linux-2.6-block.git
3 months agoio_uring/net: fix io_req_post_cqe abuse by send bundle for-6.15/io_uring-reg-vec for-6.15/io_uring-reg-vec-20250327
Pavel Begunkov [Thu, 27 Mar 2025 09:57:27 +0000 (09:57 +0000)]
io_uring/net: fix io_req_post_cqe abuse by send bundle

[  114.987980][ T5313] WARNING: CPU: 6 PID: 5313 at io_uring/io_uring.c:872 io_req_post_cqe+0x12e/0x4f0
[  114.991597][ T5313] RIP: 0010:io_req_post_cqe+0x12e/0x4f0
[  115.001880][ T5313] Call Trace:
[  115.002222][ T5313]  <TASK>
[  115.007813][ T5313]  io_send+0x4fe/0x10f0
[  115.009317][ T5313]  io_issue_sqe+0x1a6/0x1740
[  115.012094][ T5313]  io_wq_submit_work+0x38b/0xed0
[  115.013223][ T5313]  io_worker_handle_work+0x62a/0x1600
[  115.013876][ T5313]  io_wq_worker+0x34f/0xdf0

As the comment states, io_req_post_cqe() should only be used by
multishot requests, i.e. REQ_F_APOLL_MULTISHOT, which bundled sends are
not. Add a flag signifying whether a request wants to post multiple
CQEs. Eventually REQ_F_APOLL_MULTISHOT should imply the new flag, but
that's left out for simplicity.

Cc: stable@vger.kernel.org
Fixes: a05d1f625c7aa ("io_uring/net: support bundles for send")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/8b611dbb54d1cd47a88681f5d38c84d0c02bc563.1743067183.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring/net: use REQ_F_IMPORT_BUFFER for send_zc
Caleb Sander Mateos [Tue, 25 Mar 2025 14:39:42 +0000 (08:39 -0600)]
io_uring/net: use REQ_F_IMPORT_BUFFER for send_zc

Instead of a bool field in struct io_sr_msg, use REQ_F_IMPORT_BUFFER to
track whether io_send_zc() has already imported the buffer. This flag
already serves a similar purpose for sendmsg_zc and {read,write}v_fixed.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Suggested-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20250325143943.1226467-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring: move min_events sanitisation
Pavel Begunkov [Mon, 24 Mar 2025 15:32:36 +0000 (15:32 +0000)]
io_uring: move min_events sanitisation

iopoll and normal waiting already duplicate min_completion truncation,
so move them inside the corresponding routines.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/254adb289cc04638f25d746a7499260fa89a179e.1742829388.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring: rename "min" arg in io_iopoll_check()
Pavel Begunkov [Mon, 24 Mar 2025 15:32:35 +0000 (15:32 +0000)]
io_uring: rename "min" arg in io_iopoll_check()

Don't name arguments "min", it shadows the namesake function.
min_events is also more consistent.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/f52ce9d88d3bca5732a218b0da14924aa6968909.1742829388.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring: open code __io_post_aux_cqe()
Pavel Begunkov [Mon, 24 Mar 2025 15:32:34 +0000 (15:32 +0000)]
io_uring: open code __io_post_aux_cqe()

There is no reason to keep __io_post_aux_cqe() separately from
io_post_aux_cqe().

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/2c4c1f68d694deea25a212fc09bbb11f330cd82e.1742829388.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring: defer iowq cqe overflow via task_work
Pavel Begunkov [Mon, 24 Mar 2025 15:32:33 +0000 (15:32 +0000)]
io_uring: defer iowq cqe overflow via task_work

Don't handle CQE overflows in io_req_complete_post() and defer it to
flush_completions. It cuts some duplication, and I also want to limit
the number of places directly overflowing completions.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/9046410ac27e18f2baa6f7cdb363ec921cbc3b79.1742829388.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring: fix retry handling off iowq
Pavel Begunkov [Mon, 24 Mar 2025 15:32:32 +0000 (15:32 +0000)]
io_uring: fix retry handling off iowq

io_req_complete_post() doesn't handle reissue and if called with a
REQ_F_REISSUE request it might post extra unexpected completions. Fix it
by pushing into flush_completion via task work.

Fixes: d803d123948fe ("io_uring/rw: handle -EAGAIN retry at IO completion time")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/badb3d7e462881e7edbfcc2be6301090b07dbe53.1742829388.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring/net: only import send_zc buffer once
Caleb Sander Mateos [Fri, 21 Mar 2025 18:48:17 +0000 (12:48 -0600)]
io_uring/net: only import send_zc buffer once

io_send_zc() guards its call to io_send_zc_import() with if (!done_io)
in an attempt to avoid calling it redundantly on the same req. However,
if the initial non-blocking issue returns -EAGAIN, done_io will stay 0.
This causes the subsequent issue to unnecessarily re-import the buffer.

Add an explicit flag "imported" to io_sr_msg to track if its buffer has
already been imported. Clear the flag in io_send_zc_prep(). Call
io_send_zc_import() and set the flag in io_send_zc() if it is unset.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Fixes: 54cdcca05abd ("io_uring/net: switch io_send() and io_send_zc() to using io_async_msghdr")
Link: https://lore.kernel.org/r/20250321184819.3847386-2-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring/cmd: introduce io_uring_cmd_import_fixed_vec
Pavel Begunkov [Fri, 21 Mar 2025 18:04:34 +0000 (18:04 +0000)]
io_uring/cmd: introduce io_uring_cmd_import_fixed_vec

io_uring_cmd_import_fixed_vec() is a cmd helper around vectored
registered buffer import functions, which caches the memory under
the hood. The lifetime of the vectore and hence the iterator is bound to
the request. Furthermore, the user is not allowed to call it multiple
times for a single request.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/97487a80dec3fb8cf8aeedf1f9026ef6d503fe4b.1742579999.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
3 months agoio_uring/cmd: add iovec cache for commands
Pavel Begunkov [Fri, 21 Mar 2025 18:04:33 +0000 (18:04 +0000)]
io_uring/cmd: add iovec cache for commands

Add iou_vec to commands and wire caching for it, but don't expose it to
users just yet. We need the vec cleared on initial alloc, but since
we can't place it at the beginning at the moment, zero the entire
async_data. It's cached, and the performance effects only the initial
allocation, and it might be not a bad idea since we're exposing those
bits to outside drivers.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/c0f2145b75791bc6106eb4e72add2cf6a2c72a7a.1742579999.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/cmd: don't expose entire cmd async data
Pavel Begunkov [Wed, 19 Mar 2025 06:12:48 +0000 (06:12 +0000)]
io_uring/cmd: don't expose entire cmd async data

io_uring needs private bits in cmd's ->async_data, and they should never
be exposed to drivers as it'd certainly be abused. Leave struct
io_uring_cmd_data for the drivers but wrap it into a structure. It's a
prep patch and doesn't do anything useful yet.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20250319061251.21452-3-sidong.yang@furiosa.ai
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: rename the data cmd cache
Pavel Begunkov [Wed, 19 Mar 2025 06:12:47 +0000 (06:12 +0000)]
io_uring: rename the data cmd cache

Pick a more descriptive name for the cmd async data cache.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20250319061251.21452-2-sidong.yang@furiosa.ai
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: rely on io_prep_reg_vec for iovec placement
Pavel Begunkov [Sat, 8 Mar 2025 18:21:16 +0000 (18:21 +0000)]
io_uring: rely on io_prep_reg_vec for iovec placement

All vectored reg buffer users should use io_import_reg_vec() for iovec
imports, since iovec placement is the function's responsibility and
callers shouldn't know much about it, drop the offset parameter from
io_prep_reg_vec() and calculate it inside.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/08ed87ca4bbc06724373b6ce06f36b703fe60c4e.1741457480.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: introduce io_prep_reg_iovec()
Pavel Begunkov [Sat, 8 Mar 2025 18:21:15 +0000 (18:21 +0000)]
io_uring: introduce io_prep_reg_iovec()

iovecs that are turned into registered buffers are imported in a special
way with an offset, so that later we can do an in place translation. Add
a helper function taking care of it.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/7de2ecb9ed5efc3c5cf320232236966da5ad4ccc.1741457480.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: unify STOP_MULTISHOT with IOU_OK
Pavel Begunkov [Sat, 8 Mar 2025 17:19:33 +0000 (17:19 +0000)]
io_uring: unify STOP_MULTISHOT with IOU_OK

IOU_OK means that the request ownership is now handed back to core
io_uring and it has to complete it using the result provided in
req->cqe. Same is true for multishot and IOU_STOP_MULTISHOT.

Rename it into IOU_COMPLETE to avoid confusion and use for both modes.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e6a5b2edb0eb9558acb1c8f1db38ac45fee95491.1741453534.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: return -EAGAIN to continue multishot
Pavel Begunkov [Sat, 8 Mar 2025 17:19:32 +0000 (17:19 +0000)]
io_uring: return -EAGAIN to continue multishot

Multishot errors can be mapped 1:1 to normal errors, but there are not
identical. It leads to a peculiar situation where all multishot requests
has to check in what context they're run and return different codes.

Unify them starting with EAGAIN / IOU_ISSUE_SKIP_COMPLETE(EIOCBQUEUED)
pair, which mean that core io_uring still owns the request and it should
be retried. In case of multishot it's naturally just continues to poll,
otherwise it might poll, use iowq or do any other kind of allowed
blocking. Introduce IOU_RETRY aliased to -EAGAIN for that.

Apart from obvious upsides, multishot can now also check for misuse of
IOU_ISSUE_SKIP_COMPLETE.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/da117b79ce72ecc3ab488c744e29fae9ba54e23b.1741453534.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: cap cached iovec/bvec size
Pavel Begunkov [Fri, 7 Mar 2025 16:00:37 +0000 (16:00 +0000)]
io_uring: cap cached iovec/bvec size

Bvecs can be large, put an arbitrary limit on the max vector size it
can cache.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/823055fa6628daa24bbc9cd77c2da87e9a1e1e32.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/net: implement vectored reg bufs for zctx
Pavel Begunkov [Fri, 7 Mar 2025 16:00:36 +0000 (16:00 +0000)]
io_uring/net: implement vectored reg bufs for zctx

Add support for vectored registered buffers for send zc.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e484052875f862d2dca99f0f8c04407c1d51a1c1.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/net: convert to struct iou_vec
Pavel Begunkov [Fri, 7 Mar 2025 16:00:35 +0000 (16:00 +0000)]
io_uring/net: convert to struct iou_vec

Convert net.c to use struct iou_vec.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/6437b57dabed44eca708c02e390529c7ed211c78.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/net: pull vec alloc out of msghdr import
Pavel Begunkov [Fri, 7 Mar 2025 16:00:34 +0000 (16:00 +0000)]
io_uring/net: pull vec alloc out of msghdr import

I'll need more control over iovec management, move
io_net_import_vec() out of io_msg_copy_hdr().

