Re: [bug report] md: also clone new io if io accounting is disabled

Hi,

在 2024/01/31 19:17, Dan Carpenter 写道:

Hello Yu Kuai,

The patch c687297b8845: "md: also clone new io if io accounting is
disabled" from Jun 22, 2023 (linux-next), leads to the following
Smatch static checker warning:

	drivers/md/md.c:8718 md_clone_bio()
	potential NULL container_of 'clone'

drivers/md/md.c
     8711 static void md_clone_bio(struct mddev *mddev, struct bio **bio)
     8712 {
     8713         struct block_device *bdev = (*bio)->bi_bdev;
     8714         struct md_io_clone *md_io_clone;
     8715         struct bio *clone =
     8716                 bio_alloc_clone(bdev, *bio, GFP_NOIO, &mddev->io_clone_set);

If you ever look at the comment, you'll know that this will never fail.

  * If %__GFP_DIRECT_RECLAIM is set then bio_alloc will always be able to
  * allocate a bio.  This is due to the mempool guarantees.

Thanks,
Kuai

Generally in the kernel, you have to check for allocation failure.  In
this case if the allocation fails it leads to a NULL dereference.

     8717
--> 8718         md_io_clone = container_of(clone, struct md_io_clone, bio_clone);
     8719         md_io_clone->orig_bio = *bio;
     8720         md_io_clone->mddev = mddev;
     8721         if (blk_queue_io_stat(bdev->bd_disk->queue))
     8722                 md_io_clone->start_time = bio_start_io_acct(*bio);
     8723
     8724         clone->bi_end_io = md_end_clone_io;
     8725         clone->bi_private = md_io_clone;
     8726         *bio = clone;
     8727 }

regards,
dan carpenter
.