[PATCH 0/2] md: it panice after reshape from raid1 to raid5

linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed

* [PATCH 0/2] md: it panice after reshape from raid1 to raid5
@ 2021-12-10  9:31 Xiao Ni
  2021-12-10  9:31 ` [PATCH V2 1/2] md/raid0: Free r0conf memory when register integrity failed Xiao Ni
                   ` (2 more replies)
  0 siblings, 3 replies; 6+ messages in thread
From: Xiao Ni @ 2021-12-10  9:31 UTC (permalink / raw)
  To: song; +Cc: guoqing.jiang, ncroxon, linux-raid

Hi all

After reshape from raid1 to raid5, it can panice when there are I/Os

The steps can reproduce this:
mdadm -CR /dev/md0 -l1 -n2 /dev/loop0 /dev/loop1
mdadm --wait /dev/md0
mkfs.xfs /dev/md0
mdadm /dev/md0 --grow -l5
mount /dev/md0 /mnt

These two patches fix this problem.

Xiao Ni (2):
  Free r0conf memory when register integrity failed
  Move alloc/free acct bioset in to personality

 drivers/md/md.c    | 27 +++++++++++++++++----------
 drivers/md/md.h    |  2 ++
 drivers/md/raid0.c | 28 ++++++++++++++++++++++++----
 drivers/md/raid5.c | 41 ++++++++++++++++++++++++++++++-----------
 4 files changed, 73 insertions(+), 25 deletions(-)

-- 
2.31.1


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH V2 1/2] md/raid0: Free r0conf memory when register integrity failed
  2021-12-10  9:31 [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
@ 2021-12-10  9:31 ` Xiao Ni
  2021-12-10  9:31 ` [PATCH 2/2] md: Move alloc/free acct bioset in to personality Xiao Ni
  2022-01-04 23:30 ` [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
  2 siblings, 0 replies; 6+ messages in thread
From: Xiao Ni @ 2021-12-10  9:31 UTC (permalink / raw)
  To: song; +Cc: guoqing.jiang, ncroxon, linux-raid

It doesn't free memory when register integrity failed. And it will
add acct_bioset_exit in raid0_free. So split free r0conf codes
into a single function to make error handling more clear.

Signed-off-by: Xiao Ni <xni@redhat.com>
---
V2: set mddev->private to NULL and move free_conf/raid0_free
above to avoid the extra declaration
---
 drivers/md/raid0.c | 28 +++++++++++++++++++---------
 1 file changed, 19 insertions(+), 9 deletions(-)

diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 62c8b6adac70..88424d7a6ebd 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -356,7 +356,20 @@ static sector_t raid0_size(struct mddev *mddev, sector_t sectors, int raid_disks
 	return array_sectors;
 }
 
-static void raid0_free(struct mddev *mddev, void *priv);
+static void free_conf(struct mddev *mddev, struct r0conf *conf)
+{
+	kfree(conf->strip_zone);
+	kfree(conf->devlist);
+	kfree(conf);
+	mddev->private = NULL;
+}
+
+static void raid0_free(struct mddev *mddev, void *priv)
+{
+	struct r0conf *conf = priv;
+
+	free_conf(mddev, conf);
+}
 
 static int raid0_run(struct mddev *mddev)
 {
@@ -413,17 +426,14 @@ static int raid0_run(struct mddev *mddev)
 	dump_zones(mddev);
 
 	ret = md_integrity_register(mddev);
+	if (ret)
+		goto free;
 
 	return ret;
-}
-
-static void raid0_free(struct mddev *mddev, void *priv)
-{
-	struct r0conf *conf = priv;
 
-	kfree(conf->strip_zone);
-	kfree(conf->devlist);
-	kfree(conf);
+free:
+	free_conf(mddev, conf);
+	return ret;
 }
 
 static void raid0_handle_discard(struct mddev *mddev, struct bio *bio)
-- 
2.31.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* [PATCH 2/2] md: Move alloc/free acct bioset in to personality
  2021-12-10  9:31 [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
  2021-12-10  9:31 ` [PATCH V2 1/2] md/raid0: Free r0conf memory when register integrity failed Xiao Ni
@ 2021-12-10  9:31 ` Xiao Ni
  2022-01-04 23:30 ` [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
  2 siblings, 0 replies; 6+ messages in thread
From: Xiao Ni @ 2021-12-10  9:31 UTC (permalink / raw)
  To: song; +Cc: guoqing.jiang, ncroxon, linux-raid

Now it alloc acct bioset in md_run and only raid0/raid5 need acct
bioset. For example, it doesn't create acct bioset when creating
raid1. Then reshape from raid1 to raid0/raid5, it will access acct
bioset after reshaping. It can panic because of NULL pointer
reference.