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/9600ea6300f620e65d39da481c22605ddc898850.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/net: combine msghdr copy
Pavel Begunkov [Fri, 7 Mar 2025 16:00:33 +0000 (16:00 +0000)]
io_uring/net: combine msghdr copy

Call the compat version from inside of io_msg_copy_hdr() and don't
duplicate it in callers.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/25795660f7b31f9273911c99f495d9c2b169ecda.1741362889.git.asml.silence@gmail.com
[axboe: fixup msg pointer vs variable braino in io_msg_copy_hdr()]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rw: defer reg buf vec import
Pavel Begunkov [Fri, 7 Mar 2025 16:00:32 +0000 (16:00 +0000)]
io_uring/rw: defer reg buf vec import

Import registered buffers for vectored reads and writes later at issue
time as we now do for other fixed ops.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e8491c976e4ab83a4e3dc428e9fe7555e59583b8.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rw: implement vectored registered rw
Pavel Begunkov [Fri, 7 Mar 2025 16:00:31 +0000 (16:00 +0000)]
io_uring/rw: implement vectored registered rw

Implement registered buffer vectored reads with new opcodes
IORING_OP_WRITEV_FIXED and IORING_OP_READV_FIXED.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d7c89eb481e870f598edc91cc66ff4d1e4ae3788.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: add infra for importing vectored reg buffers
Pavel Begunkov [Fri, 7 Mar 2025 16:00:30 +0000 (16:00 +0000)]
io_uring: add infra for importing vectored reg buffers

Add io_import_reg_vec(), which will be responsible for importing
vectored registered buffers. The function might reallocate the vector,
but it'd try to do the conversion in place first, which is why it's
required of the user to pad the iovec to the right border of the cache.

Overlapping also depends on struct iovec being larger than bvec, which
is not the case on e.g. 32 bit architectures. Don't try to complicate
this case and make sure vectors never overlap, it'll be improved later.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/60bd246b1249476a6996407c1dbc38ef6febad14.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: introduce struct iou_vec
Pavel Begunkov [Fri, 7 Mar 2025 16:00:29 +0000 (16:00 +0000)]
io_uring: introduce struct iou_vec

I need a convenient way to pass around and work with iovec+size pair,
put them into a structure and makes use of it in rw.c

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d39fadafc9e9047b0a292e5be6db3cf2f48bb1f7.1741362889.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoMerge branch 'for-6.15/io_uring-epoll-wait' into for-6.15/io_uring-reg-vec
Jens Axboe [Fri, 7 Mar 2025 16:07:19 +0000 (09:07 -0700)]
Merge branch 'for-6.15/io_uring-epoll-wait' into for-6.15/io_uring-reg-vec

* for-6.15/io_uring-epoll-wait:
  io_uring/epoll: add support for IORING_OP_EPOLL_WAIT
  io_uring/epoll: remove CONFIG_EPOLL guards
  eventpoll: add epoll_sendevents() helper
  eventpoll: abstract out ep_try_send_events() helper
  eventpoll: abstract out parameter sanity checking

4 months agoMerge branch 'for-6.15/io_uring-rx-zc' into for-6.15/io_uring-reg-vec
Jens Axboe [Fri, 7 Mar 2025 16:07:11 +0000 (09:07 -0700)]
Merge branch 'for-6.15/io_uring-rx-zc' into for-6.15/io_uring-reg-vec

* for-6.15/io_uring-rx-zc: (80 commits)
  io_uring/zcrx: add selftest case for recvzc with read limit
  io_uring/zcrx: add a read limit to recvzc requests
  io_uring: add missing IORING_MAP_OFF_ZCRX_REGION in io_uring_mmap
  io_uring: Rename KConfig to Kconfig
  io_uring/zcrx: fix leaks on failed registration
  io_uring/zcrx: recheck ifq on shutdown
  io_uring/zcrx: add selftest
  net: add documentation for io_uring zcrx
  io_uring/zcrx: add copy fallback
  io_uring/zcrx: throttle receive requests
  io_uring/zcrx: set pp memory provider for an rx queue
  io_uring/zcrx: add io_recvzc request
  io_uring/zcrx: dma-map area for the device
  io_uring/zcrx: implement zerocopy receive pp memory provider
  io_uring/zcrx: grab a net device
  io_uring/zcrx: add io_zcrx_area
  io_uring/zcrx: add interface queue and refill queue
  net: add helpers for setting a memory provider on an rx queue
  net: page_pool: add memory provider helpers
  net: prepare for non devmem TCP memory providers
  ...

4 months agoMerge branch 'for-6.15/io_uring' into for-6.15/io_uring-reg-vec
Jens Axboe [Fri, 7 Mar 2025 16:07:07 +0000 (09:07 -0700)]
Merge branch 'for-6.15/io_uring' into for-6.15/io_uring-reg-vec

* for-6.15/io_uring: (80 commits)
  io_uring: introduce io_cache_free() helper
  io_uring/rsrc: skip NULL file/buffer checks in io_free_rsrc_node()
  io_uring/rsrc: avoid NULL node check on io_sqe_buffer_register() failure
  io_uring/rsrc: call io_free_node() on io_sqe_buffer_register() failure
  io_uring/rsrc: free io_rsrc_node using kfree()
  io_uring/rsrc: split out io_free_node() helper
  io_uring/rsrc: include io_uring_types.h in rsrc.h
  ublk: don't cast registered buffer index to int
  io_uring/nop: use io_find_buf_node()
  io_uring/rsrc: declare io_find_buf_node() in header file
  io_uring/ublk: report error when unregister operation fails
  io_uring: convert cmd_to_io_kiocb() macro to function
  io_uring/uring_cmd: specify io_uring_cmd_import_fixed() pointer type
  io_uring/rsrc: use rq_data_dir() to compute bvec dir
  selftests: ublk: add ublk zero copy test
  selftests: ublk: add file backed ublk
  selftests: ublk: add kernel selftests for ublk
  io_uring: cache nodes and mapped buffers
  ublk: zc register/unregister bvec
  io_uring: add support for kernel registered bvecs
  ...

4 months agoio_uring: introduce io_cache_free() helper
Caleb Sander Mateos [Tue, 4 Mar 2025 19:48:12 +0000 (12:48 -0700)]
io_uring: introduce io_cache_free() helper

Add a helper function io_cache_free() that returns an allocation to a
io_alloc_cache, falling back on kfree() if the io_alloc_cache is full.
This is the inverse of io_cache_alloc(), which takes an allocation from
an io_alloc_cache and falls back on kmalloc() if the cache is empty.

Convert 4 callers to use the helper.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Suggested-by: Li Zetao <lizetao1@huawei.com>
Link: https://lore.kernel.org/r/20250304194814.2346705-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: skip NULL file/buffer checks in io_free_rsrc_node()
Caleb Sander Mateos [Fri, 28 Feb 2025 23:59:14 +0000 (16:59 -0700)]
io_uring/rsrc: skip NULL file/buffer checks in io_free_rsrc_node()

io_rsrc_node's of type IORING_RSRC_FILE always have a file attached
immediately after they are allocated. IORING_RSRC_BUFFER nodes won't be
returned from io_sqe_buffer_register()/io_buffer_register_bvec() until
they have a io_mapped_ubuf attached.

So remove the checks for a NULL file/buffer in io_free_rsrc_node().

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228235916.670437-5-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: avoid NULL node check on io_sqe_buffer_register() failure
Caleb Sander Mateos [Fri, 28 Feb 2025 23:59:13 +0000 (16:59 -0700)]
io_uring/rsrc: avoid NULL node check on io_sqe_buffer_register() failure

The done: label is only reachable if node is non-NULL. So don't bother
checking, just call io_free_node().

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228235916.670437-4-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: call io_free_node() on io_sqe_buffer_register() failure
Caleb Sander Mateos [Fri, 28 Feb 2025 23:59:12 +0000 (16:59 -0700)]
io_uring/rsrc: call io_free_node() on io_sqe_buffer_register() failure

io_sqe_buffer_register() currently calls io_put_rsrc_node() if it fails
to fully set up the io_rsrc_node. io_put_rsrc_node() is more involved
than necessary, since we already know the reference count will reach 0
and no io_mapped_ubuf has been attached to the node yet.

So just call io_free_node() to release the node's memory. This also
avoids the need to temporarily set the node's buf pointer to NULL.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228235916.670437-3-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: free io_rsrc_node using kfree()
Caleb Sander Mateos [Fri, 28 Feb 2025 23:59:11 +0000 (16:59 -0700)]
io_uring/rsrc: free io_rsrc_node using kfree()

io_rsrc_node_alloc() calls io_cache_alloc(), which uses kmalloc() to
allocate the node. So it can be freed with kfree() instead of kvfree().

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228235916.670437-2-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: split out io_free_node() helper
Caleb Sander Mateos [Fri, 28 Feb 2025 23:59:10 +0000 (16:59 -0700)]
io_uring/rsrc: split out io_free_node() helper

Split the freeing of the io_rsrc_node from io_free_rsrc_node(), for use
with nodes that haven't been fully initialized.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228235916.670437-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: include io_uring_types.h in rsrc.h
Caleb Sander Mateos [Sat, 1 Mar 2025 18:36:11 +0000 (11:36 -0700)]
io_uring/rsrc: include io_uring_types.h in rsrc.h

io_uring/rsrc.h uses several types from include/linux/io_uring_types.h.
Include io_uring_types.h explicitly in rsrc.h to avoid depending on
users of rsrc.h including io_uring_types.h first.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Li Zetao <lizetao1@huawei.com>
Link: https://lore.kernel.org/r/20250301183612.937529-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoublk: don't cast registered buffer index to int
Caleb Sander Mateos [Sat, 1 Mar 2025 19:03:16 +0000 (12:03 -0700)]
ublk: don't cast registered buffer index to int

io_buffer_register_bvec() takes index as an unsigned int argument, but
ublk_register_io_buf() casts ub_cmd->addr (a u64) to int. Remove the
misleading cast and instead pass index as an unsigned value to
ublk_register_io_buf() and ublk_unregister_io_buf().

