* [PATCH 000 of 2] md: Fix a bug with backward event updates.
@ 2006-08-24 7:20 NeilBrown
2006-08-24 7:20 ` [PATCH 001 of 2] md: Avoid backward event updates in md superblock when degraded NeilBrown
2006-08-24 7:20 ` [PATCH 002 of 2] md: replace magic numbers in sb_dirty with well defined bit flags NeilBrown
0 siblings, 2 replies; 3+ messages in thread
From: NeilBrown @ 2006-08-24 7:20 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-raid, linux-kernel
Hi Andrew,
There is a bug in 2.6.18-rc which needs to be fixed before -final,
but there is a patch in -mm which is not scheduled for 2.6.18 which
touches the same code - so that patch has to change.
So:
wind back to
md-replace-magic-numbers-in-sb_dirty-with-well-defined-bit-flags.patch
discard that patch.
Apply 1 of 2 following - that needs to go to Linux (or Greg)
Then apply 2 of 2 to replace the one that was removed.
Then you should be able to wind forwards again and everything will apply.
(If there is some other way you would like me to deal with this
situation - just say so).
Thanks,
NeilBrown
[PATCH 001 of 2] md: Avoid backward event updates in md superblock when degraded.
[PATCH 002 of 2] md: replace magic numbers in sb_dirty with well defined bit flags
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 001 of 2] md: Avoid backward event updates in md superblock when degraded.
2006-08-24 7:20 [PATCH 000 of 2] md: Fix a bug with backward event updates NeilBrown
@ 2006-08-24 7:20 ` NeilBrown
2006-08-24 7:20 ` [PATCH 002 of 2] md: replace magic numbers in sb_dirty with well defined bit flags NeilBrown
1 sibling, 0 replies; 3+ messages in thread
From: NeilBrown @ 2006-08-24 7:20 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-raid, linux-kernel
If we
- shut down a clean array,
- restart with one (or more) drive(s) missing
- make some changes
- pause, so that they array gets marked 'clean',
the event count on the superblock of included drives
will be the same as that of the removed drives.
So adding the removed drive back in will cause it
to be included with no resync.
To avoid this, we only update the eventcount backwards when the array
is not degraded. In this case there can (should) be no non-connected
drives that we can get confused with, and this is the particular case
where updating-backwards is valuable.
Signed-off-by: Neil Brown <neilb@suse.de>
### Diffstat output
./drivers/md/md.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff .prev/drivers/md/md.c ./drivers/md/md.c
--- .prev/drivers/md/md.c 2006-08-24 17:03:33.000000000 +1000
+++ ./drivers/md/md.c 2006-08-24 17:03:47.000000000 +1000
@@ -1597,6 +1597,19 @@ void md_update_sb(mddev_t * mddev)
repeat:
spin_lock_irq(&mddev->write_lock);
+
+ if (mddev->degraded && mddev->sb_dirty == 3)
+ /* If the array is degraded, then skipping spares is both
+ * dangerous and fairly pointless.
+ * Dangerous because a device that was removed from the array
+ * might have a event_count that still looks up-to-date,
+ * so it can be re-added without a resync.
+ * Pointless because if there are any spares to skip,
+ * then a recovery will happen and soon that array won't
+ * be degraded any more and the spare can go back to sleep then.
+ */
+ mddev->sb_dirty = 1;
+
sync_req = mddev->in_sync;
mddev->utime = get_seconds();
if (mddev->sb_dirty == 3)
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 002 of 2] md: replace magic numbers in sb_dirty with well defined bit flags
2006-08-24 7:20 [PATCH 000 of 2] md: Fix a bug with backward event updates NeilBrown
2006-08-24 7:20 ` [PATCH 001 of 2] md: Avoid backward event updates in md superblock when degraded NeilBrown
@ 2006-08-24 7:20 ` NeilBrown
1 sibling, 0 replies; 3+ messages in thread
From: NeilBrown @ 2006-08-24 7:20 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-raid, linux-kernel
From: NeilBrown <neilb@suse.de>
Instead of magic numbers (0,1,2,3) in sb_dirty, we have
some flags instead:
MD_CHANGE_DEVS
Some device state has changed requiring superblock update
on all devices.
