public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: xfs@oss.sgi.com
Subject: [PATCH 06/12] xfs: make AIL tail pushing independent of the grant lock
Date: Mon, 13 Dec 2010 15:44:37 +1100	[thread overview]
Message-ID: <1292215483-18224-7-git-send-email-david@fromorbit.com> (raw)
In-Reply-To: <1292215483-18224-1-git-send-email-david@fromorbit.com>

From: Dave Chinner <dchinner@redhat.com>

The xlog_grant_push_ail() currently takes the grant lock internally to sample
the tail lsn, last sync lsn and the reserve grant head. Most of the callers
already hold the grant lock but have to drop it before calling
xlog_grant_push_ail(). This is a left over from when the AIL tail pushing was
done in line and hence xlog_grant_push_ail had to drop the grant lock. AIL push
is now done in another thread and hence we can safely hold the grant lock over
the entire xlog_grant_push_ail call.

Push the grant lock outside of xlog_grant_push_ail() to simplify the locking
and synchronisation needed for tail pushing.  This will reduce traffic on the
grant lock by itself, but this is only one step in preparing for the complete
removal of the grant lock.

While there, clean up the formatting of xlog_grant_push_ail() to match the
rest of the XFS code.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
---
 fs/xfs/xfs_log.c |  111 ++++++++++++++++++++++++++---------------------------
 1 files changed, 54 insertions(+), 57 deletions(-)

diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index 3656d77..d15bd5b 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -70,7 +70,7 @@ STATIC void xlog_state_want_sync(xlog_t	*log, xlog_in_core_t *iclog);
 /* local functions to manipulate grant head */
 STATIC int  xlog_grant_log_space(xlog_t		*log,
 				 xlog_ticket_t	*xtic);
-STATIC void xlog_grant_push_ail(xfs_mount_t	*mp,
+STATIC void xlog_grant_push_ail(struct log	*log,
 				int		need_bytes);
 STATIC void xlog_regrant_reserve_log_space(xlog_t	 *log,
 					   xlog_ticket_t *ticket);
@@ -318,7 +318,9 @@ xfs_log_reserve(
 
 		trace_xfs_log_reserve(log, internal_ticket);
 
-		xlog_grant_push_ail(mp, internal_ticket->t_unit_res);
+		spin_lock(&log->l_grant_lock);
+		xlog_grant_push_ail(log, internal_ticket->t_unit_res);
+		spin_unlock(&log->l_grant_lock);
 		retval = xlog_regrant_write_log_space(log, internal_ticket);
 	} else {
 		/* may sleep if need to allocate more tickets */
@@ -332,9 +334,11 @@ xfs_log_reserve(
 
 		trace_xfs_log_reserve(log, internal_ticket);
 
-		xlog_grant_push_ail(mp,
+		spin_lock(&log->l_grant_lock);
+		xlog_grant_push_ail(log,
 				    (internal_ticket->t_unit_res *
 				     internal_ticket->t_cnt));
+		spin_unlock(&log->l_grant_lock);
 		retval = xlog_grant_log_space(log, internal_ticket);
 	}
 
