io_uring: perform IOPOLL reaping if canceler is thread itself

We bypass IOPOLL completion polling (and reaping) for the SQPOLL thread,
but if it's the thread itself invoking cancelations, then we still need
to perform it or no one will.

Fixes: 9936c7c2bc76 ("io_uring: deduplicate core cancellations sequence")
Signed-off-by: Jens Axboe <[email protected]>

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 472eab7359f2a26c9d3c60e6be5e4653089f9c81..49f85f49e1c3c2aac5383f5d1dc2abf8122a238a 100644 (file)
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -8658,7 +8658,8 @@ static void io_uring_try_cancel_requests(struct io_ring_ctx *ctx,
                }
 
                /* SQPOLL thread does its own polling */
-               if (!(ctx->flags & IORING_SETUP_SQPOLL) && !files) {
+               if ((!(ctx->flags & IORING_SETUP_SQPOLL) && !files) ||
+                   (ctx->sq_data && ctx->sq_data->thread == current)) {
                        while (!list_empty_careful(&ctx->iopoll_list)) {
                                io_iopoll_try_reap_events(ctx);
                                ret = true;