From: KOSAKI Motohiro <kosaki.motoh...@jp.fujitsu.com>
During aio stress test, we observed the following lockdep warning.
This mean AIO+numa_balancing is currently deadlockable.
The problem is, aio_migratepage disable interrupt, but
__set_page_dirty_nobuffers
unintentionally enable it again.
Generally, all helper function should use spin_lock_irqsave()
instead of spin_lock_irq() because they don't know caller at all.
[ 599.843948] other info that might help us debug this:
[ 599.873748] Possible unsafe locking scenario:
[ 599.873748]
[ 599.900902] CPU0
[ 599.912701] ----
[ 599.924929] lock(&(&ctx->completion_lock)->rlock);
[ 599.950299] <Interrupt>
[ 599.962576] lock(&(&ctx->completion_lock)->rlock);
[ 599.985771]
[ 599.985771] *** DEADLOCK ***
[ 600.375623] [<ffffffff81678d3c>] dump_stack+0x19/0x1b
[ 600.398769] [<ffffffff816731aa>] print_usage_bug+0x1f7/0x208
[ 600.425092] [<ffffffff810df370>] ?
print_shortest_lock_dependencies+0x1d0/0x1d0
[ 600.458981] [<ffffffff810e08dd>] mark_lock+0x21d/0x2a0
[ 600.482910] [<ffffffff810e0a19>] mark_held_locks+0xb9/0x140
[ 600.508956] [<ffffffff8168201c>] ? _raw_spin_unlock_irq+0x2c/0x50
[ 600.536825] [<ffffffff810e0ba5>] trace_hardirqs_on_caller+0x105/0x1d0
[ 600.566861] [<ffffffff810e0c7d>] trace_hardirqs_on+0xd/0x10
[ 600.593210] [<ffffffff8168201c>] _raw_spin_unlock_irq+0x2c/0x50
[ 600.620599] [<ffffffff8117f72c>] __set_page_dirty_nobuffers+0x8c/0xf0
[ 600.649992] [<ffffffff811d1094>] migrate_page_copy+0x434/0x540
[ 600.676635] [<ffffffff8123f5b1>] aio_migratepage+0xb1/0x140
[ 600.703126] [<ffffffff811d126d>] move_to_new_page+0x7d/0x230
[ 600.729022] [<ffffffff811d1b45>] migrate_pages+0x5e5/0x700
[ 600.754705] [<ffffffff811d0070>] ? buffer_migrate_lock_buffers+0xb0/0xb0
[ 600.785784] [<ffffffff811d29cc>] migrate_misplaced_page+0xbc/0xf0
[ 600.814029] [<ffffffff8119eb62>] do_numa_page+0x102/0x190
[ 600.839182] [<ffffffff8119ee31>] handle_pte_fault+0x241/0x970
[ 600.865875] [<ffffffff811a0345>] handle_mm_fault+0x265/0x370
[ 600.892071] [<ffffffff81686d82>] __do_page_fault+0x172/0x5a0
[ 600.918065] [<ffffffff81682cd8>] ? retint_swapgs+0x13/0x1b
[ 600.943493] [<ffffffff816871ca>] do_page_fault+0x1a/0x70
[ 600.968081] [<ffffffff81682ff8>] page_fault+0x28/0x30
Signed-off-by: KOSAKI Motohiro <kosaki.motoh...@jp.fujitsu.com>
Cc: Larry Woodman <lwood...@redhat.com>
Cc: Rik van Riel <r...@redhat.com>
Cc: Johannes Weiner <jwei...@redhat.com>
Cc: sta...@vger.kernel.org
---
mm/page-writeback.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 2d30e2c..7106cb1 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -2173,11 +2173,12 @@ int __set_page_dirty_nobuffers(struct page *page)
if (!TestSetPageDirty(page)) {
struct address_space *mapping = page_mapping(page);
struct address_space *mapping2;
+ unsigned long flags;
if (!mapping)
return 1;
- spin_lock_irq(&mapping->tree_lock);
+ spin_lock_irqsave(&mapping->tree_lock, flags);
mapping2 = page_mapping(page);
if (mapping2) { /* Race with truncate? */
BUG_ON(mapping2 != mapping);
@@ -2186,7 +2187,7 @@ int __set_page_dirty_nobuffers(struct page *page)
radix_tree_tag_set(&mapping->page_tree,
page_index(page), PAGECACHE_TAG_DIRTY);
}
- spin_unlock_irq(&mapping->tree_lock);
+ spin_unlock_irqrestore(&mapping->tree_lock, flags);
if (mapping->host) {
/* !PageAnon && !swapper_space */
__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