@@ -1185,59 +1189,58 @@ xlog_commit_record(
  * water mark.  In this manner, we would be creating a low water mark.
  */
 STATIC void
-xlog_grant_push_ail(xfs_mount_t	*mp,
-		    int		need_bytes)
+xlog_grant_push_ail(
+	struct log	*log,
+	int		need_bytes)
 {
-    xlog_t	*log = mp->m_log;	/* pointer to the log */
-    xfs_lsn_t	tail_lsn;		/* lsn of the log tail */
-    xfs_lsn_t	threshold_lsn = 0;	/* lsn we'd like to be at */
-    int		free_blocks;		/* free blocks left to write to */
-    int		free_bytes;		/* free bytes left to write to */
-    int		threshold_block;	/* block in lsn we'd like to be at */
-    int		threshold_cycle;	/* lsn cycle we'd like to be at */
-    int		free_threshold;
-
-    ASSERT(BTOBB(need_bytes) < log->l_logBBsize);
-
-    spin_lock(&log->l_grant_lock);
-    free_bytes = xlog_space_left(log, &log->l_grant_reserve_head);
-    tail_lsn = log->l_tail_lsn;
-    free_blocks = BTOBBT(free_bytes);
-
-    /*
-     * Set the threshold for the minimum number of free blocks in the
-     * log to the maximum of what the caller needs, one quarter of the
-     * log, and 256 blocks.
-     */
-    free_threshold = BTOBB(need_bytes);
-    free_threshold = MAX(free_threshold, (log->l_logBBsize >> 2));
-    free_threshold = MAX(free_threshold, 256);
-    if (free_blocks < free_threshold) {
+	xfs_lsn_t	threshold_lsn = 0;
+	xfs_lsn_t	tail_lsn;
+	int		free_blocks;
+	int		free_bytes;
+	int		threshold_block;
+	int		threshold_cycle;
+	int		free_threshold;
+
+	ASSERT(BTOBB(need_bytes) < log->l_logBBsize);
+
+	tail_lsn = log->l_tail_lsn;
+	free_bytes = xlog_space_left(log, &log->l_grant_reserve_head);
+	free_blocks = BTOBBT(free_bytes);
+
+	/*
+	 * Set the threshold for the minimum number of free blocks in the
+	 * log to the maximum of what the caller needs, one quarter of the
+	 * log, and 256 blocks.
+	 */
+	free_threshold = BTOBB(need_bytes);
+	free_threshold = MAX(free_threshold, (log->l_logBBsize >> 2));
+	free_threshold = MAX(free_threshold, 256);
+	if (free_blocks >= free_threshold)
+		return;
+
 	threshold_block = BLOCK_LSN(tail_lsn) + free_threshold;
 	threshold_cycle = CYCLE_LSN(tail_lsn);
 	if (threshold_block >= log->l_logBBsize) {
-	    threshold_block -= log->l_logBBsize;
-	    threshold_cycle += 1;
+		threshold_block -= log->l_logBBsize;
+		threshold_cycle += 1;
 	}
-	threshold_lsn = xlog_assign_lsn(threshold_cycle, threshold_block);
-
-	/* Don't pass in an lsn greater than the lsn of the last
+	threshold_lsn = xlog_assign_lsn(threshold_cycle,
+					threshold_block);
+	/*
+	 * Don't pass in an lsn greater than the lsn of the last
 	 * log record known to be on disk.
 	 */
 	if (XFS_LSN_CMP(threshold_lsn, log->l_last_sync_lsn) > 0)
-	    threshold_lsn = log->l_last_sync_lsn;
-    }
-    spin_unlock(&log->l_grant_lock);
-
-    /*
-     * Get the transaction layer to kick the dirty buffers out to
-     * disk asynchronously. No point in trying to do this if
-     * the filesystem is shutting down.
-     */
-    if (threshold_lsn &&
-	!XLOG_FORCED_SHUTDOWN(log))
-	    xfs_trans_ail_push(log->l_ailp, threshold_lsn);
-}	/* xlog_grant_push_ail */
+		threshold_lsn = log->l_last_sync_lsn;
+
+	/*
+	 * Get the transaction layer to kick the dirty buffers out to
+	 * disk asynchronously. No point in trying to do this if
+	 * the filesystem is shutting down.
+	 */
+	if (!XLOG_FORCED_SHUTDOWN(log))
+		xfs_trans_ail_push(log->l_ailp, threshold_lsn);
+}
 
 /*
  * The bdstrat callback function for log bufs. This gives us a central
@@ -2543,9 +2546,7 @@ redo:
 
 		trace_xfs_log_grant_sleep2(log, tic);
 
