On Wed, Jun 24, 2020 at 07:36:47PM +0100, Matthew Wilcox wrote: > On Wed, Jun 24, 2020 at 11:33:07AM -0700, Andrew Morton wrote: > > On Wed, 24 Jun 2020 18:52:25 +0100 Catalin Marinas <catalin.marinas@xxxxxxx> wrote: > > > From: Steven Price <steven.price@xxxxxxx> > > > > > > For arm64 MTE support it is necessary to be able to mark pages that > > > contain user space visible tags that will need to be saved/restored e.g. > > > when swapped out. > > > > > > To support this add a new arch specific flag (PG_ARCH_2) that arch code > > > can opt into using ARCH_USES_PG_ARCH_2. > > > > > > ... > > > > > > --- a/fs/proc/page.c > > > +++ b/fs/proc/page.c > > > @@ -217,6 +217,9 @@ u64 stable_page_flags(struct page *page) > > > u |= kpf_copy_bit(k, KPF_PRIVATE_2, PG_private_2); > > > u |= kpf_copy_bit(k, KPF_OWNER_PRIVATE, PG_owner_priv_1); > > > u |= kpf_copy_bit(k, KPF_ARCH, PG_arch_1); > > > +#ifdef CONFIG_ARCH_USES_PG_ARCH_2 > > > + u |= kpf_copy_bit(k, KPF_ARCH_2, PG_arch_2); > > > +#endif > > > > Do we need CONFIG_ARCH_USES_PG_ARCH_2? What would be the downside to > > giving every architecture a PG_arch_2, but only arm64 uses it (at > > present)? > > 32-bit architectures don't have space for it. We could condition it on > CONFIG_64BIT instead. I'll this, though we'd still need some #ifdefs (OTOH, we get rid of the Kconfig entry). -- Catalin
