From:   linan666@huaweicloud.com
To:     song@kernel.org, ncroxon@redhat.com, vmayatskikh@digitalocean.com
Cc:     linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org,
        guoqing.jiang@linux.dev, logang@deltatee.com, axboe@kernel.dk,
        linan122@huawei.com, yukuai3@huawei.com, yi.zhang@huawei.com
Subject: [PATCH 2/2] md/raid10: fix null-ptr-deref in  raid10_sync_request
Date:   Wed, 22 Feb 2023 12:10:00 +0800
Message-Id: <20230222041000.3341651-3-linan666@huaweicloud.com>
In-Reply-To: <20230222041000.3341651-1-linan666@huaweicloud.com>
References: <20230222041000.3341651-1-linan666@huaweicloud.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Li Nan <linan122@huawei.com>

init_resync() init mempool and set conf->have_replacemnt at the begaining
of sync, close_sync() free the mempool when sync is completed.

After commit 7e83ccbecd60 ("md/raid10: Allow skipping recovery when clean
arrays are assembled"), recovery might skipped and init_resync() is called
but close_sync() is not. null-ptr-deref occurs as below:
  1) creat a array, wait for resync to complete, mddev->recovery_cp is set
     to MaxSector.
  2) recovery is woken and it is skipped. conf->have_replacement is set to
     0 in init_resync(). close_sync() not called.
  3) some io errors and rdev A is set to WantReplacement.
  4) a new device is added and set to A's replacement.
  5) recovery is woken, A have replacement, but conf->have_replacemnt is
     0. r10bio->dev[i].repl_bio will not be alloced and null-ptr-deref
     occurs.

Fix it by not init_resync() if recovery skipped.

Fixes: 7e83ccbecd60 md/raid10: Allow skipping recovery when clean arrays are assembled")
Signed-off-by: Li Nan <linan122@huawei.com>
---
 drivers/md/raid10.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index db9ee3b637d6..9e0e7bf524aa 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -3297,10 +3297,6 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr,
 	sector_t chunk_mask = conf->geo.chunk_mask;
 	int page_idx = 0;
 
-	if (!mempool_initialized(&conf->r10buf_pool))
-		if (init_resync(conf))
-			return 0;
-
 	/*
 	 * Allow skipping a full rebuild for incremental assembly
 	 * of a clean array, like RAID1 does.
@@ -3316,6 +3312,10 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr,
 		return mddev->dev_sectors - sector_nr;
 	}
 
+	if (!mempool_initialized(&conf->r10buf_pool))
+		if (init_resync(conf))
+			return 0;
+
  skipped:
 	max_sector = mddev->dev_sectors;
 	if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery) ||
-- 
2.31.1