[patch 1/5] KVM: add basic paravirt support (v2)

[Marcelo Tosatti <mtosatti@redhat.com>]

[-- Attachment #1: kvm-paravirt-core --]
[-- Type: text/plain, Size: 7131 bytes --]

Add basic KVM paravirt support. Avoid vm-exits on IO delays.

Add KVM_GET_PARA_FEATURES ioctl so paravirt features can be reported in a
single bitmask. This allows the host to disable features on runtime if 
appropriate, which would require one ioctl per feature otherwise.

The limit of 32 features can be extended to 64 if needed, beyond that a new 
MSR is required.

v1->v2:
- replace KVM_CAP_CLOCKSOURCE with KVM_CAP_PARA_FEATURES
- cover FEATURE_CLOCKSOURCE

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Anthony Liguori <aliguori@us.ibm.com>

Index: kvm.paravirt2/arch/x86/Kconfig
===================================================================
--- kvm.paravirt2.orig/arch/x86/Kconfig
+++ kvm.paravirt2/arch/x86/Kconfig
@@ -382,6 +382,14 @@ config KVM_CLOCK
 	  provides the guest with timing infrastructure such as time of day, and
 	  system time
 
+config KVM_GUEST
+ 	bool "KVM Guest support"
+ 	select PARAVIRT
+ 	depends on !(X86_VISWS || X86_VOYAGER)
+ 	help
+ 	 This option enables various optimizations for running under the KVM
+ 	 hypervisor.
+
 source "arch/x86/lguest/Kconfig"
 
 config PARAVIRT
Index: kvm.paravirt2/arch/x86/kernel/Makefile
===================================================================
--- kvm.paravirt2.orig/arch/x86/kernel/Makefile
+++ kvm.paravirt2/arch/x86/kernel/Makefile
@@ -69,6 +69,7 @@ obj-$(CONFIG_DEBUG_RODATA_TEST)	+= test_
 obj-$(CONFIG_DEBUG_NX_TEST)	+= test_nx.o
 
 obj-$(CONFIG_VMI)		+= vmi_32.o vmiclock_32.o
+obj-$(CONFIG_KVM_GUEST)		+= kvm.o
 obj-$(CONFIG_KVM_CLOCK)		+= kvmclock.o
 obj-$(CONFIG_PARAVIRT)		+= paravirt.o paravirt_patch_$(BITS).o
 
Index: kvm.paravirt2/arch/x86/kernel/kvm.c
===================================================================
--- /dev/null
+++ kvm.paravirt2/arch/x86/kernel/kvm.c
@@ -0,0 +1,52 @@
+/*
+ * KVM paravirt_ops implementation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ * Copyright (C) 2007, Red Hat, Inc., Ingo Molnar <mingo@redhat.com>
+ * Copyright IBM Corporation, 2007
+ *   Authors: Anthony Liguori <aliguori@us.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/kvm_para.h>
+#include <linux/cpu.h>
+#include <linux/mm.h>
+
+/*
+ * No need for any "IO delay" on KVM
+ */
+static void kvm_io_delay(void)
+{
+}
+
+static void paravirt_ops_setup(void)
+{
+	pv_info.name = "KVM";
+	pv_info.paravirt_enabled = 1;
+
+	if (kvm_para_has_feature(KVM_FEATURE_NOP_IO_DELAY))
+		pv_cpu_ops.io_delay = kvm_io_delay;
+
+}
+
+void __init kvm_guest_init(void)
+{
+	if (!kvm_para_available())
+		return;
+
+	paravirt_ops_setup();
+}
Index: kvm.paravirt2/arch/x86/kernel/setup_32.c
===================================================================
--- kvm.paravirt2.orig/arch/x86/kernel/setup_32.c
+++ kvm.paravirt2/arch/x86/kernel/setup_32.c
@@ -784,6 +784,7 @@ void __init setup_arch(char **cmdline_p)
 	 */
 	vmi_init();
 #endif
+	kvm_guest_init();
 
 	/*
 	 * NOTE: before this point _nobody_ is allowed to allocate
Index: kvm.paravirt2/arch/x86/kernel/setup_64.c
===================================================================
--- kvm.paravirt2.orig/arch/x86/kernel/setup_64.c
+++ kvm.paravirt2/arch/x86/kernel/setup_64.c
@@ -452,6 +452,8 @@ void __init setup_arch(char **cmdline_p)
 	init_apic_mappings();
 	ioapic_init_mappings();
 
