Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible

[PATCH 0/9] Reduce compaction scanning and lock contention · Mel Gorman <mgorman@suse.de> · 2012-09-21
[PATCH 1/9] Revert "mm: compaction: check lock contention first before taking lock" · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 1/9] Revert "mm: compaction: check lock contention first before taking lock" · Rafael Aquini <hidden> · 2012-09-21
[PATCH 3/9] Revert "mm: compaction: abort compaction loop if lock is contended or run too long" · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 3/9] Revert "mm: compaction: abort compaction loop if lock is contended or run too long" · Rafael Aquini <hidden> · 2012-09-21
[PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Rafael Aquini <hidden> · 2012-09-21
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Minchan Kim <minchan@kernel.org> · 2012-09-25
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Mel Gorman <mgorman@suse.de> · 2012-09-25
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Minchan Kim <minchan@kernel.org> · 2012-09-25
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Andrew Morton <akpm@linux-foundation.org> · 2012-09-25
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Minchan Kim <minchan@kernel.org> · 2012-09-26
Re: [PATCH 5/9] mm: compaction: Acquire the zone->lru_lock as late as possible · Mel Gorman <mgorman@suse.de> · 2012-09-26
[PATCH 4/9] mm: compaction: Abort compaction loop if lock is contended or run too long · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 4/9] mm: compaction: Abort compaction loop if lock is contended or run too long · Rafael Aquini <hidden> · 2012-09-21
Re: [PATCH 4/9] mm: compaction: Abort compaction loop if lock is contended or run too long · Andrew Morton <akpm@linux-foundation.org> · 2012-09-21
Re: [PATCH 4/9] mm: compaction: Abort compaction loop if lock is contended or run too long · Minchan Kim <minchan@kernel.org> · 2012-09-25
[PATCH 7/9] Revert "mm: have order > 0 compaction start off where it left" · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 7/9] Revert "mm: have order > 0 compaction start off where it left" · Rafael Aquini <hidden> · 2012-09-21
Re: [PATCH 7/9] Revert "mm: have order > 0 compaction start off where it left" · Minchan Kim <minchan@kernel.org> · 2012-09-25
[PATCH 6/9] mm: compaction: Acquire the zone->lock as late as possible · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 6/9] mm: compaction: Acquire the zone->lock as late as possible · Rafael Aquini <hidden> · 2012-09-21
Re: [PATCH 6/9] mm: compaction: Acquire the zone->lock as late as possible · Andrew Morton <akpm@linux-foundation.org> · 2012-09-21
Re: [PATCH 6/9] mm: compaction: Acquire the zone->lock as late as possible · Mel Gorman <mgorman@suse.de> · 2012-09-24
Re: [PATCH 6/9] mm: compaction: Acquire the zone->lock as late as possible · Minchan Kim <minchan@kernel.org> · 2012-09-25
Re: [PATCH 6/9] mm: compaction: Acquire the zone->lock as late as possible · Minchan Kim <minchan@kernel.org> · 2012-09-25
[PATCH 9/9] mm: compaction: Restart compaction from near where it left off · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 9/9] mm: compaction: Restart compaction from near where it left off · Rafael Aquini <hidden> · 2012-09-21
[PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Rafael Aquini <hidden> · 2012-09-21
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Andrew Morton <akpm@linux-foundation.org> · 2012-09-21
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Mel Gorman <mgorman@suse.de> · 2012-09-24
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Andrew Morton <akpm@linux-foundation.org> · 2012-09-24
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Mel Gorman <mgorman@suse.de> · 2012-09-25
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Andrew Morton <akpm@linux-foundation.org> · 2012-09-25
[PATCH] mm: compaction: cache if a pageblock was scanned and no pages were isolated -fix2 · Mel Gorman <mgorman@suse.de> · 2012-09-27
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Mel Gorman <mgorman@suse.de> · 2012-09-27
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Minchan Kim <minchan@kernel.org> · 2012-09-26
Re: [PATCH 8/9] mm: compaction: Cache if a pageblock was scanned and no pages were isolated · Mel Gorman <mgorman@suse.de> · 2012-09-27
[PATCH 2/9] Revert "mm-compaction-abort-compaction-loop-if-lock-is-contended-or-run-too-long-fix" · Mel Gorman <mgorman@suse.de> · 2012-09-21
Re: [PATCH 2/9] Revert "mm-compaction-abort-compaction-loop-if-lock-is-contended-or-run-too-long-fix" · Rafael Aquini <hidden> · 2012-09-21
Re: [PATCH 0/9] Reduce compaction scanning and lock contention · Rik van Riel <hidden> · 2012-09-21

From: Mel Gorman <mgorman@suse.de>
Date: 2012-09-26 10:18:19
Also in: kvm, lkml, qemu-devel

On Tue, Sep 25, 2012 at 02:39:31PM -0700, Andrew Morton wrote:

quoted hunk ↗ jump to hunk

On Tue, 25 Sep 2012 17:13:27 +0900
Minchan Kim [off-list ref] wrote:

quoted

I see. To me, your saying is better than current comment.
I hope comment could be more explicit.

diff --git a/mm/compaction.c b/mm/compaction.c
index df01b4e..f1d2cc7 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c

@@ -542,8 +542,9 @@ isolate_migratepages_range(struct zone *zone, struct compact_control *cc,
                 * splitting and collapsing (collapsing has already happened
                 * if PageLRU is set) but the lock is not necessarily taken
                 * here and it is wasteful to take it just to check transhuge.
-                * Check transhuge without lock and skip if it's either a
-                * transhuge or hugetlbfs page.
+                * Check transhuge without lock and *skip* if it's either a
+                * transhuge or hugetlbfs page because it's not safe to call
+                * compound_order.
                 */
                if (PageTransHuge(page)) {
                        if (!locked)

Going a bit further:

--- a/mm/compaction.c~mm-compaction-acquire-the-zone-lru_lock-as-late-as-possible-fix
+++ a/mm/compaction.c

@@ -415,7 +415,8 @@ isolate_migratepages_range(struct zone *
 		 * if PageLRU is set) but the lock is not necessarily taken
 		 * here and it is wasteful to take it just to check transhuge.
 		 * Check transhuge without lock and skip if it's either a
-		 * transhuge or hugetlbfs page.
+		 * transhuge or hugetlbfs page because calling compound_order()
+		 * requires lru_lock to exclude isolation and splitting.
 		 */
 		if (PageTransHuge(page)) {
 			if (!locked)

_

but...  the requirement to hold lru_lock for compound_order() is news
to me.  It doesn't seem to be written down or explained anywhere, and
one wonders why the cheerily undocumented compound_lock() doesn't have
this effect.  What's going on here??

The lru_lock is not *required* for compound_order(). Normally, users of
compound_order() know that the page is not going to collapse underneath
them. The slub allocator is not going to have a compound page it controls
disappear unexpectedly and does not need additional locking for example.

In the case where we are potentially dealing with a THP page, we have to
take into account if it can collapse underneath us. In this case, there is a
race between when when PageTransHuge is checked and compound_order is called.
The race is probably harmless but it was easy to take into account in
this case.

I think the comment saying that lru_lock is required is misleading. How
about this?

                 * Check TransHuge without lock and skip the whole
                 * pageblock if it's either a transhuge or hugetlbfs page
		 * as calling compound_order without preventing THP
		 * splitting the page underneath us may return
		 * surprising results.

-- 
Mel Gorman
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help