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, 14 insertions, 41 deletions

diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c
index 35c2aff38b2..9549ef179e0 100644
--- a/fs/xfs/xfs_extfree_item.c
+++ b/fs/xfs/xfs_extfree_item.c
@@ -147,22 +147,20 @@ xfs_efi_item_unpin(
 }
 
 /*
- * Efi items have no locking or pushing.  However, since EFIs are
- * pulled from the AIL when their corresponding EFDs are committed
- * to disk, their situation is very similar to being pinned.  Return
- * XFS_ITEM_PINNED so that the caller will eventually flush the log.
- * This should help in getting the EFI out of the AIL.
+ * Efi items have no locking or pushing.  However, since EFIs are pulled from
+ * the AIL when their corresponding EFDs are committed to disk, their situation
+ * is very similar to being pinned.  Return XFS_ITEM_PINNED so that the caller
+ * will eventually flush the log.  This should help in getting the EFI out of
+ * the AIL.
  */
 STATIC uint
-xfs_efi_item_trylock(
-        struct xfs_log_item     *lip)
+xfs_efi_item_push(
+        struct xfs_log_item     *lip,
+        struct list_head        *buffer_list)
 {
         return XFS_ITEM_PINNED;
 }
 
-/*
- * Efi items have no locking, so just return.
- */
 STATIC void
 xfs_efi_item_unlock(
         struct xfs_log_item     *lip)
@@ -190,17 +188,6 @@ xfs_efi_item_committed(
 }
 
 /*
- * There isn't much you can do to push on an efi item.  It is simply
- * stuck waiting for all of its corresponding efd items to be
- * committed to disk.
- */
-STATIC void
-xfs_efi_item_push(
-        struct xfs_log_item     *lip)
-{
-}
-
-/*
  * The EFI dependency tracking op doesn't do squat.  It can't because
  * it doesn't know where the free extent is coming from.  The dependency
  * tracking has to be handled by the "enclosing" metadata object.  For
@@ -222,7 +209,6 @@ static const struct xfs_item_ops xfs_efi_item_ops = {
         .iop_format     = xfs_efi_item_format,
         .iop_pin        = xfs_efi_item_pin,
         .iop_unpin      = xfs_efi_item_unpin,
-        .iop_trylock    = xfs_efi_item_trylock,
         .iop_unlock     = xfs_efi_item_unlock,
         .iop_committed  = xfs_efi_item_committed,
         .iop_push       = xfs_efi_item_push,
@@ -404,19 +390,17 @@ xfs_efd_item_unpin(
 }
 
 /*
- * Efd items have no locking, so just return success.
+ * There isn't much you can do to push on an efd item.  It is simply stuck
+ * waiting for the log to be flushed to disk.
  */
 STATIC uint
-xfs_efd_item_trylock(
-        struct xfs_log_item     *lip)
+xfs_efd_item_push(
+        struct xfs_log_item     *lip,
+        struct list_head        *buffer_list)
 {
-        return XFS_ITEM_LOCKED;
+        return XFS_ITEM_PINNED;
 }
 
-/*
- * Efd items have no locking or pushing, so return failure
- * so that the caller doesn't bother with us.
- */
 STATIC void
 xfs_efd_item_unlock(
         struct xfs_log_item     *lip)
@@ -451,16 +435,6 @@ xfs_efd_item_committed(
 }
 
 /*
- * There isn't much you can do to push on an efd item.  It is simply
- * stuck waiting for the log to be flushed to disk.
- */
-STATIC void
-xfs_efd_item_push(
-        struct xfs_log_item     *lip)
-{
-}
-
-/*
  * The EFD dependency tracking op doesn't do squat.  It can't because
  * it doesn't know where the free extent is coming from.  The dependency
  * tracking has to be handled by the "enclosing" metadata object.  For
@@ -482,7 +456,6 @@ static const struct xfs_item_ops xfs_efd_item_ops = {
         .iop_format     = xfs_efd_item_format,
         .iop_pin        = xfs_efd_item_pin,
         .iop_unpin      = xfs_efd_item_unpin,
-        .iop_trylock    = xfs_efd_item_trylock,
         .iop_unlock     = xfs_efd_item_unlock,
         .iop_committed  = xfs_efd_item_committed,
         .iop_push       = xfs_efd_item_push,