when the clock hardware provider is added and attempts
to access the BLK_CTL register.
Fix this by using devm_pm_runtime_enable() instead of pm_runtime_enable()
and getting rid of the pm_runtime_disable() in the cleanup path.
Fixes: 5224b189462f ("clk: imx: add i.MX95 BLK CTL clk driver") Reviewed-by: Frank Li <Frank.Li@nxp.com> Reviewed-by: Abel Vesa <abel.vesa@linaro.org> Signed-off-by: Laurentiu Palcu <laurentiu.palcu@oss.nxp.com> Signed-off-by: Peng Fan <peng.fan@nxp.com> Link: https://lore.kernel.org/r/20250707-imx95-blk-ctl-7-1-v3-2-c1b676ec13be@nxp.com Signed-off-by: Abel Vesa <abel.vesa@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
__iomem attribute is supposed to be used only with variables holding the
MMIO pointer. But here, 'mw_addr' variable is just holding a 'void *'
returned by pci_epf_alloc_space(). So annotating it with __iomem is clearly
wrong. Hence, drop the attribute.
This also fixes the below sparse warning:
drivers/pci/endpoint/functions/pci-epf-vntb.c:524:17: warning: incorrect type in assignment (different address spaces)
drivers/pci/endpoint/functions/pci-epf-vntb.c:524:17: expected void [noderef] __iomem *mw_addr
drivers/pci/endpoint/functions/pci-epf-vntb.c:524:17: got void *
drivers/pci/endpoint/functions/pci-epf-vntb.c:530:21: warning: incorrect type in assignment (different address spaces)
drivers/pci/endpoint/functions/pci-epf-vntb.c:530:21: expected unsigned int [usertype] *epf_db
drivers/pci/endpoint/functions/pci-epf-vntb.c:530:21: got void [noderef] __iomem *mw_addr
drivers/pci/endpoint/functions/pci-epf-vntb.c:542:38: warning: incorrect type in argument 2 (different address spaces)
drivers/pci/endpoint/functions/pci-epf-vntb.c:542:38: expected void *addr
drivers/pci/endpoint/functions/pci-epf-vntb.c:542:38: got void [noderef] __iomem *mw_addr
Fixes: e35f56bb0330 ("PCI: endpoint: Support NTB transfer between RC and EP") Signed-off-by: Manivannan Sadhasivam <mani@kernel.org> Reviewed-by: Frank Li <Frank.Li@nxp.com> Link: https://patch.msgid.link/20250709125022.22524-1-mani@kernel.org Signed-off-by: Sasha Levin <sashal@kernel.org>
The bus->params should be restored if the stream is failed to prepare.
The issue exists since beginning. The Fixes tag just indicates the
first commit that the commit can be applied to.
The ZUC-256 EEA (encryption) and EIA (integrity) algorithms are not
supported on QAT GEN5 devices, as their current implementation does not
align with the NIST specification. Earlier versions of the ZUC-256
specification used a different initialization scheme, which has since
been revised to comply with the 5G specification.
Due to this misalignment with the updated specification, remove support
for ZUC-256 EEA and EIA for QAT GEN5 by masking out the ZUC-256
capability.
Fixes: fcf60f4bcf549 ("crypto: qat - add support for 420xx devices") Signed-off-by: Bairavi Alagappan <bairavix.alagappan@intel.com> Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
Update the min, max ranges of the PLL clocks according to the latest
datasheet to be coherent in the driver. This patch solves the issues in
configuring the clocks related to peripherals with the desired frequency
within the range.
Currently we clear BH_New bit in case of error and also in the standard
ext4_write_end() handler (in block_commit_write()). However
ext4_journalled_write_end() misses this clearing and thus we are leaving
stale BH_New bits behind. Generally ext4_block_write_begin() clears
these bits before any harm can be done but in case blocksize < pagesize
and we hit some error when processing a page with these stale bits,
we'll try to zero buffers with these stale BH_New bits and jbd2 will
complain (as buffers were not prepared for writing in this transaction).
Fix the problem by clearing BH_New bits in ext4_journalled_write_end()
and WARN if ext4_block_write_begin() sees stale BH_New bits.
