It seems that initalising afs_vnode::lock_work a single time in the slab's
init function isn't sufficient for work_structs. This results in the
DEBUG_OBJECTS debugging stuff producing a warning occasionally when running
the generic/131 xfstest:
ODEBUG: activate not available (active state 0) object: 0000000016d8760f object type: work_struct hint: afs_lock_work+0x0/0x220
WARNING: lib/debugobjects.c:629 at debug_print_object+0x4b/0x90, CPU#3: locktest/7695
...
CPU: 3 UID: 0 PID: 7695 Comm: locktest Tainted: G S 7.1.0-build3+ #2771 PREEMPT
...
RIP: 0010:debug_print_object+0x65/0x90
...
Call Trace:
<TASK>
? __pfx_afs_lock_work+0x10/0x10
debug_object_activate+0x122/0x170
insert_work+0x25/0x60
__queue_work+0x2e0/0x340
queue_delayed_work_on+0x48/0x70
afs_fl_release_private+0x57/0x70
locks_release_private+0x5c/0xa0
locks_free_lock+0xe/0x20
posix_lock_inode+0x55f/0x5b0
locks_lock_inode_wait+0x81/0x140
? file_write_and_wait_range+0x50/0x70
afs_lock+0xcd/0x110
fcntl_setlk+0x10d/0x260
do_fcntl+0x24e/0x5b0
__do_sys_fcntl+0x6a/0x90
do_syscall_64+0x11e/0x310
entry_SYSCALL_64_after_hwframe+0x71/0x79
Fix this by reinitialising ->lock_work after allocating an inode.
Also, flush ->lock_work when the inode is being evicted to make sure it's
not still running.
Fixes: e8d6c55412 ("AFS: implement file locking")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://patch.msgid.link/20260622090856.2746629-13-dhowells@redhat.com
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Thomas Gleixner <tglx@kernel.org>
cc: linux-afs@lists.infradead.org
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
afs_do_lookup() and afs_do_lookup_one() use the same directory parsing code
as afs_readdir() and were supplying alternative dir_context actors to
retrieve dirents, but because lookup needs the vnode's uniquifier as part
of the reference, but not the DT flags, the uniquifier was being passed in
the dt flags argument to the lookup actors.
Unfortunately, commit c644bce62b, added to fix overlayfs with fuse, broke
this by masking off part of the uniquifier. This doesn't matter enough to
be directly noticeable, instead causing bulk advance inode lookups to fail
(which are retried later) and may cause dir revalidation to malfunction if
the uniquifier is changed by masking.
Fix this by making the afs directory parsing code take special ->actor
values of AFS_LOOKUP or AFS_LOOKUP_ONE instead that tell it to call
afs_lookup_filldir() or afs_lookup_one_filldir() directly rather than going
through dir_emit(). dir_emit() is still used for readdir.
Fixes: c644bce62b ("readdir: require opt-in for d_type flags")
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://patch.msgid.link/20260622090856.2746629-11-dhowells@redhat.com
cc: Amir Goldstein <amir73il@gmail.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
Regular AFS files correctly use afs_file_aops which have release_folio
set as netfs_release_folio, so AS_RELEASE_ALWAYS is valid for them
when fscache is enabled (set via afs_vnode_set_cache()).
Symlinks and mountpoints in AFS use afs_dir_aops, which does not provide
a release_folio callback. However, afs_apply_status() unconditionally
calls mapping_set_release_always() for these.
In such case when memory management code attempts to release folios,
filemap_release_folio() checks folio_needs_release() which
returns true due to AS_RELEASE_ALWAYS being set. Since there is no
release_folio callback, it falls through to try_to_free_buffers(),
which at present expects buffer_heads to be not null. For symlinks
and mountpoints without buffer_heads, this causes pointer dereference.
