KVM: VMX: Optimize %ds, %es reload

On x86_64, we can defer %ds and %es reload to the heavyweight context switch,
since nothing in the lightweight paths uses the host %ds or %es (they are
ignored by the processor).  Furthermore we can avoid the load if the segments
are null, by letting the hardware load the null segments for us.  This is the
expected case.

On i386, we could avoid the reload entirely, since the entry.S paths take care
of reload, except for the SYSEXIT path which leaves %ds and %es set to __USER_DS.
So we set them to the same values as well.

Saves about 70 cycles out of 1600 (around 4%; noisy measurements).
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

arch/x86/kvm/vmx.c

@@ -393,6 +393,9 @@ struct vcpu_vmx {
 	struct {
 		int           loaded;
 		u16           fs_sel, gs_sel, ldt_sel;
+#ifdef CONFIG_X86_64
+		u16           ds_sel, es_sel;
+#endif
 		int           gs_ldt_reload_needed;
 		int           fs_reload_needed;
 	} host_state;
@@ -1417,6 +1420,11 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu)
 		vmx->host_state.gs_ldt_reload_needed = 1;
 	}
 
+#ifdef CONFIG_X86_64
+	savesegment(ds, vmx->host_state.ds_sel);
+	savesegment(es, vmx->host_state.es_sel);
+#endif
+
 #ifdef CONFIG_X86_64
 	vmcs_writel(HOST_FS_BASE, read_msr(MSR_FS_BASE));
 	vmcs_writel(HOST_GS_BASE, read_msr(MSR_GS_BASE));
@@ -1457,6 +1465,19 @@ static void __vmx_load_host_state(struct vcpu_vmx *vmx)
 	}
 	if (vmx->host_state.fs_reload_needed)
 		loadsegment(fs, vmx->host_state.fs_sel);
+#ifdef CONFIG_X86_64
+	if (unlikely(vmx->host_state.ds_sel | vmx->host_state.es_sel)) {
+		loadsegment(ds, vmx->host_state.ds_sel);
+		loadsegment(es, vmx->host_state.es_sel);
+	}
+#else
+	/*
+	 * The sysexit path does not restore ds/es, so we must set them to
+	 * a reasonable value ourselves.
+	 */
+	loadsegment(ds, __USER_DS);
+	loadsegment(es, __USER_DS);
+#endif
 	reload_tss();
 #ifdef CONFIG_X86_64
 	wrmsrl(MSR_KERNEL_GS_BASE, vmx->msr_host_kernel_gs_base);
@@ -3640,8 +3661,18 @@ static void vmx_set_constant_host_state(void)
 	vmcs_writel(HOST_CR3, read_cr3());  /* 22.2.3  FIXME: shadow tables */
 
 	vmcs_write16(HOST_CS_SELECTOR, __KERNEL_CS);  /* 22.2.4 */
+#ifdef CONFIG_X86_64
+	/*
+	 * Load null selectors, so we can avoid reloading them in
+	 * __vmx_load_host_state(), in case userspace uses the null selectors
+	 * too (the expected case).
+	 */
+	vmcs_write16(HOST_DS_SELECTOR, 0);
+	vmcs_write16(HOST_ES_SELECTOR, 0);
+#else
 	vmcs_write16(HOST_DS_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
 	vmcs_write16(HOST_ES_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
+#endif
 	vmcs_write16(HOST_SS_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
 	vmcs_write16(HOST_TR_SELECTOR, GDT_ENTRY_TSS*8);  /* 22.2.4 */
 
@@ -6102,10 +6133,7 @@ static void atomic_switch_perf_msrs(struct vcpu_vmx *vmx)
 static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu)
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
-	u16 _ds, _es;
 
-	savesegment(ds, _ds);
-	savesegment(es, _es);
 	if (is_guest_mode(vcpu) && !vmx->nested.nested_run_pending) {
 		struct vmcs12 *vmcs12 = get_vmcs12(vcpu);
 		if (vmcs12->idt_vectoring_info_field &
@@ -6266,8 +6294,6 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu)
 		}
 	}
 
-	loadsegment(ds, _ds);
-	loadsegment(es, _es);
 	vmx->loaded_vmcs->launched = 1;
 
 	vmx->exit_reason = vmcs_read32(VM_EXIT_REASON);