Re: [syzbot] [fs?] KASAN: slab-use-after-free Read in test_bdev_super_fc

[Christoph Hellwig <hch@xxxxxx>]

On Fri, Aug 04, 2023 at 04:36:49PM +0200, Christian Brauner wrote:
> FFS

Good spot, this explains the missing dropping of s_umount.

But I don't think it's doing the right thing for MTD mount romfs,
we'll need something like this:

diff --git a/fs/romfs/super.c b/fs/romfs/super.c
index c59b230d55b435..4510a38861cfbe 100644
--- a/fs/romfs/super.c
+++ b/fs/romfs/super.c
@@ -583,16 +583,19 @@ static int romfs_init_fs_context(struct fs_context *fc)
  */
 static void romfs_kill_sb(struct super_block *sb)
 {
+	generic_shutdown_super(sb);
+
 #ifdef CONFIG_ROMFS_ON_MTD
 	if (sb->s_mtd) {
-		kill_mtd_super(sb);
-		return;
+		put_mtd_device(sb->s_mtd);
+		sb->s_mtd = NULL;
 	}
 #endif
 #ifdef CONFIG_ROMFS_ON_BLOCK
 	if (sb->s_bdev) {
-		kill_block_super(sb);
-		return;
+		sb->s_bdev->bd_super = NULL;
+		sync_blockdev(sb->s_bdev);
+		blkdev_put(sb->s_bdev, sb->s_type);
 	}
 #endif
 }