On Wed, Jun 20, 2018 at 07:49:31AM -0700, Dave Hansen wrote:
> On 06/13/2018 05:45 PM, Ram Pai wrote:
> > When a key is freed, the key is no more effective.
> > Clear the bits corresponding to the pkey in the shadow
> > register. Otherwise it will carry some spurious bits
> > which can trigger false-positive asserts.
> ...--- a/tools/testing/selftests/vm/protection_keys.c
> > +++ b/tools/testing/selftests/vm/protection_keys.c
> > @@ -556,6 +556,9 @@ int alloc_pkey(void)
> > int sys_pkey_free(unsigned long pkey)
> > {
> > int ret = syscall(SYS_pkey_free, pkey);
> > +
> > + if (!ret)
> > + shadow_pkey_reg &= clear_pkey_flags(pkey, PKEY_DISABLE_ACCESS);
> > dprintf1("%s(pkey=%ld) syscall ret: %d\n", __func__, pkey, ret);
> > return ret;
> > }
>
> This would be great code for an actual application. But, I'm not
> immediately convinced we want sane, kind behavior in our selftest. x86
> doesn't clear the hardware register at pkey_free, so wouldn't this cause
> the shadow and the hardware register to diverge?
Have deleted the code in the newer version.
RP
