| Message ID | 343c03ed-d56a-e575-bae6-9d015ab8b5e2@kernel.dk (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | io_uring: recycle provided buffer if we punt to io-wq | expand |
diff --git a/fs/io_uring.c b/fs/io_uring.c index 95a1a78d799a..d3ee4fc532fa 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -8690,6 +8690,7 @@ static void io_queue_async(struct io_kiocb *req, int ret) * Queued up for async execution, worker will release * submit reference when the iocb is actually submitted. */ + io_kbuf_recycle(req, 0); io_queue_iowq(req, NULL); break; case IO_APOLL_OK:
io_arm_poll_handler() will recycle the buffer appropriately if we end up arming poll (or if we're ready to retry), but not for the io-wq case if we have attempted poll first. Explicitly recycle the buffer to avoid both hanging on to it too long, but also to avoid multiple reads grabbing the same one. This can happen for ring mapped buffers, since it hasn't necessarily been committed. Fixes: c7fb19428d67 ("io_uring: add support for ring mapped supplied buffers") Link: https://github.com/axboe/liburing/issues/605 Signed-off-by: Jens Axboe <axboe@kernel.dk> ---