On Sun, Jun 29, 2025 at 11:41:12PM -0700, hch@xxxxxxxxxxxxx wrote: > On Fri, Jun 27, 2025 at 06:03:09PM +0100, Matthew Wilcox wrote: > > On Fri, Jun 27, 2025 at 11:03:13AM +0000, 陈涛涛 Taotao Chen wrote: > > I think this needs to be: > > > > if (iocb && iocb->ki_flags & IOCB_DONTCACHE) > > > > because it's legit to call write_begin with a NULL argument. The > > 'file' was always an optional argument, and we should preserve that > > optionality with this transformation. > > write_begin and write_end are only callbacks through helpers called > by the file system. So if the file system never passes a NULL > file/kiocb it doesn't need to check for it. Sure, but some of those helpers are non-obvious, like page_symlink().
