On 25 Nov 2024, at 16:01, Matthew Wilcox (Oracle) wrote: > We already have the concept of "frozen pages" (eg page_ref_freeze()), > so let's not complicate things by also having the concept of "unref > pages". > > Reviewed-by: David Hildenbrand <david@xxxxxxxxxx> > Reviewed-by: William Kucharski <william.kucharski@xxxxxxxxxx> > Reviewed-by: Miaohe Lin <linmiaohe@xxxxxxxxxx> > Reviewed-by: Muchun Song <songmuchun@xxxxxxxxxxxxx> > Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx> > --- > mm/internal.h | 2 +- > mm/page_alloc.c | 18 +++++++++--------- > mm/page_frag_cache.c | 6 +++--- > mm/swap.c | 2 +- > 4 files changed, 14 insertions(+), 14 deletions(-) Shouldn't the name be free_frozen_page() instead of free_frozen_pages()? It is still free a single page, right? If free_frozen_pages() is what we want, free_frozen_page_commit() should be free_frozen_pages_commit() to be consistent. Otherwise, LGTM. Reviewed-by: Zi Yan <ziy@xxxxxxxxxx> -- Best Regards, Yan, Zi
