From: Glauber Costa <glommer@redhat.com>
Date: Wed, 19 Aug 2009 06:51:58 -0400
Subject: [x86] disable kvmclock by default
Message-id: 1250679118-5730-1-git-send-email-glommer@redhat.com
O-Subject: [PATCH] RHEL5 BZ476075 Disable kvmclock by default
Bugzilla: 476075
RH-Acked-by: Prarit Bhargava <prarit@redhat.com>
RH-Acked-by: Juan Quintela <quintela@redhat.com>
RH-Acked-by: Rik van Riel <riel@redhat.com>
RH-Acked-by: Chris Lalancette <clalance@redhat.com>

This is a follow up from the last kvmclock series. Testing revealed
the kernel was crashing in baremetal. Don, I am sending this a separate
patch to ease acking. Please tell me if there is any other preferred
method.

This new patch got testing from me, jburke and clalance, and seem to be
working alright.

Changelog:
Differently from upstream, we use a use_kvm_time variable to control
whether or not kvmclock is enabled.

But if kvm is not enabled, we never get a chance to disable. So default
it to zero, instead of one, to prevent failure on baremetal.

Signed-off-by: Glauber Costa <glommer@redhat.com>

diff --git a/arch/i386/kernel/kvmclock.c b/arch/i386/kernel/kvmclock.c
index 6edd2e2..0b8f5e8 100644
--- a/arch/i386/kernel/kvmclock.c
+++ b/arch/i386/kernel/kvmclock.c
@@ -30,7 +30,7 @@
 
 static int kvmclock = 1;
 
-unsigned int use_kvm_time = 1; /* RHEL specific */
+unsigned int use_kvm_time = 0; /* RHEL specific */
 
 static int parse_no_kvmclock(char *arg)
 {
@@ -119,6 +119,7 @@ void __init kvmclock_init(void)
 		return;
 
 	if (kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE)) {
+		use_kvm_time = 1;
 		if (kvm_register_clock("boot clock")) {
 			use_kvm_time = 0;
 			return;