On 03/26/2012 07:47 AM, Jan Kiszka wrote:
On 2012-03-26 14:42, Avi Kivity wrote:
On 03/26/2012 04:06 AM, Wanpeng Li wrote:
From: Anthony Liguori<aligu...@us.ibm.com>
@@ -889,7 +900,7 @@ static DeviceState *apic_init(void *env, uint8_t apic_id)
DeviceState *dev;
static int apic_mapped;
- if (kvm_irqchip_in_kernel()) {
+ if (kvm_enabled()&& kvm_irqchip_in_kernel()) {
dev = qdev_create(NULL, "kvm-apic");
} else {
dev = qdev_create(NULL, "apic");
@@ -908,7 +919,7 @@ static DeviceState *apic_init(void *env, uint8_t apic_id)
}
/* KVM does not support MSI yet. */
- if (!kvm_irqchip_in_kernel()) {
+ if (!kvm_enabled() || !kvm_irqchip_in_kernel()) {
msi_supported = true;
Why these changes?
Yep, they are obsolete, likely related to the rebase of the original
patch. A lot of code is moved around here, and I bet there are more
artifacts...
git should have thrown a rebase error here. I think that means the conflicts
were ignored.
For something like this, if there's a rebase error, you pretty much have to
repeat the copy/paste of the code for the section that threw a rebase error.
I'd suggest rebasing again but this time, be a bit more careful when resolving
conflicts.
Regards,
Anthony Liguori
Jan