x86/HVM: avoid memory_type_changed() invocations when possible

They're expensive, and nothing changes if MTRRs are disabled and any of
the ranges gets changed, or if fixed range MTRRs are disabled and any of
them gets changed.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>

diff --git a/xen/arch/x86/hvm/mtrr.c b/xen/arch/x86/hvm/mtrr.c
index eb86cbacb4ec6cc60f871be63ee6bff8f38818af..c502dda693b62647d43eb587dfb210f9f79c9fcf 100644 (file)
--- a/xen/arch/x86/hvm/mtrr.c
+++ b/xen/arch/x86/hvm/mtrr.c
@@ -472,7 +472,9 @@ bool_t mtrr_fix_range_msr_set(struct domain *d, struct mtrr_state *m,
                 return 0;
 
         fixed_range_base[row] = msr_content;
-        memory_type_changed(d);
+
+        if ( m->enabled && m->fixed_enabled )
+            memory_type_changed(d);
     }
 
     return 1;
@@ -515,7 +517,8 @@ bool_t mtrr_var_range_msr_set(
 
     m->overlapped = is_var_mtrr_overlapped(m);
 
-    memory_type_changed(d);
+    if ( m->enabled )
+        memory_type_changed(d);
 
     return 1;
 }