summaryrefslogtreecommitdiffstats
path: root/fs/xfs/xfs_trans_ail.c
diff options
context:
space:
mode:
authorChristoph Hellwig <hch@infradead.org>2012-04-23 15:58:39 +1000
committerBen Myers <bpm@sgi.com>2012-05-14 16:20:31 -0500
commit43ff2122e6492bcc88b065c433453dce88223b30 (patch)
tree0f762cfb753edd73402b8830e0927d9efba30c61 /fs/xfs/xfs_trans_ail.c
parent960c60af8b9481595e68875e79b2602e73169c29 (diff)
downloadlinux-43ff2122e6492bcc88b065c433453dce88223b30.tar.bz2
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>
Diffstat (limited to 'fs/xfs/xfs_trans_ail.c')
-rw-r--r--fs/xfs/xfs_trans_ail.c129
1 files changed, 57 insertions, 72 deletions
diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index 0425ca16738b..49d9cde33bb3 100644
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -364,29 +364,31 @@ xfsaild_push(
xfs_log_item_t *lip;
xfs_lsn_t lsn;
xfs_lsn_t target;
- long tout = 10;
+ long tout;
int stuck = 0;
+ int flushing = 0;
int count = 0;
- int push_xfsbufd = 0;
/*
- * If last time we ran we encountered pinned items, force the log first
- * and wait for it before pushing again.
+ * If we encountered pinned items or did not finish writing out all
+ * buffers the last time we ran, force the log first and wait for it
+ * before pushing again.
*/
- spin_lock(&ailp->xa_lock);
- if (ailp->xa_last_pushed_lsn == 0 && ailp->xa_log_flush &&
- !list_empty(&ailp->xa_ail)) {
+ if (ailp->xa_log_flush && ailp->xa_last_pushed_lsn == 0 &&
+ (!list_empty_careful(&ailp->xa_buf_list) ||
+ xfs_ail_min_lsn(ailp))) {
ailp->xa_log_flush = 0;
- spin_unlock(&ailp->xa_lock);
+
XFS_STATS_INC(xs_push_ail_flush);
xfs_log_force(mp, XFS_LOG_SYNC);
- spin_lock(&ailp->xa_lock);
}
+ spin_lock(&ailp->xa_lock);
lip = xfs_trans_ail_cursor_first(ailp, &cur, ailp->xa_last_pushed_lsn);
if (!lip) {
/*
- * AIL is empty or our push has reached the end.
+ * If the AIL is empty or our push has reached the end we are
+ * done now.
*/
xfs_trans_ail_cursor_done(ailp, &cur);
spin_unlock(&ailp->xa_lock);
@@ -395,55 +397,42 @@ xfsaild_push(
XFS_STATS_INC(xs_push_ail);
- /*
- * While the item we are looking at is below the given threshold
- * try to flush it out. We'd like not to stop until we've at least
- * tried to push on everything in the AIL with an LSN less than
- * the given threshold.
- *
- * However, we will stop after a certain number of pushes and wait
- * for a reduced timeout to fire before pushing further. This
- * prevents use from spinning when we can't do anything or there is
- * lots of contention on the AIL lists.
- */
lsn = lip->li_lsn;
target = ailp->xa_target;
while ((XFS_LSN_CMP(lip->li_lsn, target) <= 0)) {
int lock_result;
+
/*
- * If we can lock the item without sleeping, unlock the AIL
- * lock and flush the item. Then re-grab the AIL lock so we
- * can look for the next item on the AIL. List changes are
- * handled by the AIL lookup functions internally
- *
- * If we can't lock the item, either its holder will flush it
- * or it is already being flushed or it is being relogged. In
- * any of these case it is being taken care of and we can just
- * skip to the next item in the list.
+ * Note that IOP_PUSH may unlock and reacquire the AIL lock. We
+ * rely on the AIL cursor implementation to be able to deal with
+ * the dropped lock.
*/
- lock_result = IOP_TRYLOCK(lip);
- spin_unlock(&ailp->xa_lock);
+ lock_result = IOP_PUSH(lip, &ailp->xa_buf_list);
switch (lock_result) {
case XFS_ITEM_SUCCESS:
XFS_STATS_INC(xs_push_ail_success);
trace_xfs_ail_push(lip);
- IOP_PUSH(lip);
ailp->xa_last_pushed_lsn = lsn;
break;
- case XFS_ITEM_PUSHBUF:
- XFS_STATS_INC(xs_push_ail_pushbuf);
- trace_xfs_ail_pushbuf(lip);
-
- if (!IOP_PUSHBUF(lip)) {
- trace_xfs_ail_pushbuf_pinned(lip);
- stuck++;
- ailp->xa_log_flush++;
- } else {
- ailp->xa_last_pushed_lsn = lsn;
- }
- push_xfsbufd = 1;
+ case XFS_ITEM_FLUSHING:
+ /*
+ * The item or its backing buffer is already beeing
+ * flushed. The typical reason for that is that an
+ * inode buffer is locked because we already pushed the
+ * updates to it as part of inode clustering.
