[PATCH v2 0/2] md/raid5: validate disk count before reshape assembly

[PATCH v2 0/2] md/raid5: validate disk count before reshape assembly

[Chen Cheng]

From: Chen Cheng <chencheng@fnnas.com>

The previous patch rejects raid4/5 arrays with too few disks in
setup_conf(). During reshape assembly, raid5_run() still divides by the
data-disk count before that check, so a crafted superblock can hit
division by zero.

Extract min_raid_disks() and reject invalid reshape geometries in
raid5_run() first.

v1 -> v2:
- Added Link

Chen Cheng (2):
  md/raid5: extract min_raid_disks()
  md/raid5: reject invalid reshape disks in raid5_run()

 drivers/md/raid5.c | 28 ++++++++++++++++++----------
 1 file changed, 18 insertions(+), 10 deletions(-)

-- 
2.55.0

[PATCH v2 1/2] md/raid5: extract min_raid_disks()

[Chen Cheng]

From: Chen Cheng <chencheng@fnnas.com>

Reuse the minimum disk count in setup_conf() and check_reshape().

Signed-off-by: Chen Cheng <chencheng@fnnas.com>
---
 drivers/md/raid5.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index d5a8428778e3..f0390918adb3 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -7630,14 +7630,19 @@ static unsigned long raid5_cache_count(struct shrinker *shrink,
 		/* unlikely, but not impossible */
 		return 0;
 	return max_stripes - min_stripes;
 }
 
+static int min_raid_disks(int level)
+{
+	return level == 6 ? 4 : 2;
+}
+
 static struct r5conf *setup_conf(struct mddev *mddev)
 {
 	struct r5conf *conf;
-	int raid_disk, memory, max_disks;
+	int raid_disk, memory, max_disks, min_disks;
 	struct md_rdev *rdev;
 	struct disk_info *disk;
 	char pers_name[6];
 	int i;
 	int group_cnt;
@@ -7657,15 +7662,15 @@ static struct r5conf *setup_conf(struct mddev *mddev)
 	     && !algorithm_valid_raid6(mddev->new_layout))) {
 		pr_warn("md/raid:%s: layout %d not supported\n",
 			mdname(mddev), mddev->new_layout);
 		return ERR_PTR(-EIO);
 	}
-	if ((mddev->new_level != 6 && mddev->raid_disks < 2) ||
-	    (mddev->new_level == 6 && mddev->raid_disks < 4)) {
+	min_disks = min_raid_disks(mddev->new_level);
+	if (mddev->raid_disks < min_disks) {
 		pr_warn("md/raid:%s: not enough configured devices (%d, minimum %d)\n",
 			mdname(mddev), mddev->raid_disks,
-			mddev->new_level == 6 ? 4 : 2);
+			min_disks);
 		return ERR_PTR(-EINVAL);
 	}
 
 	if (!mddev->new_chunk_sectors ||
 	    (mddev->new_chunk_sectors << 9) % PAGE_SIZE ||
@@ -8608,17 +8613,13 @@ static int check_reshape(struct mddev *mddev)
 	if (!mddev->new_chunk_sectors)
 		return -EINVAL;
 	if (mddev->delta_disks < 0 && mddev->reshape_position == MaxSector) {
 		/* We might be able to shrink, but the devices must
 		 * be made bigger first.
-		 * For raid6, 4 is the minimum size.
-		 * Otherwise 2 is the minimum
 		 */
-		int min = 2;
-		if (mddev->level == 6)
-			min = 4;
-		if (mddev->raid_disks + mddev->delta_disks < min)
+		if (mddev->raid_disks + mddev->delta_disks <
+		    min_raid_disks(mddev->level))
 			return -EINVAL;
 	}
 
 	if (!check_stripe_cache(mddev))
 		return -ENOSPC;
-- 
2.55.0

[PATCH v2 2/2] md/raid5: reject invalid reshape disks in raid5_run()

[Chen Cheng]

From: Chen Cheng <chencheng@fnnas.com>

Check disk count before dividing by data disks while assembling a
reshape, otherwise too few disks can cause division by zero.

Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Link: https://lore.kernel.org/r/20260911042717.487AC1F00898@smtp.kernel.org/
Signed-off-by: Chen Cheng <chencheng@fnnas.com>
---
v1 -> v2:
- Added Link
---
 drivers/md/raid5.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index f0390918adb3..5db096de8101 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -8053,10 +8053,11 @@ static int raid5_run(struct mddev *mddev)
 		 * the array in read-only mode, so we check for that.
 		 */
 		sector_t here_new, here_old;
 		int old_disks;
 		int max_degraded = (mddev->level == 6 ? 2 : 1);
+		int min_disks = min_raid_disks(mddev->level);
 		int chunk_sectors;
 		int new_data_disks;
 
 		if (journal_dev) {
 			pr_warn("md/raid:%s: don't support reshape with journal - aborting.\n",
@@ -8068,10 +8069,16 @@ static int raid5_run(struct mddev *mddev)
 			pr_warn("md/raid:%s: unsupported reshape required - aborting.\n",
 				mdname(mddev));
 			return -EINVAL;
 		}
 		old_disks = mddev->raid_disks - mddev->delta_disks;
+		if (mddev->raid_disks < min_disks || old_disks < min_disks) {
+			pr_warn("md/raid:%s: invalid reshape disks (%d -> %d, min %d)\n",
+				mdname(mddev), old_disks, mddev->raid_disks,
+				min_disks);
+			return -EINVAL;
+		}
 		/* reshape_position must be on a new-stripe boundary, and one
 		 * further up in new geometry must map after here in old
 		 * geometry.
 		 * If the chunk sizes are different, then as we perform reshape
 		 * in units of the largest of the two, reshape_position needs
-- 
2.55.0

Re: [PATCH v2 0/2] md/raid5: validate disk count before reshape assembly

[yu kuai]

在 2026/9/21 15:16, Chen Cheng 写道:

> From: Chen Cheng<chencheng@fnnas.com>
>
> The previous patch rejects raid4/5 arrays with too few disks in
> setup_conf(). During reshape assembly, raid5_run() still divides by the
> data-disk count before that check, so a crafted superblock can hit
> division by zero.
>
> Extract min_raid_disks() and reject invalid reshape geometries in
> raid5_run() first.
>
> v1 -> v2:
> - Added Link
>
> Chen Cheng (2):
>    md/raid5: extract min_raid_disks()
>    md/raid5: reject invalid reshape disks in raid5_run()
>
>   drivers/md/raid5.c | 28 ++++++++++++++++++----------
>   1 file changed, 18 insertions(+), 10 deletions(-)
Applied to md-7.4

-- 
Thanks,
Kuai