Reported-by: Baolin Liu <liubaolin12138@163.com> Reported-by: Zhi Long <longzhi@sangfor.com.cn> Fixes: 3910b513fcdf ("ext4: persist the new uptodate buffers in ext4_journalled_zero_new_buffers") Signed-off-by: Jan Kara <jack@suse.cz> Link: https://patch.msgid.link/20250709084831.23876-2-jack@suse.cz Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
The "rec->len" value comes from the firmware. We generally do
trust firmware, but it's always better to double check. If
the length value is too large it would lead to memory corruption
when we set "data[i] = ret;"
Fixes: 217209db0204 ("watchdog: ziirave_wdt: Add support to upload the firmware.") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/3b58b453f0faa8b968c90523f52c11908b56c346.1748463049.git.dan.carpenter@linaro.org Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The dma_unmap_sg() functions should be called with the same nents as the
dma_map_sg(), not the value the map function returned.
Fixes: 88a678bbc34c ("ibmvscsis: Initial commit of IBM VSCSI Tgt Driver") Signed-off-by: Thomas Fourier <fourier.thomas@gmail.com> Link: https://lore.kernel.org/r/20250630111803.94389-2-fourier.thomas@gmail.com Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The de clock is marked with CLK_SET_RATE_PARENT, which is really not
necessary (as confirmed from experimentation) and significantly
restricts flexibility for other clocks using the same parent.
In addition the source selection (parent) field is marked as using
2 bits, when it the documentation reports that it uses 3.
Fix both issues in the de clock definition.
Fixes: d0f11d14b0bc ("clk: sunxi-ng: add support for V3s CCU") Signed-off-by: Paul Kocialkowski <paulk@sys-base.io> Link: https://patch.msgid.link/20250704154008.3463257-1-paulk@sys-base.io Signed-off-by: Chen-Yu Tsai <wens@csie.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The "osc_12m" fixed factor clock refers the external oscillator by
setting clk_parent_data.fw_name to osc_24m, which is obviously wrong
since no clock-names property is allowed for compatible
thead,th1520-clk-ap.
Refer the oscillator as parent by index instead.
Fixes: ae81b69fd2b1 ("clk: thead: Add support for T-Head TH1520 AP_SUBSYS clocks") Signed-off-by: Yao Zi <ziyao@disroot.org> Reviewed-by: Drew Fustini <fustini@kernel.org> Signed-off-by: Drew Fustini <fustini@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 5a3e85c3c397 ("pinmux: Use sequential access to access
desc->pinmux data") tried to address the issue when two client of the
same gpio calls pinctrl_select_state() for the same functionality, was
resulting in NULL pointer issue while accessing desc->mux_owner.
However, issue was not completely fixed due to the way it was handled
and it can still result in the same NULL pointer.
The issue occurs due to the following interleaving:
This sequence leads to a state where the pin appears to be in use
(`mux_usecount == 1`) but has no owner (`mux_owner == NULL`), which can
cause NULL pointer on next pin_request on the same pin.
Ensure that updates to mux_usecount and mux_owner are performed
atomically under the same lock. Only clear mux_owner when mux_usecount
reaches zero and no new owner has been assigned.
Check pde->proc_ops->proc_lseek directly may cause UAF in rmmod scenario.
It's a gap in proc_reg_open() after commit 654b33ada4ab("proc: fix UAF in
proc_get_inode()"). Followed by AI Viro's suggestion, fix it in same
manner.
Link: https://lkml.kernel.org/r/20250607021353.1127963-1-wangzijie1@honor.com Fixes: 3f61631d47f1 ("take care to handle NULL ->proc_lseek()") Signed-off-by: wangzijie <wangzijie1@honor.com> Reviewed-by: Alexey Dobriyan <adobriyan@gmail.com> Cc: Alexei Starovoitov <ast@kernel.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: "Edgecombe, Rick P" <rick.p.edgecombe@intel.com> Cc: Kirill A. Shuemov <kirill.shutemov@linux.intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
A CPU mask on the stack is broken for large values of CONFIG_NR_CPUS:
kernel/trace/preemptirq_delay_test.c: In function ‘preemptirq_delay_run’:
kernel/trace/preemptirq_delay_test.c:143:1: error: the frame size of 8512 bytes is larger than 1536 bytes [-Werror=frame-larger-than=]
hr_dev->pgdir_list and hr_dev->pgdir_mutex won't be initialized if
CQ/QP record db are not enabled, but they are also needed when using
SRQ with SRQ record db enabled. Simplified the logic by always
initailizing the reosurces.
