Re: [PATCH 1/6] mm/page_io: use a folio in __end_swap_bio_read()

[zhangpeng (AS)]

On 2023/7/19 0:16, Matthew Wilcox wrote:

> On Tue, Jul 18, 2023 at 08:56:16PM +0800, zhangpeng (AS) wrote:
> > > >    	if (bio->bi_status) {
> > > > -		SetPageError(page);
> > > > -		ClearPageUptodate(page);
> > > > +		folio_set_error(folio);
> > > I appreciate this is a 1:1 conversion, but maybe we could think about
> > > this a bit.  Is there anybody who checks the
> > > PageError()/folio_test_error() for this page/folio?
> > Maybe wait_dev_supers() checks the PageError() after write_dev_supers()
> > in fs/btrfs/disk-io.c?
> How does _this_ folio end up in btrfs's write_dev_supers()?  This is a
> swap read.  The only folios which are swapped are anonymous and tmpfs.
> btrfs takes care of doing its own I/O.  wait_dev_supers() is looking
> for the error set in btrfs_end_super_write() which is the completion
> routine for write_dev_supers().  The pages involved there are attached
> to a btrfs address_space, not shmem or anon.

Thanks for your explanation!

Then I think nobody checks the PageError()/folio_test_error() for the page
in patch 1 and patch 2. I'll delete SetPageError() in a v2.

--
Best Regards,
Peng