On Fri, Oct 11, 2019 at 08:26:42PM +0200, David Sterba wrote:
> You're right, and actually I got crashes in clean_bdev_aliases when I
> supplied a NULL bdev, so I'll add it to the changelog. Thanks.
Unless there are further comments, I won't resend the whole patchset.
The changelog in this patch will be:
---
btrfs: get bdev from latest_dev for dio bh_result
To remove use of extent_map::bdev we need to find a replacement, and the
latest_bdev is the only one we can use here, because inode::i_bdev and
superblock::s_bdev are NULL.
The DIO code uses bdev in two places:
* to read blocksize to perform alignment checks in
do_blockdev_direct_IO, but we do them in btrfs code before any call to
DIO
* in the following call chain:
do_direct_IO
get_more_blocks
sdio->get_block() <-- this is btrfs_get_blocks_direct
subsequently the map_bh->b_dev member is used in clean_bdev_aliases
and dio_new_bio to set the bio's bdev to that of the buffer_head.
However, because we have provided a submit function dio_bio_submit
calls our submission function and ignores the bdev.
We can't pass NULL because there are dereferences of bdev in
__blockdev_direct_IO, even though it's not going to be used later.
So it's safe to pass any valid bdev that's used within the filesystem.
---