ACK_REQ_FREQ indicates the number of packets (after MTU fragmentation)
HW sends before setting an ACK request. When MTU is greater than or
equal to 1024, the current ACK_REQ_FREQ value causes HW to request an
ACK for every MTU fragment. The processing of a large number of ACKs
severely impacts HW performance when sending large size payloads.
Get message length of ack_req from FW so that we can adjust this
parameter according to different situations. There are several
constraints for ACK_REQ_FREQ:
1. mtu * (2 ^ ACK_REQ_FREQ) should not be too large, otherwise it may
cause some unexpected retries when sending large payload.
2. ACK_REQ_FREQ should be larger than or equal to LP_PKTN_INI.
3. ACK_REQ_FREQ must be equal to LP_PKTN_INI when using LDCP
or HC3 congestion control algorithm.
Fixes: 56518a603fd2 ("RDMA/hns: Modify the value of long message loopback slice") Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20250703113905.3597124-4-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch sets ccp_debugfs_dir to NULL after destroying it in
ccp5_debugfs_destroy, allowing the directory dentry to be
recreated when rebinding the ccp device.
Tested on AMD Ryzen 7 1700X.
Fixes: 3cdbe346ed3f ("crypto: ccp - Add debugfs entries for CCP information") Signed-off-by: Mengbiao Xiong <xisme1998@gmail.com> Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
The `dma_unmap_sg()` functions should be called with the same nents as the
`dma_map_sg()`, not the value the map function returned.
Fixes: c957f8b3e2e5 ("crypto: inside-secure - avoid unmapping DMA memory that was not mapped") Signed-off-by: Thomas Fourier <fourier.thomas@gmail.com> Reviewed-by: Antoine Tenart <atenart@kernel.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
The __snp_alloc_firmware_pages() helper allocates pages in the firmware
state (alloc + rmpupdate). In case of failed rmpupdate, it tries
reclaiming pages with already changed state. This requires calling
the PSP firmware and since there is sev_cmd_mutex to guard such calls,
the helper takes a "locked" parameter so specify if the lock needs to
be held.
Most calls happen from snp_alloc_firmware_page() which executes without
the lock. However
commit 24512afa4336 ("crypto: ccp: Handle the legacy TMR allocation when SNP is enabled")
switched sev_fw_alloc() from alloc_pages() (which does not call the PSP) to
__snp_alloc_firmware_pages() (which does) but did not account for the fact
that sev_fw_alloc() is called from __sev_platform_init_locked()
(via __sev_platform_init_handle_tmr()) and executes with the lock held.
Add a "locked" parameter to __snp_alloc_firmware_pages().
Make sev_fw_alloc() use the new parameter to prevent potential deadlock in
rmp_mark_pages_firmware() if rmpupdate() failed.
Fixes: 24512afa4336 ("crypto: ccp: Handle the legacy TMR allocation when SNP is enabled") Signed-off-by: Alexey Kardashevskiy <aik@amd.com> Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com> Reviewed-by: Pratik R. Sampat <prsampat@amd.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
hns_roce_clear_extdb_list_info() will eventually do some HW
configurations through FW, and they need to be cleared by
calling hns_roce_function_clear() when the initialization
fails.
Fixes: 7e78dd816e45 ("RDMA/hns: Clear extended doorbell info before using") Signed-off-by: wenglianfa <wenglianfa@huawei.com> Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20250703113905.3597124-3-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
rsv_qp may be double destroyed in error flow, first in free_mr_init(),
and then in hns_roce_exit(). Fix it by moving the free_mr_init() call
into hns_roce_v2_init().
The work_atoms should be freed after use. Add free_work_atoms() to
make sure to release all. It should use list_splice_init() when merging
atoms to prevent accessing invalid pointers.
