xfs: Don't issue buffer IO direct from AIL push V2
All buffers logged into the AIL are marked as delayed write.
When the AIL needs to push the buffer out, it issues an async write of the
buffer. This means that IO patterns are dependent on the order of
buffers in the AIL.
Instead of flushing the buffer, promote the buffer in the delayed
write list so that the next time the xfsbufd is run the buffer will
be flushed by the xfsbufd. Return the state to the xfsaild that the
buffer was promoted so that the xfsaild knows that it needs to cause
the xfsbufd to run to flush the buffers that were promoted.
Using the xfsbufd for issuing the IO allows us to dispatch all
buffer IO from the one queue. This means that we can make much more
enlightened decisions on what order to flush buffers to disk as
we don't have multiple places issuing IO. Optimisations to xfsbufd
will be in a future patch.
Version 2
- kill XFS_ITEM_FLUSHING as it is now unused.
Signed-off-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index e0a1158..f3c49e6 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -467,8 +467,10 @@
/*
* This is called to attempt to lock the buffer associated with this
* buf log item. Don't sleep on the buffer lock. If we can't get
- * the lock right away, return 0. If we can get the lock, pull the
- * buffer from the free list, mark it busy, and return 1.
+ * the lock right away, return 0. If we can get the lock, take a
+ * reference to the buffer. If this is a delayed write buffer that
+ * needs AIL help to be written back, invoke the pushbuf routine
+ * rather than the normal success path.
*/
STATIC uint
xfs_buf_item_trylock(
@@ -477,24 +479,18 @@
xfs_buf_t *bp;
bp = bip->bli_buf;
-
- if (XFS_BUF_ISPINNED(bp)) {
+ if (XFS_BUF_ISPINNED(bp))
return XFS_ITEM_PINNED;
- }
-
- if (!XFS_BUF_CPSEMA(bp)) {
+ if (!XFS_BUF_CPSEMA(bp))
return XFS_ITEM_LOCKED;
- }
- /*
- * Remove the buffer from the free list. Only do this
- * if it's on the free list. Private buffers like the
- * superblock buffer are not.
- */
+ /* take a reference to the buffer. */
XFS_BUF_HOLD(bp);
ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
trace_xfs_buf_item_trylock(bip);
+ if (XFS_BUF_ISDELAYWRITE(bp))
+ return XFS_ITEM_PUSHBUF;
return XFS_ITEM_SUCCESS;
}
@@ -626,11 +622,9 @@
}
/*
- * This is called to asynchronously write the buffer associated with this
- * buf log item out to disk. The buffer will already have been locked by
- * a successful call to xfs_buf_item_trylock(). If the buffer still has
- * B_DELWRI set, then get it going out to disk with a call to bawrite().
- * If not, then just release the buffer.
+ * The buffer is locked, but is not a delayed write buffer. This happens
+ * if we race with IO completion and hence we don't want to try to write it
+ * again. Just release the buffer.
*/
STATIC void
xfs_buf_item_push(
@@ -642,17 +636,29 @@
trace_xfs_buf_item_push(bip);
bp = bip->bli_buf;
+ ASSERT(!XFS_BUF_ISDELAYWRITE(bp));
+ xfs_buf_relse(bp);
+}
- if (XFS_BUF_ISDELAYWRITE(bp)) {
- int error;
- error = xfs_bawrite(bip->bli_item.li_mountp, bp);
- if (error)
- xfs_fs_cmn_err(CE_WARN, bip->bli_item.li_mountp,
- "xfs_buf_item_push: pushbuf error %d on bip %p, bp %p",
- error, bip, bp);
- } else {
- xfs_buf_relse(bp);
- }
+/*
+ * The buffer is locked and is a delayed write buffer. Promote the buffer
+ * in the delayed write queue as the caller knows that they must invoke
+ * the xfsbufd to get this buffer written. We have to unlock the buffer
+ * to allow the xfsbufd to write it, too.
+ */
+STATIC void
+xfs_buf_item_pushbuf(
+ xfs_buf_log_item_t *bip)
+{
+ xfs_buf_t *bp;
+
+ ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
+ trace_xfs_buf_item_pushbuf(bip);
+
+ bp = bip->bli_buf;
+ ASSERT(XFS_BUF_ISDELAYWRITE(bp));
+ xfs_buf_delwri_promote(bp);
+ xfs_buf_relse(bp);
}
/* ARGSUSED */
@@ -677,7 +683,7 @@
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
xfs_buf_item_committed,
.iop_push = (void(*)(xfs_log_item_t*))xfs_buf_item_push,
- .iop_pushbuf = NULL,
+ .iop_pushbuf = (void(*)(xfs_log_item_t*))xfs_buf_item_pushbuf,
.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
xfs_buf_item_committing
};