Re: [PATCH -next] mm/filemap: fix that first page is not mark accessed in filemap_read()

[Yu Kuai]

On 2022/06/03 2:22, Andrew Morton wrote:
> On Thu, 2 Jun 2022 16:21:29 +0800 Yu Kuai <yukuai3@xxxxxxxxxx> wrote:
>
> > In filemap_read(), 'ra->prev_pos' is set to 'iocb->ki_pos + copied',
> > while it should be 'iocb->ki_ops'. For consequence,
> > folio_mark_accessed() will not be called for 'fbatch.folios[0]' since
> > 'iocb->ki_pos' is always equal to 'ra->prev_pos'.
> >
> > ...
> >
> > --- a/mm/filemap.c
> > +++ b/mm/filemap.c
> > @@ -2728,10 +2728,11 @@ ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter,
> >   				flush_dcache_folio(folio);
> >   
> >   			copied = copy_folio_to_iter(folio, offset, bytes, iter);
> > -
> > -			already_read += copied;
> > -			iocb->ki_pos += copied;
> > -			ra->prev_pos = iocb->ki_pos;
> > +			if (copied) {
> > +				ra->prev_pos = iocb->ki_pos;
> > +				already_read += copied;
> > +				iocb->ki_pos += copied;
> > +			}
> >   
> >   			if (copied < bytes) {
> >   				error = -EFAULT;
>
> It seems tidier, but does it matter?  If copied==0 we're going to break
> out and return -EFAULT anyway?
Hi,

Please notice that I set 'prev_ops' to 'ki_pos' first here, instead of
'ki_pos + copied'.

Thanks,
Kuai
> .