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250301190317.950208-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoLinux 6.14-rc5 v6.14-rc5
Linus Torvalds [Sun, 2 Mar 2025 19:48:20 +0000 (11:48 -0800)]
Linux 6.14-rc5

4 months agoMerge tag 'dmaengine-fix-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul...
Linus Torvalds [Sun, 2 Mar 2025 18:08:45 +0000 (10:08 -0800)]
Merge tag 'dmaengine-fix-6.14' of git://git./linux/kernel/git/vkoul/dmaengine

Pull dmaengine fixes from Vinod Koul:

 - tegra210 div_u64 divison and max page fixes

 - revert Qualcomm unavailable register workaround which is causing
   regression, fixes have been proposed but still gaps are present so
   revert this for now

* tag 'dmaengine-fix-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine:
  dmaengine: Revert "dmaengine: qcom: bam_dma: Avoid writing unavailable register"
  dmaengine: tegra210-adma: check for adma max page
  dmaengine: tegra210-adma: Use div_u64 for 64 bit division

4 months agoMerge tag 'phy-fixes-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy
Linus Torvalds [Sun, 2 Mar 2025 17:54:25 +0000 (09:54 -0800)]
Merge tag 'phy-fixes-6.14' of git://git./linux/kernel/git/phy/linux-phy

Pull phy fixes from Vinod Koul:

 - rockchip phy kconfig dependency fix with USB_COMMON and regression
   fix for old DT

 - stm32 phy overflow assertion fix

 - exonysfs phy refclk masks fix and power gate on exit fix

 - freescale fix for clock dividor valid range

 - TI regmap syscon register fix

 - tegra reset registers on init fix

* tag 'phy-fixes-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy:
  phy: tegra: xusb: reset VBUS & ID OVERRIDE
  phy: ti: gmii-sel: Do not use syscon helper to build regmap
  phy: exynos5-usbdrd: gs101: ensure power is gated to SS phy in phy_exit()
  phy: freescale: fsl-samsung-hdmi: Limit PLL lock detection clock divider to valid range
  phy: exynos5-usbdrd: fix MPLL_MULTIPLIER and SSC_REFCLKSEL masks in refclk
  phy: stm32: Fix constant-value overflow assertion
  phy: rockchip: naneng-combphy: compatible reset with old DT
  phy: rockchip: fix Kconfig dependency more

4 months agoMerge tag 'gpio-fixes-for-v6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 2 Mar 2025 12:35:39 +0000 (04:35 -0800)]
Merge tag 'gpio-fixes-for-v6.14-rc5' of git://git./linux/kernel/git/brgl/linux

Pull gpio fix from Bartosz Golaszewski:

 - fix a buggy get_direction() retval check

* tag 'gpio-fixes-for-v6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpiolib: Fix Oops in gpiod_direction_input_nonotify()

4 months agoMerge tag 'mips-fixes_6.14_2' of git://git.kernel.org/pub/scm/linux/kernel/git/mips...
Linus Torvalds [Sun, 2 Mar 2025 12:33:15 +0000 (04:33 -0800)]
Merge tag 'mips-fixes_6.14_2' of git://git./linux/kernel/git/mips/linux

Pull MIPS fix from Thomas Bogendoerfer:
 "Fix fallout of /scripts/sorttable cleanup"

* tag 'mips-fixes_6.14_2' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
  MIPS: Ignore relocs against __ex_table for relocatable kernel

4 months agoMerge tag 'v6.14-rc4-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 2 Mar 2025 00:10:31 +0000 (16:10 -0800)]
Merge tag 'v6.14-rc4-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fix from Steve French:
 "Fix SMB1 netfs client regression"

* tag 'v6.14-rc4-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: Fix the smb1 readv callback to correctly call netfs

4 months agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Sat, 1 Mar 2025 21:44:51 +0000 (13:44 -0800)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes from Will Deacon:
 "Ryan's been hard at work finding and fixing mm bugs in the arm64 code,
  so here's a small crop of fixes for -rc5.

  The main changes are to fix our zapping of non-present PTEs for
  hugetlb entries created using the contiguous bit in the page-table
  rather than a block entry at the level above. Prior to these fixes, we
  were pulling the contiguous bit back out of the PTE in order to
  determine the size of the hugetlb page but this is clearly bogus if
  the thing isn't present and consequently both the clearing of the
  PTE(s) and the TLB invalidation were unreliable.

  Although the problem was found by code inspection, we really don't
  want this sitting around waiting to trigger and the changes are CC'd
  to stable accordingly.

  Note that the diffstat looks a lot worse than it really is;
  huge_ptep_get_and_clear() now takes a size argument from the core code
  and so all the arch implementations of that have been updated in a
  pretty mechanical fashion.

   - Fix a sporadic boot failure due to incorrect randomization of the
     linear map on systems that support it

   - Fix the zapping (both clearing the entries *and* invalidating the
     TLB) of hugetlb PTEs constructed using the contiguous bit"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: hugetlb: Fix flush_hugetlb_tlb_range() invalidation level
  arm64: hugetlb: Fix huge_ptep_get_and_clear() for non-present ptes
  mm: hugetlb: Add huge page size param to huge_ptep_get_and_clear()
  arm64/mm: Fix Boot panic on Ampere Altra

4 months agoMerge tag 'i2c-for-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sat, 1 Mar 2025 20:43:11 +0000 (12:43 -0800)]
Merge tag 'i2c-for-6.14-rc5' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "All driver fixes this time:

   - fix interrupt initialization sequence (npcm)

   - fix frequency setting (ls2x)

   - re-enable interrupts properly at irq handler's exit (amd-asf)"

* tag 'i2c-for-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: amd-asf: Fix EOI register write to enable successive interrupts
  i2c: ls2x: Fix frequency division register access
  i2c: npcm: disable interrupt enable bit before devm_request_irq

4 months agoMerge tag 'ata-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/libata...
Linus Torvalds [Sat, 1 Mar 2025 16:59:29 +0000 (08:59 -0800)]
Merge tag 'ata-6.14-rc5' of git://git./linux/kernel/git/libata/linux

Pull ata fixes from Niklas Cassel:

 - Fix a regression where the enablement of the PHYs would be skipped
   for device trees without any port child nodes (me)

 - Revert ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives, as it stops
   systems from entering lower package states. LPM works on newer
   firmware versions. We will need a more refined quirk that only
   targets the older firmware versions (me)

* tag 'ata-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux:
  Revert "ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives"
  ata: ahci: Make ahci_ignore_port() handle empty mask_port_map

4 months agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Sat, 1 Mar 2025 16:48:53 +0000 (08:48 -0800)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "ARM:

   - Fix TCR_EL2 configuration to not use the ASID in TTBR1_EL2 and not
     mess-up T1SZ/PS by using the HCR_EL2.E2H==0 layout.

   - Bring back the VMID allocation to the vcpu_load phase, ensuring
     that we only setup VTTBR_EL2 once on VHE. This cures an ugly race
     that would lead to running with an unallocated VMID.

  RISC-V:

   - Fix hart status check in SBI HSM extension

   - Fix hart suspend_type usage in SBI HSM extension

   - Fix error returned by SBI IPI and TIME extensions for unsupported
     function IDs

   - Fix suspend_type usage in SBI SUSP extension

   - Remove unnecessary vcpu kick after injecting interrupt via IMSIC
     guest file

  x86:

   - Fix an nVMX bug where KVM fails to detect that, after nested
     VM-Exit, L1 has a pending IRQ (or NMI).

   - To avoid freeing the PIC while vCPUs are still around, which would
     cause a NULL pointer access with the previous patch, destroy vCPUs
     before any VM-level destruction.

   - Handle failures to create vhost_tasks"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  kvm: retry nx_huge_page_recovery_thread creation
  vhost: return task creation error instead of NULL
  KVM: nVMX: Process events on nested VM-Exit if injectable IRQ or NMI is pending
  KVM: x86: Free vCPUs before freeing VM state
  riscv: KVM: Remove unnecessary vcpu kick
  KVM: arm64: Ensure a VMID is allocated before programming VTTBR_EL2
  KVM: arm64: Fix tcr_el2 initialisation in hVHE mode
  riscv: KVM: Fix SBI sleep_type use
  riscv: KVM: Fix SBI TIME error generation
  riscv: KVM: Fix SBI IPI error generation
  riscv: KVM: Fix hart suspend_type use
  riscv: KVM: Fix hart suspend status check

4 months agoRevert "ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives"
Niklas Cassel [Fri, 28 Feb 2025 12:26:04 +0000 (13:26 +0100)]
Revert "ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives"

This reverts commit cc77e2ce187d26cc66af3577bf896d7410eb25ab.

It was reported that adding ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives
breaks entering lower package states for certain systems.

It turns out that Samsung SSD 870 QVO actually has working LPM when using
a recent SSD firmware version.

The author of commit cc77e2ce187d ("ata: libata-core: Add ATA_QUIRK_NOLPM
for Samsung SSD 870 QVO drives") reported himself that only older SSD
firmware versions have broken LPM:
https://lore.kernel.org/stable/93c10d38-718c-459d-84a5-4d87680b4da7@debian.org/

Unfortunately, he did not specify which older firmware version he was using
which had broken LPM.

Let's revert this quirk, which has FW version field specified as NULL
(which means that it applies for all Samsung SSD 870 QVO firmware versions)
for now. Once the author reports which older firmware version(s) that are
broken, we can create a more fine grained quirk, which populates the FW
version field accordingly.