We can move alloc/free jobs to personality. pers->run alloc acct
bioset and pers->clean free it.

Fixes: daee2024715d (md: check level before create and exit io_acct_set)
Signed-off-by: Xiao Ni <xni@redhat.com>
---
 drivers/md/md.c    | 27 +++++++++++++++++----------
 drivers/md/md.h    |  2 ++
 drivers/md/raid0.c | 10 +++++++++-
 drivers/md/raid5.c | 41 ++++++++++++++++++++++++++++++-----------
 4 files changed, 58 insertions(+), 22 deletions(-)

diff --git a/drivers/md/md.c b/drivers/md/md.c
index e8666bdc0d28..0fc34a05a655 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -5878,13 +5878,6 @@ int md_run(struct mddev *mddev)
 		if (err)
 			goto exit_bio_set;
 	}
-	if (mddev->level != 1 && mddev->level != 10 &&
-	    !bioset_initialized(&mddev->io_acct_set)) {
-		err = bioset_init(&mddev->io_acct_set, BIO_POOL_SIZE,
-				  offsetof(struct md_io_acct, bio_clone), 0);
-		if (err)
-			goto exit_sync_set;
-	}
 
 	spin_lock(&pers_lock);
 	pers = find_pers(mddev->level, mddev->clevel);
@@ -6061,9 +6054,6 @@ int md_run(struct mddev *mddev)
 	module_put(pers->owner);
 	md_bitmap_destroy(mddev);
 abort:
-	if (mddev->level != 1 && mddev->level != 10)
-		bioset_exit(&mddev->io_acct_set);
-exit_sync_set:
 	bioset_exit(&mddev->sync_set);
 exit_bio_set:
 	bioset_exit(&mddev->bio_set);
@@ -8596,6 +8586,23 @@ void md_submit_discard_bio(struct mddev *mddev, struct md_rdev *rdev,
 }
 EXPORT_SYMBOL_GPL(md_submit_discard_bio);
 
+int acct_bioset_init(struct mddev *mddev)
+{
+	int err = 0;
+
+	if (!bioset_initialized(&mddev->io_acct_set))
+		err = bioset_init(&mddev->io_acct_set, BIO_POOL_SIZE,
+			offsetof(struct md_io_acct, bio_clone), 0);
+	return err;
+}
+EXPORT_SYMBOL_GPL(acct_bioset_init);
+
+void acct_bioset_exit(struct mddev *mddev)
+{
+	bioset_exit(&mddev->io_acct_set);
+}
+EXPORT_SYMBOL_GPL(acct_bioset_exit);
+
 static void md_end_io_acct(struct bio *bio)
 {
 	struct md_io_acct *md_io_acct = bio->bi_private;
diff --git a/drivers/md/md.h b/drivers/md/md.h
index 53ea7a6961de..f1bf3625ef4c 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -721,6 +721,8 @@ extern void md_error(struct mddev *mddev, struct md_rdev *rdev);
 extern void md_finish_reshape(struct mddev *mddev);
 void md_submit_discard_bio(struct mddev *mddev, struct md_rdev *rdev,
 			struct bio *bio, sector_t start, sector_t size);
+int acct_bioset_init(struct mddev *mddev);
+void acct_bioset_exit(struct mddev *mddev);
 void md_account_bio(struct mddev *mddev, struct bio **bio);
 
 extern bool __must_check md_flush_request(struct mddev *mddev, struct bio *bio);
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 88424d7a6ebd..b59a77b31b90 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -369,6 +369,7 @@ static void raid0_free(struct mddev *mddev, void *priv)
 	struct r0conf *conf = priv;
 
 	free_conf(mddev, conf);
+	acct_bioset_exit(mddev);
 }
 
 static int raid0_run(struct mddev *mddev)
@@ -383,11 +384,16 @@ static int raid0_run(struct mddev *mddev)
 	if (md_check_no_bitmap(mddev))
 		return -EINVAL;
 
+	if (acct_bioset_init(mddev)) {
+		pr_err("md/raid0:%s: alloc acct bioset failed.\n", mdname(mddev));
+		return -ENOMEM;
+	}
+
 	/* if private is not null, we are here after takeover */
 	if (mddev->private == NULL) {
 		ret = create_strip_zones(mddev, &conf);
 		if (ret < 0)
-			return ret;
+			goto exit_acct_set;
 		mddev->private = conf;
 	}
 	conf = mddev->private;
@@ -433,6 +439,8 @@ static int raid0_run(struct mddev *mddev)
 
 free:
 	free_conf(mddev, conf);
+exit_acct_set:
+	acct_bioset_exit(mddev);
 	return ret;
 }
 
