On 4/9/23 15:40, Paolo Bonzini wrote:
On 9/4/23 14:43, Philippe Mathieu-Daudé wrote:
Call kvm_enabled() before kvm_enable_x2apic() to
let the compiler elide its call.
Suggested-by: Daniel Henrique Barboza <dbarboza@xxxxxxxxxxxxxxxx>
Signed-off-by: Philippe Mathieu-Daudé <philmd@xxxxxxxxxx>
---
hw/i386/intel_iommu.c | 2 +-
hw/i386/x86.c | 2 +-
target/i386/kvm/kvm-stub.c | 7 -------
3 files changed, 2 insertions(+), 9 deletions(-)
diff --git a/hw/i386/x86.c b/hw/i386/x86.c
index a88a126123..d2920af792 100644
--- a/hw/i386/x86.c
+++ b/hw/i386/x86.c
@@ -136,7 +136,7 @@ void x86_cpus_init(X86MachineState *x86ms, int
default_cpu_version)
* With KVM's in-kernel lapic: only if X2APIC API is enabled.
*/
if (x86ms->apic_id_limit > 255 && !xen_enabled() &&
- (!kvm_irqchip_in_kernel() || !kvm_enable_x2apic())) {
+ kvm_enabled() && (!kvm_irqchip_in_kernel() ||
!kvm_enable_x2apic())) {
This "!xen && kvm" expression can be simplified.
I am queuing the series with this squashed in:
diff --git a/hw/i386/x86.c b/hw/i386/x86.c
index d2920af792d..3e86cf3060f 100644
--- a/hw/i386/x86.c
+++ b/hw/i386/x86.c
@@ -129,14 +129,11 @@ void x86_cpus_init(X86MachineState *x86ms, int
default_cpu_version)
ms->smp.max_cpus
- 1) + 1;
/*
- * Can we support APIC ID 255 or higher?
- *
- * Under Xen: yes.
- * With userspace emulated lapic: no
- * With KVM's in-kernel lapic: only if X2APIC API is enabled.
+ * Can we support APIC ID 255 or higher? With KVM, that requires
+ * both in-kernel lapic and X2APIC userspace API.
*/
- if (x86ms->apic_id_limit > 255 && !xen_enabled() &&
- kvm_enabled() && (!kvm_irqchip_in_kernel() ||
!kvm_enable_x2apic())) {
+ if (x86ms->apic_id_limit > 255 && kvm_enabled() &&
+ (!kvm_irqchip_in_kernel() || !kvm_enable_x2apic())) {
error_report("current -smp configuration requires kernel "
"irqchip and X2APIC API support.");
exit(EXIT_FAILURE);
Paolo
Thank you!