Fixes: cc77e2ce187d ("ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives")
Reported-by: Dieter Mummenschanz <dmummenschanz@web.de>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219747
Link: https://lore.kernel.org/r/20250228122603.91814-2-cassel@kernel.org
Signed-off-by: Niklas Cassel <cassel@kernel.org>
4 months agokvm: retry nx_huge_page_recovery_thread creation
Keith Busch [Thu, 27 Feb 2025 23:06:31 +0000 (15:06 -0800)]
kvm: retry nx_huge_page_recovery_thread creation

A VMM may send a non-fatal signal to its threads, including vCPU tasks,
at any time, and thus may signal vCPU tasks during KVM_RUN.  If a vCPU
task receives the signal while its trying to spawn the huge page recovery
vhost task, then KVM_RUN will fail due to copy_process() returning
-ERESTARTNOINTR.

Rework call_once() to mark the call complete if and only if the called
function succeeds, and plumb the function's true error code back to the
call_once() invoker.  This provides userspace with the correct, non-fatal
error code so that the VMM doesn't terminate the VM on -ENOMEM, and allows
subsequent KVM_RUN a succeed by virtue of retrying creation of the NX huge
page task.

Co-developed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
[implemented the kvm user side]
Signed-off-by: Keith Busch <kbusch@kernel.org>
Message-ID: <20250227230631.303431-3-kbusch@meta.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 months agovhost: return task creation error instead of NULL
Keith Busch [Thu, 27 Feb 2025 23:06:30 +0000 (15:06 -0800)]
vhost: return task creation error instead of NULL

Lets callers distinguish why the vhost task creation failed. No one
currently cares why it failed, so no real runtime change from this
patch, but that will not be the case for long.

Signed-off-by: Keith Busch <kbusch@kernel.org>
Message-ID: <20250227230631.303431-2-kbusch@meta.com>
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 months agoio_uring/nop: use io_find_buf_node()
Caleb Sander Mateos [Sat, 1 Mar 2025 00:16:08 +0000 (17:16 -0700)]
io_uring/nop: use io_find_buf_node()

Call io_find_buf_node() to avoid duplicating it in io_nop().

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250301001610.678223-2-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: declare io_find_buf_node() in header file
Caleb Sander Mateos [Sat, 1 Mar 2025 00:16:07 +0000 (17:16 -0700)]
io_uring/rsrc: declare io_find_buf_node() in header file

Declare io_find_buf_node() in io_uring/rsrc.h so it can be called from
other files.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250301001610.678223-1-csander@purestorage.com
[axboe: keep the inline for local hot path usage]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/ublk: report error when unregister operation fails
Caleb Sander Mateos [Fri, 28 Feb 2025 23:14:31 +0000 (16:14 -0700)]
io_uring/ublk: report error when unregister operation fails

Indicate to userspace applications if a UBLK_IO_UNREGISTER_IO_BUF
command specifies an invalid buffer index by returning an error code.
Return -EINVAL if no buffer is registered with the given index, and
-EBUSY if the registered buffer is not a kernel bvec.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228231432.642417-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: convert cmd_to_io_kiocb() macro to function
Caleb Sander Mateos [Fri, 28 Feb 2025 23:03:04 +0000 (16:03 -0700)]
io_uring: convert cmd_to_io_kiocb() macro to function

The cmd_to_io_kiocb() macro applies a pointer cast to its input without
parenthesizing it. Currently all inputs are variable names, so this has
the intended effect. But since casts have relatively high precedence,
the macro would apply the cast to the wrong value if the input was a
pointer addition, for example.

Turn the macro into a static inline function to ensure the pointer cast
is applied to the full input value.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228230305.630885-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/uring_cmd: specify io_uring_cmd_import_fixed() pointer type
Caleb Sander Mateos [Fri, 28 Feb 2025 22:15:13 +0000 (15:15 -0700)]
io_uring/uring_cmd: specify io_uring_cmd_import_fixed() pointer type

io_uring_cmd_import_fixed() takes a struct io_uring_cmd *, but the type
of the ioucmd parameter is void *. Make the pointer type explicit so the
compiler can type check it.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228221514.604350-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rsrc: use rq_data_dir() to compute bvec dir
Caleb Sander Mateos [Fri, 28 Feb 2025 22:30:56 +0000 (15:30 -0700)]
io_uring/rsrc: use rq_data_dir() to compute bvec dir

The macro rq_data_dir() already computes a request's data direction.
Use it in place of the if-else to set imu->dir.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Link: https://lore.kernel.org/r/20250228223057.615284-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoMerge tag 'thermal-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 1 Mar 2025 02:02:23 +0000 (18:02 -0800)]
Merge tag 'thermal-6.14-rc5' of git://git./linux/kernel/git/rafael/linux-pm

Pull thermal control fixes from Rafael Wysocki:
 "These fix the processing of DT thermal properties and the Power
  Allocator thermal governor:

   - Fix parsing cooling-maps in DT for trip points with more than one
     cooling device (Rafael Wysocki)

   - Fix granted_power computation in the Power Allocator thermal
     governor and make it update total_weight on configuration changes
     after the thermal zone has been registered (Yu-Che Cheng)"

* tag 'thermal-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  thermal: gov_power_allocator: Update total_weight on bind and cdev updates
  thermal/of: Fix cdev lookup in thermal_of_should_bind()
  thermal: gov_power_allocator: Fix incorrect calculation in divvy_up_power()

4 months agoMerge tag 'pm-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Linus Torvalds [Sat, 1 Mar 2025 01:40:27 +0000 (17:40 -0800)]
Merge tag 'pm-6.14-rc5' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fix from Rafael Wysocki:
 "Fix the handling of processors that stop the TSC in deeper C-states in
  the intel_idle driver (Thomas Gleixner)"

* tag 'pm-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  intel_idle: Handle older CPUs, which stop the TSC in deeper C states, correctly

4 months agoMerge tag 'x86-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 1 Mar 2025 01:05:22 +0000 (17:05 -0800)]
Merge tag 'x86-urgent-2025-02-28' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:

 - Fix conflicts between devicetree and ACPI SMP discovery & setup

 - Fix a warm-boot lockup on AMD SC1100 SoC systems

 - Fix a W=1 build warning related to x86 IRQ trace event setup

 - Fix a kernel-doc warning

* tag 'x86-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/entry: Fix kernel-doc warning
  x86/irq: Define trace events conditionally
  x86/CPU: Fix warm boot hang regression on AMD SC1100 SoC systems
  x86/of: Don't use DTB for SMP setup if ACPI is enabled

4 months agoMerge tag 'sched-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 1 Mar 2025 01:00:16 +0000 (17:00 -0800)]
Merge tag 'sched-urgent-2025-02-28' of git://git./linux/kernel/git/tip/tip

Pull scheduler fix from Ingo Molnar:
 "Prevent cond_resched() based preemption when interrupts are disabled,
  on PREEMPT_NONE and PREEMPT_VOLUNTARY kernels"

* tag 'sched-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/core: Prevent rescheduling when interrupts are disabled

4 months agoMerge tag 'perf-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 1 Mar 2025 00:52:10 +0000 (16:52 -0800)]
Merge tag 'perf-urgent-2025-02-28' of git://git./linux/kernel/git/tip/tip

Pull perf event fixes from Ingo Molnar:
 "Miscellaneous perf events fixes and a minor HW enablement change:

   - Fix missing RCU protection in perf_iterate_ctx()

   - Fix pmu_ctx_list ordering bug

   - Reject the zero page in uprobes

   - Fix a family of bugs related to low frequency sampling

   - Add Intel Arrow Lake U CPUs to the generic Arrow Lake RAPL support
     table

   - Fix a lockdep-assert false positive in uretprobes"

* tag 'perf-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  uprobes: Remove too strict lockdep_assert() condition in hprobe_expire()
  perf/x86/rapl: Add support for Intel Arrow Lake U
  perf/x86/intel: Use better start period for frequency mode
  perf/core: Fix low freq setting via IOC_PERIOD
  perf/x86: Fix low freqency setting issue
  uprobes: Reject the shared zeropage in uprobe_write_opcode()
  perf/core: Order the PMU list to fix warning about unordered pmu_ctx_list
  perf/core: Add RCU read lock protection to perf_iterate_ctx()

4 months agoMerge tag 'objtool-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 1 Mar 2025 00:45:36 +0000 (16:45 -0800)]
Merge tag 'objtool-urgent-2025-02-28' of git://git./linux/kernel/git/tip/tip

Pull objtool fixes from Ingo Molnar:
 "Fix an objtool false positive, and objtool related build warnings that
  happens on PIE-enabled architectures such as LoongArch"

* tag 'objtool-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  objtool: Add bch2_trans_unlocked_or_in_restart_error() to bcachefs noreturns
  objtool: Fix C jump table annotations for Clang
  vmlinux.lds: Ensure that const vars with relocations are mapped R/O

4 months agoMerge tag 'locking-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 1 Mar 2025 00:07:18 +0000 (16:07 -0800)]
Merge tag 'locking-urgent-2025-02-28' of git://git./linux/kernel/git/tip/tip

Pull locking fix from Ingo Molnar:
 "Fix an rcuref_put() slowpath race"

* tag 'locking-urgent-2025-02-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  rcuref: Plug slowpath race in rcuref_put()

4 months agoMerge tag 'trace-v6.14-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace...
Linus Torvalds [Fri, 28 Feb 2025 23:43:32 +0000 (15:43 -0800)]
Merge tag 'trace-v6.14-rc4' of git://git./linux/kernel/git/trace/linux-trace

Pull tracing fixes from Steven Rostedt:

 - Fix crash from bad histogram entry

   An error path in the histogram creation could leave an entry in a
   link list that gets freed. Then when a new entry is added it can
   cause a u-a-f bug. This is fixed by restructuring the code so that
   the histogram is consistent on failure and everything is cleaned up
   appropriately.

 - Fix fprobe self test

   The fprobe self test relies on no function being attached by ftrace.
   BPF programs can attach to functions via ftrace and systemd now does
   so. This causes those functions to appear in the enabled_functions
   list which holds all functions attached by ftrace. The selftest also
   uses that file to see if functions are being connected correctly. It
   counts the functions in the file, but if there's already functions in
   the file, it fails. Instead, add the number of functions in the file
   at the start of the test to all the calculations during the test.

