xfs: revert AIL TASK_KILLABLE threshold

In commit 9adf40249e6c, we changed the behavior of the AIL thread to
set its own task state to KILLABLE whenever the timeout value is
nonzero.  Unfortunately, this missed the fact that xfsaild_push will
return 50ms (aka a longish sleep) when we reach the push target or the
AIL becomes empty, so xfsaild goes to sleep for a long period of time in
uninterruptible D state.

This results in artificially high load averages because KILLABLE
processes are UNINTERRUPTIBLE, which contributes to load average even
though the AIL is asleep waiting for someone to interrupt it.  It's not
blocked on IOs or anything, but people scrap ps for processes that look
like they're stuck in D state, so restore the previous threshold.

Fixes: 9adf40249e6c ("xfs: AIL doesn't need manual pushing")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>

diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index 0fafcc9f3dbe4447525bb0fc964d27089a791331..8ede9d099d1fea6b6399b5a0617246baee057c07 100644 (file)
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -644,7 +644,12 @@ xfsaild(
        set_freezable();
 
        while (1) {
-               if (tout)
+               /*
+                * Long waits of 50ms or more occur when we've run out of items
+                * to push, so we only want uninterruptible state if we're
+                * actually blocked on something.
+                */
+               if (tout && tout <= 20)
                        set_current_state(TASK_KILLABLE|TASK_FREEZABLE);
                else
                        set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);