Pull NVMe fix from Christoph:
"nvme fixes for Linux 6.15
- fix an out-of-bounds access in nvmet_enable_port (Richard Weinberger)"
* tag 'nvme-6.15-2025-04-24' of git://git.infradead.org/nvme:
nvmet: fix out-of-bounds access in nvmet_enable_port
The current way of reading a timestamp snapshot in stmmac can lead to
integer overflow, as the computation is done on 32 bits. The issue has
been observed on a dwmac-socfpga platform returning chaotic timestamp
values due to this overflow. The corresponding multiplication is done
with a MUL instruction, which returns 32 bit values. Explicitly casting
the value to 64 bits replaced the MUL with a UMLAL, which computes and
returns the result on 64 bits, and so returns correctly the timestamps.
Prevent this overflow by explicitly casting the intermediate value to
u64 to make sure that the whole computation is made on u64. While at it,
apply the same cast on the other dwmac variant (GMAC4) method for
snapshot retrieval.
Fixes: 477c3e1f63 ("net: stmmac: Introduce dwmac1000 timestamping operations")
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20250423-stmmac_ts-v2-2-e2cf2bbd61b1@bootlin.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
When a PTP interrupt occurs, the driver accesses the wrong offset to
learn about the number of available snapshots in the FIFO for dwmac1000:
it should be accessing bits 29..25, while it is currently reading bits
19..16 (those are bits about the auxiliary triggers which have generated
the timestamps). As a consequence, it does not compute correctly the
number of available snapshots, and so possibly do not generate the
corresponding clock events if the bogus value ends up being 0.
Fix clock events generation by reading the correct bits in the timestamp
register for dwmac1000.
Fixes: 477c3e1f63 ("net: stmmac: Introduce dwmac1000 timestamping operations")
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20250423-stmmac_ts-v2-1-e2cf2bbd61b1@bootlin.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Make the wait_context a full part of the q_info struct rather
than a stack variable that goes away after pdsc_adminq_post()
is done so that the context is still available after the wait
loop has given up.
There was a case where a slow development firmware caused
the adminq request to time out, but then later the FW finally
finished the request and sent the interrupt. The handler tried
to complete_all() the completion context that had been created
on the stack in pdsc_adminq_post() but no longer existed.
This caused bad pointer usage, kernel crashes, and much wailing
and gnashing of teeth.
Fixes: 01ba61b55b ("pds_core: Add adminq processing and commands")
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20250421174606.3892-5-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the pds_core driver was first created there were some race
conditions around using the adminq, especially for client drivers.
To reduce the possibility of a race condition there's a check
against pf->state in pds_client_adminq_cmd(). This is problematic
for a couple of reasons:
1. The PDSC_S_INITING_DRIVER bit is set during probe, but not
cleared until after everything in probe is complete, which
includes creating the auxiliary devices. For pds_fwctl this
means it can't make any adminq commands until after pds_core's
probe is complete even though the adminq is fully up by the
time pds_fwctl's auxiliary device is created.
2. The race conditions around using the adminq have been fixed
and this path is already protected against client drivers
calling pds_client_adminq_cmd() if the adminq isn't ready,
i.e. see pdsc_adminq_post() -> pdsc_adminq_inc_if_up().
Fix this by removing the pf->state check in pds_client_adminq_cmd()
because invalid accesses to pds_core's adminq is already handled by
pdsc_adminq_post()->pdsc_adminq_inc_if_up().
Fixes: 10659034c6 ("pds_core: add the aux client API")
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Brett Creeley <brett.creeley@amd.com>
Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20250421174606.3892-4-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If the FW doesn't support the PDS_CORE_CMD_FW_CONTROL command
the driver might at the least print garbage and at the worst
crash when the user runs the "devlink dev info" devlink command.
This happens because the stack variable fw_list is not 0
initialized which results in fw_list.num_fw_slots being a
garbage value from the stack. Then the driver tries to access
fw_list.fw_names[i] with i >= ARRAY_SIZE and runs off the end
of the array.
Fix this by initializing the fw_list and by not failing
completely if the devcmd fails because other useful information
is printed via devlink dev info even if the devcmd fails.