 - Fix potential division by zero of the function profiler stddev

   The calculated divisor that calculates the standard deviation of the
   function times can overflow. If the overflow happens to land on zero,
   that can cause a division by zero. Check for zero from the
   calculation before doing the division.

   TODO: Catch when it ever overflows and report it accordingly. For
   now, just prevent the system from crashing.

* tag 'trace-v6.14-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  ftrace: Avoid potential division by zero in function_stat_show()
  selftests/ftrace: Let fprobe test consider already enabled functions
  tracing: Fix bad hist from corrupting named_triggers list

4 months agoMerge tag 'iommu-fixes-v6.14-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 28 Feb 2025 23:39:17 +0000 (15:39 -0800)]
Merge tag 'iommu-fixes-v6.14-rc4' of git://git./linux/kernel/git/iommu/linux

Pull iommu fixes from Joerg Roedel:

 - Intel VT-d fixes:
     - Fix suspicious RCU usage splat
     - Fix passthrough for devices under PCIe-PCI bridge

 - AMD-Vi fix:
     - Fix to preserve bits when updating device table entries

* tag 'iommu-fixes-v6.14-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux:
  iommu/vt-d: Fix suspicious RCU usage
  iommu/vt-d: Remove device comparison in context_setup_pass_through_cb
  iommu/amd: Preserve default DTE fields when updating Host Page Table Root

4 months agointel_idle: Handle older CPUs, which stop the TSC in deeper C states, correctly
Thomas Gleixner [Tue, 25 Feb 2025 22:37:08 +0000 (23:37 +0100)]
intel_idle: Handle older CPUs, which stop the TSC in deeper C states, correctly

The Intel idle driver is preferred over the ACPI processor idle driver,
but fails to implement the work around for Core2 generation CPUs, where
the TSC stops in C2 and deeper C-states. This causes stalls and boot
delays, when the clocksource watchdog does not catch the unstable TSC
before the CPU goes deep idle for the first time.

The ACPI driver marks the TSC unstable when it detects that the CPU
supports C2 or deeper and the CPU does not have a non-stop TSC.

Add the equivivalent work around to the Intel idle driver to cure that.

Fixes: 18734958e9bf ("intel_idle: Use ACPI _CST for processor models without C-state tables")
Reported-by: Fab Stz <fabstz-it@yahoo.fr>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Fab Stz <fabstz-it@yahoo.fr>
Cc: All applicable <stable@vger.kernel.org>
Closes: https://lore.kernel.org/all/10cf96aa-1276-4bd4-8966-c890377030c3@yahoo.fr
Link: https://patch.msgid.link/87bjupfy7f.ffs@tglx
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
4 months agoMerge tag 'block-6.14-20250228' of git://git.kernel.dk/linux
Linus Torvalds [Fri, 28 Feb 2025 17:43:46 +0000 (09:43 -0800)]
Merge tag 'block-6.14-20250228' of git://git.kernel.dk/linux

Pull block fixes from Jens Axboe:

 - Fix plugging for native zone writes

 - Fix segment limit settings for != 4K page size archs

 - Fix for slab names overflowing

* tag 'block-6.14-20250228' of git://git.kernel.dk/linux:
  block: fix 'kmem_cache of name 'bio-108' already exists'
  block: Remove zone write plugs when handling native zone append writes
  block: make segment size limit workable for > 4K PAGE_SIZE

4 months agoMerge tag 'io_uring-6.14-20250228' of git://git.kernel.dk/linux
Linus Torvalds [Fri, 28 Feb 2025 17:11:15 +0000 (09:11 -0800)]
Merge tag 'io_uring-6.14-20250228' of git://git.kernel.dk/linux

Pull io_uring fix from Jens Axboe:
 "Just a single fix headed for stable, ensuring that msg_control is
  properly saved in compat mode as well"

* tag 'io_uring-6.14-20250228' of git://git.kernel.dk/linux:
  io_uring/net: save msg_control for compat

4 months agoMerge tag 'efi-fixes-for-v6.14-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 28 Feb 2025 16:47:21 +0000 (08:47 -0800)]
Merge tag 'efi-fixes-for-v6.14-2' of git://git./linux/kernel/git/efi/efi

Pull EFI fixes from Ard Biesheuvel:
 "Another couple of EFI fixes for v6.14.

  Only James's patch stands out, as it implements a workaround for odd
  behavior in fwupd in user space, which creates EFI variables by
  touching a file in efivarfs, clearing the immutable bit (which gets
  set automatically for $reasons) and then opening it again for writing,
  none of which is really necessary.

  The fwupd author and LVFS maintainer is already rolling out a fix for
  this on the fwupd side, and suggested that the workaround in this PR
  could be backed out again during the next cycle.

  (There is a semantic mismatch in efivarfs where some essential
  variable attributes are stored in the first 4 bytes of the file, and
  so zero length files cannot exist, as they cannot be written back to
  the underlying variable store. So now, they are dropped once the last
  reference is released.)

  Summary:

   - Fix CPER error record parsing bugs

   - Fix a couple of efivarfs issues that were introduced in the merge
     window

   - Fix an issue in the early remapping code of the MOKvar table"

* tag 'efi-fixes-for-v6.14-2' of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi:
  efi/mokvar-table: Avoid repeated map/unmap of the same page
  efi: Don't map the entire mokvar table to determine its size
  efivarfs: allow creation of zero length files
  efivarfs: Defer PM notifier registration until .fill_super
  efi/cper: Fix cper_arm_ctx_info alignment
  efi/cper: Fix cper_ia_proc_ctx alignment

4 months agoselftests: ublk: add ublk zero copy test
Ming Lei [Fri, 28 Feb 2025 16:19:16 +0000 (00:19 +0800)]
selftests: ublk: add ublk zero copy test

Enable zero copy on file backed target, meantime add one fio test for
covering write verify, another test for mkfs/mount/umount.

Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250228161919.2869102-4-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoselftests: ublk: add file backed ublk
Ming Lei [Fri, 28 Feb 2025 16:19:15 +0000 (00:19 +0800)]
selftests: ublk: add file backed ublk

Add file backed ublk target code, meantime add one fio test for
covering write verify, another test for mkfs/mount/umount.

Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250228161919.2869102-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoselftests: ublk: add kernel selftests for ublk
Ming Lei [Fri, 28 Feb 2025 16:19:14 +0000 (00:19 +0800)]
selftests: ublk: add kernel selftests for ublk

Both ublk driver and userspace heavily depends on io_uring subsystem,
and tools/testing/selftests/ should be the best place for holding this
cross-subsystem tests.

Add basic read/write IO test over this ublk null disk, and make sure ublk
working.

More tests will be added.

Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250228161919.2869102-2-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoMerge tag 'i2c-host-fixes-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel...
Wolfram Sang [Fri, 28 Feb 2025 16:27:44 +0000 (17:27 +0100)]
Merge tag 'i2c-host-fixes-6.14-rc5' of git://git./linux/kernel/git/andi.shyti/linux into i2c/for-current

i2c-host-fixes for v6.14-rc5

- npcm fixes interrupt initialization sequence.
- ls2x fixes frequency setting.
- amd-asf re-enables interrupts properly at irq handler's exit.

4 months agogpiolib: Fix Oops in gpiod_direction_input_nonotify()
Dan Carpenter [Fri, 28 Feb 2025 13:03:47 +0000 (16:03 +0300)]
gpiolib: Fix Oops in gpiod_direction_input_nonotify()

The gpiod_direction_input_nonotify() function is supposed to return zero
if the direction for the pin is input.  But instead it accidentally
returns GPIO_LINE_DIRECTION_IN (1) which will be cast into an ERR_PTR()
in gpiochip_request_own_desc().  The callers dereference it and it leads
to a crash.

I changed gpiod_direction_output_raw_commit() just for consistency but
returning GPIO_LINE_DIRECTION_OUT (0) is fine.

Cc: stable@vger.kernel.org
Fixes: 9d846b1aebbe ("gpiolib: check the return value of gpio_chip::get_direction()")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://lore.kernel.org/r/254f3925-3015-4c9d-aac5-bb9b4b2cd2c5@stanley.mountain
[Bartosz: moved the variable declarations to the top of the functions]
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
4 months agoblock: fix 'kmem_cache of name 'bio-108' already exists' block-6.14-20250228
Ming Lei [Fri, 28 Feb 2025 13:26:56 +0000 (21:26 +0800)]
block: fix 'kmem_cache of name 'bio-108' already exists'

Device mapper bioset often has big bio_slab size, which can be more than
1000, then 8byte can't hold the slab name any more, cause the kmem_cache
allocation warning of 'kmem_cache of name 'bio-108' already exists'.

Fix the warning by extending bio_slab->name to 12 bytes, but fix output
of /proc/slabinfo

Reported-by: Guangwu Zhang <guazhang@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250228132656.2838008-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: cache nodes and mapped buffers
Keith Busch [Thu, 27 Feb 2025 22:39:16 +0000 (14:39 -0800)]
io_uring: cache nodes and mapped buffers

Frequent alloc/free cycles on these is pretty costly. Use an io cache to
more efficiently reuse these buffers.

Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20250227223916.143006-7-kbusch@meta.com
[axboe: fix imu leak]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoublk: zc register/unregister bvec
Keith Busch [Thu, 27 Feb 2025 22:39:15 +0000 (14:39 -0800)]
ublk: zc register/unregister bvec

Provide new operations for the user to request mapping an active request
to an io uring instance's buf_table. The user has to provide the index
it wants to install the buffer.

A reference count is taken on the request to ensure it can't be
completed while it is active in a ring's buf_table.

Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20250227223916.143006-6-kbusch@meta.com
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring: add support for kernel registered bvecs
Keith Busch [Thu, 27 Feb 2025 22:39:14 +0000 (14:39 -0800)]
io_uring: add support for kernel registered bvecs

Provide an interface for the kernel to leverage the existing
pre-registered buffers that io_uring provides. User space can reference
these later to achieve zero-copy IO.

User space must register an empty fixed buffer table with io_uring in
order for the kernel to make use of it.

Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20250227223916.143006-5-kbusch@meta.com
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agonvme: map uring_cmd data even if address is 0
Xinyu Zhang [Thu, 27 Feb 2025 22:39:13 +0000 (14:39 -0800)]
nvme: map uring_cmd data even if address is 0

When using kernel registered bvec fixed buffers, the "address" is
actually the offset into the bvec rather than userspace address.
Therefore it can be 0.

