From: Pavel Begunkov Date: Sun, 3 Dec 2023 15:37:53 +0000 (+0000) Subject: io_uring: fix mutex_unlock with unreferenced ctx X-Git-Tag: v5.15.143~62 X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=ea3291cb9075254ae9db999b1f3f89f42ad1009b;p=thirdparty%2Fkernel%2Fstable.git io_uring: fix mutex_unlock with unreferenced ctx commit f7b32e785042d2357c5abc23ca6db1b92c91a070 upstream. Callers of mutex_unlock() have to make sure that the mutex stays alive for the whole duration of the function call. For io_uring that means that the following pattern is not valid unless we ensure that the context outlives the mutex_unlock() call. mutex_lock(&ctx->uring_lock); req_put(req); // typically via io_req_task_submit() mutex_unlock(&ctx->uring_lock); Most contexts are fine: io-wq pins requests, syscalls hold the file, task works are taking ctx references and so on. However, the task work fallback path doesn't follow the rule. Cc: Fixes: 04fc6c802d ("io_uring: save ctx put/get for task_work submit") Reported-by: Jann Horn Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/io-uring/CAG48ez3xSoYb+45f1RLtktROJrpiDQ1otNvdR+YLQf7m+Krj5Q@mail.gmail.com/ Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman --- diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 5b2dbd3dc2deb..95e7ce84fa2ac 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -2601,6 +2601,7 @@ static void io_iopoll_try_reap_events(struct io_ring_ctx *ctx) if (!(ctx->flags & IORING_SETUP_IOPOLL)) return; + percpu_ref_get(&ctx->refs); mutex_lock(&ctx->uring_lock); while (!list_empty(&ctx->iopoll_list)) { unsigned int nr_events = 0; @@ -2622,6 +2623,7 @@ static void io_iopoll_try_reap_events(struct io_ring_ctx *ctx) } } mutex_unlock(&ctx->uring_lock); + percpu_ref_put(&ctx->refs); } static int io_iopoll_check(struct io_ring_ctx *ctx, long min) @@ -9771,12 +9773,7 @@ static void io_ring_exit_work(struct work_struct *work) init_completion(&exit.completion); init_task_work(&exit.task_work, io_tctx_exit_cb); exit.ctx = ctx; - /* - * Some may use context even when all refs and requests have been put, - * and they are free to do so while still holding uring_lock or - * completion_lock, see io_req_task_submit(). Apart from other work, - * this lock/unlock section also waits them to finish. - */ + mutex_lock(&ctx->uring_lock); while (!list_empty(&ctx->tctx_list)) { WARN_ON_ONCE(time_after(jiffies, timeout));