MD_CHANGE_CLEAN
The array has transitions from 'clean' to 'dirty' or back,
requiring a superblock update on active devices, but possibly
not on spares
MD_CHANGE_PENDING
A superblock update is underway.
We wait for an update to complete by waiting for all flags to be clear. A
flag can be set at any time, even during an update, without risk that the
change will be lost.
Stop exporting md_update_sb - isn't needed.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Avoid backward event updates in md superblock when degraded.
If we
- shut down a clean array,
- restart with one (or more) drive(s) missing
- make some changes
- pause, so that they array gets marked 'clean',
the event count on the superblock of included drives
will be the same as that of the removed drives.
So adding the removed drive back in will cause it
to be included with no resync.
To avoid this, we only update the eventcount backwards when the array
is not degraded. In this case there can (should) be no non-connected
drives that we can get confused with, and this is the particular case
where updating-backwards is valuable.
Signed-off-by: Neil Brown <neilb@suse.de>
### Diffstat output
./drivers/md/md.c | 76 +++++++++++++++++++++++---------------------
./drivers/md/multipath.c | 3 -
./drivers/md/raid1.c | 2 -
./drivers/md/raid10.c | 2 -
./drivers/md/raid5.c | 8 ++--
./include/linux/raid/md.h | 1
./include/linux/raid/md_k.h | 6 ++-
7 files changed, 52 insertions(+), 46 deletions(-)
diff .prev/drivers/md/md.c ./drivers/md/md.c
--- .prev/drivers/md/md.c 2006-08-24 17:03:47.000000000 +1000
+++ ./drivers/md/md.c 2006-08-24 17:06:12.000000000 +1000
@@ -1587,7 +1587,7 @@ static void sync_sbs(mddev_t * mddev, in
}
}
-void md_update_sb(mddev_t * mddev)
+static void md_update_sb(mddev_t * mddev, int force_change)
{
int err;
struct list_head *tmp;
@@ -1598,7 +1598,18 @@ void md_update_sb(mddev_t * mddev)
repeat:
spin_lock_irq(&mddev->write_lock);
- if (mddev->degraded && mddev->sb_dirty == 3)
+ set_bit(MD_CHANGE_PENDING, &mddev->flags);
+ if (test_and_clear_bit(MD_CHANGE_DEVS, &mddev->flags))
+ force_change = 1;
+ if (test_and_clear_bit(MD_CHANGE_CLEAN, &mddev->flags))
+ /* just a clean<-> dirty transition, possibly leave spares alone,
+ * though if events isn't the right even/odd, we will have to do
+ * spares after all
+ */
+ nospares = 1;
+ if (force_change)
+ nospares = 0;
+ if (mddev->degraded)
/* If the array is degraded, then skipping spares is both
* dangerous and fairly pointless.
* Dangerous because a device that was removed from the array
@@ -1608,20 +1619,14 @@ repeat:
* then a recovery will happen and soon that array won't
* be degraded any more and the spare can go back to sleep then.
