summaryrefslogtreecommitdiffstats
path: root/drivers/md
diff options
context:
space:
mode:
authorNeilBrown <neilb@suse.de>2014-09-30 08:10:42 +1000
committerNeilBrown <neilb@suse.de>2014-10-14 13:08:28 +1100
commitac05f256691fe427a3e84c19261adb0b67dd73c0 (patch)
tree2f7254e7117eac0fc0d974029e5bb5422866a73c /drivers/md
parent8b1afc3d6751063d3f0cdefe55719b1cd2f7edcc (diff)
downloadop-kernel-dev-ac05f256691fe427a3e84c19261adb0b67dd73c0.zip
op-kernel-dev-ac05f256691fe427a3e84c19261adb0b67dd73c0.tar.gz
md: don't start resync thread directly from md thread.
The main 'md' thread is needed for processing writes, so if it blocks write requests could be delayed. Starting a new thread requires some GFP_KERNEL allocations and so can wait for writes to complete. This can deadlock. So instead, ask a workqueue to start the sync thread. There is no particular rush for this to happen, so any work queue will do. MD_RECOVERY_RUNNING is used to ensure only one thread is started. Reported-by: BillStuff <billstuff2001@sbcglobal.net> Signed-off-by: NeilBrown <neilb@suse.de>
Diffstat (limited to 'drivers/md')
-rw-r--r--drivers/md/md.c59
1 files changed, 36 insertions, 23 deletions
diff --git a/drivers/md/md.c b/drivers/md/md.c
index a7e9fae..19171c5 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -7767,6 +7767,33 @@ no_add:
return spares;
}
+static void md_start_sync(struct work_struct *ws)
+{
+ struct mddev *mddev = container_of(ws, struct mddev, del_work);
+
+ mddev->sync_thread = md_register_thread(md_do_sync,
+ mddev,
+ "resync");
+ if (!mddev->sync_thread) {
+ printk(KERN_ERR "%s: could not start resync"
+ " thread...\n",
+ mdname(mddev));
+ /* leave the spares where they are, it shouldn't hurt */
+ clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
+ clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
+ clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
+ clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
+ clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
+ if (test_and_clear_bit(MD_RECOVERY_RECOVER,
+ &mddev->recovery))
+ if (mddev->sysfs_action)
+ sysfs_notify_dirent_safe(mddev->sysfs_action);
+ } else
+ md_wakeup_thread(mddev->sync_thread);
+ sysfs_notify_dirent_safe(mddev->sysfs_action);
+ md_new_event(mddev);
+}
+
/*
* This routine is regularly called by all per-raid-array threads to
* deal with generic issues like resync and super-block update.
@@ -7883,7 +7910,7 @@ void md_check_recovery(struct mddev *mddev)
if (!test_and_clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||
test_bit(MD_RECOVERY_FROZEN, &mddev->recovery))
- goto unlock;
+ goto not_running;
/* no recovery is running.
* remove any failed drives, then
* add spares if possible.
@@ -7895,7 +7922,7 @@ void md_check_recovery(struct mddev *mddev)
if (mddev->pers->check_reshape == NULL ||
mddev->pers->check_reshape(mddev) != 0)
/* Cannot proceed */
- goto unlock;
+ goto not_running;
set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
} else if ((spares = remove_and_add_spares(mddev, NULL))) {
@@ -7908,7 +7935,7 @@ void md_check_recovery(struct mddev *mddev)
clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
} else if (!test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
/* nothing to be done ... */
- goto unlock;
+ goto not_running;
if (mddev->pers->sync_request) {
if (spares) {
@@ -7918,27 +7945,11 @@ void md_check_recovery(struct mddev *mddev)
*/
bitmap_write_all(mddev->bitmap);
}
- mddev->sync_thread = md_register_thread(md_do_sync,
- mddev,
- "resync");
- if (!mddev->sync_thread) {
- printk(KERN_ERR "%s: could not start resync"
- " thread...\n",
- mdname(mddev));
- /* leave the spares where they are, it shouldn't hurt */
- clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
- clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
- clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
- clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
- clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
- } else
- md_wakeup_thread(mddev->sync_thread);
- sysfs_notify_dirent_safe(mddev->sysfs_action);
- md_new_event(mddev);
+ INIT_WORK(&mddev->del_work, md_start_sync);
+ queue_work(md_misc_wq, &mddev->del_work);
+ goto unlock;
}
- unlock:
- wake_up(&mddev->sb_wait);
-
+ not_running:
if (!mddev->sync_thread) {
clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
if (test_and_clear_bit(MD_RECOVERY_RECOVER,
@@ -7946,6 +7957,8 @@ void md_check_recovery(struct mddev *mddev)
if (mddev->sysfs_action)
sysfs_notify_dirent_safe(mddev->sysfs_action);
}
+ unlock:
+ wake_up(&mddev->sb_wait);
mddev_unlock(mddev);
}
}
OpenPOWER on IntegriCloud