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 1240a5c16af8..13afa8c5cc8a 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -7447,12 +7447,19 @@ static int raid5_run(struct mddev *mddev)
 	struct md_rdev *rdev;
 	struct md_rdev *journal_dev = NULL;
 	sector_t reshape_offset = 0;
-	int i;
+	int i, ret = 0;
 	long long min_offset_diff = 0;
 	int first = 1;
 
-	if (mddev_init_writes_pending(mddev) < 0)
+	if (acct_bioset_init(mddev)) {
+		pr_err("md/raid456:%s: alloc acct bioset failed.\n", mdname(mddev));
 		return -ENOMEM;
+	}
+
+	if (mddev_init_writes_pending(mddev) < 0) {
+		ret = -ENOMEM;
+		goto exit_acct_set;
+	}
 
 	if (mddev->recovery_cp != MaxSector)
 		pr_notice("md/raid:%s: not clean -- starting background reconstruction\n",
@@ -7483,7 +7490,8 @@ static int raid5_run(struct mddev *mddev)
 	    (mddev->bitmap_info.offset || mddev->bitmap_info.file)) {
 		pr_notice("md/raid:%s: array cannot have both journal and bitmap\n",
 			  mdname(mddev));
-		return -EINVAL;
+		ret = -EINVAL;
+		goto exit_acct_set;
 	}
 
 	if (mddev->reshape_position != MaxSector) {
@@ -7508,13 +7516,15 @@ static int raid5_run(struct mddev *mddev)
 		if (journal_dev) {
 			pr_warn("md/raid:%s: don't support reshape with journal - aborting.\n",
 				mdname(mddev));
-			return -EINVAL;
+			ret = -EINVAL;
+			goto exit_acct_set;
 		}
 
 		if (mddev->new_level != mddev->level) {
 			pr_warn("md/raid:%s: unsupported reshape required - aborting.\n",
 				mdname(mddev));
-			return -EINVAL;
+			ret = -EINVAL;
+			goto exit_acct_set;
 		}
 		old_disks = mddev->raid_disks - mddev->delta_disks;
 		/* reshape_position must be on a new-stripe boundary, and one
@@ -7530,7 +7540,8 @@ static int raid5_run(struct mddev *mddev)
 		if (sector_div(here_new, chunk_sectors * new_data_disks)) {
 			pr_warn("md/raid:%s: reshape_position not on a stripe boundary\n",
 				mdname(mddev));
-			return -EINVAL;
+			ret = -EINVAL;
+			goto exit_acct_set;
 		}
 		reshape_offset = here_new * chunk_sectors;
 		/* here_new is the stripe we will write to */
@@ -7552,7 +7563,8 @@ static int raid5_run(struct mddev *mddev)
 			else if (mddev->ro == 0) {
 				pr_warn("md/raid:%s: in-place reshape must be started in read-only mode - aborting\n",
 					mdname(mddev));
-				return -EINVAL;
+				ret = -EINVAL;
+				goto exit_acct_set;
 			}
 		} else if (mddev->reshape_backwards
 		    ? (here_new * chunk_sectors + min_offset_diff <=
@@ -7562,7 +7574,8 @@ static int raid5_run(struct mddev *mddev)
 			/* Reading from the same stripe as writing to - bad */
 			pr_warn("md/raid:%s: reshape_position too early for auto-recovery - aborting.\n",
 				mdname(mddev));
-			return -EINVAL;
+			ret = -EINVAL;
+			goto exit_acct_set;
 		}
 		pr_debug("md/raid:%s: reshape will continue\n", mdname(mddev));
 		/* OK, we should be able to continue; */
@@ -7586,8 +7599,10 @@ static int raid5_run(struct mddev *mddev)
 	else
 		conf = mddev->private;
 
-	if (IS_ERR(conf))
-		return PTR_ERR(conf);
+	if (IS_ERR(conf)) {
+		ret = PTR_ERR(conf);
+		goto exit_acct_set;
+	}
 
 	if (test_bit(MD_HAS_JOURNAL, &mddev->flags)) {
 		if (!journal_dev) {
@@ -7784,7 +7799,10 @@ static int raid5_run(struct mddev *mddev)
 	free_conf(conf);
 	mddev->private = NULL;
 	pr_warn("md/raid:%s: failed to run raid set.\n", mdname(mddev));
-	return -EIO;
+	ret = -EIO;
+exit_acct_set:
+	acct_bioset_exit(mddev);
+	return ret;
 }
 
 static void raid5_free(struct mddev *mddev, void *priv)
@@ -7792,6 +7810,7 @@ static void raid5_free(struct mddev *mddev, void *priv)
 	struct r5conf *conf = priv;
 
 	free_conf(conf);
+	acct_bioset_exit(mddev);
 	mddev->to_remove = &raid5_attrs_group;
 }
 