*/
- mddev->sb_dirty = 1;
+ nospares = 0;
sync_req = mddev->in_sync;
mddev->utime = get_seconds();
- if (mddev->sb_dirty == 3)
- /* just a clean<-> dirty transition, possibly leave spares alone,
- * though if events isn't the right even/odd, we will have to do
- * spares after all
- */
- nospares = 1;
/* If this is just a dirty<->clean transition, and the array is clean
* and 'events' is odd, we can roll back to the previous clean state */
- if (mddev->sb_dirty == 3
+ if (nospares
&& (mddev->in_sync && mddev->recovery_cp == MaxSector)
&& (mddev->events & 1))
mddev->events--;
@@ -1652,7 +1657,6 @@ repeat:
MD_BUG();
mddev->events --;
}
- mddev->sb_dirty = 2;
sync_sbs(mddev, nospares);
/*
@@ -1660,7 +1664,7 @@ repeat:
* nonpersistent superblocks
*/
if (!mddev->persistent) {
- mddev->sb_dirty = 0;
+ clear_bit(MD_CHANGE_PENDING, &mddev->flags);
spin_unlock_irq(&mddev->write_lock);
wake_up(&mddev->sb_wait);
return;
@@ -1697,20 +1701,20 @@ repeat:
break;
}
md_super_wait(mddev);
- /* if there was a failure, sb_dirty was set to 1, and we re-write super */
+ /* if there was a failure, MD_CHANGE_DEVS was set, and we re-write super */
spin_lock_irq(&mddev->write_lock);
- if (mddev->in_sync != sync_req|| mddev->sb_dirty == 1) {
+ if (mddev->in_sync != sync_req ||
+ test_bit(MD_CHANGE_DEVS, &mddev->flags)) {
/* have to write it out again */
spin_unlock_irq(&mddev->write_lock);
goto repeat;
}
- mddev->sb_dirty = 0;
+ clear_bit(MD_CHANGE_PENDING, &mddev->flags);
spin_unlock_irq(&mddev->write_lock);
wake_up(&mddev->sb_wait);
}
-EXPORT_SYMBOL_GPL(md_update_sb);
/* words written to sysfs files may, or my not, be \n terminated.
* We want to accept with case. For this we use cmd_match.
@@ -1783,7 +1787,7 @@ state_store(mdk_rdev_t *rdev, const char
else {
mddev_t *mddev = rdev->mddev;
kick_rdev_from_array(rdev);
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
md_new_event(mddev);
err = 0;
}
@@ -2426,7 +2430,7 @@ array_state_store(mddev_t *mddev, const
spin_lock_irq(&mddev->write_lock);
if (atomic_read(&mddev->writes_pending) == 0) {
mddev->in_sync = 1;
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_CLEAN, &mddev->flags);
}
spin_unlock_irq(&mddev->write_lock);
} else {
@@ -2438,7 +2442,7 @@ array_state_store(mddev_t *mddev, const
case active:
if (mddev->pers) {
restart_array(mddev);
- mddev->sb_dirty = 0;
+ clear_bit(MD_CHANGE_CLEAN, &mddev->flags);
wake_up(&mddev->sb_wait);
err = 0;
} else {
@@ -2543,7 +2547,7 @@ size_store(mddev_t *mddev, const char *b
if (mddev->pers) {
err = update_size(mddev, size);
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
} else {
if (mddev->size == 0 ||
mddev->size > size)
@@ -3111,8 +3115,8 @@ static int do_md_run(mddev_t * mddev)
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
- if (mddev->sb_dirty)
- md_update_sb(mddev);
+ if (mddev->flags)
+ md_update_sb(mddev, 0);
set_capacity(disk, mddev->array_size<<1);
@@ -3275,10 +3279,10 @@ static int do_md_stop(mddev_t * mddev, i
if (mddev->ro)
mddev->ro = 0;
}
- if (!mddev->in_sync || mddev->sb_dirty) {
+ if (!mddev->in_sync || mddev->flags) {
/* mark array as shutdown cleanly */
mddev->in_sync = 1;
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
}
if (mode == 1)
set_disk_ro(disk, 1);
@@ -3747,7 +3751,7 @@ static int hot_remove_disk(mddev_t * mdd
goto busy;
kick_rdev_from_array(rdev);
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