We can skip checking whether the address is NULL before mapping
uring_cmd data. Bad userspace address will be handled properly later when
the user buffer is imported.

With this patch, we will be able to use the kernel registered bvec fixed
buffers in io_uring NVMe passthru with ublk zero-copy support.

Reviewed-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Xinyu Zhang <xizhang@purestorage.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20250227223916.143006-4-kbusch@meta.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rw: move fixed buffer import to issue path
Keith Busch [Thu, 27 Feb 2025 22:39:12 +0000 (14:39 -0800)]
io_uring/rw: move fixed buffer import to issue path

Registered buffers may depend on a linked command, which makes the prep
path too early to import. Move to the issue path when the node is
actually needed like all the other users of fixed buffers.

Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20250227223916.143006-3-kbusch@meta.com
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoio_uring/rw: move buffer_select outside generic prep
Keith Busch [Thu, 27 Feb 2025 22:39:11 +0000 (14:39 -0800)]
io_uring/rw: move buffer_select outside generic prep

Cleans up the generic rw prep to not require the do_import flag. Use a
different prep function for callers that might need buffer select.

Based-on-a-patch-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20250227223916.143006-2-kbusch@meta.com
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 months agoiommu/vt-d: Fix suspicious RCU usage
Lu Baolu [Fri, 28 Feb 2025 10:27:26 +0000 (18:27 +0800)]
iommu/vt-d: Fix suspicious RCU usage

Commit <d74169ceb0d2> ("iommu/vt-d: Allocate DMAR fault interrupts
locally") moved the call to enable_drhd_fault_handling() to a code
path that does not hold any lock while traversing the drhd list. Fix
it by ensuring the dmar_global_lock lock is held when traversing the
drhd list.

Without this fix, the following warning is triggered:
 =============================
 WARNING: suspicious RCU usage
 6.14.0-rc3 #55 Not tainted
 -----------------------------
 drivers/iommu/intel/dmar.c:2046 RCU-list traversed in non-reader section!!
               other info that might help us debug this:
               rcu_scheduler_active = 1, debug_locks = 1
 2 locks held by cpuhp/1/23:
 #0: ffffffff84a67c50 (cpu_hotplug_lock){++++}-{0:0}, at: cpuhp_thread_fun+0x87/0x2c0
 #1: ffffffff84a6a380 (cpuhp_state-up){+.+.}-{0:0}, at: cpuhp_thread_fun+0x87/0x2c0
 stack backtrace:
 CPU: 1 UID: 0 PID: 23 Comm: cpuhp/1 Not tainted 6.14.0-rc3 #55
 Call Trace:
  <TASK>
  dump_stack_lvl+0xb7/0xd0
  lockdep_rcu_suspicious+0x159/0x1f0
  ? __pfx_enable_drhd_fault_handling+0x10/0x10
  enable_drhd_fault_handling+0x151/0x180
  cpuhp_invoke_callback+0x1df/0x990
  cpuhp_thread_fun+0x1ea/0x2c0
  smpboot_thread_fn+0x1f5/0x2e0
  ? __pfx_smpboot_thread_fn+0x10/0x10
  kthread+0x12a/0x2d0
  ? __pfx_kthread+0x10/0x10
  ret_from_fork+0x4a/0x60
  ? __pfx_kthread+0x10/0x10
  ret_from_fork_asm+0x1a/0x30
  </TASK>

Holding the lock in enable_drhd_fault_handling() triggers a lockdep splat
about a possible deadlock between dmar_global_lock and cpu_hotplug_lock.
This is avoided by not holding dmar_global_lock when calling
iommu_device_register(), which initiates the device probe process.

Fixes: d74169ceb0d2 ("iommu/vt-d: Allocate DMAR fault interrupts locally")
Reported-and-tested-by: Ido Schimmel <idosch@nvidia.com>
Closes: https://lore.kernel.org/linux-iommu/Zx9OwdLIc_VoQ0-a@shredder.mtl.com/
Tested-by: Breno Leitao <leitao@debian.org>
Cc: stable@vger.kernel.org
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20250218022422.2315082-1-baolu.lu@linux.intel.com
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
4 months agoiommu/vt-d: Remove device comparison in context_setup_pass_through_cb
Jerry Snitselaar [Fri, 28 Feb 2025 10:27:25 +0000 (18:27 +0800)]
iommu/vt-d: Remove device comparison in context_setup_pass_through_cb

Remove the device comparison check in context_setup_pass_through_cb.
pci_for_each_dma_alias already makes a decision on whether the
callback function should be called for a device. With the check
in place it will fail to create context entries for aliases as
it walks up to the root bus.

Fixes: 2031c469f816 ("iommu/vt-d: Add support for static identity domain")
Closes: https://lore.kernel.org/linux-iommu/82499eb6-00b7-4f83-879a-e97b4144f576@linux.intel.com/
Cc: stable@vger.kernel.org
Signed-off-by: Jerry Snitselaar <jsnitsel@redhat.com>
Link: https://lore.kernel.org/r/20250224180316.140123-1-jsnitsel@redhat.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
4 months agoiommu/amd: Preserve default DTE fields when updating Host Page Table Root
Alejandro Jimenez [Mon, 6 Jan 2025 19:14:13 +0000 (19:14 +0000)]
iommu/amd: Preserve default DTE fields when updating Host Page Table Root

When updating the page table root field on the DTE, avoid overwriting any
bits that are already set. The earlier call to make_clear_dte() writes
default values that all DTEs must have set (currently DTE[V]), and those
must be preserved.

Currently this doesn't cause problems since the page table root update is
the first field that is set after make_clear_dte() is called, and
DTE_FLAG_V is set again later along with the permission bits (IR/IW).
Remove this redundant assignment too.

Fixes: fd5dff9de4be ("iommu/amd: Modify set_dte_entry() to use 256-bit DTE helpers")
Signed-off-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Reviewed-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
Reviewed-by: Vasant Hegde <vasant.hegde@amd.com>
Link: https://lore.kernel.org/r/20250106191413.3107140-1-alejandro.j.jimenez@oracle.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
4 months agoMerge tag 'drm-fixes-2025-02-28' of https://gitlab.freedesktop.org/drm/kernel
Linus Torvalds [Fri, 28 Feb 2025 05:17:58 +0000 (21:17 -0800)]
Merge tag 'drm-fixes-2025-02-28' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "This week's fixes pull, amdgpu mostly, with some xe and a few misc
  others, the fb defio fix is bit of a change, but it avoids some nasty
  NULL pointer crashes due to defio assuming page backing in places it
  didn't have pages.

  amdgpu:
   - Legacy dpm suspend/resume fix
   - Runtime PM fix for DELL G5 SE
   - MAINTAINERS updates
   - Enforce Isolation fixes
   - mailmap update
   - EDID reading i2c fix
   - PSR fix
   - eDP fix
   - HPD interrupt handling fix
   - Clear memory fix

  amdkfd:
   - MQD handling fix

  vkms:
   - fix rounding error

  imagination:
   - header fix

  nouveau:
   - connector status fix

  fb/defio:
   - NULL ptr fix for defio drivers

  i915:
   - Fix encoder HW state readout for DP UHBR MST

  xe:
   - OA uapi fix (Umesh)
   - Userptr related fixes
   - Remove a duplicated register entry
   - Scheduler related fix to prevent exec races when freeing it"

* tag 'drm-fixes-2025-02-28' of https://gitlab.freedesktop.org/drm/kernel: (25 commits)
  drm/fbdev-dma: Add shadow buffering for deferred I/O
  drm/nouveau: Do not override forced connector status
  drm/i915/dp_mst: Fix encoder HW state readout for UHBR MST
  drm/xe: cancel pending job timer before freeing scheduler
  drm/xe/regs: remove a duplicate definition for RING_CTL_SIZE(size)
  drm/imagination: remove unnecessary header include path
  drm/amdgpu: init return value in amdgpu_ttm_clear_buffer
  drm/amd/display: Fix HPD after gpu reset
  drm/amd/display: add a quirk to enable eDP0 on DP1
  drm/amd/display: Disable PSR-SU on eDP panels
  MAINTAINERS: Update AMDGPU DML maintainers info
  drm/amd/display: restore edid reading from a given i2c adapter
  mailmap: Add entry for Rodrigo Siqueira
  MAINTAINERS: Change my role from Maintainer to Reviewer
  drm/amdgpu/mes: keep enforce isolation up to date
  drm/amdgpu/gfx: only call mes for enforce isolation if supported
  MAINTAINERS: update amdgpu maintainers list
  drm/amdgpu: disable BAR resize on Dell G5 SE
  drm/amdkfd: Preserve cp_hqd_pq_control on update_mqd
  amdgpu/pm/legacy: fix suspend/resume issues
  ...

4 months agoftrace: Avoid potential division by zero in function_stat_show()
Nikolay Kuratov [Thu, 6 Feb 2025 09:01:56 +0000 (12:01 +0300)]
ftrace: Avoid potential division by zero in function_stat_show()

Check whether denominator expression x * (x - 1) * 1000 mod {2^32, 2^64}
produce zero and skip stddev computation in that case.

For now don't care about rec->counter * rec->counter overflow because
rec->time * rec->time overflow will likely happen earlier.

Cc: stable@vger.kernel.org
Cc: Wen Yang <wenyang@linux.alibaba.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/20250206090156.1561783-1-kniv@yandex-team.ru
Fixes: e31f7939c1c27 ("ftrace: Avoid potential division by zero in function profiler")
Signed-off-by: Nikolay Kuratov <kniv@yandex-team.ru>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
4 months agoselftests/ftrace: Let fprobe test consider already enabled functions
Heiko Carstens [Wed, 26 Feb 2025 14:27:03 +0000 (15:27 +0100)]
selftests/ftrace: Let fprobe test consider already enabled functions

The fprobe test fails on Fedora 41 since the fprobe test assumption that
the number of enabled_functions is zero before the test starts is not
necessarily true. Some user space tools, like systemd, add BPF programs
that attach to functions. Those will show up in the enabled_functions table
and must be taken into account by the fprobe test.

Therefore count the number of lines of enabled_functions before tests
start, and use that as base when comparing expected results.

