From 7a6144076166944655d86f1449be8566d1a3c71a Mon Sep 17 00:00:00 2001
From: George Wilson <george.wilson@delphix.com>
Date: Thu, 21 Feb 2013 13:58:29 -0800
Subject: [PATCH] Illumos #3552

3552 condensing one space map burns 3 seconds of CPU in spa_sync()
     thread (fix race condition)

References:
  https://www.illumos.org/issues/3552
  illumos/illumos-gate@03f8c366886542ed249a15d755ae78ea4e775d9d

Ported-by: Richard Yao <ryao@gentoo.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

Porting notes:

This fixes an upstream regression that was introduced in commit
zfsonlinux/zfs@e51be06697762215dc3b679f8668987034a5a048, which
ported the Illumos 3552 changes. This fix was added to upstream
rather quickly, but at the time of the port, no one spotted it and
the race was rare enough that it passed our regression tests. I
discovered this when comparing our metaslab.c to the illumos
metaslab.c.

Without this change it is possible for metaslab_group_alloc() to
consume a large amount of cpu time.  Since this occurs under a
mutex in a rcu critical section the kernel will log this to the
console as a self-detected cpu stall as follows:

  INFO: rcu_sched self-detected stall on CPU { 0}
  (t=60000 jiffies g=11431890 c=11431889 q=18271)

Closes #1687
Closes #1720
Closes #1731
Closes #1747
---
 module/zfs/metaslab.c | 27 +++++++++++++++++----------
 1 file changed, 17 insertions(+), 10 deletions(-)

diff --git a/module/zfs/metaslab.c b/module/zfs/metaslab.c
index f9cb8cead..7837822e4 100644
--- a/module/zfs/metaslab.c
+++ b/module/zfs/metaslab.c
@@ -1388,6 +1388,13 @@ metaslab_group_alloc(metaslab_group_t *mg, uint64_t psize, uint64_t asize,
 				mutex_exit(&mg->mg_lock);
 				return (-1ULL);
 			}
+
+			/*
+			 * If the selected metaslab is condensing, skip it.
+			 */
+			if (msp->ms_map->sm_condensing)
+				continue;
+
 			was_active = msp->ms_weight & METASLAB_ACTIVE_MASK;
 			if (activation_weight == METASLAB_WEIGHT_PRIMARY)
 				break;
@@ -1427,16 +1434,6 @@ metaslab_group_alloc(metaslab_group_t *mg, uint64_t psize, uint64_t asize,
 
 		mutex_enter(&msp->ms_lock);
 
-		/*
-		 * If this metaslab is currently condensing then pick again as
-		 * we can't manipulate this metaslab until it's committed
-		 * to disk.
-		 */
-		if (msp->ms_map->sm_condensing) {
-			mutex_exit(&msp->ms_lock);
-			continue;
-		}
-
 		/*
 		 * Ensure that the metaslab we have selected is still
 		 * capable of handling our request. It's possible that
@@ -1463,6 +1460,16 @@ metaslab_group_alloc(metaslab_group_t *mg, uint64_t psize, uint64_t asize,
 			continue;
 		}
 
+		/*
+		 * If this metaslab is currently condensing then pick again as
+		 * we can't manipulate this metaslab until it's committed
+		 * to disk.
+		 */
+		if (msp->ms_map->sm_condensing) {
+			mutex_exit(&msp->ms_lock);
+			continue;
+		}
+
 		if ((offset = space_map_alloc(msp->ms_map, asize)) != -1ULL)
 			break;