[dh: Added more bits that were missed]
Fixes: eae9e78951 ("afs: Use netfslib for symlinks, allowing them to be cached")
Signed-off-by: Deepakkumar Karn <dkarn@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://patch.msgid.link/20260622090856.2746629-6-dhowells@redhat.com
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
Add a regression test for the fsuidgid_has_mapping() check in
vfs_tmpfile(). It idmaps a detached tmpfs mount so that the
caller-visible id range [0, 10000) maps onto the on-disk range
[10000, 20000) and checks that:
- a caller whose fsuid/fsgid fall outside that range cannot create an
O_TMPFILE through the mount and gets -EOVERFLOW instead of an inode
owned by (uid_t)-1;
- a mapped caller can create an O_TMPFILE, link it into the namespace,
and the ownership round-trips through the mount idmap: it is reported
as 0 through the mount and stored as 10000 on the underlying tmpfs.
The test runs entirely as root and uses setfsuid()/setfsgid() to become
the unmapped caller, so it needs no helper user. The layer directory is
world-writable so that an unmapped caller still clears the directory
permission check and reaches the fsuidgid_has_mapping() test.
Link: https://patch.msgid.link/20260615-work-idmapped-tmpfile-v1-2-754a94d81f83@kernel.org
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
vfs_tmpfile() never checked that the caller's fsuid and fsgid map into
the filesystem. On an idmapped mount whose idmapping does not cover the
caller's fs{u,g}id, the ->tmpfile() instance initializes the new inode
through inode_init_owner(), where mapped_fsuid()/mapped_fsgid() return
INVALID_UID/INVALID_GID, and the tmpfile ends up owned by (uid_t)-1.
Every other creation path already refuses this: may_o_create() (O_CREAT)
and may_create_dentry() (mkdir, mknod, symlink, link) bail out with
-EOVERFLOW via fsuidgid_has_mapping() precisely so that an object cannot
be created with an owner the filesystem cannot represent. An O_TMPFILE
is no exception: it is created I_LINKABLE and linkat(2) can splice it
into the namespace afterwards, so the same guarantee must hold.
Add the missing fsuidgid_has_mapping() check to vfs_tmpfile(). On a
non-idmapped mount the caller's fs{u,g}id always map in the superblock's
user namespace, so this is a no-op there and only takes effect on an
idmapped mount that does not map the caller. It applies to every
filesystem that sets FS_ALLOW_IDMAP and implements ->tmpfile() (tmpfs,
ext4, btrfs, xfs, f2fs, ...), and to overlayfs, whose upper-layer
tmpfile creation funnels through vfs_tmpfile() via backing_tmpfile_open().
Fixes: 8e5389132a ("fs: introduce fsuidgid_has_mapping() helper")
Link: https://patch.msgid.link/20260615-work-idmapped-tmpfile-v1-1-754a94d81f83@kernel.org
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
The cache manager callback path now attaches the server record to an
incoming call through the rxrpc peer's app data. That association is
not guaranteed to exist for every callback request, and most callback
handlers already tolerate that case.
Make CB.InitCallBackState3 follow the same pattern by checking whether a
server record was attached before using it. If the peer is not mapped
to a server record, trace the request and ignore it, matching the
existing behaviour for other unmatched callback requests.
This keeps the callback handler consistent with the rest of the cache
manager service and avoids depending on peer state that may not be
available for a given request.
Fixes: 40e8b52fe8 ("afs: Use the per-peer app data provided by rxrpc")
Cc: stable@kernel.org
Reported-by: Yuan Tan <yuantan098@gmail.com>
Reported-by: Yifan Wu <yifanwucs@gmail.com>
Reported-by: Juefei Pu <tomapufckgml@gmail.com>
Reported-by: Xin Liu <bird@lzu.edu.cn>
Signed-off-by: Nan Li <tonanli66@gmail.com>
Signed-off-by: Ren Wei <n05ec@lzu.edu.cn>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://patch.msgid.link/20260622090856.2746629-2-dhowells@redhat.com
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
On i.MX8MQ, the MIPI CSI reset lines are active-low and not self-clearing.
Writing '0' asserts reset and it remains asserted until explicitly
deasserted by software.
This driver previously treated the MIPI CSI reset signals as active-high,
which led to incorrect reset assert/deassert sequencing. This issue was
exposed by commit 6d79bb8fd2 ("media: imx8mq-mipi-csi2: Explicitly
release reset").