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Link: https://lore.kernel.org/20250226142703.910860-1-hca@linux.ibm.com
Fixes: e85c5e9792b9 ("selftests/ftrace: Update fprobe test to check enabled_functions file")
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
4 months agotracing: Fix bad hist from corrupting named_triggers list
Steven Rostedt [Thu, 27 Feb 2025 21:39:44 +0000 (16:39 -0500)]
tracing: Fix bad hist from corrupting named_triggers list

The following commands causes a crash:

 ~# cd /sys/kernel/tracing/events/rcu/rcu_callback
 ~# echo 'hist:name=bad:keys=common_pid:onmax(bogus).save(common_pid)' > trigger
 bash: echo: write error: Invalid argument
 ~# echo 'hist:name=bad:keys=common_pid' > trigger

Because the following occurs:

event_trigger_write() {
  trigger_process_regex() {
    event_hist_trigger_parse() {

      data = event_trigger_alloc(..);

      event_trigger_register(.., data) {
        cmd_ops->reg(.., data, ..) [hist_register_trigger()] {
          data->ops->init() [event_hist_trigger_init()] {
            save_named_trigger(name, data) {
              list_add(&data->named_list, &named_triggers);
            }
          }
        }
      }

      ret = create_actions(); (return -EINVAL)
      if (ret)
        goto out_unreg;
[..]
      ret = hist_trigger_enable(data, ...) {
        list_add_tail_rcu(&data->list, &file->triggers); <<<---- SKIPPED!!! (this is important!)
[..]
 out_unreg:
      event_hist_unregister(.., data) {
        cmd_ops->unreg(.., data, ..) [hist_unregister_trigger()] {
          list_for_each_entry(iter, &file->triggers, list) {
            if (!hist_trigger_match(data, iter, named_data, false))   <- never matches
                continue;
            [..]
            test = iter;
          }
          if (test && test->ops->free) <<<-- test is NULL

            test->ops->free(test) [event_hist_trigger_free()] {
              [..]
              if (data->name)
                del_named_trigger(data) {
                  list_del(&data->named_list);  <<<<-- NEVER gets removed!
                }
              }
           }
         }

         [..]
         kfree(data); <<<-- frees item but it is still on list

The next time a hist with name is registered, it causes an u-a-f bug and
the kernel can crash.

Move the code around such that if event_trigger_register() succeeds, the
next thing called is hist_trigger_enable() which adds it to the list.

A bunch of actions is called if get_named_trigger_data() returns false.
But that doesn't need to be called after event_trigger_register(), so it
can be moved up, allowing event_trigger_register() to be called just
before hist_trigger_enable() keeping them together and allowing the
file->triggers to be properly populated.

Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/20250227163944.1c37f85f@gandalf.local.home
Fixes: 067fe038e70f6 ("tracing: Add variable reference handling to hist triggers")
Reported-by: Tomas Glozar <tglozar@redhat.com>
Tested-by: Tomas Glozar <tglozar@redhat.com>
Reviewed-by: Tom Zanussi <zanussi@kernel.org>
Closes: https://lore.kernel.org/all/CAP4=nvTsxjckSBTz=Oe_UYh8keD9_sZC4i++4h72mJLic4_W4A@mail.gmail.com/
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
4 months agoMerge tag 'drm-xe-fixes-2025-02-27' of https://gitlab.freedesktop.org/drm/xe/kernel...
Dave Airlie [Fri, 28 Feb 2025 00:45:58 +0000 (10:45 +1000)]
Merge tag 'drm-xe-fixes-2025-02-27' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

uAPI:
- OA uapi fix (Umesh)

Driver:
- Userptr related fixes (Auld)
- Remove a duplicated register entry (Mingong)
- Scheduler related fix to prevent exec races when freeing it (Tejas)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/Z8CSqJre1VCjPXt2@intel.com
4 months agoMerge tag 'drm-intel-fixes-2025-02-27' of https://gitlab.freedesktop.org/drm/i915...
Dave Airlie [Fri, 28 Feb 2025 00:34:06 +0000 (10:34 +1000)]
Merge tag 'drm-intel-fixes-2025-02-27' of https://gitlab.freedesktop.org/drm/i915/kernel into drm-fixes

- Fix encoder HW state readout for DP UHBR MST (Imre)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/Z8CRM7XzlerbWSJy@intel.com
4 months agoMerge tag 'drm-misc-fixes-2025-02-27' of https://gitlab.freedesktop.org/drm/misc...
Dave Airlie [Thu, 27 Feb 2025 21:50:47 +0000 (07:50 +1000)]
Merge tag 'drm-misc-fixes-2025-02-27' of https://gitlab.freedesktop.org/drm/misc/kernel into drm-fixes

Fix a rounding error in vkms, a header fix for img, a connector status
fix for nouveau, and a NULL pointer dereference fix for deferred IO
drivers.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maxime Ripard <mripard@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250227-antique-robust-earthworm-09dfd1@houat
4 months agoMerge tag 'amd-drm-fixes-6.14-2025-02-26' of https://gitlab.freedesktop.org/agd5f...
Dave Airlie [Thu, 27 Feb 2025 20:18:26 +0000 (06:18 +1000)]
Merge tag 'amd-drm-fixes-6.14-2025-02-26' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-6.14-2025-02-26:

amdgpu:
- Legacy dpm suspend/resume fix
- Runtime PM fix for DELL G5 SE
- MAINTAINERS updates
- Enforce Isolation fixes
- mailmap update
- EDID reading i2c fix
- PSR fix
- eDP fix
- HPD interrupt handling fix
- Clear memory fix

amdkfd:
- MQD handling fix

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250226200342.3685347-1-alexander.deucher@amd.com
4 months agosched/core: Prevent rescheduling when interrupts are disabled
Thomas Gleixner [Mon, 16 Dec 2024 13:20:56 +0000 (14:20 +0100)]
sched/core: Prevent rescheduling when interrupts are disabled

David reported a warning observed while loop testing kexec jump:

  Interrupts enabled after irqrouter_resume+0x0/0x50
  WARNING: CPU: 0 PID: 560 at drivers/base/syscore.c:103 syscore_resume+0x18a/0x220
   kernel_kexec+0xf6/0x180
   __do_sys_reboot+0x206/0x250
   do_syscall_64+0x95/0x180

The corresponding interrupt flag trace:

  hardirqs last  enabled at (15573): [<ffffffffa8281b8e>] __up_console_sem+0x7e/0x90
  hardirqs last disabled at (15580): [<ffffffffa8281b73>] __up_console_sem+0x63/0x90

That means __up_console_sem() was invoked with interrupts enabled. Further
instrumentation revealed that in the interrupt disabled section of kexec
jump one of the syscore_suspend() callbacks woke up a task, which set the
NEED_RESCHED flag. A later callback in the resume path invoked
cond_resched() which in turn led to the invocation of the scheduler:

  __cond_resched+0x21/0x60
  down_timeout+0x18/0x60
  acpi_os_wait_semaphore+0x4c/0x80
  acpi_ut_acquire_mutex+0x3d/0x100
  acpi_ns_get_node+0x27/0x60
  acpi_ns_evaluate+0x1cb/0x2d0
  acpi_rs_set_srs_method_data+0x156/0x190
  acpi_pci_link_set+0x11c/0x290
  irqrouter_resume+0x54/0x60
  syscore_resume+0x6a/0x200
  kernel_kexec+0x145/0x1c0
  __do_sys_reboot+0xeb/0x240
  do_syscall_64+0x95/0x180

This is a long standing problem, which probably got more visible with
the recent printk changes. Something does a task wakeup and the
scheduler sets the NEED_RESCHED flag. cond_resched() sees it set and
invokes schedule() from a completely bogus context. The scheduler
enables interrupts after context switching, which causes the above
warning at the end.

Quite some of the code paths in syscore_suspend()/resume() can result in
triggering a wakeup with the exactly same consequences. They might not
have done so yet, but as they share a lot of code with normal operations
it's just a question of time.

The problem only affects the PREEMPT_NONE and PREEMPT_VOLUNTARY scheduling
models. Full preemption is not affected as cond_resched() is disabled and
the preemption check preemptible() takes the interrupt disabled flag into
account.

Cure the problem by adding a corresponding check into cond_resched().

Reported-by: David Woodhouse <dwmw@amazon.co.uk>
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: David Woodhouse <dwmw@amazon.co.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: stable@vger.kernel.org
Closes: https://lore.kernel.org/all/7717fe2ac0ce5f0a2c43fdab8b11f4483d54a2a4.camel@infradead.org

4 months agoarm64: hugetlb: Fix flush_hugetlb_tlb_range() invalidation level
Ryan Roberts [Wed, 26 Feb 2025 12:06:53 +0000 (12:06 +0000)]
arm64: hugetlb: Fix flush_hugetlb_tlb_range() invalidation level

commit c910f2b65518 ("arm64/mm: Update tlb invalidation routines for
FEAT_LPA2") changed the "invalidation level unknown" hint from 0 to
TLBI_TTL_UNKNOWN (INT_MAX). But the fallback "unknown level" path in
flush_hugetlb_tlb_range() was not updated. So as it stands, when trying
to invalidate CONT_PMD_SIZE or CONT_PTE_SIZE hugetlb mappings, we will
spuriously try to invalidate at level 0 on LPA2-enabled systems.

Fix this so that the fallback passes TLBI_TTL_UNKNOWN, and while we are
at it, explicitly use the correct stride and level for CONT_PMD_SIZE and
CONT_PTE_SIZE, which should provide a minor optimization.

Cc: stable@vger.kernel.org
Fixes: c910f2b65518 ("arm64/mm: Update tlb invalidation routines for FEAT_LPA2")
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Link: https://lore.kernel.org/r/20250226120656.2400136-4-ryan.roberts@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
4 months agoarm64: hugetlb: Fix huge_ptep_get_and_clear() for non-present ptes
Ryan Roberts [Wed, 26 Feb 2025 12:06:52 +0000 (12:06 +0000)]
arm64: hugetlb: Fix huge_ptep_get_and_clear() for non-present ptes

arm64 supports multiple huge_pte sizes. Some of the sizes are covered by
a single pte entry at a particular level (PMD_SIZE, PUD_SIZE), and some
are covered by multiple ptes at a particular level (CONT_PTE_SIZE,
CONT_PMD_SIZE). So the function has to figure out the size from the
huge_pte pointer. This was previously done by walking the pgtable to
determine the level and by using the PTE_CONT bit to determine the
number of ptes at the level.

