On Tue, Jan 14, 2020 at 11:20:04PM -0800, Christoph Hellwig wrote: > On Tue, Jan 14, 2020 at 06:38:43PM -0800, Matthew Wilcox wrote: > > From: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx> > > > > We already have various bits of add_to_page_cache() executed conditionally > > on !PageHuge(page); add the add_to_page_cache_lru() pieces as some > > more code which isn't executed for huge pages. This lets us remove > > the old add_to_page_cache() and rename __add_to_page_cache_locked() to > > add_to_page_cache(). Include a compatibility define so we don't have > > to change all 20+ callers of add_to_page_cache_lru(). > > I'd rather change them. 20ish isn't that much after all, and not > keeping pointless aliases around keeps the code easier to read. Almost all of them are called in the ->readpages() function, so they'll go away as filesystems are converted to ->readahead(). I'd rather not introduce something that makes patches harder to reorder.