md_new_event(mddev);
return 0;
@@ -3824,7 +3828,7 @@ static int hot_add_disk(mddev_t * mddev,
rdev->raid_disk = -1;
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
/*
* Kick recovery, maybe this spare has to be added to the
@@ -3955,7 +3959,8 @@ static int set_array_info(mddev_t * mdde
mddev->max_disks = MD_SB_DISKS;
- mddev->sb_dirty = 1;
+ mddev->flags = 0;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
mddev->default_bitmap_offset = MD_SB_BYTES >> 9;
mddev->bitmap_offset = 0;
@@ -4124,7 +4129,7 @@ static int update_array_info(mddev_t *md
mddev->bitmap_offset = 0;
}
}
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
return rv;
}
@@ -4960,12 +4965,12 @@ void md_write_start(mddev_t *mddev, stru
spin_lock_irq(&mddev->write_lock);
if (mddev->in_sync) {
mddev->in_sync = 0;
- mddev->sb_dirty = 3;
+ set_bit(MD_CHANGE_CLEAN, &mddev->flags);
md_wakeup_thread(mddev->thread);
}
spin_unlock_irq(&mddev->write_lock);
}
- wait_event(mddev->sb_wait, mddev->sb_dirty==0);
+ wait_event(mddev->sb_wait, mddev->flags==0);
}
void md_write_end(mddev_t *mddev)
@@ -5235,7 +5240,6 @@ void md_do_sync(mddev_t *mddev)
!test_bit(In_sync, &rdev->flags) &&
rdev->recovery_offset < mddev->curr_resync)
rdev->recovery_offset = mddev->curr_resync;
- mddev->sb_dirty = 1;
}
}
@@ -5292,7 +5296,7 @@ void md_check_recovery(mddev_t *mddev)
}
if ( ! (
- mddev->sb_dirty ||
+ mddev->flags ||
test_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||
test_bit(MD_RECOVERY_DONE, &mddev->recovery) ||
(mddev->safemode == 1) ||
@@ -5308,14 +5312,14 @@ void md_check_recovery(mddev_t *mddev)
if (mddev->safemode && !atomic_read(&mddev->writes_pending) &&
!mddev->in_sync && mddev->recovery_cp == MaxSector) {
mddev->in_sync = 1;
- mddev->sb_dirty = 3;
+ set_bit(MD_CHANGE_CLEAN, &mddev->flags);
}
if (mddev->safemode == 1)
mddev->safemode = 0;
spin_unlock_irq(&mddev->write_lock);
- if (mddev->sb_dirty)
- md_update_sb(mddev);
+ if (mddev->flags)
+ md_update_sb(mddev, 0);
if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
@@ -5334,7 +5338,7 @@ void md_check_recovery(mddev_t *mddev)
/* activate any spares */
mddev->pers->spare_active(mddev);
}
- md_update_sb(mddev);
+ md_update_sb(mddev, 1);
/* if array is no-longer degraded, then any saved_raid_disk
* information must be scrapped
diff .prev/drivers/md/multipath.c ./drivers/md/multipath.c
--- .prev/drivers/md/multipath.c 2006-08-24 17:03:33.000000000 +1000
+++ ./drivers/md/multipath.c 2006-08-24 17:04:03.000000000 +1000
@@ -253,7 +253,7 @@ static void multipath_error (mddev_t *md
char b[BDEVNAME_SIZE];
clear_bit(In_sync, &rdev->flags);
set_bit(Faulty, &rdev->flags);
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
conf->working_disks--;
printk(KERN_ALERT "multipath: IO failure on %s,"
" disabling IO path. \n Operation continuing"
@@ -470,7 +470,6 @@ static int multipath_run (mddev_t *mddev
}
conf->raid_disks = mddev->raid_disks;
- mddev->sb_dirty = 1;
conf->mddev = mddev;
spin_lock_init(&conf->device_lock);
INIT_LIST_HEAD(&conf->retry_list);
diff .prev/drivers/md/raid1.c ./drivers/md/raid1.c
--- .prev/drivers/md/raid1.c 2006-08-24 17:03:33.000000000 +1000
+++ ./drivers/md/raid1.c 2006-08-24 17:04:03.000000000 +1000
@@ -966,7 +966,7 @@ static void error(mddev_t *mddev, mdk_rd
}
clear_bit(In_sync, &rdev->flags);
set_bit(Faulty, &rdev->flags);
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