Fix this by reflecting the correct reset polarity and ensuring proper
reset handling.
Fixes: c979dbf599 ("reset: imx7: Add support for i.MX8MQ IP block variant")
Cc: stable@vger.kernel.org # 6d79bb8fd2: media: imx8mq-mipi-csi2: Explicitly release reset
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Robby Cai <robby.cai@nxp.com>
Reviewed-by: Guoniu Zhou <guoniu.zhou@oss.nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
In sunxi_reset_init(), when ioremap() fails, the memory region obtained
via request_mem_region() is not released, leading to a resource leak.
Add an err_mem_region label to properly release the memory region before
freeing the data structure.
Fixes: 8f1ae77f46 ("reset: Add Allwinner SoCs Reset Controller Driver")
Cc: stable@vger.kernel.org
Signed-off-by: Zhao Dongdong <zhaodongdong@kylinos.cn>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Add COMBOPHY_RESET definition at index 38 for the combo PHY reset
control on Altera Agilex5 SoCs. This reset is used by peripherals
such as the SD/eMMC controller that share the combo PHY.
Signed-off-by: Tanmay Kathpalia <tanmay.kathpalia@altera.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
According to SpacemiT K3's updated docs, the USB2 ahb reset and USB2 bus
clock enable bit was wrongly swapped, the correct one should be:
Register : APMU_USB_CLK_RES_CTRL
bit[1] : usb2_port_bus_clk_en
bit[0] : usb2_port_ahb_rstn
Fixes: a0e0c2f8c5 ("reset: spacemit: k3: Decouple composite reset lines")
Reported-by: Junzhong Pan <panjunzhong@linux.spacemit.com>
Signed-off-by: Yixun Lan <dlan@kernel.org>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Building mlx5 on s390 shows a rather high stack usage that can exceed
the warning limit when that is set to a lower but still reasonable value:
drivers/infiniband/hw/mlx5/wr.c:1051:5: error: stack frame size (1328) exceeds limit (1280) in 'mlx5_ib_post_send' [-Werror,-Wframe-larger-than]
The problem here is 'struct ib_reg_wr' on the stack of
handle_reg_mr_integrity(), which gets inlined into mlx5_ib_post_send()
along with a number of smaller functions.
Keeping the inner function out of line like gcc does avoids the
warning and reduces the total stack usage in other functions called
from mlx5_ib_post_send(), though handle_reg_mr_integrity() itself
still has the same problem as before.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://patch.msgid.link/20260612201611.4127750-1-arnd@kernel.org
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Probe information replies contain a firmware-provided element count. IPC3
uses that count to copy an array, then returns the unchecked count to its
caller. A short reply can therefore make the caller walk beyond the copied
array.
IPC4 similarly uses the count both to allocate the destination array and
to walk the reply. On 32-bit systems the allocation size can wrap, while on
all systems an excessive count reads beyond the reply payload.
Validate each count against the actual reply size before copying or
allocating the array, and use kcalloc() for the IPC4 allocation.
Signed-off-by: Yousef Alhouseen <alhouseenyousef@gmail.com>
Link: https://patch.msgid.link/20260628000329.18606-1-alhouseenyousef@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
tascam_disconnect() cancels capture_work and midi_in_work before
usb_kill_anchored_urbs() kills the capture/MIDI-in URBs. Those URBs
self-resubmit, and their completion handlers reschedule the work.
A URB that completes in the small window between cancel_work_sync() and
usb_kill_anchored_urbs() therefore re-arms the work after its only
cancel. Nothing cancels it again before snd_card_free() frees the
card-private tascam structure, so the work handler then runs on freed
memory.
Kill the anchored URBs before cancelling the work; once the work is
cancelled no remaining URB can complete to re-arm it.
