提交 b589334c 编写于 作者: D David Chinner 提交者: Lachlan McIlroy

[XFS] Prevent AIL lock contention during transaction completion

When hundreds of processors attempt to commit transactions at the same
time, they can contend on the AIL lock when updating the tail LSN held in
the in-core log structure.

At the moment, the tail LSN is only needed when actually writing out an
iclog, so it really does not need to be updated on every single
transaction completion - only those that result in switching iclogs and
flushing them to disk.

The result is that we reduce the number of times we need to grab the AIL
lock and the log grant lock by up to two orders of magnitude on large
processor count machines. The problem has previously been hidden by AIL
lock contention walking the AIL list which was recently solved and
uncovered this issue.

SGI-PV: 975671
SGI-Modid: xfs-linux-melb:xfs-kern:30504a
Signed-off-by: NDavid Chinner <dgc@sgi.com>
Signed-off-by: NTim Shimmin <tes@sgi.com>
Signed-off-by: NLachlan McIlroy <lachlan@sgi.com>
上级 33540408
...@@ -2813,15 +2813,13 @@ xlog_state_put_ticket(xlog_t *log, ...@@ -2813,15 +2813,13 @@ xlog_state_put_ticket(xlog_t *log,
* *
*/ */
STATIC int STATIC int
xlog_state_release_iclog(xlog_t *log, xlog_state_release_iclog(
xlog_in_core_t *iclog) xlog_t *log,
xlog_in_core_t *iclog)
{ {
int sync = 0; /* do we sync? */ int sync = 0; /* do we sync? */
xlog_assign_tail_lsn(log->l_mp);
spin_lock(&log->l_icloglock); spin_lock(&log->l_icloglock);
if (iclog->ic_state & XLOG_STATE_IOERROR) { if (iclog->ic_state & XLOG_STATE_IOERROR) {
spin_unlock(&log->l_icloglock); spin_unlock(&log->l_icloglock);
return XFS_ERROR(EIO); return XFS_ERROR(EIO);
...@@ -2833,13 +2831,14 @@ xlog_state_release_iclog(xlog_t *log, ...@@ -2833,13 +2831,14 @@ xlog_state_release_iclog(xlog_t *log,
if (--iclog->ic_refcnt == 0 && if (--iclog->ic_refcnt == 0 &&
iclog->ic_state == XLOG_STATE_WANT_SYNC) { iclog->ic_state == XLOG_STATE_WANT_SYNC) {
/* update tail before writing to iclog */
xlog_assign_tail_lsn(log->l_mp);
sync++; sync++;
iclog->ic_state = XLOG_STATE_SYNCING; iclog->ic_state = XLOG_STATE_SYNCING;
iclog->ic_header.h_tail_lsn = cpu_to_be64(log->l_tail_lsn); iclog->ic_header.h_tail_lsn = cpu_to_be64(log->l_tail_lsn);
xlog_verify_tail_lsn(log, iclog, log->l_tail_lsn); xlog_verify_tail_lsn(log, iclog, log->l_tail_lsn);
/* cycle incremented when incrementing curr_block */ /* cycle incremented when incrementing curr_block */
} }
spin_unlock(&log->l_icloglock); spin_unlock(&log->l_icloglock);
/* /*
...@@ -2849,11 +2848,9 @@ xlog_state_release_iclog(xlog_t *log, ...@@ -2849,11 +2848,9 @@ xlog_state_release_iclog(xlog_t *log,
* this iclog has consistent data, so we ignore IOERROR * this iclog has consistent data, so we ignore IOERROR
* flags after this point. * flags after this point.
*/ */
if (sync) { if (sync)
return xlog_sync(log, iclog); return xlog_sync(log, iclog);
}
return 0; return 0;
} /* xlog_state_release_iclog */ } /* xlog_state_release_iclog */
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册