Fixes: b1ffe8f3e0c96f552 ("perf sched: Finish latency => atom rename and misc cleanups") Reviewed-by: Ian Rogers <irogers@google.com> Tested-by: Ian Rogers <irogers@google.com> Link: https://lore.kernel.org/r/20250703014942.1369397-8-namhyung@kernel.org Signed-off-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In many perf sched subcommand saves priv data structure in the thread
but it forgot to free them. As it's an opaque type with 'void *', it
needs to register that knows how to free the data. In this case, just
regular 'free()' is fine.
Fixes: 04cb4fc4d40a5bf1 ("perf thread: Allow tools to register a thread->priv destructor") Reviewed-by: Ian Rogers <irogers@google.com> Tested-by: Ian Rogers <irogers@google.com> Link: https://lore.kernel.org/r/20250703014942.1369397-3-namhyung@kernel.org Signed-off-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The parse_options_subcommand() allocates the usage string based on the
given subcommands. So it should reach the end of the function to free
the string to prevent memory leaks.
Fixes: 1a5efc9e13f357ab ("libsubcmd: Don't free the usage string") Reviewed-by: Ian Rogers <irogers@google.com> Tested-by: Ian Rogers <irogers@google.com> Link: https://lore.kernel.org/r/20250703014942.1369397-2-namhyung@kernel.org Signed-off-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Infineon SEMPER flash family does not support E9h opcode as Exit 4-byte
mode (EX4B). Therefore, params->set_4byte_addr_mode is not determined by
BFPT parse. Fixup it up by introducing vendor specific EX4B opcode (B8h)
and function.
The kcore loading creates a set of list nodes that have reference
counted references to maps of the kcore. The list node freeing in the
success path wasn't releasing the maps, add the missing puts. It is
unclear why this leak was being missed by leak sanitizer.
Fixes: 83720209961f ("perf map: Move map list node into symbol") Signed-off-by: Ian Rogers <irogers@google.com> Link: https://lore.kernel.org/r/20250624190326.2038704-2-irogers@google.com Signed-off-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently perf aborts when it finds an invalid command. I guess it
depends on the environment as I have some custom commands in the path.
$ perf bad-command
perf: 'bad-command' is not a perf-command. See 'perf --help'.
Aborted (core dumped)
It's because the exclude_cmds() in libsubcmd has a use-after-free when
it removes some entries. After copying one to another entry, it keeps
the pointer in the both position. And the next copy operation will free
the later one but it's the same entry in the previous one.
For example, let's say cmds = { A, B, C, D, E } and excludes = { B, E }.
ci cj ei cmds-name excludes
-----------+--------------------
0 0 0 | A B : cmp < 0, ci == cj
1 1 0 | B B : cmp == 0
2 1 1 | C E : cmp < 0, ci != cj
At this point, it frees cmds->names[1] and cmds->names[1] is assigned to
cmds->names[2].
3 2 1 | D E : cmp < 0, ci != cj
Now it frees cmds->names[2] but it's the same as cmds->names[1]. So
accessing cmds->names[1] will be invalid.
This makes the subcmd tests succeed.
$ perf test subcmd
69: libsubcmd help tests :
69.1: Load subcmd names : Ok
69.2: Uniquify subcmd names : Ok
69.3: Exclude duplicate subcmd names : Ok
Fixes: 4b96679170c6 ("libsubcmd: Avoid SEGV/use-after-free when commands aren't excluded") Reviewed-by: Ian Rogers <irogers@google.com> Link: https://lore.kernel.org/r/20250701201027.1171561-3-namhyung@kernel.org Signed-off-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The fpfd_max frequency should be set to 450 MHz instead of 300 MHz.
Well, it actually depends on the platform speed grade but we are being
conservative for ultrascale so let's be consistent. In a following
change we will set these limits at runtime.
Unlike file_handle, type and len of struct fanotify_fh are u8.
Traditionally, filesystem return handle_type < 0xff, but there
is no enforecement for that in vfs.
Add a sanity check in fanotify to avoid truncating handle_type
if its value is > 0xff.