Fixes: c1bb0c13e4 ("ALSA: usb-audio: us144mkii: Implement audio capture and decoding")
Cc: stable@vger.kernel.org
Assisted-by: Claude:claude-opus-4-8
Signed-off-by: HyeongJun An <sammiee5311@gmail.com>
Link: https://patch.msgid.link/20260701095231.1020811-1-sammiee5311@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
IORING_OP_MSG_RING with IORING_MSG_RING_FLAGS_PASS allows a sender to
pass completion flags through sqe->file_index. If the sender sets
IORING_CQE_F_32 in file_index, the target-side completion path treats it
as a 32b CQE and writes big_cqe[0] and big_cqe[1] into the CQ ring
regardless of whether the target ring was created with
IORING_SETUP_CQE32 or IORING_SETUP_CQE_MIXED.
On a normal 16b CQE ring, this writes 16 extra bytes (two u64 big_cqe
fields) into the next CQE slot in the ring buffer. As the receiving ring
doesn't understand 32b CQEs, this is incorrect and they should be
rejected.
Fixes: cbeb47a7b5 ("io_uring/msg_ring: Pass custom flags to the cqe")
Signed-off-by: Melbin K Mathew <mlbnkm1@gmail.com>
Link: https://patch.msgid.link/20260701081145.196730-1-mlbnkm1@gmail.com
[axboe: edit commit message]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
__xfs_buf_ioend can only resubmit the buffer for asynchronous
writes, which means the retry handling xfs_buf_iowait is not needed.
Because of this can stop returning a value from __xfs_buf_ioend and
just release the buffer for async I/O that does not require retries.
Also drop the __-prefix now that the semantics are straight forward.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Synchronous readers and writers already run __xfs_buf_ioend from
xfs_buf_iowait after being woken through bp->b_iowait, so we
should not call it here, which can lead to double completions.
Fixes: 4b90de5bc0 ("xfs: reduce context switches for synchronous buffered I/O")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
There are two callers of xfs_buf_ioend, one of which always has the
XBF_ASYNC flag set. Open code the logic in both callers to prepare for a
bug fix.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Move setting the ASYNC flag into xfs_buf_ioend_fail, assert that the
buffer is locked as expected, and drop the confusing _ioend in the
name.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Commit 292db66afd ("ACPICA: Unbreak tools build after switching over
to strscpy_pad()") added an #ifdef based on a __KERNEL__ check which is
sort of nasty to the acpi_ut_safe_strncpy() definition to unbreak ACPICA
tools builds broken by commit 97f7d3f9c9 ("ACPICA: Replace strncpy()
with strscpy_pad() in acpi_ut_safe_strncpy()"). However, that #ifdef
effectively produces dead code when tools are built because they don't
call acpi_ut_safe_strncpy().
Accordingly, drop the existing definition of acpi_ut_safe_strncpy() and
define it as a strscpy_pad() alias.
Fixes: 292db66afd ("ACPICA: Unbreak tools build after switching over to strscpy_pad()")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
[ rjw: Tweak the changelog ]
Link: https://patch.msgid.link/12941764.O9o76ZdvQC@rafael.j.wysocki
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
A hotplug or link-loss event can tear down the MST topology
(setting mgr->mst_state = false and mgr->mst_primary = NULL) concurrently
with a caller invoking drm_dp_mst_topology_queue_probe(). Since the check
is already performed under mgr->lock, the condition is not a programming
error but a valid race -- the topology was valid when the caller decided
to call this function, but was torn down before the lock was acquired.
Replace the drm_WARN_ON() with a graceful early return. This eliminates
spurious kernel warnings and the resulting compositor crashes observed
when connecting/disconnecting DP MST monitors, while keeping the correct
behavior of doing nothing when MST is not active. A drm_dbg_mst() trace
is added so the skipped probe remains observable under MST debug logging.
The existing WARN_ON(mgr->mst_primary) in drm_dp_mst_topology_mgr_set_mst()
already catches the case where the topology is initialized twice, so no
diagnostic coverage is lost.
Fixes: dbaeef363e ("drm/dp_mst: Add a helper to queue a topology probe")
Cc: Imre Deak <imre.deak@intel.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: stable@vger.kernel.org
Cc: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Jonas Emilsson <jonas.emilsson@gmail.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/all/20260503034533.1023686-1-jonas.emilsson@gmail.com
Acked-by: Imre Deak <imre.deak@intel.com>
Link: https://patch.msgid.link/20260622140532.526722-1-luciano.coelho@intel.com
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
If xfs_metadir_create() fails in xfs_dqinode_metadir_create(), the current
code returns directly, leaking the allocated update and transaction state.
