xfs: on-stack delayed write buffer lists

Queue delwri buffers on a local on-stack list instead of a per-buftarg one,
and write back the buffers per-process instead of by waking up xfsbufd.

This is now easily doable given that we have very few places left that write
delwri buffers:

 - log recovery:
	Only done at mount time, and already forcing out the buffers
	synchronously using xfs_flush_buftarg

 - quotacheck:
	Same story.

 - dquot reclaim:
	Writes out dirty dquots on the LRU under memory pressure.  We might
	want to look into doing more of this via xfsaild, but it's already
	more optimal than the synchronous inode reclaim that writes each
	buffer synchronously.

 - xfsaild:
	This is the main beneficiary of the change.  By keeping a local list
	of buffers to write we reduce latency of writing out buffers, and
	more importably we can remove all the delwri list promotions which
	were hitting the buffer cache hard under sustained metadata loads.

The implementation is very straight forward - xfs_buf_delwri_queue now gets
a new list_head pointer that it adds the delwri buffers to, and all callers
need to eventually submit the list using xfs_buf_delwi_submit or
xfs_buf_delwi_submit_nowait.  Buffers that already are on a delwri list are
skipped in xfs_buf_delwri_queue, assuming they already are on another delwri
list.  The biggest change to pass down the buffer list was done to the AIL
pushing. Now that we operate on buffers the trylock, push and pushbuf log
item methods are merged into a single push routine, which tries to lock the
item, and if possible add the buffer that needs writeback to the buffer list.
This leads to much simpler code than the previous split but requires the
individual IOP_PUSH instances to unlock and reacquire the AIL around calls
to blocking routines.

Given that xfsailds now also handle writing out buffers, the conditions for
log forcing and the sleep times needed some small changes.  The most
important one is that we consider an AIL busy as long we still have buffers
to push, and the other one is that we do increment the pushed LSN for
buffers that are under flushing at this moment, but still count them towards
the stuck items for restart purposes.  Without this we could hammer on stuck
items without ever forcing the log and not make progress under heavy random
delete workloads on fast flash storage devices.

[ Dave Chinner:
	- rebase on previous patches.
	- improved comments for XBF_DELWRI_Q handling
	- fix XBF_ASYNC handling in queue submission (test 106 failure)
	- rename delwri submit function buffer list parameters for clarity
	- xfs_efd_item_push() should return XFS_ITEM_PINNED ]

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

1 files changed, 26 insertions, 70 deletions

diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 3a0bc38f1859..fb20f384b566 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -418,7 +418,6 @@ xfs_buf_item_unpin(
         if (freed && stale) {
                 ASSERT(bip->bli_flags & XFS_BLI_STALE);
                 ASSERT(xfs_buf_islocked(bp));
-                ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
                 ASSERT(XFS_BUF_ISSTALE(bp));
                 ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
 
@@ -469,34 +468,28 @@ xfs_buf_item_unpin(
         }
 }
 
-/*
- * 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, 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(
-        struct xfs_log_item     *lip)
+xfs_buf_item_push(
+        struct xfs_log_item     *lip,
+        struct list_head        *buffer_list)
 {
         struct xfs_buf_log_item *bip = BUF_ITEM(lip);
         struct xfs_buf          *bp = bip->bli_buf;
+        uint                    rval = XFS_ITEM_SUCCESS;
 
         if (xfs_buf_ispinned(bp))
                 return XFS_ITEM_PINNED;
         if (!xfs_buf_trylock(bp))
                 return XFS_ITEM_LOCKED;
 
-        /* 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;
+
+        trace_xfs_buf_item_push(bip);
+
+        if (!xfs_buf_delwri_queue(bp, buffer_list))
+                rval = XFS_ITEM_FLUSHING;
+        xfs_buf_unlock(bp);
+        return rval;
 }
 
 /*
@@ -609,48 +602,6 @@ xfs_buf_item_committed(
         return lsn;
 }
 
-/*
- * The buffer is locked, but is not a delayed write buffer.
- */
-STATIC void
-xfs_buf_item_push(
-        struct xfs_log_item     *lip)
-{
-        struct xfs_buf_log_item *bip = BUF_ITEM(lip);
-        struct xfs_buf          *bp = bip->bli_buf;
-
-        ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-        ASSERT(!XFS_BUF_ISDELAYWRITE(bp));
-
-        trace_xfs_buf_item_push(bip);
-
-        xfs_buf_delwri_queue(bp);
-        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 bool
-xfs_buf_item_pushbuf(
-        struct xfs_log_item     *lip)
-{
-        struct xfs_buf_log_item *bip = BUF_ITEM(lip);
-        struct xfs_buf          *bp = bip->bli_buf;
-
-        ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-        ASSERT(XFS_BUF_ISDELAYWRITE(bp));
-
-        trace_xfs_buf_item_pushbuf(bip);
-
-        xfs_buf_delwri_promote(bp);
-        xfs_buf_relse(bp);
-        return true;
-}
-
 STATIC void
 xfs_buf_item_committing(
         struct xfs_log_item     *lip,
@@ -666,11 +617,9 @@ static const struct xfs_item_ops xfs_buf_item_ops = {
         .iop_format     = xfs_buf_item_format,
         .iop_pin        = xfs_buf_item_pin,
         .iop_unpin      = xfs_buf_item_unpin,
-        .iop_trylock    = xfs_buf_item_trylock,
         .iop_unlock     = xfs_buf_item_unlock,
         .iop_committed  = xfs_buf_item_committed,
         .iop_push       = xfs_buf_item_push,
-        .iop_pushbuf    = xfs_buf_item_pushbuf,
         .iop_committing = xfs_buf_item_committing
 };
 
@@ -989,20 +938,27 @@ xfs_buf_iodone_callbacks(
          * If the write was asynchronous then no one will be looking for the
          * error.  Clear the error state and write the buffer out again.
          *
-         * During sync or umount we'll write all pending buffers again
-         * synchronous, which will catch these errors if they keep hanging
-         * around.
+         * XXX: This helps against transient write errors, but we need to find
+         * a way to shut the filesystem down if the writes keep failing.
+         *
+         * In practice we'll shut the filesystem down soon as non-transient
+         * erorrs tend to affect the whole device and a failing log write
+         * will make us give up.  But we really ought to do better here.
          */
         if (XFS_BUF_ISASYNC(bp)) {
+                ASSERT(bp->b_iodone != NULL);
+
+                trace_xfs_buf_item_iodone_async(bp, _RET_IP_);
+
                 xfs_buf_ioerror(bp, 0); /* errno of 0 unsets the flag */
 
                 if (!XFS_BUF_ISSTALE(bp)) {
-                        xfs_buf_delwri_queue(bp);
-                        XFS_BUF_DONE(bp);
+                        bp->b_flags |= XBF_WRITE | XBF_ASYNC | XBF_DONE;
+                        xfs_bdstrat_cb(bp);
+                } else {
+                        xfs_buf_relse(bp);
                 }
-                ASSERT(bp->b_iodone != NULL);
-                trace_xfs_buf_item_iodone_async(bp, _RET_IP_);
-                xfs_buf_relse(bp);
+
                 return;
         }