On Wed, Apr 21, 2021 at 11:34:44AM +0100, David Howells wrote:
> Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
>
> > On Tue, Apr 20, 2021 at 04:12:57PM -0400, Jeff Layton wrote:
> > > > @@ -210,6 +208,7 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
> > > > * not worth getting one just for that.
> > > > */
> > > > read_pages(ractl, &page_pool, true);
> > > > + i = ractl->_index + ractl->_nr_pages - index;
> >
> > i = ractl->_index + ractl->_nr_pages - index - 1;
> >
> > > > @@ -223,6 +222,7 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
> > > > gfp_mask) < 0) {
> > > > put_page(page);
> > > > read_pages(ractl, &page_pool, true);
> > > > + i = ractl->_index + ractl->_nr_pages - index;
> >
> > i = ractl->_index + ractl->_nr_pages - index - 1;
> >
> > > Thanks Willy, but I think this may not be quite right. A kernel with
> > > this patch failed to boot for me:
> >
> > Silly off-by-one errors. xfstests running against xfs is up to generic/278
> > with the off-by-one fixed.
>
> You can add my Tested-by - or do you want me to add it to my patchset?
I think you need it as part of your patchset, ordered before
readahead_expand(). It probably needs a rewritten description ...