But the PTE_CONT bit is only valid when the pte is present. For
non-present pte values (e.g. markers, migration entries), the previous
implementation was therefore erroneously determining the size. There is
at least one known caller in core-mm, move_huge_pte(), which may call
huge_ptep_get_and_clear() for a non-present pte. So we must be robust to
this case. Additionally the "regular" ptep_get_and_clear() is robust to
being called for non-present ptes so it makes sense to follow the
behavior.

Fix this by using the new sz parameter which is now provided to the
function. Additionally when clearing each pte in a contig range, don't
gather the access and dirty bits if the pte is not present.

An alternative approach that would not require API changes would be to
store the PTE_CONT bit in a spare bit in the swap entry pte for the
non-present case. But it felt cleaner to follow other APIs' lead and
just pass in the size.

As an aside, PTE_CONT is bit 52, which corresponds to bit 40 in the swap
entry offset field (layout of non-present pte). Since hugetlb is never
swapped to disk, this field will only be populated for markers, which
always set this bit to 0 and hwpoison swap entries, which set the offset
field to a PFN; So it would only ever be 1 for a 52-bit PVA system where
memory in that high half was poisoned (I think!). So in practice, this
bit would almost always be zero for non-present ptes and we would only
clear the first entry if it was actually a contiguous block. That's
probably a less severe symptom than if it was always interpreted as 1
and cleared out potentially-present neighboring PTEs.

Cc: stable@vger.kernel.org
Fixes: 66b3923a1a0f ("arm64: hugetlb: add support for PTE contiguous bit")
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Link: https://lore.kernel.org/r/20250226120656.2400136-3-ryan.roberts@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
4 months agomm: hugetlb: Add huge page size param to huge_ptep_get_and_clear()
Ryan Roberts [Wed, 26 Feb 2025 12:06:51 +0000 (12:06 +0000)]
mm: hugetlb: Add huge page size param to huge_ptep_get_and_clear()

In order to fix a bug, arm64 needs to be told the size of the huge page
for which the huge_pte is being cleared in huge_ptep_get_and_clear().
Provide for this by adding an `unsigned long sz` parameter to the
function. This follows the same pattern as huge_pte_clear() and
set_huge_pte_at().

This commit makes the required interface modifications to the core mm as
well as all arches that implement this function (arm64, loongarch, mips,
parisc, powerpc, riscv, s390, sparc). The actual arm64 bug will be fixed
in a separate commit.

Cc: stable@vger.kernel.org
Fixes: 66b3923a1a0f ("arm64: hugetlb: add support for PTE contiguous bit")
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Alexandre Ghiti <alexghiti@rivosinc.com> # riscv
Reviewed-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com> # s390
Link: https://lore.kernel.org/r/20250226120656.2400136-2-ryan.roberts@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
4 months agoMerge tag 'net-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 27 Feb 2025 17:32:42 +0000 (09:32 -0800)]
Merge tag 'net-6.14-rc5' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from bluetooth.

  We didn't get netfilter or wireless PRs this week, so next week's PR
  is probably going to be bigger. A healthy dose of fixes for bugs
  introduced in the current release nonetheless.

  Current release - regressions:

   - Bluetooth: always allow SCO packets for user channel

   - af_unix: fix memory leak in unix_dgram_sendmsg()

   - rxrpc:
       - remove redundant peer->mtu_lock causing lockdep splats
       - fix spinlock flavor issues with the peer record hash

   - eth: iavf: fix circular lock dependency with netdev_lock

   - net: use rtnl_net_dev_lock() in
     register_netdevice_notifier_dev_net() RDMA driver register notifier
     after the device

  Current release - new code bugs:

   - ethtool: fix ioctl confusing drivers about desired HDS user config

   - eth: ixgbe: fix media cage present detection for E610 device

  Previous releases - regressions:

   - loopback: avoid sending IP packets without an Ethernet header

   - mptcp: reset connection when MPTCP opts are dropped after join

  Previous releases - always broken:

   - net: better track kernel sockets lifetime

   - ipv6: fix dst ref loop on input in seg6 and rpl lw tunnels

   - phy: qca807x: use right value from DTS for DAC_DSP_BIAS_CURRENT

   - eth: enetc: number of error handling fixes

   - dsa: rtl8366rb: reshuffle the code to fix config / build issue with
     LED support"

* tag 'net-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (53 commits)
  net: ti: icss-iep: Reject perout generation request
  idpf: fix checksums set in idpf_rx_rsc()
  selftests: drv-net: Check if combined-count exists
  net: ipv6: fix dst ref loop on input in rpl lwt
  net: ipv6: fix dst ref loop on input in seg6 lwt
  usbnet: gl620a: fix endpoint checking in genelink_bind()
  net/mlx5: IRQ, Fix null string in debug print
  net/mlx5: Restore missing trace event when enabling vport QoS
  net/mlx5: Fix vport QoS cleanup on error
  net: mvpp2: cls: Fixed Non IP flow, with vlan tag flow defination.
  af_unix: Fix memory leak in unix_dgram_sendmsg()
  net: Handle napi_schedule() calls from non-interrupt
  net: Clear old fragment checksum value in napi_reuse_skb
  gve: unlink old napi when stopping a queue using queue API
  net: Use rtnl_net_dev_lock() in register_netdevice_notifier_dev_net().
  tcp: Defer ts_recent changes until req is owned
  net: enetc: fix the off-by-one issue in enetc_map_tx_tso_buffs()
  net: enetc: remove the mm_lock from the ENETC v4 driver
  net: enetc: add missing enetc4_link_deinit()
  net: enetc: update UDP checksum when updating originTimestamp field
  ...

4 months agoefi/mokvar-table: Avoid repeated map/unmap of the same page
Ard Biesheuvel [Thu, 27 Feb 2025 13:30:22 +0000 (14:30 +0100)]
efi/mokvar-table: Avoid repeated map/unmap of the same page

Tweak the logic that traverses the MOKVAR UEFI configuration table to
only unmap the entry header and map the next one if they don't live in
the same physical page.

Link: https://lore.kernel.org/all/8f085931-3e9d-4386-9209-1d6c95616327@uncooperative.org/
Tested-By: Peter Jones <pjones@redhat.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
4 months agoefi: Don't map the entire mokvar table to determine its size
Peter Jones [Wed, 26 Feb 2025 20:18:39 +0000 (15:18 -0500)]
efi: Don't map the entire mokvar table to determine its size

Currently, when validating the mokvar table, we (re)map the entire table
on each iteration of the loop, adding space as we discover new entries.
If the table grows over a certain size, this fails due to limitations of
early_memmap(), and we get a failure and traceback:

  ------------[ cut here ]------------
  WARNING: CPU: 0 PID: 0 at mm/early_ioremap.c:139 __early_ioremap+0xef/0x220
  ...
  Call Trace:
   <TASK>
   ? __early_ioremap+0xef/0x220
   ? __warn.cold+0x93/0xfa
   ? __early_ioremap+0xef/0x220
   ? report_bug+0xff/0x140
   ? early_fixup_exception+0x5d/0xb0
   ? early_idt_handler_common+0x2f/0x3a
   ? __early_ioremap+0xef/0x220
   ? efi_mokvar_table_init+0xce/0x1d0
   ? setup_arch+0x864/0xc10
   ? start_kernel+0x6b/0xa10
   ? x86_64_start_reservations+0x24/0x30
   ? x86_64_start_kernel+0xed/0xf0
   ? common_startup_64+0x13e/0x141
   </TASK>
  ---[ end trace 0000000000000000 ]---
  mokvar: Failed to map EFI MOKvar config table pa=0x7c4c3000, size=265187.

Mapping the entire structure isn't actually necessary, as we don't ever
need more than one entry header mapped at once.

Changes efi_mokvar_table_init() to only map each entry header, not the
entire table, when determining the table size.  Since we're not mapping
any data past the variable name, it also changes the code to enforce
that each variable name is NUL terminated, rather than attempting to
verify it in place.

Cc: <stable@vger.kernel.org>
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
4 months agoMerge tag 'sound-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Thu, 27 Feb 2025 16:41:19 +0000 (08:41 -0800)]
Merge tag 'sound-6.14-rc5' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of fixes. The only slightly large change is for ASoC
  Cirrus codec, but that's still in a normal range. All the rest are
  small device-specific fixes and should be fairly safe to take"

* tag 'sound-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda/realtek: Fix microphone regression on ASUS N705UD
  ALSA: hda/realtek: Fix wrong mic setup for ASUS VivoBook 15
  ASoC: cs35l56: Prevent races when soft-resetting using SPI control
  firmware: cs_dsp: Remove async regmap writes
  ASoC: Intel: sof_sdw: warn both sdw and pch dmic are used
  ASoC: SOF: Intel: don't check number of sdw links when set dmic_fixup
  ASoC: dapm-graph: set fill colour of turned on nodes
  ASoC: fsl: Rename stream name of SAI DAI driver
  ASoC: es8328: fix route from DAC to output
  ALSA: usb-audio: Re-add sample rate quirk for Pioneer DJM-900NXS2
  ASoC: tas2764: Set the SDOUT polarity correctly
  ASoC: tas2764: Fix power control mask
  ALSA: usb-audio: Avoid dropping MIDI events at closing multiple ports
  ASoC: tas2770: Fix volume scale

4 months agonet: ti: icss-iep: Reject perout generation request
Meghana Malladi [Thu, 27 Feb 2025 09:24:41 +0000 (14:54 +0530)]
net: ti: icss-iep: Reject perout generation request

IEP driver supports both perout and pps signal generation
but perout feature is faulty with half-cooked support
due to some missing configuration. Remove perout
support from the driver and reject perout requests with
"not supported" error code.

Fixes: c1e0230eeaab2 ("net: ti: icss-iep: Add IEP driver")
Signed-off-by: Meghana Malladi <m-malladi@ti.com>
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Link: https://patch.msgid.link/20250227092441.1848419-1-m-malladi@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>