From: Jan Beulich <jbeulich@suse.com>
Date: Mon, 22 Jul 2019 09:32:20 +0000 (+0200)
Subject: x86/cpuidle: really use C1 for "urgent" CPUs
X-Git-Url: http://xenbits.xensource.com/gitweb?a=commitdiff_plain;h=1bce2a94e929ddb0aa0ba8222cdaf0fd10ef1821;p=people%2Fdariof%2Fxen.git

x86/cpuidle: really use C1 for "urgent" CPUs

For one on recent AMD CPUs entering C1 (if available at all) requires
use of MWAIT, while HLT (i.e. default_idle()) would put the processor
into as deep as CC6. And then even on other vendors' CPUs we should
avoid entering default_idle() when the intended state can be reached
by using the active idle driver's facilities.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
---

diff --git a/xen/arch/x86/acpi/cpu_idle.c b/xen/arch/x86/acpi/cpu_idle.c
index 6896cfa875..ff945da7a1 100644
--- a/xen/arch/x86/acpi/cpu_idle.c
+++ b/xen/arch/x86/acpi/cpu_idle.c
@@ -582,12 +582,15 @@ static void acpi_processor_idle(void)
     u32 exp = 0, pred = 0;
     u32 irq_traced[4] = { 0 };
 
-    if ( max_cstate > 0 && power && !sched_has_urgent_vcpu() &&
+    if ( max_cstate > 0 && power &&
          (next_state = cpuidle_current_governor->select(power)) > 0 )
     {
+        unsigned int max_state = sched_has_urgent_vcpu() ? ACPI_STATE_C1
+                                                         : max_cstate;
+
         do {
             cx = &power->states[next_state];
-        } while ( cx->type > max_cstate && --next_state );
+        } while ( cx->type > max_state && --next_state );
         if ( next_state )
         {
             if ( cx->type == ACPI_STATE_C3 && power->flags.bm_check &&
diff --git a/xen/arch/x86/cpu/mwait-idle.c b/xen/arch/x86/cpu/mwait-idle.c
index 361de1bb35..9e7c9bc28e 100644
--- a/xen/arch/x86/cpu/mwait-idle.c
+++ b/xen/arch/x86/cpu/mwait-idle.c
@@ -724,11 +724,14 @@ static void mwait_idle(void)
 	u64 before, after;
 	u32 exp = 0, pred = 0, irq_traced[4] = { 0 };
 
-	if (max_cstate > 0 && power && !sched_has_urgent_vcpu() &&
+	if (max_cstate > 0 && power &&
 	    (next_state = cpuidle_current_governor->select(power)) > 0) {
+		unsigned int max_state = sched_has_urgent_vcpu() ? ACPI_STATE_C1
+								 : max_cstate;
+
 		do {
 			cx = &power->states[next_state];
-		} while (cx->type > max_cstate && --next_state);
+		} while (cx->type > max_state && --next_state);
 		if (!next_state)
 			cx = NULL;
 		else if (tb_init_done)