+ *
+ * We do not want to to stop flushing just because lots
+ * of items are already beeing flushed, but we need to
+ * re-try the flushing relatively soon if most of the
+ * AIL is beeing flushed.
+ */
+ XFS_STATS_INC(xs_push_ail_flushing);
+ trace_xfs_ail_flushing(lip);
+
+ flushing++;
+ ailp->xa_last_pushed_lsn = lsn;
break;
case XFS_ITEM_PINNED:
@@ -453,23 +442,22 @@ xfsaild_push(
stuck++;
ailp->xa_log_flush++;
break;
-
case XFS_ITEM_LOCKED:
XFS_STATS_INC(xs_push_ail_locked);
trace_xfs_ail_locked(lip);
+
stuck++;
break;
-
default:
ASSERT(0);
break;
}
- spin_lock(&ailp->xa_lock);
count++;
/*
* Are there too many items we can't do anything with?
+ *
* If we we are skipping too many items because we can't flush
* them or they are already being flushed, we back off and
* given them time to complete whatever operation is being
@@ -491,42 +479,36 @@ xfsaild_push(
xfs_trans_ail_cursor_done(ailp, &cur);
spin_unlock(&ailp->xa_lock);
- if (push_xfsbufd) {
- /* we've got delayed write buffers to flush */
- wake_up_process(mp->m_ddev_targp->bt_task);
- }
+ if (xfs_buf_delwri_submit_nowait(&ailp->xa_buf_list))
+ ailp->xa_log_flush++;
- /* assume we have more work to do in a short while */
+ if (!count || XFS_LSN_CMP(lsn, target) >= 0) {
out_done:
- if (!count) {
- /* We're past our target or empty, so idle */
- ailp->xa_last_pushed_lsn = 0;
- ailp->xa_log_flush = 0;
-
- tout = 50;
- } else if (XFS_LSN_CMP(lsn, target) >= 0) {
/*
- * We reached the target so wait a bit longer for I/O to
- * complete and remove pushed items from the AIL before we
- * start the next scan from the start of the AIL.
+ * We reached the target or the AIL is empty, so wait a bit
+ * longer for I/O to complete and remove pushed items from the
+ * AIL before we start the next scan from the start of the AIL.
*/
tout = 50;
ailp->xa_last_pushed_lsn = 0;
- } else if ((stuck * 100) / count > 90) {
+ } else if (((stuck + flushing) * 100) / count > 90) {
/*
- * Either there is a lot of contention on the AIL or we
- * are stuck due to operations in progress. "Stuck" in this
- * case is defined as >90% of the items we tried to push
- * were stuck.
+ * Either there is a lot of contention on the AIL or we are
+ * stuck due to operations in progress. "Stuck" in this case
+ * is defined as >90% of the items we tried to push were stuck.
*
* Backoff a bit more to allow some I/O to complete before
- * restarting from the start of the AIL. This prevents us
- * from spinning on the same items, and if they are pinned will
- * all the restart to issue a log force to unpin the stuck
- * items.
+ * restarting from the start of the AIL. This prevents us from
+ * spinning on the same items, and if they are pinned will all
+ * the restart to issue a log force to unpin the stuck items.
*/
tout = 20;
ailp->xa_last_pushed_lsn = 0;
+ } else {
+ /*
+ * Assume we have more work to do in a short while.
+ */
+ tout = 10;
}
return tout;
@@ -539,6 +521,8 @@ xfsaild(
struct xfs_ail *ailp = data;
long tout = 0; /* milliseconds */
+ current->flags |= PF_MEMALLOC;
+
while (!kthread_should_stop()) {
if (tout && tout <= 20)
__set_current_state(TASK_KILLABLE);
@@ -794,6 +778,7 @@ xfs_trans_ail_init(
INIT_LIST_HEAD(&ailp->xa_ail);
INIT_LIST_HEAD(&ailp->xa_cursors);
spin_lock_init(&ailp->xa_lock);
+ INIT_LIST_HEAD(&ailp->xa_buf_list);
init_waitqueue_head(&ailp->xa_empty);
ailp->xa_task = kthread_run(xfsaild, ailp, "xfsaild/%s",