Skip to contentSkip to conversation

ocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio

PATCH New Ryan Ding 3 participants 3 messages Aug 24, 2015 OCFS2

lore
Share & tools
lore

Ryan Ding

author
#0 lore raw

PATCH 1/1 ocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio

Orabug: 21612107

Use wrong return value in ocfs2_file_write_iter(). This will cause ocfs2_rw_unlock() be called both in write_iter & end_io, and trigger a BUG_ON.

This issue exist since commit 7da839c475894ea872ec909a5d2e83dddccff5be.

Signed-off-by Ryan Ding <[email protected]>
---
 fs/ocfs2/file.c |   28 ++++++++++++++--------------
 1 files changed, 14 insertions(+), 14 deletions(-)
fs/ocfs2/file.c +14 -14
index 719f7f4..33efa33 100644
@@ -2372,6 +2372,20 @@ relock:
 	/* buffered aio wouldn't have proper lock coverage today */
 	BUG_ON(written == -EIOCBQUEUED && !(iocb->ki_flags & IOCB_DIRECT));
 
+	/*
+	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
+	 * function pointer which is called when o_direct io completes so that
+	 * it can unlock our rw lock.
+	 * Unfortunately there are error cases which call end_io and others
+	 * that don't.  so we don't have to unlock the rw_lock if either an
+	 * async dio is going to do it in the future or an end_io after an
+	 * error has already done it.
+	 */
+	if ((written == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
+		rw_level = -1;
+		unaligned_dio = 0;
+	}
+
 	if (unlikely(written <= 0))
 		goto no_sync;
 

@@ -2396,20 +2410,6 @@ relock:
 	}
 
 no_sync:
-	/*
-	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
-	 * function pointer which is called when o_direct io completes so that
-	 * it can unlock our rw lock.
-	 * Unfortunately there are error cases which call end_io and others
-	 * that don't.  so we don't have to unlock the rw_lock if either an
-	 * async dio is going to do it in the future or an end_io after an
-	 * error has already done it.
-	 */
-	if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
-		rw_level = -1;
-		unaligned_dio = 0;
-	}
-
 	if (unaligned_dio) {
 		ocfs2_iocb_clear_unaligned_aio(iocb);
 		mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio);
Signature · 1 line
1.7.1
#1 lore raw

PATCH 1/1 ocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio

Quoted 6 lines · Ryan Ding

Orabug: 21612107

Use wrong return value in ocfs2_file_write_iter(). This will cause ocfs2_rw_unlock() be called both in write_iter & end_io, and trigger a BUG_ON.

This issue exist since commit 7da839c475894ea872ec909a5d2e83dddccff5be.

Better say:
This issue is introduced by commit 7da839c47589 ("ocfs2: use __generic_file_write_iter()") , or checkpatch will report a style error. Other looks good.

Reviewed-by Junxiao Bi <[email protected]>

Thanks,
Junxiao.

Quoted 53 lines
Signed-off-by Ryan Ding <[email protected]>
---
 fs/ocfs2/file.c |   28 ++++++++++++++--------------
 1 files changed, 14 insertions(+), 14 deletions(-)
fs/ocfs2/file.c +14 -14
index 719f7f4..33efa33 100644
@@ -2372,6 +2372,20 @@ relock:
 	/* buffered aio wouldn't have proper lock coverage today */
 	BUG_ON(written == -EIOCBQUEUED && !(iocb->ki_flags & IOCB_DIRECT));
 
+	/*
+	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
+	 * function pointer which is called when o_direct io completes so that
+	 * it can unlock our rw lock.
+	 * Unfortunately there are error cases which call end_io and others
+	 * that don't.  so we don't have to unlock the rw_lock if either an
+	 * async dio is going to do it in the future or an end_io after an
+	 * error has already done it.
+	 */
+	if ((written == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
+		rw_level = -1;
+		unaligned_dio = 0;
+	}
+
 	if (unlikely(written <= 0))
 		goto no_sync;
 

