On Fri, Jan 12, 2024 at 02:44:35PM +0000, Matthew Wilcox wrote: > On Fri, Jan 12, 2024 at 02:21:55PM +0000, David Howells wrote: > > Jitindar Singh, Suraj <surajjs@xxxxxxxxxx> wrote: > > > > > When testing the v6.1.69 kernel I bisected an issue to the below commit > > > which was added in v6.1.68. When running the xfstests[1] on cifs I > > > observe a null pointer dereference in cifs_flush_folio() because folio > > > is null and dereferenced in size = folio_size(folio). > > > > Yeah. __filemap_get_folio() works differently in v6.1.y. There it returns a > > Let's cc hch so he's aware of how much pain this change caused. > > Never change function semantics without changing its name or arguments. Maybe the maintainer should have pushed back if he feels strongly about that? That being said backporting any fs code without first having done a QA run just feels horribly dangerous to start with. People need to stop doing that.