Fixes: 45d76f4929 ("pds_core: set up device and adminq")
Signed-off-by: Brett Creeley <brett.creeley@amd.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20250421174606.3892-3-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The pds_core's adminq is protected by the adminq_lock, which prevents
more than 1 command to be posted onto it at any one time. This makes it
so the client drivers cannot simultaneously post adminq commands.
However, the completions happen in a different context, which means
multiple adminq commands can be posted sequentially and all waiting
on completion.
On the FW side, the backing adminq request queue is only 16 entries
long and the retry mechanism and/or overflow/stuck prevention is
lacking. This can cause the adminq to get stuck, so commands are no
longer processed and completions are no longer sent by the FW.
As an initial fix, prevent more than 16 outstanding adminq commands so
there's no way to cause the adminq from getting stuck. This works
because the backing adminq request queue will never have more than 16
pending adminq commands, so it will never overflow. This is done by
reducing the adminq depth to 16.
Fixes: 45d76f4929 ("pds_core: set up device and adminq")
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Brett Creeley <brett.creeley@amd.com>
Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20250421174606.3892-2-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MT7531 standalone and MMIO variants found in MT7988 and EN7581 share
most basic properties. Despite that, assisted_learning_on_cpu_port and
mtu_enforcement_ingress were only applied for MT7531 but not for MT7988
or EN7581, causing the expected issues on MMIO devices.
Apply both settings equally also for MT7988 and EN7581 by moving both
assignments form mt7531_setup() to mt7531_setup_common().
This fixes unwanted flooding of packets due to unknown unicast
during DA lookup, as well as issues with heterogenous MTU settings.
Fixes: 7f54cc9772 ("net: dsa: mt7530: split-off common parts from mt7531_setup")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Reviewed-by: Chester A. Unal <chester.a.unal@arinc9.com>
Link: https://patch.msgid.link/89ed7ec6d4fa0395ac53ad2809742bb1ce61ed12.1745290867.git.daniel@makrotopia.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Cong Wang says:
====================
net_sched: Fix UAF vulnerability in HFSC qdisc
This patchset contains two bug fixes and a selftest for the first one
which we have a reliable reproducer, please check each patch
description for details.
====================
Link: https://patch.msgid.link/20250417184732.943057-1-xiyou.wangcong@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch fixes a Use-After-Free vulnerability in the HFSC qdisc class
handling. The issue occurs due to a time-of-check/time-of-use condition
in hfsc_change_class() when working with certain child qdiscs like netem
or codel.
The vulnerability works as follows:
1. hfsc_change_class() checks if a class has packets (q.qlen != 0)
2. It then calls qdisc_peek_len(), which for certain qdiscs (e.g.,
codel, netem) might drop packets and empty the queue
3. The code continues assuming the queue is still non-empty, adding
the class to vttree
4. This breaks HFSC scheduler assumptions that only non-empty classes
are in vttree
5. Later, when the class is destroyed, this can lead to a Use-After-Free
The fix adds a second queue length check after qdisc_peek_len() to verify
the queue wasn't emptied.
Fixes: 21f4d5cc25 ("net_sched/hfsc: fix curve activation in hfsc_change_class()")
Reported-by: Gerrard Tai <gerrard.tai@starlabs.sg>
Reviewed-by: Konstantin Khlebnikov <koct9i@gmail.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Link: https://patch.msgid.link/20250417184732.943057-2-xiyou.wangcong@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Matthieu Baerts says:
====================
mptcp: pm: Defer freeing userspace pm entries
Here are two unrelated fixes for MPTCP:
- Patch 1: free userspace PM entry with RCU helpers. A fix for v6.14.
- Patch 2: avoid a warning when running diag.sh selftest. A fix for
v6.15-rc1.
====================
Link: https://patch.msgid.link/20250421-net-mptcp-pm-defer-freeing-v1-0-e731dc6e86b9@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When running diag.sh in a loop, chk_dump_one will report the following
"grep: write error":
13 ....chk 2 cestab [ OK ]
grep: write error
14 ....chk dump_one [ OK ]
15 ....chk 2->0 msk in use after flush [ OK ]
16 ....chk 2->0 cestab after flush [ OK ]
This error is caused by a broken pipe. When the output of 'ss' is processed
by grep, 'head -n 1' will exit immediately after getting the first line,
causing the subsequent pipe to close. At this time, if 'grep' is still
trying to write data to the closed pipe, it will trigger a SIGPIPE signal,
causing a write error.
