Re: [PATCHv5 07/19] x86/mm: Mask out KeyID bits from page table entry pfn

On Thu, Jul 19, 2018 at 07:19:01AM -0700, Dave Hansen wrote:

On 07/19/2018 02:54 AM, Kirill A. Shutemov wrote:

quoted

On Wed, Jul 18, 2018 at 04:13:20PM -0700, Dave Hansen wrote:

quoted

On 07/17/2018 04:20 AM, Kirill A. Shutemov wrote:

quoted

+	} else {
+		/*
+		 * Reset __PHYSICAL_MASK.
+		 * Maybe needed if there's inconsistent configuation
+		 * between CPUs.
+		 */
+		physical_mask = (1ULL << __PHYSICAL_MASK_SHIFT) - 1;
+	}

This seems like an appropriate place for a WARN_ON().  Either that, or
axe this code.

There's pr_err_once() above in the function.

Do you mean for the (tme_activate != tme_activate_cpu0) check?

But that's about double-activating this feature.  This check is about an
inconsistent configuration between two CPUs which seems totally different.

Could you explain?

(tme_activate != tme_activate_cpu0) check is about inconsistent
configuration. It checks if MSR's content on the given CPU matches MSR on
CPU0.

-- 
 Kirill A. Shutemov