Message ID | e6dd13c5-8db0-4392-6e78-a42ee5d2a1c4@i-love.sakura.ne.jp (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | [v2,(RESEND)] block: genhd: fix double kfree() in __alloc_disk_node() | expand |
On 10/2/21 3:23 AM, Tetsuo Handa wrote: > syzbot is reporting use-after-free read at bdev_free_inode() [1], for > kfree() from __alloc_disk_node() is called before bdev_free_inode() > (which is called after RCU grace period) reads bdev->bd_disk and calls > kfree(bdev->bd_disk). > > Fix use-after-free read followed by double kfree() problem > by making sure that bdev->bd_disk is NULL when calling iput(). Applied for 5.15, thanks.
diff --git a/block/bdev.c b/block/bdev.c index cf2780cb44a7..485a258b0ab3 100644 --- a/block/bdev.c +++ b/block/bdev.c @@ -490,7 +490,6 @@ struct block_device *bdev_alloc(struct gendisk *disk, u8 partno) bdev = I_BDEV(inode); mutex_init(&bdev->bd_fsfreeze_mutex); spin_lock_init(&bdev->bd_size_lock); - bdev->bd_disk = disk; bdev->bd_partno = partno; bdev->bd_inode = inode; bdev->bd_stats = alloc_percpu(struct disk_stats); @@ -498,6 +497,7 @@ struct block_device *bdev_alloc(struct gendisk *disk, u8 partno) iput(inode); return NULL; } + bdev->bd_disk = disk; return bdev; } diff --git a/block/genhd.c b/block/genhd.c index 7b6e5e1cf956..496e8458c357 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -1268,6 +1268,7 @@ struct gendisk *__alloc_disk_node(struct request_queue *q, int node_id, out_destroy_part_tbl: xa_destroy(&disk->part_tbl); + disk->part0->bd_disk = NULL; iput(disk->part0->bd_inode); out_free_bdi: bdi_put(disk->bdi);