aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorJun'ichi Nomura <j-nomura@ce.jp.nec.com>2007-01-26 03:57:07 -0500
committerLinus Torvalds <torvalds@woody.linux-foundation.org>2007-01-26 16:51:00 -0500
commitbfa152fa5e4d328fe3ebf15908ee8ec20a0ce6dc (patch)
tree1695481328f884639d6237eebfa48f4c29f9ce7c
parente540eb45a5254873245fd377f2fe3afc47bd33c1 (diff)
[PATCH] dm-multipath: fix stall on noflush suspend/resume
Allow noflush suspend/resume of device-mapper device only for the case where the device size is unchanged. Otherwise, dm-multipath devices can stall when resumed if noflush was used when suspending them, all paths have failed and queue_if_no_path is set. Explanation: 1. Something is doing fsync() on the block dev, holding inode->i_sem 2. The fsync write is blocked by all-paths-down and queue_if_no_path 3. Someone requests to suspend the dm device with noflush. Pending writes are left in queue. 4. In the middle of dm_resume(), __bind() tries to get inode->i_sem to do __set_size() and waits forever. 'noflush suspend' is a new device-mapper feature introduced in early 2.6.20. So I hope the fix being included before 2.6.20 is released. Example of reproducer: 1. Create a multipath device by dmsetup 2. Fail all paths during mkfs 3. Do dmsetup suspend --noflush and load new map with healthy paths 4. Do dmsetup resume Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Acked-by: Alasdair G Kergon <agk@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-rw-r--r--drivers/md/dm.c27
1 files changed, 19 insertions, 8 deletions
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index fe7c56e10435..3668b170ea68 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -1116,7 +1116,8 @@ static int __bind(struct mapped_device *md, struct dm_table *t)
1116 if (size != get_capacity(md->disk)) 1116 if (size != get_capacity(md->disk))
1117 memset(&md->geometry, 0, sizeof(md->geometry)); 1117 memset(&md->geometry, 0, sizeof(md->geometry));
1118 1118
1119 __set_size(md, size); 1119 if (md->suspended_bdev)
1120 __set_size(md, size);
1120 if (size == 0) 1121 if (size == 0)
1121 return 0; 1122 return 0;
1122 1123
@@ -1264,6 +1265,11 @@ int dm_swap_table(struct mapped_device *md, struct dm_table *table)
1264 if (!dm_suspended(md)) 1265 if (!dm_suspended(md))
1265 goto out; 1266 goto out;
1266 1267
1268 /* without bdev, the device size cannot be changed */
1269 if (!md->suspended_bdev)
1270 if (get_capacity(md->disk) != dm_table_get_size(table))
1271 goto out;
1272
1267 __unbind(md); 1273 __unbind(md);
1268 r = __bind(md, table); 1274 r = __bind(md, table);
1269 1275
@@ -1341,11 +1347,14 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags)
1341 /* This does not get reverted if there's an error later. */ 1347 /* This does not get reverted if there's an error later. */
1342 dm_table_presuspend_targets(map); 1348 dm_table_presuspend_targets(map);
1343 1349
1344 md->suspended_bdev = bdget_disk(md->disk, 0); 1350 /* bdget() can stall if the pending I/Os are not flushed */
1345 if (!md->suspended_bdev) { 1351 if (!noflush) {
1346 DMWARN("bdget failed in dm_suspend"); 1352 md->suspended_bdev = bdget_disk(md->disk, 0);
1347 r = -ENOMEM; 1353 if (!md->suspended_bdev) {
1348 goto flush_and_out; 1354 DMWARN("bdget failed in dm_suspend");
1355 r = -ENOMEM;
1356 goto flush_and_out;
1357 }
1349 } 1358 }
1350 1359
1351 /* 1360 /*
@@ -1473,8 +1482,10 @@ int dm_resume(struct mapped_device *md)
1473 1482
1474 unlock_fs(md); 1483 unlock_fs(md);
1475 1484
1476 bdput(md->suspended_bdev); 1485 if (md->suspended_bdev) {
1477 md->suspended_bdev = NULL; 1486 bdput(md->suspended_bdev);
1487 md->suspended_bdev = NULL;
1488 }
1478 1489
1479 clear_bit(DMF_SUSPENDED, &md->flags); 1490 clear_bit(DMF_SUSPENDED, &md->flags);
1480 1491