One solution is not to use this problematic "head -n 1" command, but to use
mptcp_lib_get_info_value() helper defined in mptcp_lib.sh to get the value
of 'token'.
Fixes: ba24001665 ("selftests: mptcp: add a test for mptcp_diag_dump_one")
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Tested-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20250421-net-mptcp-pm-defer-freeing-v1-2-e731dc6e86b9@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When path manager entries are deleted from the local address list, they
are first unlinked from the address list using list_del_rcu(). The
entries must not be freed until after the RCU grace period, but the
existing code immediately frees the entry.
Use kfree_rcu_mightsleep() and adjust sk_omem_alloc in open code instead
of using the sock_kfree_s() helper. This code path is only called in a
netlink handler, so the "might sleep" function is preferable to adding
a rarely-used rcu_head member to struct mptcp_pm_addr_entry.
Fixes: 88d0973163 ("mptcp: drop free_list for deleting entries")
Cc: stable@vger.kernel.org
Signed-off-by: Mat Martineau <martineau@kernel.org>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20250421-net-mptcp-pm-defer-freeing-v1-1-e731dc6e86b9@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit a402006d48 ("misc: pci_endpoint_test: Remove global 'irq_type'
and 'no_msi'") changed so that the default IRQ vector requested by
pci_endpoint_test_probe() was no longer the module param 'irq_type', but
instead test->irq_type. test->irq_type is by default IRQ_TYPE_UNDEFINED
(until someone calls ioctl(PCITEST_SET_IRQTYPE)).
However, the commit also changed so that after initializing test->irq_type
to IRQ_TYPE_UNDEFINED, it also overrides it with driver_data->irq_type, if
the PCI device and vendor ID provides driver_data.
This causes a regression for PCI device and vendor IDs that do not provide
driver_data, and the host side pci_endpoint_test_driver driver failed to
probe on such platforms:
pci-endpoint-test 0001:01:00.0: Invalid IRQ type selected
pci-endpoint-test 0001:01:00.0: probe with driver pci-endpoint-test failed with error -22
Considering that the pci endpoint selftests and the old pcitest.sh always
call ioctl(PCITEST_SET_IRQTYPE) before performing any test that requires
IRQs, fix the regression by removing the allocation of IRQs in
pci_endpoint_test_probe(). The IRQ allocation will occur when
ioctl(PCITEST_SET_IRQTYPE) is called.
A positive side effect of this is that even if the endpoint controller has
issues with IRQs, the user can do still do all the tests/ioctls() that do
not require working IRQs, e.g. PCITEST_BAR and PCITEST_BARS.
This also means that we can remove the now unused irq_type from
driver_data. The irq_type will always be the one configured by the user
using ioctl(PCITEST_SET_IRQTYPE). (A user that does not know, or care
which irq_type that is used, can use PCITEST_IRQ_TYPE_AUTO. This has
superseded the need for a default irq_type in driver_data.)
[bhelgaas: add probe failure details]
Fixes: a402006d48 ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20250416142825.336554-2-cassel@kernel.org
With the new large sector size support, it's now the case that
set_blocksize can change i_blksize and the folio order in a manner that
conflicts with a concurrent reader and causes a kernel crash.
Specifically, let's say that udev-worker calls libblkid to detect the
labels on a block device. The read call can create an order-0 folio to
read the first 4096 bytes from the disk. But then udev is preempted.
Next, someone tries to mount an 8k-sectorsize filesystem from the same
block device. The filesystem calls set_blksize, which sets i_blksize to
8192 and the minimum folio order to 1.
Now udev resumes, still holding the order-0 folio it allocated. It then
tries to schedule a read bio and do_mpage_readahead tries to create
bufferheads for the folio. Unfortunately, blocks_per_folio == 0 because
the page size is 4096 but the blocksize is 8192 so no bufferheads are
attached and the bh walk never sets bdev. We then submit the bio with a
NULL block device and crash.
