Re: [PATCH v3] cifs: fix pagecache leak when do writepages

[Matthew Wilcox <willy@infradead.org>]

On Thu, Sep 11, 2025 at 05:25:06PM +0100, David Howells wrote:
> yangerkun <yangerkun@huawei.com> wrote:
> 
> > >     	if (folio->mapping != mapping ||
> > >   	    !folio_test_dirty(folio)) {
> > >   		start += folio_size(folio);
> > > +		folio_put(folio);
> > >   		folio_unlock(folio);
> > >   		goto search_again;
> 
> I wonder if the put should be prior to the unlock.  It probably doesn't matter
> as we keep control of the folio until both have happened.

Well, folio->mapping != mapping is the condition for 'this folio has
been truncated', so this folio_put() may well be the last one.  I'd
put it after the folio_unlock() for safety.
>