Fixes: 7cdafe6cc4a6 ("exportfs: check for error return value from exportfs_encode_*()") Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz> Link: https://patch.msgid.link/20250627104835.184495-1-amir73il@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
Zeroing out registers does not happen in the downstream kernel, and will
"tune" the repeater in surely unexpected ways since most registers don't
have a reset value of 0x0.
Stop doing that and instead just set the registers that are in the init
sequence (though long term I don't think there's actually PMIC-specific
init sequences, there's board specific tuning, but that's a story for
another day).
Fixes: 99a517a582fc ("phy: qualcomm: phy-qcom-eusb2-repeater: Zero out untouched tuning regs") Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com> Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org> Signed-off-by: Luca Weiss <luca.weiss@fairphone.com> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com> Reviewed-by: Abel Vesa <abel.vesa@linaro.org> Link: https://lore.kernel.org/r/20250617-eusb2-repeater-tuning-v2-2-ed6c484f18ee@fairphone.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the original implementation, krealloc() failure handling incorrectly
assigned the original memory pointer to NULL after kfree(), causing a
memory leak when reallocation failed.
In sunxi_pctrl_dt_node_to_map(), when krealloc() fails to resize
the pinctrl_map array, the function returns -ENOMEM directly
without freeing the previously allocated *map buffer. This results
in a memory leak of the original kmalloc_array allocation.
According the function documentation of epf_ntb_init_epc_bar(), the
function should return an error code on error. However, it returns -1 when
no BAR is available i.e., when pci_epc_get_next_free_bar() fails.
Return -ENOENT instead.
Fixes: e35f56bb0330 ("PCI: endpoint: Support NTB transfer between RC and EP") Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
[mani: changed err code to -ENOENT] Signed-off-by: Manivannan Sadhasivam <mani@kernel.org> Reviewed-by: Frank Li <Frank.Li@nxp.com> Link: https://patch.msgid.link/20250603-pci-vntb-bar-mapping-v2-1-fc685a22ad28@baylibre.com Signed-off-by: Sasha Levin <sashal@kernel.org>
I get a very rare -Wstringop-overread warning with gcc-15 for one function
in aesbs_ctr_encrypt():
arch/arm/crypto/aes-neonbs-glue.c: In function 'ctr_encrypt':
arch/arm/crypto/aes-neonbs-glue.c:212:1446: error: '__builtin_memcpy' offset [17, 2147483647] is out of the bounds [0, 16] of object 'buf' with type 'u8[16]' {aka 'unsigned char[16]'} [-Werror=array-bounds=]
212 | src = dst = memcpy(buf + sizeof(buf) - bytes,
arch/arm/crypto/aes-neonbs-glue.c: In function 'ctr_encrypt':
arch/arm/crypto/aes-neonbs-glue.c:218:17: error: 'aesbs_ctr_encrypt' reading 1 byte from a region of size 0 [-Werror=stringop-overread]
218 | aesbs_ctr_encrypt(dst, src, ctx->rk, ctx->rounds, bytes, walk.iv);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
arch/arm/crypto/aes-neonbs-glue.c:218:17: note: referencing argument 2 of type 'const u8[0]' {aka 'const unsigned char[]'}
arch/arm/crypto/aes-neonbs-glue.c:218:17: note: referencing argument 3 of type 'const u8[0]' {aka 'const unsigned char[]'}
arch/arm/crypto/aes-neonbs-glue.c:218:17: note: referencing argument 6 of type 'u8[0]' {aka 'unsigned char[]'}
arch/arm/crypto/aes-neonbs-glue.c:36:17: note: in a call to function 'aesbs_ctr_encrypt'
36 | asmlinkage void aesbs_ctr_encrypt(u8 out[], u8 const in[], u8 const rk[],
This could happen in theory if walk.nbytes is larger than INT_MAX and gets
converted to a negative local variable.
Keep the type unsigned like the orignal nbytes to be sure there is no
integer overflow.
