On Mon, Nov 23, 2020 at 06:09:56PM +0100, David Sterba wrote:
> On Wed, Nov 18, 2020 at 11:18:12AM -0800, Omar Sandoval wrote:
> > From: Omar Sandoval <osandov@xxxxxx>
> >
> > Commit 1dae796aabf6 ("btrfs: inode: sink parameter start and len to
> > check_data_csum()") replaced the start parameter to check_data_csum()
> > with page_offset(), but page_offset() is not meaningful for direct I/O
> > pages. Bring back the start parameter.
> >
> > Fixes: 1dae796aabf6 ("btrfs: inode: sink parameter start and len to check_data_csum()")
>
> This is part of the subpage preparatory patches still in misc-next , I
> can drop the part that removes the start parameter if you're going to
> use it.
To be clear, the original patch is buggy. It causes check_data_csum() to
print nonsense for checksum errors encountered during direct I/O. So,
this should be probably be folded in to the original patch.
