From 54559642b96116b45e4b5ca7fd9f7835b8561272 Mon Sep 17 00:00:00 2001 From: Pavel Begunkov Date: Wed, 12 Jun 2024 13:56:38 +0100 Subject: [PATCH 1/2] io_uring/rsrc: don't lock while !TASK_RUNNING There is a report of io_rsrc_ref_quiesce() locking a mutex while not TASK_RUNNING, which is due to forgetting restoring the state back after io_run_task_work_sig() and attempts to break out of the waiting loop. do not call blocking ops when !TASK_RUNNING; state=1 set at [] prepare_to_wait+0xa4/0x380 kernel/sched/wait.c:237 WARNING: CPU: 2 PID: 397056 at kernel/sched/core.c:10099 __might_sleep+0x114/0x160 kernel/sched/core.c:10099 RIP: 0010:__might_sleep+0x114/0x160 kernel/sched/core.c:10099 Call Trace: __mutex_lock_common kernel/locking/mutex.c:585 [inline] __mutex_lock+0xb4/0x940 kernel/locking/mutex.c:752 io_rsrc_ref_quiesce+0x590/0x940 io_uring/rsrc.c:253 io_sqe_buffers_unregister+0xa2/0x340 io_uring/rsrc.c:799 __io_uring_register io_uring/register.c:424 [inline] __do_sys_io_uring_register+0x5b9/0x2400 io_uring/register.c:613 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xd8/0x270 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x6f/0x77 Reported-by: Li Shi Fixes: 4ea15b56f0810 ("io_uring/rsrc: use wq for quiescing") Cc: stable@vger.kernel.org Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/77966bc104e25b0534995d5dbb152332bc8f31c0.1718196953.git.asml.silence@gmail.com Signed-off-by: Jens Axboe --- io_uring/rsrc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c index 65417c9553b1..edb9c5baf2e2 100644 --- a/io_uring/rsrc.c +++ b/io_uring/rsrc.c @@ -249,6 +249,7 @@ __cold static int io_rsrc_ref_quiesce(struct io_rsrc_data *data, ret = io_run_task_work_sig(ctx); if (ret < 0) { + __set_current_state(TASK_RUNNING); mutex_lock(&ctx->uring_lock); if (list_empty(&ctx->rsrc_ref_list)) ret = 0; From f4a1254f2a076afb0edd473589bf40f9b4d36b41 Mon Sep 17 00:00:00 2001 From: Pavel Begunkov Date: Fri, 14 Jun 2024 01:04:29 +0100 Subject: [PATCH 2/2] io_uring: fix cancellation overwriting req->flags Only the current owner of a request is allowed to write into req->flags. Hence, the cancellation path should never touch it. Add a new field instead of the flag, move it into the 3rd cache line because it should always be initialised. poll_refs can move further as polling is an involved process anyway. It's a minimal patch, in the future we can and should find a better place for it and remove now unused REQ_F_CANCEL_SEQ. Fixes: 521223d7c229f ("io_uring/cancel: don't default to setting req->work.cancel_seq") Cc: stable@vger.kernel.org Reported-by: Li Shi Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/6827b129f8f0ad76fa9d1f0a773de938b240ffab.1718323430.git.asml.silence@gmail.com Signed-off-by: Jens Axboe --- include/linux/io_uring_types.h | 3 ++- io_uring/cancel.h | 4 ++-- io_uring/io_uring.c | 1 + 3 files changed, 5 insertions(+), 3 deletions(-) diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h index 7a6b190c7da7..b48570eaa449 100644 --- a/include/linux/io_uring_types.h +++ b/include/linux/io_uring_types.h @@ -648,7 +648,7 @@ struct io_kiocb { struct io_rsrc_node *rsrc_node; atomic_t refs; - atomic_t poll_refs; + bool cancel_seq_set; struct io_task_work io_task_work; /* for polled requests, i.e. IORING_OP_POLL_ADD and async armed poll */ struct hlist_node hash_node; @@ -657,6 +657,7 @@ struct io_kiocb { /* opcode allocated if it needs to store data for async defer */ void *async_data; /* linked requests, IFF REQ_F_HARDLINK or REQ_F_LINK are set */ + atomic_t poll_refs; struct io_kiocb *link; /* custom credentials, valid IFF REQ_F_CREDS is set */ const struct cred *creds; diff --git a/io_uring/cancel.h b/io_uring/cancel.h index 76b32e65c03c..b33995e00ba9 100644 --- a/io_uring/cancel.h +++ b/io_uring/cancel.h @@ -27,10 +27,10 @@ bool io_cancel_req_match(struct io_kiocb *req, struct io_cancel_data *cd); static inline bool io_cancel_match_sequence(struct io_kiocb *req, int sequence) { - if ((req->flags & REQ_F_CANCEL_SEQ) && sequence == req->work.cancel_seq) + if (req->cancel_seq_set && sequence == req->work.cancel_seq) return true; - req->flags |= REQ_F_CANCEL_SEQ; + req->cancel_seq_set = true; req->work.cancel_seq = sequence; return false; } diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 816e93e7f949..154b25b8a613 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -2058,6 +2058,7 @@ static int io_init_req(struct io_ring_ctx *ctx, struct io_kiocb *req, req->file = NULL; req->rsrc_node = NULL; req->task = current; + req->cancel_seq_set = false; if (unlikely(opcode >= IORING_OP_LAST)) { req->opcode = 0;