Therefore, truncate the page cache after flushing but before updating
i_blksize. However, that's not enough -- we also need to lock out file
IO and page faults during the update. Take both the i_rwsem and the
invalidate_lock in exclusive mode for invalidations, and in shared mode
for read/write operations.
I don't know if this is the correct fix, but xfs/259 found it.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
Tested-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Link: https://lore.kernel.org/r/174543795699.4139148.2086129139322431423.stgit@frogsfrogsfrogs
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This was triggered by one of my mis-uses causing odd build warnings on
sparc in linux-next, but while figuring out why the "obviously correct"
use of cc-option caused such odd breakage, I found eight other cases of
the same thing in the tree.
The root cause is that 'cc-option' doesn't work for checking negative
warning options (ie things like '-Wno-stringop-overflow') because gcc
will silently accept options it doesn't recognize, and so 'cc-option'
ends up thinking they are perfectly fine.
And it all works, until you have a situation where _another_ warning is
emitted. At that point the compiler will go "Hmm, maybe the user
intended to disable this warning but used that wrong option that I
didn't recognize", and generate a warning for the unrecognized negative
option.
Which explains why we have several cases of this in the tree: the
'cc-option' test really doesn't work for this situation, but most of the
time it simply doesn't matter that ity doesn't work.
The reason my recently added case caused problems on sparc was pointed
out by Thomas Weißschuh: the sparc build had a previous explicit warning
that then triggered the new one.
I think the best fix for this would be to make 'cc-option' a bit smarter
about this sitation, possibly by adding an intentional warning to the
test case that then triggers the unrecognized option warning reliably.
But the short-term fix is to replace 'cc-option' with an existing helper
designed for this exact case: 'cc-disable-warning', which picks the
negative warning but uses the positive form for testing the compiler
support.
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Link: https://lore.kernel.org/all/20250422204718.0b4e3f81@canb.auug.org.au/
Explained-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Michael Larabel reported [1] a nginx performance regression in v6.15-rc3
and bisected it to commit 51339d99c0 ("locking/local_lock, mm: replace
localtry_ helpers with local_trylock_t type")
The problem is the _Generic() usage with a default association that
masks the fact that "local_trylock_t *" association is not being
selected as expected. Replacing the default with the only other
expected type "local_lock_t *" reveals the underlying problem:
include/linux/local_lock_internal.h:174:26: error: ‘_Generic’ selector of type ‘__seg_gs local_lock_t *’ is not compatible with any association
The local_locki's are part of __percpu structures and thus the __percpu
attribute is needed to associate the type properly. Add the attribute
and keep the default replaced to turn any further mismatches into
compile errors.
The failure to recognize local_try_lock_t in __local_lock_release()
means that a local_trylock[_irqsave]() operation will set tl->acquired
to 1 (there's no _Generic() part in the trylock code), but then
local_unlock[_irqrestore]() will not set tl->acquired back to 0, so
further trylock operations will always fail on the same cpu+lock, while
non-trylock operations continue to work - a lockdep_assert() is also not
being executed in the _Generic() part of local_lock() code.
This means consume_stock() and refill_stock() operations will fail
deterministically, resulting in taking the slow paths and worse
performance.
Fixes: 51339d99c0 ("locking/local_lock, mm: replace localtry_ helpers with local_trylock_t type")
Reported-by: Michael Larabel <Michael@phoronix.com>
Closes: https://www.phoronix.com/review/linux-615-nginx-regression/2 [1]
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull virtio fixes from Michael Tsirkin:
"A small number of fixes:
- virtgpu is exempt from reset shutdown fow now - a more complete fix
is in the works
- spec compliance fixes in:
- virtio-pci cap commands
- vhost_scsi_send_bad_target
- virtio console resize
- missing locking fix in vhost-scsi
- virtio ring - a KCSAN false positive fix
- VHOST_*_OWNER documentation fix"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
vhost-scsi: Fix vhost_scsi_send_status()
vhost-scsi: Fix vhost_scsi_send_bad_target()
vhost-scsi: protect vq->log_used with vq->mutex
vhost_task: fix vhost_task_create() documentation
virtio_console: fix order of fields cols and rows
virtio_console: fix missing byte order handling for cols and rows
virtgpu: don't reset on shutdown
virtio_ring: Fix data race by tagging event_triggered as racy for KCSAN
vhost: fix VHOST_*_OWNER documentation
virtio_pci: Use self group type for cap commands
Recently _pgd_alloc() was switched from using __get_free_pages() to
pagetable_alloc_noprof(), which might return a compound page in case
the allocation order is larger than 0.