Fixes: c8bf850e991a ("crypto: arm/aes-neonbs-ctr - deal with non-multiples of AES block size") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Ard Biesheuvel <ardb@kernel.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
When the kernel is not configured CONFIG_OF, the max14577_charger_dt_init
function returns NULL. Fix the max14577_charger_probe functionby returning
-ENODATA instead of potentially passing a NULL pointer to PTR_ERR.
This fixes the below smatch warning:
max14577_charger_probe() warn: passing zero to 'PTR_ERR'
Fixes: e30110e9c96f ("charger: max14577: Configure battery-dependent settings from DTS and sysfs") Signed-off-by: Charles Han <hanchunchao@inspur.com> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Link: https://lore.kernel.org/r/20250519061601.8755-1-hanchunchao@inspur.com Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the cpcap_usb_detect() function, the power_supply_get_by_name()
function may return `NULL` instead of an error pointer.
To prevent potential null pointer dereferences, Added a null check.
Fixes: eab4e6d953c1 ("power: supply: cpcap-charger: get the battery inserted infomation from cpcap-battery") Signed-off-by: Charles Han <hanchunchao@inspur.com> Link: https://lore.kernel.org/r/20250519024741.5846-1-hanchunchao@inspur.com Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The `separate_colour_plane_flag` element is only present in the SPS if
`chroma_format_idc == 3`, so the corresponding flag should be disabled
whenever that is not the case and not just on profiles where
`chroma_format_idc` is not present.
Fixes: b32e48503df0 ("media: controls: Validate H264 stateless controls") Signed-off-by: James Cowgill <james.cowgill@blaize.com> Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Sasha Levin <sashal@kernel.org>
devm_kasprintf() returns NULL when memory allocation fails. Currently,
davinci_lpsc_clk_register() does not check for this case, which results
in a NULL pointer dereference.
Add NULL check after devm_kasprintf() to prevent this issue and ensuring
no resources are left allocated.
Fixes: c6ed4d734bc7 ("clk: davinci: New driver for davinci PSC clocks") Signed-off-by: Henry Martin <bsdhenrymartin@gmail.com> Link: https://lore.kernel.org/r/20250401131341.26800-1-bsdhenrymartin@gmail.com Reviewed-by: David Lechner <david@lechnology.com> Signed-off-by: Stephen Boyd <sboyd@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The expression '1 << EraseUnitSize' is evaluated in int, which causes
a negative result when shifting by 31 - the upper bound of the valid
range [10, 31], enforced by scan_header(). This leads to incorrect
extension when storing the result in 'erase->len' (uint64_t), producing
a large unexpected value.
Found by Linux Verification Center (linuxtesting.org) with Svace.
Change the logic in the restore function to properly handle bank
exceptions.
The check for exceptions in the saved state should be performed before
conducting any other ringstat register checks.
If a bank was saved with an exception, the ringstat will have the
appropriate rp_halt/rp_exception bits set, causing the driver to exit
the restore process with an error. Instead, the restore routine should
first check the ringexpstat register, and if any exception was raised,
it should stop further checks and return without any error. In other
words, if a ring pair is in an exception state at the source, it should
be restored the same way at the destination but without raising an error.
Even though this approach might lead to losing the exception state
during migration, the driver will log the exception from the saved state
during the restore process.
Signed-off-by: Svyatoslav Pankratov <svyatoslav.pankratov@intel.com> Fixes: bbfdde7d195f ("crypto: qat - add bank save and restore flows") Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
The commit ca88a2bdd4dd ("crypto: qat - allow disabling SR-IOV VFs")
introduced an unnecessary change that prevented enabling SR-IOV when
IOMMU is disabled. In certain scenarios, it is desirable to enable
SR-IOV even in the absence of IOMMU. Thus, restoring the previous
functionality to allow VFs to be enumerated in the absence of IOMMU.
Fixes: ca88a2bdd4dd ("crypto: qat - allow disabling SR-IOV VFs") Signed-off-by: Ahsan Atta <ahsan.atta@intel.com> Reviewed-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Reviewed-by: Michal Witwicki <michal.witwicki@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
The dc_data structure holds data required for handling compression
operations, such as overflow buffers. In this context, the use of
managed memory allocation APIs (devm_kzalloc() and devm_kfree())
is not necessary, as these data structures are freed and
re-allocated when a device is restarted in adf_dev_down() and
adf_dev_up().
