diff options
| author | Jens Axboe <axboe@kernel.dk> | 2021-01-16 19:52:11 +0100 |
|---|---|---|
| committer | Jens Axboe <axboe@kernel.dk> | 2021-01-16 20:13:59 +0100 |
| commit | c93cc9e16d88e0f5ea95d2d65d58a8a4dab258bc (patch) | |
| tree | 475e92df72996c3077abb8c0ab41e243dc2b85bc | |
| parent | io_uring: ensure finish_wait() is always called in __io_uring_task_cancel() (diff) | |
| download | linux-c93cc9e16d88e0f5ea95d2d65d58a8a4dab258bc.tar.xz linux-c93cc9e16d88e0f5ea95d2d65d58a8a4dab258bc.zip | |
io_uring: iopoll requests should also wake task ->in_idle state
If we're freeing/finishing iopoll requests, ensure we check if the task
is in idling in terms of cancelation. Otherwise we could end up waiting
forever in __io_uring_task_cancel() if the task has active iopoll
requests that need cancelation.
Cc: stable@vger.kernel.org # 5.9+
Signed-off-by: Jens Axboe <axboe@kernel.dk>
| -rw-r--r-- | fs/io_uring.c | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/fs/io_uring.c b/fs/io_uring.c index 985a9e3f976d..5cda878b69cf 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -2270,6 +2270,8 @@ static void io_req_free_batch_finish(struct io_ring_ctx *ctx, struct io_uring_task *tctx = rb->task->io_uring; percpu_counter_sub(&tctx->inflight, rb->task_refs); + if (atomic_read(&tctx->in_idle)) + wake_up(&tctx->wait); put_task_struct_many(rb->task, rb->task_refs); rb->task = NULL; } @@ -2288,6 +2290,8 @@ static void io_req_free_batch(struct req_batch *rb, struct io_kiocb *req) struct io_uring_task *tctx = rb->task->io_uring; percpu_counter_sub(&tctx->inflight, rb->task_refs); + if (atomic_read(&tctx->in_idle)) + wake_up(&tctx->wait); put_task_struct_many(rb->task, rb->task_refs); } rb->task = req->task; |