printk(KERN_ALERT "raid1: Disk failure on %s, disabling device. \n"
" Operation continuing on %d devices\n",
bdevname(rdev->bdev,b), conf->working_disks);
diff .prev/drivers/md/raid10.c ./drivers/md/raid10.c
--- .prev/drivers/md/raid10.c 2006-08-24 17:03:33.000000000 +1000
+++ ./drivers/md/raid10.c 2006-08-24 17:04:03.000000000 +1000
@@ -960,7 +960,7 @@ static void error(mddev_t *mddev, mdk_rd
}
clear_bit(In_sync, &rdev->flags);
set_bit(Faulty, &rdev->flags);
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
printk(KERN_ALERT "raid10: Disk failure on %s, disabling device. \n"
" Operation continuing on %d devices\n",
bdevname(rdev->bdev,b), conf->working_disks);
diff .prev/drivers/md/raid5.c ./drivers/md/raid5.c
--- .prev/drivers/md/raid5.c 2006-08-24 17:03:33.000000000 +1000
+++ ./drivers/md/raid5.c 2006-08-24 17:04:03.000000000 +1000
@@ -696,7 +696,7 @@ static void error(mddev_t *mddev, mdk_rd
PRINTK("raid5: error called\n");
if (!test_bit(Faulty, &rdev->flags)) {
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
if (test_bit(In_sync, &rdev->flags)) {
conf->working_disks--;
mddev->degraded++;
@@ -2781,9 +2781,9 @@ static sector_t reshape_request(mddev_t
wait_event(conf->wait_for_overlap,
atomic_read(&conf->reshape_stripes)==0);
mddev->reshape_position = conf->expand_progress;
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
md_wakeup_thread(mddev->thread);
- wait_event(mddev->sb_wait, mddev->sb_dirty == 0 ||
+ wait_event(mddev->sb_wait, mddev->flags == 0 ||
kthread_should_stop());
spin_lock_irq(&conf->device_lock);
conf->expand_lo = mddev->reshape_position;
@@ -3605,7 +3605,7 @@ static int raid5_start_reshape(mddev_t *
mddev->degraded = (conf->raid_disks - conf->previous_raid_disks) - added_devices;
mddev->raid_disks = conf->raid_disks;
mddev->reshape_position = 0;
- mddev->sb_dirty = 1;
+ set_bit(MD_CHANGE_DEVS, &mddev->flags);
clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
diff .prev/include/linux/raid/md.h ./include/linux/raid/md.h
--- .prev/include/linux/raid/md.h 2006-08-24 17:03:33.000000000 +1000
+++ ./include/linux/raid/md.h 2006-08-24 17:04:03.000000000 +1000
@@ -93,7 +93,6 @@ extern int sync_page_io(struct block_dev
extern void md_do_sync(mddev_t *mddev);
extern void md_new_event(mddev_t *mddev);
-extern void md_update_sb(mddev_t * mddev);
#endif
diff .prev/include/linux/raid/md_k.h ./include/linux/raid/md_k.h
--- .prev/include/linux/raid/md_k.h 2006-08-24 17:03:33.000000000 +1000
+++ ./include/linux/raid/md_k.h 2006-08-24 17:04:03.000000000 +1000
@@ -114,7 +114,11 @@ struct mddev_s
dev_t unit;
int md_minor;
struct list_head disks;
- int sb_dirty;
+ unsigned long flags;
+#define MD_CHANGE_DEVS 0 /* Some device status has changed */
+#define MD_CHANGE_CLEAN 1 /* transition to or from 'clean' */
+#define MD_CHANGE_PENDING 2 /* superblock update in progress */
+
int ro;
struct gendisk *gendisk;
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2006-08-24 7:20 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-08-24 7:20 [PATCH 000 of 2] md: Fix a bug with backward event updates NeilBrown
2006-08-24 7:20 ` [PATCH 001 of 2] md: Avoid backward event updates in md superblock when degraded NeilBrown
2006-08-24 7:20 ` [PATCH 002 of 2] md: replace magic numbers in sb_dirty with well defined bit flags NeilBrown
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).