If the subsequent commit fails, the caller-owned inode reference is left
behind.
Fix this memory leak by routing the create failure path through
xfs_metadir_cancel(). For both create and commit failures, finish and
release any inode returned to the caller, mirroring the unwind pattern in
xfs_metadir_mkdir().
The bug was first flagged by an experimental analysis tool we are
developing for kernel memory-management bugs while analyzing
v6.13-rc1. The tool is still under development and is not yet publicly
available. Manual inspection confirms that the bug is still
present in v7.1.1.
An x86_64 allyesconfig build showed no new warnings. Runtime validation
used kprobe fault injection during `mount -o uquota` on a metadir XFS
image. Injecting xfs_metadir_create() reproduced the old active-update path
that left mount stuck later in mount setup; after this change, the same
injection reported cancel_hits=1 and irele_hits=1. Injecting
xfs_metadir_commit() exercised the old inode-reference leak path; after
this change, it reported irele_hits=1.
Fixes: e80fbe1ad8 ("xfs: use metadir for quota inodes")
Cc: stable@vger.kernel.org # v6.13
Signed-off-by: Dawei Feng <dawei.feng@seu.edu.cn>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
xfs_qm_dqpurge() gets a locked buffer from xfs_dquot_use_attached_buf().
If xfs_qm_dqflush() fails, the error path skips xfs_buf_relse() and then
calls xfs_dquot_detach_buf(), which tries to lock the same buffer again.
Release the buffer after xfs_qm_dqflush() returns so the error path drops
the caller hold and unlocks the buffer before the dquot is detached,
matching the other dqflush callers.
Fixes: a40fe30868 ("xfs: separate dquot buffer reads from xfs_dqflush")
Cc: stable@vger.kernel.org # v6.13+
Signed-off-by: Yingjie Gao <gaoyingjie@uniontech.com>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
We should treat the buffer that caused a shutdown the same as handling
buffers after a shutdown, so use the same stale && !DONE logic here.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
This better integrates with the other failure handling in xfs_buf_submit,
and prepares for a better API in xfs_buf_ioend_fail.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
The pack allocator only flushes predictors when reusing a dirty pack for
cBPF, eBPF allocations never trigger a flush. Currently, eBPF picks the
first free pack, which could be a clean pack. As an optimization, leaving
a clean pack for cBPF can avoid flushes.
Prefer dirty packs for eBPF and keep clean packs free for cBPF. This
mirrors the existing cBPF preference for clean packs: each program kind
prefers the pack that avoids an extra flush, and falls back to the other
kind only when no preferred pack has room. eBPF reuse of a dirty pack is
harmless since eBPF being privileged does not flush.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Currently BPF pack allocator picks the chunks from the first available
pack. While this is okay, it naturally leads to more frequent flushes
when there are multiple packs in the system that weren't used since the
last flush.
As an optimization prefer allocating the new programs from packs that
are unused since last flush. When all packs are dirty, allocation forces
a flush and marks all packs clean.
Below are some future optimizations ideas:
1. Currently, the "dirty" tracking is only done at the pack-level.
Flush frequency can further be reduced with chunk-level tracking.
This requires a new bitmap per-pack to track the dirty state.
2. IBPB flush is done on all CPUs, even if only a single CPU ran the
BPF program. On a system with hundreds of CPUs this could be a
major bottleneck forcing hundreds of IPIs to deliver the flush.
The solution is to track the CPUs where a BPF program ran, and
issue IBPB only on those CPUs.
3. Avoid IBPB when flush is already done at other sources (e.g.
context switch).
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
bpf_prog_pack_alloc() issues IBPB on all CPUs on every cBPF allocation,
even when reusing chunks from an existing pack where no new memory was
touched since the last IBPB.
Since IBPB on all CPUs is heavy, Dave Hansen suggested to track allocation
since last IBPB, and only issue IBPB at reuse for the chunks that have not
seen an IBPB since they were last freed.
