[XFS] Delay I/O completion for unwritten extents after conversion
SGI-PV: 936584
SGI-Modid: xfs-linux:xfs-kern:196886a
Signed-off-by: Christoph Hellwig <hch@sgi.com>
Signed-off-by: Nathan Scott <nathans@sgi.com>
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index 2add9a8..ea615e2 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -136,10 +136,21 @@
vnode_t *vp = ioend->io_vnode;
xfs_off_t offset = ioend->io_offset;
size_t size = ioend->io_size;
+ struct buffer_head *bh, *next;
int error;
if (ioend->io_uptodate)
VOP_BMAP(vp, offset, size, BMAPI_UNWRITTEN, NULL, NULL, error);
+
+ /* ioend->io_buffer_head is only non-NULL for buffered I/O */
+ for (bh = ioend->io_buffer_head; bh; bh = next) {
+ next = bh->b_private;
+
+ bh->b_end_io = NULL;
+ clear_buffer_unwritten(bh);
+ end_buffer_async_write(bh, ioend->io_uptodate);
+ }
+
xfs_destroy_ioend(ioend);
}
@@ -165,6 +176,7 @@
atomic_set(&ioend->io_remaining, 1);
ioend->io_uptodate = 1; /* cleared if any I/O fails */
ioend->io_vnode = LINVFS_GET_VP(inode);
+ ioend->io_buffer_head = NULL;
atomic_inc(&ioend->io_vnode->v_iocount);
ioend->io_offset = 0;
ioend->io_size = 0;
@@ -180,15 +192,26 @@
int uptodate)
{
xfs_ioend_t *ioend = bh->b_private;
+ static spinlock_t unwritten_done_lock = SPIN_LOCK_UNLOCKED;
+ unsigned long flags;
ASSERT(buffer_unwritten(bh));
bh->b_end_io = NULL;
- clear_buffer_unwritten(bh);
+
if (!uptodate)
ioend->io_uptodate = 0;
+ /*
+ * Deep magic here. We reuse b_private in the buffer_heads to build
+ * a chain for completing the I/O from user context after we've issued
+ * a transaction to convert the unwritten extent.
+ */
+ spin_lock_irqsave(&unwritten_done_lock, flags);
+ bh->b_private = ioend->io_buffer_head;
+ ioend->io_buffer_head = bh;
+ spin_unlock_irqrestore(&unwritten_done_lock, flags);
+
xfs_finish_ioend(ioend);
- end_buffer_async_write(bh, uptodate);
}
STATIC int