+	kvm_guest_init();
+
 	/*
 	 * We trust e820 completely. No explicit ROM probing in memory.
 	 */
Index: kvm.paravirt2/arch/x86/kvm/x86.c
===================================================================
--- kvm.paravirt2.orig/arch/x86/kvm/x86.c
+++ kvm.paravirt2/arch/x86/kvm/x86.c
@@ -788,7 +788,7 @@ int kvm_dev_ioctl_check_extension(long e
 	case KVM_CAP_USER_MEMORY:
 	case KVM_CAP_SET_TSS_ADDR:
 	case KVM_CAP_EXT_CPUID:
-	case KVM_CAP_CLOCKSOURCE:
+	case KVM_CAP_PARA_FEATURES:
 		r = 1;
 		break;
 	case KVM_CAP_VAPIC:
@@ -854,6 +854,15 @@ long kvm_arch_dev_ioctl(struct file *fil
 		r = 0;
 		break;
 	}
+	case KVM_GET_PARA_FEATURES: {
+		__u32 para_features = KVM_PARA_FEATURES;
+
+		r = -EFAULT;
+		if (copy_to_user(argp, &para_features, sizeof para_features))
+			goto out;
+		r = 0;
+		break;
+	}
 	default:
 		r = -EINVAL;
 	}
Index: kvm.paravirt2/include/asm-x86/kvm_para.h
===================================================================
--- kvm.paravirt2.orig/include/asm-x86/kvm_para.h
+++ kvm.paravirt2/include/asm-x86/kvm_para.h
@@ -10,7 +10,8 @@
  * paravirtualization, the appropriate feature bit should be checked.
  */
 #define KVM_CPUID_FEATURES	0x40000001
-#define KVM_FEATURE_CLOCKSOURCE 0
+#define KVM_FEATURE_CLOCKSOURCE		0
+#define KVM_FEATURE_NOP_IO_DELAY	1
 
 #define MSR_KVM_WALL_CLOCK  0x11
 #define MSR_KVM_SYSTEM_TIME 0x12
@@ -39,6 +40,9 @@ struct kvm_wall_clock {
 extern void kvmclock_init(void);
 
 
+#define KVM_PARA_FEATURES ((1UL << KVM_FEATURE_NOP_IO_DELAY)	|	\
+			   (1UL << KVM_FEATURE_CLOCKSOURCE))
+
 /* This instruction is vmcall.  On non-VT architectures, it will generate a
  * trap that we will then rewrite to the appropriate instruction.
  */
Index: kvm.paravirt2/include/linux/kvm.h
===================================================================
--- kvm.paravirt2.orig/include/linux/kvm.h
+++ kvm.paravirt2/include/linux/kvm.h
@@ -222,6 +222,7 @@ struct kvm_vapic_addr {
  */
 #define KVM_GET_VCPU_MMAP_SIZE    _IO(KVMIO,   0x04) /* in bytes */
 #define KVM_GET_SUPPORTED_CPUID   _IOWR(KVMIO, 0x05, struct kvm_cpuid2)
+#define KVM_GET_PARA_FEATURES     _IOR(KVMIO,  0x05, __u32)
 
 /*
  * Extension capability list.
@@ -233,7 +234,7 @@ struct kvm_vapic_addr {
 #define KVM_CAP_SET_TSS_ADDR 4
 #define KVM_CAP_VAPIC 6
 #define KVM_CAP_EXT_CPUID 7
-#define KVM_CAP_CLOCKSOURCE 8
+#define KVM_CAP_PARA_FEATURES 8
 
 /*
  * ioctls for VM fds
Index: kvm.paravirt2/include/linux/kvm_para.h
===================================================================
--- kvm.paravirt2.orig/include/linux/kvm_para.h
+++ kvm.paravirt2/include/linux/kvm_para.h
@@ -20,6 +20,12 @@
 #include <asm/kvm_para.h>
 
 #ifdef __KERNEL__
+#ifdef CONFIG_KVM_GUEST
+void __init kvm_guest_init(void);
+#else
+#define kvm_guest_init() do { } while (0)
+#endif
+
 static inline int kvm_para_has_feature(unsigned int feature)
 {
 	if (kvm_arch_para_features() & (1UL << feature))

-- 


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/