On x86 this will be the case if CONFIG_MITIGATION_PAGE_TABLE_ISOLATION
is set, even if PTI has been disabled at runtime.
When running as a Xen PV guest (this will always disable PTI), using
a compound page for a PGD will result in VM_BUG_ON_PGFLAGS being
triggered when the Xen code tries to pin the PGD.
Fix the Xen issue together with the not needed 8k allocation for a
PGD with PTI disabled by replacing PGD_ALLOCATION_ORDER with an
inline helper returning the needed order for PGD allocations.
Fixes: a9b3c355c2 ("asm-generic: pgalloc: provide generic __pgd_{alloc,free}")
Reported-by: Petr Vaněk <arkamar@atlas.cz>
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Tested-by: Petr Vaněk <arkamar@atlas.cz>
Cc:stable@vger.kernel.org
Link: https://lore.kernel.org/all/20250422131717.25724-1-jgross%40suse.com
There is a spelling mistake in a DRM_DEV_DEBUG_KMS message. Fix it.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Corrected a spelling mistake in the exynos_drm_fimd driver to improve code
readability. No functional changes were made.
Signed-off-by: Anindya Sundar Gayen <anindya.sg@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
In the vidi_get_modes() function, if either drm_edid_dup() or
drm_edid_alloc() fails, the function will immediately return 0,
indicating that no display modes can be retrieved. However, in
the event of failure in these two functions, it is still necessary
to call the subsequent drm_edid_connector_update() function with
a NULL drm_edid as an argument. This ensures that operations such
as connector settings are performed in its callee function,
_drm_edid_connector_property_update. To maintain the integrity of
the operation, redundant error handling needs to be removed.
Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Merge ARM cpufreq fixes for 6.15-rc from Viresh Kumar:
"- Fix possible out-of-bound / null-ptr-deref in drivers (Andre Przywara
and Henry Martin).
- Fix Kconfig issues with compile-test (Johan Hovold and Krzysztof
Kozlowski).
- Fix invalid return value in .get() (Marc Zyngier).
- Add SM8650 to cpufreq-dt-platdev blocklist (Pengyu Luo)."
* tag 'cpufreq-arm-fixes-6.15-rc' of git://git.kernel.org/pub/scm/linux/kernel/git/vireshk/pm:
cpufreq: fix compile-test defaults
cpufreq: cppc: Fix invalid return value in .get() callback
cpufreq: scpi: Fix null-ptr-deref in scpi_cpufreq_get_rate()
cpufreq: scmi: Fix null-ptr-deref in scmi_cpufreq_get_rate()
cpufreq: apple-soc: Fix null-ptr-deref in apple_soc_cpufreq_get_rate()
cpufreq: Do not enable by default during compile testing
cpufreq: Add SM8650 to cpufreq-dt-platdev blocklist
cpufreq: sun50i: prevent out-of-bounds access
Normally do_lock_mount(path, _) is locking a mountpoint pinned by
*path and at the time when matching unlock_mount() unlocks that
location it is still pinned by the same thing.
Unfortunately, for 'beneath' case it's no longer that simple -
the object being locked is not the one *path points to. It's the
mountpoint of path->mnt. The thing is, without sufficient locking
->mnt_parent may change under us and none of the locks are held
at that point. The rules are
* mount_lock stabilizes m->mnt_parent for any mount m.
* namespace_sem stabilizes m->mnt_parent, provided that
m is mounted.
* if either of the above holds and refcount of m is positive,
we are guaranteed the same for refcount of m->mnt_parent.
namespace_sem nests inside inode_lock(), so do_lock_mount() has
to take inode_lock() before grabbing namespace_sem. It does
recheck that path->mnt is still mounted in the same place after
getting namespace_sem, and it does take care to pin the dentry.