Track per-pack whether an IBPB is needed via arch_flush_needed. Set it when
allocating a chunk, reset on IBPB flush. On reuse, conditionally issue the
flush. Since IBPB invalidates all BTB entries, clear the flag on all packs
after flushing.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Currently predictor flush on memory reuse is done for all BPF JIT
allocations, but only cBPF programs can be loaded by an unprivileged user.
eBPF is privileged by default, and flushing predictors for all CPUs on
every eBPF reuse penalizes the common case for no security benefit.
eBPF allocations can be frequent on busy systems, only flush predictors
for cBPF programs. Trampoline and dispatcher allocations also skip the
flush as they are eBPF-only.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Enable hardening against JIT spraying when Spectre-v2 mitigations are in
use. Specifically, issue an IBPB flush on BPF JIT memory reuse. Skip
enabling the IBPB flush if the BPF dispatcher is already using a retpoline
sequence.
This hardening applies only when BPF-JIT is in use. Guard the enabling
under CONFIG_BPF_JIT so that bugs.c still builds with CONFIG_BPF_JIT=n.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
The BPF JIT allocator packs many small programs into larger executable
allocations and reuses space within those allocations as programs are
loaded and freed. When fresh code is written into space that a previous
program occupied, an indirect jump into the new program can reuse a branch
prediction left behind by the old one.
Flush the indirect branch predictors before reusing JIT memory so that
indirect jumps into a newly written program don't reuse predictions from an
old program that occupied the same space.
Introduce bpf_arch_pred_flush_enabled static key and bpf_arch_pred_flush
static call for flushing the branch predictors on JIT memory reuse.
Architectures that need a flush, can update it to a predictor flush
function. By default, its a NOP and does not emit any CALL.
Allocations larger than a pack are not covered by this flush. That is safe
because cBPF programs (the unprivileged attack surface) are bounded well
below a pack size. Issue a warning if this assumption is ever violated
while the flush is active.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
HHF reset does not clear the classifier state used to identify heavy
hitters. Packets after reset can therefore be scheduled using flow
history from before the reset.
The reset operation should return the qdisc to an empty state.
Clear the heavy-hitter classifier tables when HHF is reset.
Fixes: 10239edf86 ("net-qdisc-hhf: Heavy-Hitter Filter (HHF) qdisc")
Assisted-by: Codex:gpt-5.5-cyber-preview
Signed-off-by: Samuel Moelius <sam.moelius@trailofbits.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
split_strings() walks each directory entry with strlen(). Although the
transport adds a terminator after the reply buffer, a malformed reply
without a final NUL inside its advertised length would let that walk
cross the protocol payload boundary.
Reject such replies before counting the strings. Report the protocol
violation once and return -EIO to the caller.
Signed-off-by: Yousef Alhouseen <alhouseenyousef@gmail.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Message-ID: <20260626223738.43742-1-alhouseenyousef@gmail.com>
DualPI2 leaves previous classification state attached to an skb when
filter classification returns no match. The enqueue path can then act
on stale state from an earlier classification attempt.
A filter miss should fall back to the default class without reusing old
per-packet classification data.
Initialize the classification result to CLASSIC before running the
classifier. Explicit L4S, priority, and successful filter
classification can still override that default.
Fixes: 8f9516daed ("sched: Add enqueue/dequeue of dualpi2 qdisc")
Assisted-by: Codex:gpt-5.5-cyber-preview
Signed-off-by: Samuel Moelius <sam.moelius@trailofbits.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David Howells
Yuto Ohnuki
Li RongQing
Zilin Guan
Matvey Kovalev
Christian Brauner
Dan Carpenter
Nan Li
Robby Cai
Zhao Dongdong
Tanmay Kathpalia
Yixun Lan
Arnd Bergmann
Yousef Alhouseen
HyeongJun An
Melbin K Mathew
Yi Xie
Damien Le Moal
Christoph Hellwig
Rafael J. Wysocki
Luca Coelho
Dawei Feng
Yingjie Gao
Pawan Gupta
Samuel Moelius