ceph: increment/decrement dio counter on async requests

Ceph can in some cases issue an async DIO request, in which case we can
end up calling ceph_end_io_direct before the I/O is actually complete.
That may allow buffered operations to proceed while DIO requests are
still in flight.

Fix this by incrementing the i_dio_count when issuing an async DIO
request, and decrement it when tearing down the aio_req.

Fixes: 321fe13c93 ("ceph: add buffered/direct exclusionary locking for reads and writes")
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
This commit is contained in:
Jeff Layton 2019-11-13 09:56:06 -05:00 committed by Ilya Dryomov
parent a81bc3102b
commit 6a81749ebe

View File

@ -753,6 +753,9 @@ static void ceph_aio_complete(struct inode *inode,
if (!atomic_dec_and_test(&aio_req->pending_reqs))
return;
if (aio_req->iocb->ki_flags & IOCB_DIRECT)
inode_dio_end(inode);
ret = aio_req->error;
if (!ret)
ret = aio_req->total_len;
@ -1091,6 +1094,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
CEPH_CAP_FILE_RD);
list_splice(&aio_req->osd_reqs, &osd_reqs);
inode_dio_begin(inode);
while (!list_empty(&osd_reqs)) {
req = list_first_entry(&osd_reqs,
struct ceph_osd_request,