On Thu, Nov 12, 2020 at 09:45:45AM +0000, Vincenzo Frascino wrote:
> On 11/12/20 9:39 AM, Catalin Marinas wrote:
> > On Tue, Nov 10, 2020 at 11:10:29PM +0100, Andrey Konovalov wrote:
> >> diff --git a/arch/arm64/kernel/mte.c b/arch/arm64/kernel/mte.c
> >> index 664c968dc43c..dbda6598c19d 100644
> >> --- a/arch/arm64/kernel/mte.c
> >> +++ b/arch/arm64/kernel/mte.c
> >> @@ -129,6 +131,26 @@ void *mte_set_mem_tag_range(void *addr, size_t size, u8 tag)
> >> return ptr;
> >> }
> >>
> >> +void mte_init_tags(u64 max_tag)
> >> +{
> >> + static bool gcr_kernel_excl_initialized = false;
> >> +
> >> + if (!gcr_kernel_excl_initialized) {
> >> + /*
> >> + * The format of the tags in KASAN is 0xFF and in MTE is 0xF.
> >> + * This conversion extracts an MTE tag from a KASAN tag.
> >> + */
> >> + u64 incl = GENMASK(FIELD_GET(MTE_TAG_MASK >> MTE_TAG_SHIFT,
> >> + max_tag), 0);
> >> +
> >> + gcr_kernel_excl = ~incl & SYS_GCR_EL1_EXCL_MASK;
> >> + gcr_kernel_excl_initialized = true;
> >> + }
> >> +
> >> + /* Enable the kernel exclude mask for random tags generation. */
> >> + write_sysreg_s(SYS_GCR_EL1_RRND | gcr_kernel_excl, SYS_GCR_EL1);
> >> +}
> >
> > I don't think this function belongs to this patch. There is an earlier
> > patch that talks about mte_init_tags() but no trace of it until this
> > patch.
>
> Could you please point out to which patch are you referring to?
I replied to it already (or you can search ;)). But this patch is about
switching GCR_EL1 on exception entry/exit rather than setting up the
initial kernel GCR_EL1 value.
--
Catalin
