On Fri, 13 Jul 2018 22:47:13 +0200 Richard Weinberger <richard at nod.at> wrote: > Am Freitag, 13. Juli 2018, 16:47:16 CEST schrieb Arnd Bergmann: > > The VFS now uses timespec64 timestamps consistently, but jffs2 still > > converts them to 32-bit numbers on the storage medium. As the helper > > functions for the conversion (get_seconds() and timespec_to_timespec64()) > > are now deprecated, let's change them over to the more modern > > replacements. > > > > This keeps the traditional interpretation of those values, where > > the on-disk 32-bit numbers are taken to be negative numbers, i.e. > > dates before 1970, on 32-bit machines, but future numbers past 2038 > > on 64-bit machines. > > > > Signed-off-by: Arnd Bergmann <arnd at arndb.de> > > --- > > I originally sent these two patches on Jun 19, but got no reply aside > > from a harmless sparse warning reported by the kbuild test robot. > > > > Looking at the git history for jffs2, it seems that David Woodhouse > > hasn't applied any patches for over two years, so I suppose he's > > not taking these either. > > > > Al or Andrew, could you pick these up instead? > > We can carry it also via the MTD tree. I'll queue them to mtd/next. Thanks, Boris