Additionally, managed APIs automatically handle memory cleanup when the
device is detached, which can lead to conflicts with manual cleanup
processes. Specifically, if a device driver invokes the adf_dev_down()
function as part of the cleanup registered with
devm_add_action_or_reset(), it may attempt to free memory that is also
managed by the device's resource management system, potentially leading
to a double-free.
This might result in a warning similar to the following when unloading
the device specific driver, for example qat_6xxx.ko:
Use unmanaged memory allocation APIs (kzalloc_node() and kfree()) for
the dc_data structure. This ensures that memory is explicitly allocated
and freed under the control of the driver code, preventing manual
deallocation from interfering with automatic cleanup.
Fixes: 1198ae56c9a5 ("crypto: qat - expose deflate through acomp api for QAT GEN2") Signed-off-by: Suman Kumar Chakraborty <suman.kumar.chakraborty@intel.com> Reviewed-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
In sun8i_ce_cipher_unprepare(), dma_unmap_sg() is incorrectly called with
the number of entries returned by dma_map_sg(), rather than using the
original number of entries passed when mapping the scatterlist.
To fix this, stash the original number of entries passed to dma_map_sg()
in the request context.
Commit bc4d25fdfadf ("clk: renesas: rzv2h: Add support for dynamic
switching divider clocks") missed setting the `CLK_SET_RATE_PARENT`
flag when registering ddiv clocks.
Without this flag, rate changes to the divider clock do not propagate
to its parent, potentially resulting in incorrect clock configurations.
Fix this by setting `CLK_SET_RATE_PARENT` in the clock init data.
When FORTIFY_SOURCE reports about a run-time buffer overread, the wrong
buffer size was being shown in the error message. (The bounds checking
was correct.)
When gmin_get_config_var() calls efi.get_variable() and the EFI variable
is larger than the expected buffer size, two behaviors combine to create
a stack buffer overflow:
1. gmin_get_config_var() does not return the proper error code when
efi.get_variable() fails. It returns the stale 'ret' value from
earlier operations instead of indicating the EFI failure.
2. When efi.get_variable() returns EFI_BUFFER_TOO_SMALL, it updates
*out_len to the required buffer size but writes no data to the output
buffer. However, due to bug #1, gmin_get_var_int() believes the call
succeeded.
The caller gmin_get_var_int() then performs:
- Allocates val[CFG_VAR_NAME_MAX + 1] (65 bytes) on stack
- Calls gmin_get_config_var(dev, is_gmin, var, val, &len) with len=64
- If EFI variable is >64 bytes, efi.get_variable() sets len=required_size
- Due to bug #1, thinks call succeeded with len=required_size
- Executes val[len] = 0, writing past end of 65-byte stack buffer
This creates a stack buffer overflow when EFI variables are larger than
64 bytes. Since EFI variables can be controlled by firmware or system
configuration, this could potentially be exploited for code execution.
Fix the bug by returning proper error codes from gmin_get_config_var()
based on EFI status instead of stale 'ret' value.
The gmin_get_var_int() function is called during device initialization
for camera sensor configuration on Intel Bay Trail and Cherry Trail
platforms using the atomisp camera stack.
In the ARM64 BPF JIT when prog->aux->exception_boundary is set for a BPF
program, find_used_callee_regs() is not called because for a program
acting as exception boundary, all callee saved registers are saved.
find_used_callee_regs() sets `ctx->fp_used = true;` when it sees FP
being used in any of the instructions.
For programs acting as exception boundary, ctx->fp_used remains false
even if frame pointer is used by the program and therefore, FP is not
set-up for such programs in the prologue. This can cause the kernel to
crash due to a pagefault.
Fix it by setting ctx->fp_used = true for exception boundary programs as
fp is always saved in such programs.
rt->fib6_nsiblings can be read locklessly, add corresponding
READ_ONCE() and WRITE_ONCE() annotations.
Fixes: 66f5d6ce53e6 ("ipv6: replace rwlock with rcu and spinlock in fib6_table") Signed-off-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20250725140725.3626540-5-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
fib6_info_uses_dev() seems to rely on RCU without an explicit
protection.