It is needed, since otherwise we might end up with racing mount --move
(or umount) happening while we were getting locks; in that case
dentry would no longer be a mountpoint and could've been evicted
on memory pressure along with its inode - not something you want
when grabbing lock on that inode.
However, pinning a dentry is not enough - the matching mount is
also pinned only by the fact that path->mnt is mounted on top it
and at that point we are not holding any locks whatsoever, so
the same kind of races could end up with all references to
that mount gone just as we are about to enter inode_lock().
If that happens, we are left with filesystem being shut down while
we are holding a dentry reference on it; results are not pretty.
What we need to do is grab both dentry and mount at the same time;
that makes inode_lock() safe *and* avoids the problem with fs getting
shut down under us. After taking namespace_sem we verify that
path->mnt is still mounted (which stabilizes its ->mnt_parent) and
check that it's still mounted at the same place. From that point
on to the matching namespace_unlock() we are guaranteed that
mount/dentry pair we'd grabbed are also pinned by being the mountpoint
of path->mnt, so we can quietly drop both the dentry reference (as
the current code does) and mnt one - it's OK to do under namespace_sem,
since we are not dropping the final refs.
That solves the problem on do_lock_mount() side; unlock_mount()
also has one, since dentry is guaranteed to stay pinned only until
the namespace_unlock(). That's easy to fix - just have inode_unlock()
done earlier, while it's still pinned by mp->m_dentry.
Fixes: 6ac3928156 "fs: allow to mount beneath top mount" # v6.5+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Fix off-by-one bug in the last page calculation for src and dst.
Reported-by: Nhat Pham <nphamcs@gmail.com>
Fixes: 2d3553ecb4 ("crypto: scomp - Remove support for some non-trivial SG lists")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When pausing rx (e.g. set up xdp, xsk pool, rx resize), we call
napi_disable() on the receive queue's napi. In delayed refill_work, it
also calls napi_disable() on the receive queue's napi. When
napi_disable() is called on an already disabled napi, it will sleep in
napi_disable_locked while still holding the netdev_lock. As a result,
later napi_enable gets stuck too as it cannot acquire the netdev_lock.
This leads to refill_work and the pause-then-resume tx are stuck
altogether.
This scenario can be reproducible by binding a XDP socket to virtio-net
interface without setting up the fill ring. As a result, try_fill_recv
will fail until the fill ring is set up and refill_work is scheduled.
This commit adds virtnet_rx_(pause/resume)_all helpers and fixes up the
virtnet_rx_resume to disable future and cancel all inflights delayed
refill_work before calling napi_disable() to pause the rx.
Fixes: 413f0271f3 ("net: protect NAPI enablement with netdev_lock()")
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Bui Quang Minh <minhquangbui99@gmail.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://patch.msgid.link/20250417072806.18660-2-minhquangbui99@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A network restart test on a router led to an out-of-memory condition,
which was traced to a memory leak in the PHY LED trigger code.
The root cause is misuse of the devm API. The registration function
(phy_led_triggers_register) is called from phy_attach_direct, not
phy_probe, and the unregister function (phy_led_triggers_unregister)
is called from phy_detach, not phy_remove. This means the register and
unregister functions can be called multiple times for the same PHY
device, but devm-allocated memory is not freed until the driver is
unbound.
This also prevents kmemleak from detecting the leak, as the devm API
internally stores the allocated pointer.
Fix this by replacing devm_kzalloc/devm_kcalloc with standard
kzalloc/kcalloc, and add the corresponding kfree calls in the unregister
path.
Fixes: 3928ee6485 ("net: phy: leds: Add support for "link" trigger")
Fixes: 2e0bc452f4 ("net: phy: leds: add support for led triggers on phy link state change")
Signed-off-by: Hao Guan <hao.guan@siflower.com.cn>
Signed-off-by: Qingfang Deng <qingfang.deng@siflower.com.cn>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250417032557.2929427-1-dqfext@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
As a result of an email from the fbnic author, I reviewed the phylink
documentation, and I have decided to clarify the wording in the
mac_link_(up|down)() kernel documentation as this was written from the
point of view of mvneta/mvpp2 and is misleading.
