Re: [PATCH v5 11/12] filemap: Add support for authoritative mappings

[Matthew Wilcox <willy@infradead.org>]

Sashiko:
> --- Patch [11]: [PATCH v5 11/12] filemap: Add support for authoritative mappings ---
> commit 7b270a787d853107f7a64dd3127d799a28d3142e
> Author: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
> filemap: Add support for authoritative mappings
> 
> An authoritative mapping knows about all the folios in the mapping.
> If read() finds a missing folio, there's no reason to allocate one and
> try to read it because we know it's a zero region of the file.  We can
> just call iov_iter_zero() instead.
> 
> > diff --git a/mm/filemap.c b/mm/filemap.c
> > index 26a5f18121f90..5a8cc20e624e4 100644
> > --- a/mm/filemap.c
> > +++ b/mm/filemap.c
> > @@ -2697,6 +2697,8 @@ static int filemap_get_pages(struct kiocb *iocb, size_t count,
> >  	if (!folio_batch_count(fbatch)) {
> >  		DEFINE_READAHEAD(ractl, filp, &filp->f_ra, mapping, index);
> >  
> > +		if (mapping_is_authoritative(mapping))
> > +			return 0;
> 
> Does returning 0 here cause an infinite loop in filemap_splice_read()?

As before, hugetlb doesn't use splice_read() so adding support would be
dead untestable code.  We'll add support here when we need it.

> > @@ -2853,6 +2855,22 @@ ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter,
> >  			goto put_folios;
> >  		end_offset = min_t(loff_t, isize, iocb->ki_pos + iter->count);
> >  
> > +		if (!folio_batch_count(&fbatch)) {
> > +			size_t fsize = mapping_min_folio_nrbytes(mapping);
> > +			size_t offset = iocb->ki_pos & (fsize - 1);
> > +			size_t bytes = min_t(loff_t, end_offset - iocb->ki_pos,
> > +					     fsize - offset);
> > +			size_t copied = iov_iter_zero(bytes, iter);
> > +
> > +			already_read += copied;
> > +			iocb->ki_pos += copied;
> > +			last_pos = iocb->ki_pos;
> > +
> > +			if (copied < bytes)
> > +				error = -EFAULT;
> > +			continue;
> > +		}
> 
> Should filemap_fault() also be updated to handle authoritative mappings?

Eventually, if hugetlbfs is ever converted to use filemap_fault().
As above, it'd be dead code.


That's it for interesting Sashiko reviews.  I've omitted some which
are suplicates, either within this patch series or have been reported
against previous patch series.  Feels like we're close!