Like the prior fix in rt6_nlmsg_size(),
we need to make sure fib6_del_route() or fib6_add_rt2node()
have not removed the anchor from the list, or we risk an infinite loop.
Fixes: d9ccb18f83ea ("ipv6: Fix soft lockups in fib6_select_path under high next hop churn") Signed-off-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20250725140725.3626540-4-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This is because fib6_del_route() and fib6_add_rt2node()
uses list_del_rcu(), which can confuse rcu readers,
because they might no longer see the head of the list.
Restart the loop if f6i->fib6_nsiblings is zero.
Fixes: d9ccb18f83ea ("ipv6: Fix soft lockups in fib6_select_path under high next hop churn") Signed-off-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20250725140725.3626540-3-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The esp4_offload module, loaded during IPsec offload tests, should
be reset to its default settings after testing.
Otherwise, leaving it enabled could unintentionally affect subsequence
test cases by keeping offload active.
A negative overflow can happen when the budget number of descs are
consumed. as long as the budget is decreased to zero, it will again go
into while (budget-- > 0) statement and get decreased by one, so the
overflow issue can happen. It will lead to returning true whereas the
expected value should be false.
In this case where all the budget is used up, it means zc function
should return false to let the poll run again because normally we
might have more data to process. Without this patch, zc function would
return true instead.
Fixes: 132c32ee5bc0 ("net: stmmac: Add TX via XDP zero-copy socket") Signed-off-by: Jason Xing <kernelxing@tencent.com> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com> Link: https://patch.msgid.link/20250723142327.85187-2-kerneljasonxing@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When KSZ8863 support was first added to KSZ driver the RX drop MIB
counter was somehow defined as 0x105. The TX drop MIB counter
starts at 0x100 for port 1, 0x101 for port 2, and 0x102 for port 3, so
the RX drop MIB counter should start at 0x103 for port 1, 0x104 for
port 2, and 0x105 for port 3.
There are 5 ports for KSZ8895, so its RX drop MIB counter starts at
0x105.
Fixes: 4b20a07e103f ("net: dsa: microchip: ksz8795: add support for ksz88xx chips") Signed-off-by: Tristram Ha <tristram.ha@microchip.com> Reviewed-by: Oleksij Rempel <o.rempel@pengutronix.de> Link: https://patch.msgid.link/20250723030403.56878-1-Tristram.Ha@microchip.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Hardware returns a unique identifier for a decrypted packet's xfrm
state, this state is looked up in an xarray. However, the state might
have been freed by the time of this lookup.
Currently, if the state is not found, only a counter is incremented.
The secpath (sp) extension on the skb is not removed, resulting in
sp->len becoming 0.
Subsequently, functions like __xfrm_policy_check() attempt to access
fields such as xfrm_input_state(skb)->xso.type (which dereferences
sp->xvec[sp->len - 1]) without first validating sp->len. This leads to
a crash when dereferencing an invalid state pointer.
This patch prevents the crash by explicitly removing the secpath
extension from the skb if the xfrm state is not found after hardware
decryption. This ensures downstream functions do not operate on a
zero-length secpath.
When updating the PBMC register, we read its current value,
modify desired fields, then write it back.
The port_buffer_size field within PBMC is Read-Only (RO).
If this RO field contains a non-zero value when read,
attempting to write it back will cause the entire PBMC
register update to fail.
This commit ensures port_buffer_size is explicitly cleared
to zero after reading the PBMC register but before writing
back the modified value.
This allows updates to other fields in the PBMC register to succeed.
Assign netdev.dev_port based on the device channel index, to indicate the
port number of the network device.
While this driver already uses netdev.dev_id for that purpose, dev_port is
more appropriate. However, retain dev_id to avoid potential regressions.
Fixes: 3e66d0138c05 ("can: populate netdev::dev_id for udev discrimination") Reviewed-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr> Signed-off-by: Jimmy Assarsson <extja@kvaser.com> Link: https://patch.msgid.link/20250725123452.41-4-extja@kvaser.com Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
projects / thirdparty / kernel / stable.git / log
