From: Hollis Blanchard <[EMAIL PROTECTED]> We're saving the host TLB state to memory on every exit, but never using it. Originally I had thought that we'd want to restore host TLB for heavyweight exits, but that could actually hurt when context switching to an unrelated host process (i.e. not qemu).
Since this decreases the performance penalty of all exits, this patch improves guest boot time by about 15%. Signed-off-by: Hollis Blanchard <[EMAIL PROTECTED]> Signed-off-by: Avi Kivity <[EMAIL PROTECTED]> --- arch/powerpc/include/asm/kvm_host.h | 2 -- arch/powerpc/kernel/asm-offsets.c | 1 - arch/powerpc/kvm/booke_interrupts.S | 17 +++-------------- 3 files changed, 3 insertions(+), 17 deletions(-) diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h index 23bad40..dc3a756 100644 --- a/arch/powerpc/include/asm/kvm_host.h +++ b/arch/powerpc/include/asm/kvm_host.h @@ -81,8 +81,6 @@ struct kvm_vcpu_arch { struct tlbe shadow_tlb[PPC44x_TLB_SIZE]; /* Pages which are referenced in the shadow TLB. */ struct page *shadow_pages[PPC44x_TLB_SIZE]; - /* Copy of the host's TLB. */ - struct tlbe host_tlb[PPC44x_TLB_SIZE]; u32 host_stack; u32 host_pid; diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 92768d3..5940649 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -356,7 +356,6 @@ int main(void) DEFINE(VCPU_HOST_STACK, offsetof(struct kvm_vcpu, arch.host_stack)); DEFINE(VCPU_HOST_PID, offsetof(struct kvm_vcpu, arch.host_pid)); - DEFINE(VCPU_HOST_TLB, offsetof(struct kvm_vcpu, arch.host_tlb)); DEFINE(VCPU_SHADOW_TLB, offsetof(struct kvm_vcpu, arch.shadow_tlb)); DEFINE(VCPU_GPRS, offsetof(struct kvm_vcpu, arch.gpr)); DEFINE(VCPU_LR, offsetof(struct kvm_vcpu, arch.lr)); diff --git a/arch/powerpc/kvm/booke_interrupts.S b/arch/powerpc/kvm/booke_interrupts.S index 8eaba26..3e88dfa 100644 --- a/arch/powerpc/kvm/booke_interrupts.S +++ b/arch/powerpc/kvm/booke_interrupts.S @@ -342,26 +342,15 @@ lightweight_exit: andc r6, r5, r6 mtmsr r6 - /* Save the host's non-pinned TLB mappings, and load the guest mappings - * over them. Leave the host's "pinned" kernel mappings in place. */ - /* XXX optimization: use generation count to avoid swapping unmodified - * entries. */ + /* Load the guest mappings, leaving the host's "pinned" kernel mappings + * in place. */ + /* XXX optimization: load only modified guest entries. */ mfspr r10, SPRN_MMUCR /* Save host MMUCR. */ lis r8, [EMAIL PROTECTED] lwz r8, [EMAIL PROTECTED](r8) - addi r3, r4, VCPU_HOST_TLB - 4 addi r9, r4, VCPU_SHADOW_TLB - 4 li r6, 0 1: - /* Save host entry. */ - tlbre r7, r6, PPC44x_TLB_PAGEID - mfspr r5, SPRN_MMUCR - stwu r5, 4(r3) - stwu r7, 4(r3) - tlbre r7, r6, PPC44x_TLB_XLAT - stwu r7, 4(r3) - tlbre r7, r6, PPC44x_TLB_ATTRIB - stwu r7, 4(r3) /* Load guest entry. */ lwzu r7, 4(r9) mtspr SPRN_MMUCR, r7 -- 1.6.0.1 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html