The documentation talks about forcing the link - indeed, this is what
is done in the mvneta and mvpp2 drivers but not at the physical layer
but the MACs idea, which has the effect of only allowing or stopping
packet flow at the MAC. This "link" needs to be controlled when using
a PHY or fixed link to start or stop packet flow at the MAC. However,
as the MAC and PCS are tightly integrated, if the MACs idea of the
link is forced down, it has the side effect that there is no way to
determine that the media link has come up - in this mode, the MAC must
be allowed to follow its built-in PCS so we can read the link state.
Frame the documentation in more generic terms, to avoid the thought
that the physical media link to the partner needs in some way to be
forced up or down with these calls; it does not. If that were to be
done, it would be a self-fulfilling prophecy - e.g. if the media link
goes down, then mac_link_down() will be called, and if the media link
is then placed into a forced down state, there is no possibility
that the media link will ever come up again - clearly this is a wrong
interpretation.
These methods are notifications to the MAC about what has happened to
the media link state - either from the PHY, or a PCS, or whatever
mechanism fixed-link is using. Thus, reword them to get away from
talking about changing link state to avoid confusion with media link
state.
This is not a change of any requirements of these methods.
Also, remove the obsolete references to EEE for these methods, we now
have the LPI functions for configuring the EEE parameters which
renders this redundant, and also makes the passing of "phy" to the
mac_link_up() function obsolete.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/E1u5Ah5-001GO1-7E@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When WoL is enabled, we update the software state in phylink to
indicate that the link is down, and disable the resolver from
bringing the link back up.
On resume, we attempt to bring the overall state into consistency
by calling the .mac_link_down() method, but this is wrong if the
link was already down, as phylink strictly orders the .mac_link_up()
and .mac_link_down() methods - and this would break that ordering.
Fixes: f97493657c ("net: phylink: add suspend/resume support")
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Tested-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/E1u55Qf-0016RN-PA@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
[Why]
Recent findings show negligible power savings between IPS2 and RCG
during static desktop. In fact, DCN related clocks are higher
when IPS2 is enabled vs RCG.
RCG_IN_ACTIVE is also the default policy for another OS supported by
DC, and it has faster entry/exit.
[How]
Remove previous logic that checked for IPS2 support, and just default
to `DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF`.
Fixes: 199888aa25 ("drm/amd/display: Update IPS default mode for DCN35/DCN351")
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 8f772d79ef)
Cc: stable@vger.kernel.org
[Why/How]
LTTPR are required to program DPCD 0000Eh to 0x4 (16ms) upon AUX read
reply to this register. Since old Sinks witih DPCD rev 1.1 and earlier
may not support this register, assume the mandatory value is programmed
by the LTTPR to avoid AUX timeout issues.
Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 1594b60d74)
[Why]
The ACPI EDID in the BIOS of a Lenovo laptop includes 3 blocks, but
dm_helpers_probe_acpi_edid() has a start that is 'char'. The 3rd
block index starts after 255, so it can't be indexed properly.
This leads to problems with the display when the EDID is parsed.
[How]
Change the variable type to 'short' so that larger values can be indexed.
Cc: Renjith Pananchikkal <renjith.pananchikkal@amd.com>
Reported-by: Mark Pearson <mpearson@lenovo.com>
Suggested-by: David Ober <dober@lenovo.com>
Fixes: c6a837088b ("drm/amd/display: Fetch the EDID from _DDC if available for eDP")
Reviewed-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit a918bb4a90)
Cc: stable@vger.kernel.org
[Why]
Urgent latency adjustment was disabled on DCN35 due to issues with P0
enablement on some platforms. Without urgent latency, underflows occur
when doing certain high timing configurations. After testing, we found
that reenabling urgent latency didn't reintroduce p0 support on multiple
platforms.
[How]
renable urgent latency on DCN35 and setting it to 3000 Mhz.
This reverts commit 3412860cc4.
Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Signed-off-by: Nicholas Susanto <nsusanto@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit cd74ce1f0c)
