x86/ucode: Fix stability of the raw CPU Policy rescan

Always run microcode_update_helper() on the BSP, so the the updated Raw CPU
policy doesn't get non-BSP topology details included.

Have calculate_raw_cpu_policy() clear the instantanious XSTATE sizes.  The
value XCR0 | MSR_XSS had when we scanned the policy isn't terribly interesting
to report.

When CPUID Masking is active, it affects CPUID instructions issued by Xen
too.  Transiently disable masking to get a clean scan.

Fixes: 694d79ed5a ("x86/ucode: Refresh raw CPU policy after microcode load")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>

xen/arch/x86/cpu-policy.c

@@ -353,6 +353,13 @@ void calculate_raw_cpu_policy(void)
     /* Nothing good will come from Xen and libx86 disagreeing on vendor. */
     ASSERT(p->x86_vendor == boot_cpu_data.x86_vendor);
 
+    /*
+     * Clear the truly dynamic fields.  These vary with the in-context XCR0
+     * and MSR_XSS, and aren't interesting fields in the raw policy.
+     */
+    p->xstate.raw[0].b = 0;
+    p->xstate.raw[1].b = 0;
+
     /* 0x000000ce  MSR_INTEL_PLATFORM_INFO */
     /* Was already added by probe_cpuid_faulting() */
 }

xen/arch/x86/cpu/microcode/core.c

@@ -680,8 +680,18 @@ static long cf_check microcode_update_helper(void *data)
         microcode_update_cache(patch);
         spin_unlock(&microcode_mutex);
 
-        /* Refresh the raw CPU policy, in case the features have changed. */
+        /*
+         * Refresh the raw CPU policy, in case the features have changed.
+         * Disable CPUID masking if in use, to avoid having current's
+         * cpu_policy affect the rescan.
+         */
+	if ( ctxt_switch_masking )
+            alternative_vcall(ctxt_switch_masking, NULL);
+
         calculate_raw_cpu_policy();
+
+	if ( ctxt_switch_masking )
+            alternative_vcall(ctxt_switch_masking, current);
     }
     else
         microcode_free_patch(patch);
@@ -721,8 +731,12 @@ int microcode_update(XEN_GUEST_HANDLE(const_void) buf, unsigned long len)
     }
     buffer->len = len;
 
-    return continue_hypercall_on_cpu(smp_processor_id(),
-                                     microcode_update_helper, buffer);
+    /*
+     * Always queue microcode_update_helper() on CPU0.  Most of the logic
+     * won't care, but the update of the Raw CPU policy wants to (re)run on
+     * the BSP.
+     */
+    return continue_hypercall_on_cpu(0, microcode_update_helper, buffer);
 }
 
 static int __init cf_check microcode_init(void)