[28/32] aio: simplify cancellation

Commit Message

Christoph Hellwig Jan. 10, 2018, 3:58 p.m. UTC

With the current aio code there is no need for the magic KIOCB_CANCELLED
value, as a cancelation just kicks the driver to queue the completion
ASAP, with all actual completion handling done in another thread. Given
that both the completion path and cancelation take the context lock there
is no need for magic cmpxchg loops either.

Signed-off-by: Christoph Hellwig <[e-mail]>
---
 fs/aio.c | 37 +++++++++----------------------------
 1 file changed, 9 insertions(+), 28 deletions(-)

Comments

Jeff Moyer Jan. 10, 2018, 10:50 p.m. UTC | #1

Christoph Hellwig <[e-mail]> writes:

> With the current aio code there is no need for the magic KIOCB_CANCELLED
> value, as a cancelation just kicks the driver to queue the completion
> ASAP, with all actual completion handling done in another thread. Given
> that both the completion path and cancelation take the context lock there
> is no need for magic cmpxchg loops either.

The cmpxchg was made unnecessary when the complementary xchg was removed
in commit bec68faaf ("aio: io_cancel() no longer returns the io_event").
There was no longer a need to ensure only the completion path or the
cancellation path delivered the completion event.  Now, the completion
is always delivered via aio_complete.

So yeah, we can get rid of this.

Acked-by: Jeff Moyer <[e-mail]>

Patch

diff --git a/fs/aio.c b/fs/aio.c
index fe241b5b44b2..6204c1b81a36 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -156,19 +156,6 @@  struct kioctx {
 	unsigned		id;
 };
 
-/*
- * We use ki_cancel == KIOCB_CANCELLED to indicate that a kiocb has been either
- * cancelled or completed (this makes a certain amount of sense because
- * successful cancellation - io_cancel() - does deliver the completion to
- * userspace).
- *
- * And since most things don't implement kiocb cancellation and we'd really like
- * kiocb completion to be lockless when possible, we use ki_cancel to
- * synchronize cancellation and completion - we only set it to KIOCB_CANCELLED
- * with xchg() or cmpxchg(), see batch_complete_aio() and kiocb_cancel().
- */
-#define KIOCB_CANCELLED		((void *) (~0ULL))
-
 struct aio_kiocb {
 	union {
 		struct kiocb		rw;
@@ -565,24 +552,18 @@  void kiocb_set_cancel_fn(struct kiocb *iocb, kiocb_cancel_fn *cancel)
 }
 EXPORT_SYMBOL(kiocb_set_cancel_fn);
 
+/*
+ * Only cancel if there ws a ki_cancel function to start with, and we
+ * are the one how managed to clear it (to protect against simulatinious
+ * cancel calls).
+ */
 static int kiocb_cancel(struct aio_kiocb *kiocb)
 {
-	kiocb_cancel_fn *old, *cancel;
-
-	/*
-	 * Don't want to set kiocb->ki_cancel = KIOCB_CANCELLED unless it
-	 * actually has a cancel function, hence the cmpxchg()
-	 */
-
-	cancel = READ_ONCE(kiocb->ki_cancel);
-	do {
-		if (!cancel || cancel == KIOCB_CANCELLED)
-			return -EINVAL;
-
-		old = cancel;
-		cancel = cmpxchg(&kiocb->ki_cancel, old, KIOCB_CANCELLED);
-	} while (cancel != old);
+	kiocb_cancel_fn *cancel = kiocb->ki_cancel;
 
+	if (!cancel)
+		return -EINVAL;
+	kiocb->ki_cancel = NULL;
 	return cancel(&kiocb->rw);
 }