-- 
2.31.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH 0/2] md: it panice after reshape from raid1 to raid5
  2021-12-10  9:31 [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
  2021-12-10  9:31 ` [PATCH V2 1/2] md/raid0: Free r0conf memory when register integrity failed Xiao Ni
  2021-12-10  9:31 ` [PATCH 2/2] md: Move alloc/free acct bioset in to personality Xiao Ni
@ 2022-01-04 23:30 ` Xiao Ni
  2022-01-05 18:59   ` Song Liu
  2 siblings, 1 reply; 6+ messages in thread
From: Xiao Ni @ 2022-01-04 23:30 UTC (permalink / raw)
  To: song; +Cc: guoqing.jiang, ncroxon, linux-raid

Hi Song

Ping. Do I still change something else?

Regards

Xiao

在 2021/12/10 17:31, Xiao Ni 写道:
> Hi all
>
> After reshape from raid1 to raid5, it can panice when there are I/Os
>
> The steps can reproduce this:
> mdadm -CR /dev/md0 -l1 -n2 /dev/loop0 /dev/loop1
> mdadm --wait /dev/md0
> mkfs.xfs /dev/md0
> mdadm /dev/md0 --grow -l5
> mount /dev/md0 /mnt
>
> These two patches fix this problem.
>
> Xiao Ni (2):
>    Free r0conf memory when register integrity failed
>    Move alloc/free acct bioset in to personality
>
>   drivers/md/md.c    | 27 +++++++++++++++++----------
>   drivers/md/md.h    |  2 ++
>   drivers/md/raid0.c | 28 ++++++++++++++++++++++++----
>   drivers/md/raid5.c | 41 ++++++++++++++++++++++++++++++-----------
>   4 files changed, 73 insertions(+), 25 deletions(-)
>


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH 0/2] md: it panice after reshape from raid1 to raid5
  2022-01-04 23:30 ` [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
@ 2022-01-05 18:59   ` Song Liu
  2022-01-06  1:53     ` Xiao Ni
  0 siblings, 1 reply; 6+ messages in thread
From: Song Liu @ 2022-01-05 18:59 UTC (permalink / raw)
  To: Xiao Ni; +Cc: Guoqing Jiang, Nigel Croxon, linux-raid

On Tue, Jan 4, 2022 at 3:30 PM Xiao Ni <xni@redhat.com> wrote:
>
> Hi Song
>
> Ping. Do I still change something else?

I merged the two patches into one, rewrote the commit log, added
Guoqing's Acked-by, and applied it to md-next.

For future patches, please write the commit log according to the
guidance in Documentation/process/submitting-patches.rst.

Thanks,
Song

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH 0/2] md: it panice after reshape from raid1 to raid5
  2022-01-05 18:59   ` Song Liu
@ 2022-01-06  1:53     ` Xiao Ni
  0 siblings, 0 replies; 6+ messages in thread
From: Xiao Ni @ 2022-01-06  1:53 UTC (permalink / raw)
  To: Song Liu; +Cc: Guoqing Jiang, Nigel Croxon, linux-raid

On Thu, Jan 6, 2022 at 2:59 AM Song Liu <song@kernel.org> wrote:
>
> On Tue, Jan 4, 2022 at 3:30 PM Xiao Ni <xni@redhat.com> wrote:
> >
> > Hi Song
> >
> > Ping. Do I still change something else?
>
> I merged the two patches into one, rewrote the commit log, added
> Guoqing's Acked-by, and applied it to md-next.
>
> For future patches, please write the commit log according to the
> guidance in Documentation/process/submitting-patches.rst.
>
> Thanks,
> Song
>

Thanks. I'll read this doc and follow the instructions.

Regards
Xiao


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2022-01-06  1:53 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-12-10  9:31 [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
2021-12-10  9:31 ` [PATCH V2 1/2] md/raid0: Free r0conf memory when register integrity failed Xiao Ni
2021-12-10  9:31 ` [PATCH 2/2] md: Move alloc/free acct bioset in to personality Xiao Ni
2022-01-04 23:30 ` [PATCH 0/2] md: it panice after reshape from raid1 to raid5 Xiao Ni
2022-01-05 18:59   ` Song Liu
2022-01-06  1:53     ` Xiao Ni

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).