On Thu, Mar 05, 2020 at 04:21:34PM +0000, Steven Price wrote:
> On 26/02/2020 18:05, Catalin Marinas wrote:
> > + if (system_supports_mte()) {
> > + /*
> > + * Changing the memory type between Normal and Normal-Tagged
> > + * is safe since Tagged is considered a permission attribute
> > + * from the mismatched attribute aliases perspective.
> > + */
> > + if ((old & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL) ||
> > + (old & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL_TAGGED) ||
> > + (new & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL) ||
> > + (new & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL_TAGGED))
> > + mask |= PTE_ATTRINDX_MASK;
> > + }
> > +
> > return ((old ^ new) & ~mask) == 0;
> > }
>
> This is much more permissive than I would expect. If either the old or
> new memory type is NORMAL (or NORMAL_TAGGED) then the memory type is
> ignored altogether.
>
> Should this check be:
>
> if (((old & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL) ||
> (old & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL_TAGGED)) &&
> ((new & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL) ||
> (new & PTE_ATTRINDX_MASK) == PTE_ATTRINDX(MT_NORMAL_TAGGED)))
You are right, I think my patch allows many other memory type
combinations. Thanks.
--
Catalin