@@ -2396,20 +2410,6 @@ relock:
 	}
 
 no_sync:
-	/*
-	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
-	 * function pointer which is called when o_direct io completes so that
-	 * it can unlock our rw lock.
-	 * Unfortunately there are error cases which call end_io and others
-	 * that don't.  so we don't have to unlock the rw_lock if either an
-	 * async dio is going to do it in the future or an end_io after an
-	 * error has already done it.
-	 */
-	if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
-		rw_level = -1;
-		unaligned_dio = 0;
-	}
-
 	if (unaligned_dio) {
 		ocfs2_iocb_clear_unaligned_aio(iocb);
 		mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio);
#2 lore raw

PATCH 1/1 ocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio

Quoted 6 lines · On Mon, 24 Aug 2015 15:39:04 +0800 Junxiao Bi <[email protected]> wrote:
Quoted 5 lines · Ryan Ding

Orabug: 21612107

Use wrong return value in ocfs2_file_write_iter(). This will cause ocfs2_rw_unlock() be called both in write_iter & end_io, and trigger a BUG_ON.

Quoted 8 lines
Quoted 1 line

This issue exist since commit 7da839c475894ea872ec909a5d2e83dddccff5be.

Better say:
This issue is introduced by commit 7da839c47589 ("ocfs2: use __generic_file_write_iter()") , or checkpatch will report a style error. Other looks good.

Reviewed-by Junxiao Bi <[email protected]>

Also we've recently adopted the convention of using the "Fixes:" tag:

Fixes 7da839c47589 ("ocfs2: use __generic_file_write_iter()")

And as the patch fixes a regression we should add the "Cc:
<[email protected]>" tag.

And we should the author of the bad patch for review (hi, Al).

I've made these changes and I'll get this patch into Linus this week, for 4.2.

From: Ryan Ding <[email protected]>
Subject: ocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio

Orabug: 21612107

ocfs2_file_write_iter() is usng the wrong return value ('written'). This will cause ocfs2_rw_unlock() be called both in write_iter & end_io, triggering a BUG_ON.

This issue was introduced by commit 7da839c47589 ("ocfs2: use __generic_file_write_iter()").

Fixes 7da839c47589 ("ocfs2: use __generic_file_write_iter()")
Signed-off-by Ryan Ding <[email protected]>
Reviewed-by Junxiao Bi <[email protected]>
Cc Al Viro <[email protected]>
Cc Mark Fasheh <[email protected]>
Cc Joel Becker <[email protected]>
Cc <[email protected]>
Signed-off-by Andrew Morton <[email protected]>

---

 fs/ocfs2/file.c |   28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)
fs/ocfs2/file.c renamed from fs/ocfs2/file.c~ocfs2-direct-write-will-call-ocfs2_rw_unlock-twice-when-doing-aiodio +14 -14
diff -puN fs/ocfs2/file.c~ocfs2-direct-write-will-call-ocfs2_rw_unlock-twice-when-doing-aiodio fs/ocfs2/file.c
@@ -2366,6 +2366,20 @@ relock:
 	/* buffered aio wouldn't have proper lock coverage today */
 	BUG_ON(written == -EIOCBQUEUED && !(iocb->ki_flags & IOCB_DIRECT));
 
+	/*
+	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
+	 * function pointer which is called when o_direct io completes so that
+	 * it can unlock our rw lock.
+	 * Unfortunately there are error cases which call end_io and others
+	 * that don't.  so we don't have to unlock the rw_lock if either an
+	 * async dio is going to do it in the future or an end_io after an
+	 * error has already done it.
+	 */
+	if ((written == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
+		rw_level = -1;
+		unaligned_dio = 0;
+	}
+
 	if (unlikely(written <= 0))
 		goto no_sync;
 

@@ -2390,20 +2404,6 @@ relock:
 	}
 
 no_sync:
-	/*
-	 * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io
-	 * function pointer which is called when o_direct io completes so that
-	 * it can unlock our rw lock.
-	 * Unfortunately there are error cases which call end_io and others
-	 * that don't.  so we don't have to unlock the rw_lock if either an
-	 * async dio is going to do it in the future or an end_io after an
-	 * error has already done it.
-	 */
-	if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) {
-		rw_level = -1;
-		unaligned_dio = 0;
-	}
-
 	if (unaligned_dio && ocfs2_iocb_is_unaligned_aio(iocb)) {
 		ocfs2_iocb_clear_unaligned_aio(iocb);
 		mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio);

_