summaryrefslogtreecommitdiffstats
path: root/drivers/md/dm-raid.c
diff options
context:
space:
mode:
authorHeinz Mauelshagen <heinzm@redhat.com>2016-06-14 15:23:13 -0400
committerMike Snitzer <snitzer@redhat.com>2016-06-14 18:52:14 -0400
commit6e20902e8f9e1551afa75bd499be853a95745b9f (patch)
tree7aab7c01aae2cf81f3da0fb987673a0b8d4d8bf0 /drivers/md/dm-raid.c
parent4257e085e26edaba0bf516ea231bd5122e3f3e6f (diff)
downloadlinux-6e20902e8f9e1551afa75bd499be853a95745b9f.tar.gz
linux-6e20902e8f9e1551afa75bd499be853a95745b9f.tar.bz2
linux-6e20902e8f9e1551afa75bd499be853a95745b9f.zip
dm raid: fix failed takeover/reshapes by keeping raid set frozen
Superblock updates where bogus causing some takovers/reshapes to fail. Introduce new runtime flag (RT_FLAG_KEEP_RS_FROZEN) to keep a raid set frozen when a layout change was requested. Userpace will immediately reload the table w/o the flags requesting such change once they made it to the superblocks and any change of recovery/reshape offsets has to be avoided until after read. Signed-off-by: Heinz Mauelshagen <heinzm@redhat.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Diffstat (limited to 'drivers/md/dm-raid.c')
-rw-r--r--drivers/md/dm-raid.c85
1 files changed, 56 insertions, 29 deletions
diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index 7df450877423..8d4865184b96 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -190,6 +190,7 @@ struct raid_dev {
#define RT_FLAG_RS_BITMAP_LOADED 2
#define RT_FLAG_UPDATE_SBS 3
#define RT_FLAG_RESHAPE_RS 4
+#define RT_FLAG_KEEP_RS_FROZEN 5
/* Array elements of 64 bit needed for rebuild/write_mostly bits */
#define DISKS_ARRAY_ELEMS ((MAX_RAID_DEVICES + (sizeof(uint64_t) * 8 - 1)) / sizeof(uint64_t) / 8)
@@ -2727,6 +2728,7 @@ static int raid_ctr(struct dm_target *ti, unsigned argc, char **argv)
return r;
set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
+ set_bit(RT_FLAG_KEEP_RS_FROZEN, &rs->runtime_flags);
rs_set_new(rs);
} else if (rs_reshape_requested(rs)) {
if (rs_is_reshaping(rs)) {
@@ -2767,13 +2769,19 @@ static int raid_ctr(struct dm_target *ti, unsigned argc, char **argv)
* Would cause allocations in raid1->check_reshape
* though, thus more issues with potential failures
*/
- else if (rs_is_raid1(rs))
+ else if (rs_is_raid1(rs)) {
+ set_bit(RT_FLAG_KEEP_RS_FROZEN, &rs->runtime_flags);
rs->md.raid_disks = rs->raid_disks;
+ }
+
+ if (test_bit(RT_FLAG_RESHAPE_RS, &rs->runtime_flags)) {
+ set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
+ set_bit(RT_FLAG_KEEP_RS_FROZEN, &rs->runtime_flags);
+ }
if (rs->md.raid_disks < rs->raid_disks)
set_bit(MD_ARRAY_FIRST_USE, &rs->md.flags);
- set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
rs_set_cur(rs);
} else
rs_set_cur(rs);
@@ -3231,9 +3239,11 @@ static void raid_postsuspend(struct dm_target *ti)
{
struct raid_set *rs = ti->private;
- mddev_suspend(&rs->md);
- rs->md.ro = 1;
- clear_bit(RT_FLAG_RS_RESUMED, &rs->runtime_flags);
+ if (test_and_clear_bit(RT_FLAG_RS_RESUMED, &rs->runtime_flags)) {
+ if (!rs->md.suspended)
+ mddev_suspend(&rs->md);
+ rs->md.ro = 1;
+ }
}
static void attempt_restore_of_faulty_devices(struct raid_set *rs)
@@ -3308,6 +3318,18 @@ static int __load_dirty_region_bitmap(struct raid_set *rs)
return r;
}
+/* Enforce updating all superblocks */
+static void rs_update_sbs(struct raid_set *rs)
+{
+ struct mddev *mddev = &rs->md;
+ int ro = mddev->ro;
+
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
+ mddev->ro = 0;
+ md_update_sb(mddev, 1);
+ mddev->ro = ro;
+}
+
/*
* Reshape changes raid algorithm of @rs to new one within personality
* (e.g. raid6_zr -> raid6_nc), changes stripe size, adds/removes
@@ -3356,9 +3378,12 @@ static int rs_start_reshape(struct raid_set *rs)
if (!mddev->suspended)
mddev_suspend(mddev);
- mddev->ro = 0;
- md_update_sb(mddev, 1);
- mddev->ro = 1;
+ /*
+ * Now reshape got set up, update superblocks to
+ * reflect the fact so that a table reload will
+ * access proper superblock content in the ctr.
+ */
+ rs_update_sbs(rs);
return 0;
}
@@ -3375,22 +3400,12 @@ static int raid_preresume(struct dm_target *ti)
/*
* The superblocks need to be updated on disk if the
- * array is new or __load_dirty_region_bitmap will overwrite them
- * in core with old data.
- *
- * In case the array got modified (takeover/reshape/resize)
- * or the data offsets on the component devices changed, they
- * have to be updated as well.
- *
- * Have to switch to readwrite and back in order to
- * allow for the superblock updates.
+ * array is new or new devices got added (thus zeroed
+ * out by userspace) or __load_dirty_region_bitmap
+ * will overwrite them in core with old data or fail.
*/
- if (test_and_clear_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags)) {
- set_bit(MD_CHANGE_DEVS, &mddev->flags);
- mddev->ro = 0;
- md_update_sb(mddev, 1);
- mddev->ro = 1;
- }
+ if (test_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags))
+ rs_update_sbs(rs);
/*
* Disable/enable discard support on raid set after any
@@ -3449,14 +3464,26 @@ static void raid_resume(struct dm_target *ti)
* devices are reachable again.
*/
attempt_restore_of_faulty_devices(rs);
- }
+ } else {
+ mddev->ro = 0;
+ mddev->in_sync = 0;
- mddev->ro = 0;
- mddev->in_sync = 0;
- clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+ /*
+ * When passing in flags to the ctr, we expect userspace
+ * to reset them because they made it to the superblocks
+ * and reload the mapping anyway.
+ *
+ * -> only unfreeze recovery in case of a table reload or
+ * we'll have a bogus recovery/reshape position
+ * retrieved from the superblock by the ctr because
+ * the ongoing recovery/reshape will change it after read.
+ */
+ if (!test_bit(RT_FLAG_KEEP_RS_FROZEN, &rs->runtime_flags))
+ clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
- if (mddev->suspended)
- mddev_resume(mddev);
+ if (mddev->suspended)
+ mddev_resume(mddev);
+ }
}
static struct target_type raid_target = {