-		spin_unlock(&log->l_grant_lock);
-		xlog_grant_push_ail(log->l_mp, need_bytes);
-		spin_lock(&log->l_grant_lock);
+		xlog_grant_push_ail(log, need_bytes);
 
 		XFS_STATS_INC(xs_sleep_logspace);
 		xlog_wait(&tic->t_wait, &log->l_grant_lock);
@@ -2641,9 +2642,7 @@ xlog_regrant_write_log_space(xlog_t	   *log,
 
 			trace_xfs_log_regrant_write_sleep1(log, tic);
 
-			spin_unlock(&log->l_grant_lock);
-			xlog_grant_push_ail(log->l_mp, need_bytes);
-			spin_lock(&log->l_grant_lock);
+			xlog_grant_push_ail(log, need_bytes);
 
 			XFS_STATS_INC(xs_sleep_logspace);
 			xlog_wait(&tic->t_wait, &log->l_grant_lock);
@@ -2666,9 +2665,7 @@ redo:
 	if (free_bytes < need_bytes) {
 		if (list_empty(&tic->t_queue))
 			list_add_tail(&tic->t_queue, &log->l_writeq);
-		spin_unlock(&log->l_grant_lock);
-		xlog_grant_push_ail(log->l_mp, need_bytes);
-		spin_lock(&log->l_grant_lock);
+		xlog_grant_push_ail(log, need_bytes);
 
 		XFS_STATS_INC(xs_sleep_logspace);
 		trace_xfs_log_regrant_write_sleep2(log, tic);
-- 
1.7.2.3

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  parent reply	other threads:[~2010-12-13  5:01 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-12-13  4:44 xfs: grant lock scaling and removal V3 Dave Chinner
2010-12-13  4:44 ` [PATCH 01/12] xfs: convert log grant ticket queues to list heads Dave Chinner
2010-12-20 11:34   ` Christoph Hellwig
2010-12-21  0:55     ` Dave Chinner
2010-12-13  4:44 ` [PATCH 02/12] xfs: fact out common grant head/log tail verification code Dave Chinner
2010-12-20 11:51   ` Christoph Hellwig
2010-12-13  4:44 ` [PATCH 03/12] xfs: rework log grant space calculations Dave Chinner
2010-12-20 11:37   ` Christoph Hellwig
2010-12-13  4:44 ` [PATCH 04/12] xfs: combine grant heads into a single 64 bit integer Dave Chinner
2010-12-20 11:40   ` Christoph Hellwig
2010-12-13  4:44 ` [PATCH 05/12] xfs: use wait queues directly for the log wait queues Dave Chinner
2010-12-20 11:40   ` Christoph Hellwig
2010-12-13  4:44 ` Dave Chinner [this message]
2010-12-20 11:41   ` [PATCH 06/12] xfs: make AIL tail pushing independent of the grant lock Christoph Hellwig
2010-12-13  4:44 ` [PATCH 07/12] xfs: convert l_last_sync_lsn to an atomic variable Dave Chinner
2010-12-13  4:44 ` [PATCH 08/12] xfs: convert l_tail_lsn " Dave Chinner
2010-12-20 12:06   ` Christoph Hellwig
2010-12-21  1:27     ` Dave Chinner
2010-12-13  4:44 ` [PATCH 09/12] xfs: convert log grant heads to atomic variables Dave Chinner
2010-12-13  4:44 ` [PATCH 10/12] xfs: introduce new locks for the log grant ticket wait queues Dave Chinner
2010-12-20 11:49   ` Christoph Hellwig
2010-12-13  4:44 ` [PATCH 11/12] xfs: convert grant head manipulations to lockless algorithm Dave Chinner
2010-12-20 11:50   ` Christoph Hellwig
2010-12-13  4:44 ` [PATCH 12/12] xfs: kill useless spinlock_destroy macro Dave Chinner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1292215483-18224-7-git-send-email-david@fromorbit.com \
    --to=david@fromorbit.com \
    --cc=xfs@oss.sgi.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox