kvm tools: Split x86 arch-specific bits into x86/

Create a new arch-specific subdirectory to contain architecture-specific code and includes.

The Makefile now adds various arch-specific objects based on detected architecture. That aside, this patch should only contain code moves. These include:

- x86-specific kvm_cpu setup, kernel loading, memory setup etc. now in x86/kvm{-cpu}.c
- BIOS now lives in x86/bios/
- ioport setup
- KVM extensions are asserted in arch-specific kvm.c now, so each architecture can manage its own dependencies.
- Various architecture-specific #defines are moved into $(ARCH)/include/kvm{-cpu}.h such as struct kvm_cpu, KVM_NR_CPUS, KVM_32BIT_GAP_SIZE.

Signed-off-by Matt Evans <[email protected]>

---
 tools/kvm/Makefile                              |   96 ++++---
 tools/kvm/builtin-run.c                         |    6 +-
 tools/kvm/include/kvm/ioport.h                  |    2 +-
 tools/kvm/include/kvm/kvm-cpu.h                 |   27 +--
 tools/kvm/include/kvm/kvm.h                     |   58 +---
 tools/kvm/ioport.c                              |   54 ----
 tools/kvm/kvm-cpu.c                             |  372 ----------------------
 tools/kvm/kvm.c                                 |  323 +-------------------
 tools/kvm/{ => x86}/bios.c                      |    0
 tools/kvm/{ => x86}/bios/.gitignore             |    0
 tools/kvm/{ => x86}/bios/bios-rom.S             |    2 +-
 tools/kvm/{ => x86}/bios/e820.c                 |    0
 tools/kvm/{ => x86}/bios/entry.S                |    0
 tools/kvm/{ => x86}/bios/gen-offsets.sh         |    0
 tools/kvm/{ => x86}/bios/int10.c                |    0
 tools/kvm/{ => x86}/bios/int15.c                |    0
 tools/kvm/{ => x86}/bios/local.S                |    0
 tools/kvm/{ => x86}/bios/macro.S                |    0
 tools/kvm/{ => x86}/bios/memcpy.c               |    0
 tools/kvm/{ => x86}/bios/rom.ld.S               |    0
 tools/kvm/{ => x86}/cpuid.c                     |    0
 tools/kvm/{ => x86}/include/kvm/assembly.h      |    0
 tools/kvm/{ => x86}/include/kvm/barrier.h       |    0
 tools/kvm/{ => x86}/include/kvm/bios-export.h   |    0
 tools/kvm/{ => x86}/include/kvm/bios.h          |    0
 tools/kvm/{ => x86}/include/kvm/boot-protocol.h |    0
 tools/kvm/{ => x86}/include/kvm/cpufeature.h    |    0
 tools/kvm/{ => x86}/include/kvm/interrupt.h     |    0
 tools/kvm/x86/include/kvm/kvm-arch.h            |   59 ++++
 tools/kvm/x86/include/kvm/kvm-cpu-arch.h        |   33 ++
 tools/kvm/{ => x86}/include/kvm/mptable.h       |    0
 tools/kvm/{ => x86}/interrupt.c                 |    0
 tools/kvm/x86/ioport.c                          |   59 ++++
 tools/kvm/{ => x86}/irq.c                       |    0
 tools/kvm/x86/kvm-cpu.c                         |  383 +++++++++++++++++++++++
 tools/kvm/x86/kvm.c                             |  330 +++++++++++++++++++
 tools/kvm/{ => x86}/mptable.c                   |    0
 37 files changed, 951 insertions(+), 853 deletions(-)
 rename tools/kvm/{ => x86}/bios.c (100%)
 rename tools/kvm/{ => x86}/bios/.gitignore (100%)
 rename tools/kvm/{ => x86}/bios/bios-rom.S (80%)
 rename tools/kvm/{ => x86}/bios/e820.c (100%)
 rename tools/kvm/{ => x86}/bios/entry.S (100%)
 rename tools/kvm/{ => x86}/bios/gen-offsets.sh (100%)
 rename tools/kvm/{ => x86}/bios/int10.c (100%)
 rename tools/kvm/{ => x86}/bios/int15.c (100%)
 rename tools/kvm/{ => x86}/bios/local.S (100%)
 rename tools/kvm/{ => x86}/bios/macro.S (100%)
 rename tools/kvm/{ => x86}/bios/memcpy.c (100%)
 rename tools/kvm/{ => x86}/bios/rom.ld.S (100%)
 rename tools/kvm/{ => x86}/cpuid.c (100%)
 rename tools/kvm/{ => x86}/include/kvm/assembly.h (100%)
 rename tools/kvm/{ => x86}/include/kvm/barrier.h (100%)
 rename tools/kvm/{ => x86}/include/kvm/bios-export.h (100%)
 rename tools/kvm/{ => x86}/include/kvm/bios.h (100%)
 rename tools/kvm/{ => x86}/include/kvm/boot-protocol.h (100%)
 rename tools/kvm/{ => x86}/include/kvm/cpufeature.h (100%)
 rename tools/kvm/{ => x86}/include/kvm/interrupt.h (100%)
 create mode 100644 tools/kvm/x86/include/kvm/kvm-arch.h
 create mode 100644 tools/kvm/x86/include/kvm/kvm-cpu-arch.h
 rename tools/kvm/{ => x86}/include/kvm/mptable.h (100%)
 rename tools/kvm/{ => x86}/interrupt.c (100%)
 create mode 100644 tools/kvm/x86/ioport.c
 rename tools/kvm/{ => x86}/irq.c (100%)
 create mode 100644 tools/kvm/x86/kvm-cpu.c
 create mode 100644 tools/kvm/x86/kvm.c
 rename tools/kvm/{ => x86}/mptable.c (100%)

tools/kvm/Makefile +60 -36

index bb5f6b0..243886e 100644
@@ -33,13 +33,11 @@ OBJS	+= builtin-run.o
 OBJS	+= builtin-setup.o
 OBJS	+= builtin-stop.o
 OBJS	+= builtin-version.o
-OBJS	+= cpuid.o
 OBJS	+= disk/core.o
 OBJS	+= framebuffer.o
 OBJS	+= guest_compat.o
 OBJS	+= hw/rtc.o
 OBJS	+= hw/serial.o
-OBJS	+= interrupt.o
 OBJS	+= ioport.o
 OBJS	+= kvm-cpu.o
 OBJS	+= kvm.o

@@ -61,7 +59,6 @@ OBJS	+= disk/blk.o
 OBJS	+= disk/qcow.o
 OBJS	+= disk/raw.o
 OBJS	+= ioeventfd.o
-OBJS	+= irq.o
 OBJS	+= net/uip/core.o
 OBJS	+= net/uip/arp.o
 OBJS	+= net/uip/icmp.o

@@ -72,7 +69,6 @@ OBJS	+= net/uip/buf.o
 OBJS	+= net/uip/csum.o
 OBJS	+= net/uip/dhcp.o
 OBJS	+= kvm-cmd.o
-OBJS	+= mptable.o
 OBJS	+= rbtree.o
 OBJS	+= threadpool.o
 OBJS	+= util/parse-options.o

@@ -123,12 +119,6 @@ ifeq ($(has_AIO),y)
 	LIBS	+= -laio
 endif
 
-DEPS	:= $(patsubst %.o,%.d,$(OBJS))
-
-# Exclude BIOS object files from header dependencies.
-OBJS	+= bios.o
-OBJS	+= bios/bios-rom.o
-
 LIBS	+= -lrt
 LIBS	+= -lpthread
 LIBS	+= -lutil

@@ -150,12 +140,43 @@ ifeq ($(uname_M),x86_64)
 	DEFINES      += -DCONFIG_X86_64
 endif
 
+
+### Arch-specific stuff
+
+#x86
+ifeq ($(ARCH),x86)
+	DEFINES += -DCONFIG_X86
+	OBJS	+= x86/cpuid.o
+	OBJS	+= x86/interrupt.o
+	OBJS	+= x86/ioport.o
+	OBJS	+= x86/irq.o
+	OBJS	+= x86/kvm.o
+	OBJS	+= x86/kvm-cpu.o
+	OBJS	+= x86/mptable.o
+# Exclude BIOS object files from header dependencies.
+	OTHEROBJS	+= x86/bios.o
+	OTHEROBJS	+= x86/bios/bios-rom.o
+	ARCH_INCLUDE := x86/include
+endif
+
+###
+
+ifeq (,$(ARCH_INCLUDE))
+	UNSUPP_ERR = @echo "This architecture is not supported in kvmtool." && exit 1
+else
+	UNSUPP_ERR =
+endif
+
+DEPS	:= $(patsubst %.o,%.d,$(OBJS))
+OBJS	+= $(OTHEROBJS)
+
 DEFINES	+= -D_FILE_OFFSET_BITS=64
 DEFINES	+= -D_GNU_SOURCE
 DEFINES	+= -DKVMTOOLS_VERSION='"$(KVMTOOLS_VERSION)"'
+DEFINES	+= -DBUILD_ARCH='"$(ARCH)"'
 
 KVM_INCLUDE := include
-CFLAGS	+= $(CPPFLAGS) $(DEFINES) -I$(KVM_INCLUDE) -I../../include -I../../arch/$(ARCH)/include/ -Os -g
+CFLAGS	+= $(CPPFLAGS) $(DEFINES) -I$(KVM_INCLUDE) -I$(ARCH_INCLUDE) -I../../include -I../../arch/$(ARCH)/include/ -Os -g
 
 ifneq ($(WERROR),0)
 	WARNINGS += -Werror

@@ -179,7 +200,10 @@ WARNINGS += -Wwrite-strings
 
 CFLAGS	+= $(WARNINGS)
 
-all: $(PROGRAM) $(GUEST_INIT)
+all: arch_support_check $(PROGRAM) $(GUEST_INIT)
+
+arch_support_check:
+	$(UNSUPP_ERR)
 
 KVMTOOLS-VERSION-FILE:
 	@$(SHELL_PATH) util/KVMTOOLS-VERSION-GEN $(OUTPUT)

@@ -227,33 +251,33 @@ BIOS_CFLAGS += -mregparm=3
 BIOS_CFLAGS += -fno-stack-protector
 BIOS_CFLAGS += -I../../arch/$(ARCH)
 
-bios.o: bios/bios.bin bios/bios-rom.h
-
-bios/bios.bin.elf: bios/entry.S bios/e820.c bios/int10.c bios/int15.c bios/rom.ld.S
-	$(E) "  CC       bios/memcpy.o"
-	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s bios/memcpy.c -o bios/memcpy.o
-	$(E) "  CC       bios/e820.o"
-	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s bios/e820.c -o bios/e820.o
-	$(E) "  CC       bios/int10.o"
-	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s bios/int10.c -o bios/int10.o
-	$(E) "  CC       bios/int15.o"
-	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s bios/int15.c -o bios/int15.o
-	$(E) "  CC       bios/entry.o"
-	$(Q) $(CC) $(CFLAGS) $(BIOS_CFLAGS) -c -s bios/entry.S -o bios/entry.o
+x86/bios.o: x86/bios/bios.bin x86/bios/bios-rom.h
+
+x86/bios/bios.bin.elf: x86/bios/entry.S x86/bios/e820.c x86/bios/int10.c x86/bios/int15.c x86/bios/rom.ld.S
+	$(E) "  CC       x86/bios/memcpy.o"
+	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s x86/bios/memcpy.c -o x86/bios/memcpy.o
+	$(E) "  CC       x86/bios/e820.o"
+	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s x86/bios/e820.c -o x86/bios/e820.o
+	$(E) "  CC       x86/bios/int10.o"
+	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s x86/bios/int10.c -o x86/bios/int10.o
+	$(E) "  CC       x86/bios/int15.o"
+	$(Q) $(CC) -include code16gcc.h $(CFLAGS) $(BIOS_CFLAGS) -c -s x86/bios/int15.c -o x86/bios/int15.o
+	$(E) "  CC       x86/bios/entry.o"
+	$(Q) $(CC) $(CFLAGS) $(BIOS_CFLAGS) -c -s x86/bios/entry.S -o x86/bios/entry.o
 	$(E) "  LD      " $@
-	$(Q) ld -T bios/rom.ld.S -o bios/bios.bin.elf bios/memcpy.o bios/entry.o bios/e820.o bios/int10.o bios/int15.o
+	$(Q) ld -T x86/bios/rom.ld.S -o x86/bios/bios.bin.elf x86/bios/memcpy.o x86/bios/entry.o x86/bios/e820.o x86/bios/int10.o x86/bios/int15.o
 
-bios/bios.bin: bios/bios.bin.elf
+x86/bios/bios.bin: x86/bios/bios.bin.elf
 	$(E) "  OBJCOPY " $@
-	$(Q) objcopy -O binary -j .text bios/bios.bin.elf bios/bios.bin
+	$(Q) objcopy -O binary -j .text x86/bios/bios.bin.elf x86/bios/bios.bin
 
-bios/bios-rom.o: bios/bios-rom.S bios/bios.bin bios/bios-rom.h
+x86/bios/bios-rom.o: x86/bios/bios-rom.S x86/bios/bios.bin x86/bios/bios-rom.h
 	$(E) "  CC      " $@
-	$(Q) $(CC) -c $(CFLAGS) bios/bios-rom.S -o bios/bios-rom.o
+	$(Q) $(CC) -c $(CFLAGS) x86/bios/bios-rom.S -o x86/bios/bios-rom.o
 
-bios/bios-rom.h: bios/bios.bin.elf
+x86/bios/bios-rom.h: x86/bios/bios.bin.elf
 	$(E) "  NM      " $@
-	$(Q) cd bios && sh gen-offsets.sh > bios-rom.h && cd ..
+	$(Q) cd x86/bios && sh gen-offsets.sh > bios-rom.h && cd ..
 
 check: $(PROGRAM)
 	$(MAKE) -C tests

@@ -263,10 +287,10 @@ check: $(PROGRAM)
 
 clean:
 	$(E) "  CLEAN"
-	$(Q) rm -f bios/*.bin
-	$(Q) rm -f bios/*.elf
-	$(Q) rm -f bios/*.o
-	$(Q) rm -f bios/bios-rom.h
+	$(Q) rm -f x86/bios/*.bin
+	$(Q) rm -f x86/bios/*.elf
+	$(Q) rm -f x86/bios/*.o
+	$(Q) rm -f x86/bios/bios-rom.h
 	$(Q) rm -f tests/boot/boot_test.iso
 	$(Q) rm -rf tests/boot/rootfs/
 	$(Q) rm -f $(DEPS) $(OBJS) $(PROGRAM) $(GUEST_INIT)

tools/kvm/builtin-run.c +3 -3

index 33de4f6..9148d83 100644
@@ -568,7 +568,7 @@ static const char *host_kernels[] = {
 
 static const char *default_kernels[] = {
 	"./bzImage",
-	"../../arch/x86/boot/bzImage",
+	"../../arch/" BUILD_ARCH "/boot/bzImage",
 	NULL
 };
 

@@ -886,7 +886,7 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 
 	kvm->vmlinux		= vmlinux_filename;
 
-	ioport__setup_legacy();
+	ioport__setup_arch();
 
 	rtc__init();
 

@@ -931,7 +931,7 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 
 	kvm__start_timer(kvm);
 
-	kvm__setup_bios(kvm);
+	kvm__arch_setup_firmware(kvm);
 
 	for (i = 0; i < nrcpus; i++) {
 		kvm_cpus[i] = kvm_cpu__init(kvm, i);

tools/kvm/include/kvm/ioport.h +1 -1

index 5b857dd..61a70ec 100644
@@ -28,7 +28,7 @@ struct ioport_operations {
 	bool (*io_out)(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size);
 };
 
-void ioport__setup_legacy(void);
+void ioport__setup_arch(void);
 
 u16 ioport__register(u16 port, struct ioport_operations *ops, int count, void *param);
 

tools/kvm/include/kvm/kvm-cpu.h +1 -26

index 01540ac..719e286 100644
@@ -1,32 +1,7 @@
 #ifndef KVM__KVM_CPU_H
 #define KVM__KVM_CPU_H
 
-#include <linux/kvm.h>	/* for struct kvm_regs */
-
-#include <pthread.h>
-
-struct kvm;
-
-struct kvm_cpu {
-	pthread_t		thread;		/* VCPU thread */
-
-	unsigned long		cpu_id;
-
-	struct kvm		*kvm;		/* parent KVM */
-	int			vcpu_fd;	/* For VCPU ioctls() */
-	struct kvm_run		*kvm_run;
-
-	struct kvm_regs		regs;
-	struct kvm_sregs	sregs;
-	struct kvm_fpu		fpu;
-
-	struct kvm_msrs		*msrs;		/* dynamically allocated */
-
-	u8			is_running;
-	u8			paused;
-
-	struct kvm_coalesced_mmio_ring	*ring;
-};
+#include "kvm/kvm-cpu-arch.h"
 
 struct kvm_cpu *kvm_cpu__init(struct kvm *kvm, unsigned long cpu_id);
 void kvm_cpu__delete(struct kvm_cpu *vcpu);

tools/kvm/include/kvm/kvm.h +15 -43

index 2b3024a..ca1acc0 100644
@@ -1,22 +1,13 @@
 #ifndef KVM__KVM_H
 #define KVM__KVM_H
 
-#include "kvm/interrupt.h"
-#include "kvm/segment.h"
+#include "kvm/kvm-arch.h"
 
 #include <stdbool.h>
 #include <linux/types.h>
 #include <time.h>
 #include <signal.h>
 
-#define KVM_NR_CPUS		(255)
-
-/*
- * The hole includes VESA framebuffer and PCI memory.
- */
-#define KVM_32BIT_GAP_SIZE	(768 << 20)
-#define KVM_32BIT_GAP_START	((1ULL << 32) - KVM_32BIT_GAP_SIZE)
-
 #define SIGKVMEXIT		(SIGRTMIN + 0)
 #define SIGKVMPAUSE		(SIGRTMIN + 1)
 #define SIGKVMSTOP		(SIGRTMIN + 4)

@@ -25,33 +16,15 @@
 #define KVM_PID_FILE_PATH	"/.kvm-tools/"
 #define HOME_DIR		getenv("HOME")
 
-struct kvm {
-	int			sys_fd;		/* For system ioctls(), i.e. /dev/kvm */
-	int			vm_fd;		/* For VM ioctls() */
-	timer_t			timerid;	/* Posix timer for interrupts */
-
-	int			nrcpus;		/* Number of cpus to run */
-
-	u32			mem_slots;	/* for KVM_SET_USER_MEMORY_REGION */
-
-	u64			ram_size;
-	void			*ram_start;
-
-	bool			nmi_disabled;
-
-	bool			single_step;
+#define PAGE_SIZE (sysconf(_SC_PAGE_SIZE))
 
-	u16			boot_selector;
-	u16			boot_ip;
-	u16			boot_sp;
+#define DEFINE_KVM_EXT(ext)		\
+	.name = #ext,			\
+	.code = ext
 
-	struct interrupt_table	interrupt_table;
-
-	const char		*vmlinux;
-	struct disk_image       **disks;
-	int                     nr_disks;
-
-	const char		*name;
+struct kvm_ext {
+	const char *name;
+	int code;
 };
 
 void kvm__set_dir(const char *fmt, ...);

@@ -64,7 +37,6 @@ void kvm__init_ram(struct kvm *kvm);
 void kvm__delete(struct kvm *kvm);
 bool kvm__load_kernel(struct kvm *kvm, const char *kernel_filename,
 			const char *initrd_filename, const char *kernel_cmdline, u16 vidmode);
-void kvm__setup_bios(struct kvm *kvm);
 void kvm__start_timer(struct kvm *kvm);
 void kvm__stop_timer(struct kvm *kvm);
 void kvm__irq_line(struct kvm *kvm, int irq, int level);

@@ -81,6 +53,13 @@ int kvm__get_sock_by_instance(const char *name);
 int kvm__enumerate_instances(int (*callback)(const char *name, int pid));
 void kvm__remove_socket(const char *name);
 
+void kvm__arch_init(struct kvm *kvm, const char *kvm_dev, u64 ram_size, const char *name);
+void kvm__arch_setup_firmware(struct kvm *kvm);
+bool kvm__arch_cpu_supports_vm(void);
+
+int load_flat_binary(struct kvm *kvm, int fd);
+bool load_bzimage(struct kvm *kvm, int fd_kernel, int fd_initrd, const char *kernel_cmdline, u16 vidmode);
+
 /*
  * Debugging
  */

@@ -98,11 +77,4 @@ static inline void *guest_flat_to_host(struct kvm *kvm, unsigned long offset)
 	return kvm->ram_start + offset;
 }
 
-static inline void *guest_real_to_host(struct kvm *kvm, u16 selector, u16 offset)
-{
-	unsigned long flat = segment_to_flat(selector, offset);
-
-	return guest_flat_to_host(kvm, flat);
-}
-
 #endif /* KVM__KVM_H */

tools/kvm/ioport.c +0 -54

index 7cbc44e..965cfc2 100644
@@ -52,34 +52,6 @@ static int ioport_insert(struct rb_root *root, struct ioport *data)
 	return rb_int_insert(root, &data->node);
 }
 
-static bool debug_io_out(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size)
-{
-	exit(EXIT_SUCCESS);
-}
-
-static struct ioport_operations debug_ops = {
-	.io_out		= debug_io_out,
-};
-
-static bool dummy_io_in(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size)
-{
-	return true;
-}
-
-static bool dummy_io_out(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size)
-{
-	return true;
-}
-
-static struct ioport_operations dummy_read_write_ioport_ops = {
-	.io_in		= dummy_io_in,
-	.io_out		= dummy_io_out,
-};
-
-static struct ioport_operations dummy_write_only_ioport_ops = {
-	.io_out		= dummy_io_out,
-};
-
 u16 ioport__register(u16 port, struct ioport_operations *ops, int count, void *param)
 {
 	struct ioport *entry;

@@ -164,29 +136,3 @@ error:
 
 	return !ioport_debug;
 }
-
-void ioport__setup_legacy(void)
-{
-	/* 0x0020 - 0x003F - 8259A PIC 1 */
-	ioport__register(0x0020, &dummy_read_write_ioport_ops, 2, NULL);
-
-	/* PORT 0040-005F - PIT - PROGRAMMABLE INTERVAL TIMER (8253, 8254) */
-	ioport__register(0x0040, &dummy_read_write_ioport_ops, 4, NULL);
-
-	/* 0x00A0 - 0x00AF - 8259A PIC 2 */
-	ioport__register(0x00A0, &dummy_read_write_ioport_ops, 2, NULL);
-
-	/* PORT 00E0-00EF are 'motherboard specific' so we use them for our
-	   internal debugging purposes.  */
-	ioport__register(IOPORT_DBG, &debug_ops, 1, NULL);
-
-	/* PORT 00ED - DUMMY PORT FOR DELAY??? */
-	ioport__register(0x00ED, &dummy_write_only_ioport_ops, 1, NULL);
-
-	/* 0x00F0 - 0x00FF - Math co-processor */
-	ioport__register(0x00F0, &dummy_write_only_ioport_ops, 2, NULL);
-
-	/* PORT 03D4-03D5 - COLOR VIDEO - CRT CONTROL REGISTERS */
-	ioport__register(0x03D4, &dummy_read_write_ioport_ops, 1, NULL);
-	ioport__register(0x03D5, &dummy_write_only_ioport_ops, 1, NULL);
-}

tools/kvm/kvm-cpu.c +0 -372

index 0ad6f3b..5aba3bb 100644
@@ -4,8 +4,6 @@
 #include "kvm/util.h"
 #include "kvm/kvm.h"
 
-#include <asm/msr-index.h>
-
 #include <sys/ioctl.h>
 #include <sys/mman.h>
 #include <signal.h>

@@ -14,106 +12,9 @@
 #include <errno.h>
 #include <stdio.h>
 
-#define PAGE_SIZE (sysconf(_SC_PAGE_SIZE))
-
 extern struct kvm_cpu *kvm_cpus[KVM_NR_CPUS];
 extern __thread struct kvm_cpu *current_kvm_cpu;
 
-static int debug_fd;
-
-void kvm_cpu__set_debug_fd(int fd)
-{
-	debug_fd = fd;
-}
-
-int kvm_cpu__get_debug_fd(void)
-{
-	return debug_fd;
-}
-
-static inline bool is_in_protected_mode(struct kvm_cpu *vcpu)
-{
-	return vcpu->sregs.cr0 & 0x01;
-}
-
-static inline u64 ip_to_flat(struct kvm_cpu *vcpu, u64 ip)
-{
-	u64 cs;
-
-	/*
-	 * NOTE! We should take code segment base address into account here.
-	 * Luckily it's usually zero because Linux uses flat memory model.
-	 */
-	if (is_in_protected_mode(vcpu))
-		return ip;
-
-	cs = vcpu->sregs.cs.selector;
-
-	return ip + (cs << 4);
-}
-
-static inline u32 selector_to_base(u16 selector)
-{
-	/*
-	 * KVM on Intel requires 'base' to be 'selector * 16' in real mode.
-	 */
-	return (u32)selector * 16;
-}
-
-static struct kvm_cpu *kvm_cpu__new(struct kvm *kvm)
-{
-	struct kvm_cpu *vcpu;
-
-	vcpu		= calloc(1, sizeof *vcpu);
-	if (!vcpu)
-		return NULL;
-
-	vcpu->kvm	= kvm;
-
-	return vcpu;
-}
-
-void kvm_cpu__delete(struct kvm_cpu *vcpu)
-{
-	if (vcpu->msrs)
-		free(vcpu->msrs);
-
-	free(vcpu);
-}
-
-struct kvm_cpu *kvm_cpu__init(struct kvm *kvm, unsigned long cpu_id)
-{
-	struct kvm_cpu *vcpu;
-	int mmap_size;
-	int coalesced_offset;
-
-	vcpu		= kvm_cpu__new(kvm);
-	if (!vcpu)
-		return NULL;
-
-	vcpu->cpu_id	= cpu_id;
-
-	vcpu->vcpu_fd = ioctl(vcpu->kvm->vm_fd, KVM_CREATE_VCPU, cpu_id);
-	if (vcpu->vcpu_fd < 0)
-		die_perror("KVM_CREATE_VCPU ioctl");
-
-	mmap_size = ioctl(vcpu->kvm->sys_fd, KVM_GET_VCPU_MMAP_SIZE, 0);
-	if (mmap_size < 0)
-		die_perror("KVM_GET_VCPU_MMAP_SIZE ioctl");
-
-	vcpu->kvm_run = mmap(NULL, mmap_size, PROT_RW, MAP_SHARED, vcpu->vcpu_fd, 0);
-	if (vcpu->kvm_run == MAP_FAILED)
-		die("unable to mmap vcpu fd");
-
-	coalesced_offset = ioctl(kvm->sys_fd, KVM_CHECK_EXTENSION, KVM_CAP_COALESCED_MMIO);
-	if (coalesced_offset)
-		vcpu->ring = (void *)vcpu->kvm_run + (coalesced_offset * PAGE_SIZE);
-
-	vcpu->is_running = true;
-
-	return vcpu;
-}
-
 void kvm_cpu__enable_singlestep(struct kvm_cpu *vcpu)
 {
 	struct kvm_guest_debug debug = {

@@ -124,278 +25,6 @@ void kvm_cpu__enable_singlestep(struct kvm_cpu *vcpu)
 		pr_warning("KVM_SET_GUEST_DEBUG failed");
 }
 
-static struct kvm_msrs *kvm_msrs__new(size_t nmsrs)
-{
-	struct kvm_msrs *vcpu = calloc(1, sizeof(*vcpu) + (sizeof(struct kvm_msr_entry) * nmsrs));
-
-	if (!vcpu)
-		die("out of memory");
-
-	return vcpu;
-}
-
-#define KVM_MSR_ENTRY(_index, _data)	\
-	(struct kvm_msr_entry) { .index = _index, .data = _data }
-
-static void kvm_cpu__setup_msrs(struct kvm_cpu *vcpu)
-{
-	unsigned long ndx = 0;
-
-	vcpu->msrs = kvm_msrs__new(100);
-
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_SYSENTER_CS,	0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_SYSENTER_ESP,	0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_SYSENTER_EIP,	0x0);
-#ifdef CONFIG_X86_64
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_STAR,			0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_CSTAR,			0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_KERNEL_GS_BASE,		0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_SYSCALL_MASK,		0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_LSTAR,			0x0);
-#endif
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_TSC,		0x0);
-	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_MISC_ENABLE,
-						MSR_IA32_MISC_ENABLE_FAST_STRING);
-
-	vcpu->msrs->nmsrs	= ndx;
-
-	if (ioctl(vcpu->vcpu_fd, KVM_SET_MSRS, vcpu->msrs) < 0)
-		die_perror("KVM_SET_MSRS failed");
-}
-
-static void kvm_cpu__setup_fpu(struct kvm_cpu *vcpu)
-{
-	vcpu->fpu = (struct kvm_fpu) {
-		.fcw		= 0x37f,
-		.mxcsr		= 0x1f80,
-	};
-
-	if (ioctl(vcpu->vcpu_fd, KVM_SET_FPU, &vcpu->fpu) < 0)
-		die_perror("KVM_SET_FPU failed");
-}
-
-static void kvm_cpu__setup_regs(struct kvm_cpu *vcpu)
-{
-	vcpu->regs = (struct kvm_regs) {
-		/* We start the guest in 16-bit real mode  */
-		.rflags		= 0x0000000000000002ULL,
-
-		.rip		= vcpu->kvm->boot_ip,
-		.rsp		= vcpu->kvm->boot_sp,
-		.rbp		= vcpu->kvm->boot_sp,
-	};
-
-	if (vcpu->regs.rip > USHRT_MAX)
-		die("ip 0x%llx is too high for real mode", (u64) vcpu->regs.rip);
-
-	if (ioctl(vcpu->vcpu_fd, KVM_SET_REGS, &vcpu->regs) < 0)
-		die_perror("KVM_SET_REGS failed");
-}
-
-static void kvm_cpu__setup_sregs(struct kvm_cpu *vcpu)
-{
-
-	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &vcpu->sregs) < 0)
-		die_perror("KVM_GET_SREGS failed");
-
-	vcpu->sregs.cs.selector	= vcpu->kvm->boot_selector;
-	vcpu->sregs.cs.base	= selector_to_base(vcpu->kvm->boot_selector);
-	vcpu->sregs.ss.selector	= vcpu->kvm->boot_selector;
-	vcpu->sregs.ss.base	= selector_to_base(vcpu->kvm->boot_selector);
-	vcpu->sregs.ds.selector	= vcpu->kvm->boot_selector;
-	vcpu->sregs.ds.base	= selector_to_base(vcpu->kvm->boot_selector);
-	vcpu->sregs.es.selector	= vcpu->kvm->boot_selector;
-	vcpu->sregs.es.base	= selector_to_base(vcpu->kvm->boot_selector);
-	vcpu->sregs.fs.selector	= vcpu->kvm->boot_selector;
-	vcpu->sregs.fs.base	= selector_to_base(vcpu->kvm->boot_selector);
-	vcpu->sregs.gs.selector	= vcpu->kvm->boot_selector;
-	vcpu->sregs.gs.base	= selector_to_base(vcpu->kvm->boot_selector);
-
-	if (ioctl(vcpu->vcpu_fd, KVM_SET_SREGS, &vcpu->sregs) < 0)
-		die_perror("KVM_SET_SREGS failed");
-}
-
-/**
- * kvm_cpu__reset_vcpu - reset virtual CPU to a known state
- */
-void kvm_cpu__reset_vcpu(struct kvm_cpu *vcpu)
-{
-	kvm_cpu__setup_sregs(vcpu);
-	kvm_cpu__setup_regs(vcpu);
-	kvm_cpu__setup_fpu(vcpu);
-	kvm_cpu__setup_msrs(vcpu);
-}
-
-static void print_dtable(const char *name, struct kvm_dtable *dtable)
-{
-	dprintf(debug_fd, " %s                 %016llx  %08hx\n",
-		name, (u64) dtable->base, (u16) dtable->limit);
-}
-
-static void print_segment(const char *name, struct kvm_segment *seg)
-{
-	dprintf(debug_fd, " %s       %04hx      %016llx  %08x  %02hhx    %x %x   %x  %x %x %x %x\n",
-		name, (u16) seg->selector, (u64) seg->base, (u32) seg->limit,
-		(u8) seg->type, seg->present, seg->dpl, seg->db, seg->s, seg->l, seg->g, seg->avl);
-}
-
-void kvm_cpu__show_registers(struct kvm_cpu *vcpu)
-{
-	unsigned long cr0, cr2, cr3;
-	unsigned long cr4, cr8;
-	unsigned long rax, rbx, rcx;
-	unsigned long rdx, rsi, rdi;
-	unsigned long rbp,  r8,  r9;
-	unsigned long r10, r11, r12;
-	unsigned long r13, r14, r15;
-	unsigned long rip, rsp;
-	struct kvm_sregs sregs;
-	unsigned long rflags;
-	struct kvm_regs regs;
-	int i;
-
-	if (ioctl(vcpu->vcpu_fd, KVM_GET_REGS, &regs) < 0)
-		die("KVM_GET_REGS failed");
-
-	rflags = regs.rflags;
-
-	rip = regs.rip; rsp = regs.rsp;
-	rax = regs.rax; rbx = regs.rbx; rcx = regs.rcx;
-	rdx = regs.rdx; rsi = regs.rsi; rdi = regs.rdi;
-	rbp = regs.rbp; r8  = regs.r8;  r9  = regs.r9;
-	r10 = regs.r10; r11 = regs.r11; r12 = regs.r12;
-	r13 = regs.r13; r14 = regs.r14; r15 = regs.r15;
-
-	dprintf(debug_fd, "\n Registers:\n");
-	dprintf(debug_fd,   " ----------\n");
-	dprintf(debug_fd, " rip: %016lx   rsp: %016lx flags: %016lx\n", rip, rsp, rflags);
-	dprintf(debug_fd, " rax: %016lx   rbx: %016lx   rcx: %016lx\n", rax, rbx, rcx);
-	dprintf(debug_fd, " rdx: %016lx   rsi: %016lx   rdi: %016lx\n", rdx, rsi, rdi);
-	dprintf(debug_fd, " rbp: %016lx    r8: %016lx    r9: %016lx\n", rbp, r8,  r9);
-	dprintf(debug_fd, " r10: %016lx   r11: %016lx   r12: %016lx\n", r10, r11, r12);
-	dprintf(debug_fd, " r13: %016lx   r14: %016lx   r15: %016lx\n", r13, r14, r15);
-
-	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &sregs) < 0)
-		die("KVM_GET_REGS failed");
-
-	cr0 = sregs.cr0; cr2 = sregs.cr2; cr3 = sregs.cr3;
-	cr4 = sregs.cr4; cr8 = sregs.cr8;
-
-	dprintf(debug_fd, " cr0: %016lx   cr2: %016lx   cr3: %016lx\n", cr0, cr2, cr3);
-	dprintf(debug_fd, " cr4: %016lx   cr8: %016lx\n", cr4, cr8);
-	dprintf(debug_fd, "\n Segment registers:\n");
-	dprintf(debug_fd,   " ------------------\n");
-	dprintf(debug_fd, " register  selector  base              limit     type  p dpl db s l g avl\n");
-	print_segment("cs ", &sregs.cs);
-	print_segment("ss ", &sregs.ss);
-	print_segment("ds ", &sregs.ds);
-	print_segment("es ", &sregs.es);
-	print_segment("fs ", &sregs.fs);
-	print_segment("gs ", &sregs.gs);
-	print_segment("tr ", &sregs.tr);
-	print_segment("ldt", &sregs.ldt);
-	print_dtable("gdt", &sregs.gdt);
-	print_dtable("idt", &sregs.idt);
-
-	dprintf(debug_fd, "\n APIC:\n");
-	dprintf(debug_fd,   " -----\n");
-	dprintf(debug_fd, " efer: %016llx  apic base: %016llx  nmi: %s\n",
-		(u64) sregs.efer, (u64) sregs.apic_base,
-		(vcpu->kvm->nmi_disabled ? "disabled" : "enabled"));
-
-	dprintf(debug_fd, "\n Interrupt bitmap:\n");
-	dprintf(debug_fd,   " -----------------\n");
-	for (i = 0; i < (KVM_NR_INTERRUPTS + 63) / 64; i++)
-		dprintf(debug_fd, " %016llx", (u64) sregs.interrupt_bitmap[i]);
-	dprintf(debug_fd, "\n");
-}
-
-#define MAX_SYM_LEN		128
-
-void kvm_cpu__show_code(struct kvm_cpu *vcpu)
-{
-	unsigned int code_bytes = 64;
-	unsigned int code_prologue = code_bytes * 43 / 64;
-	unsigned int code_len = code_bytes;
-	char sym[MAX_SYM_LEN];
-	unsigned char c;
-	unsigned int i;
-	u8 *ip;
-
-	if (ioctl(vcpu->vcpu_fd, KVM_GET_REGS, &vcpu->regs) < 0)
-		die("KVM_GET_REGS failed");
-
-	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &vcpu->sregs) < 0)
-		die("KVM_GET_SREGS failed");
-
-	ip = guest_flat_to_host(vcpu->kvm, ip_to_flat(vcpu, vcpu->regs.rip) - code_prologue);
-
-	dprintf(debug_fd, "\n Code:\n");
-	dprintf(debug_fd,   " -----\n");
-
-	symbol__lookup(vcpu->kvm, vcpu->regs.rip, sym, MAX_SYM_LEN);
-
-	dprintf(debug_fd, " rip: [<%016lx>] %s\n\n", (unsigned long) vcpu->regs.rip, sym);
-
-	for (i = 0; i < code_len; i++, ip++) {
-		if (!host_ptr_in_ram(vcpu->kvm, ip))
-			break;
-
-		c = *ip;
-
-		if (ip == guest_flat_to_host(vcpu->kvm, ip_to_flat(vcpu, vcpu->regs.rip)))
-			dprintf(debug_fd, " <%02x>", c);
-		else
-			dprintf(debug_fd, " %02x", c);
-	}
-
-	dprintf(debug_fd, "\n");
-
-	dprintf(debug_fd, "\n Stack:\n");
-	dprintf(debug_fd,   " ------\n");
-	kvm__dump_mem(vcpu->kvm, vcpu->regs.rsp, 32);
-}
-
-void kvm_cpu__show_page_tables(struct kvm_cpu *vcpu)
-{
-	u64 *pte1;
-	u64 *pte2;
-	u64 *pte3;
-	u64 *pte4;
-
-	if (!is_in_protected_mode(vcpu))
-		return;
-
-	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &vcpu->sregs) < 0)
-		die("KVM_GET_SREGS failed");
-
-	pte4	= guest_flat_to_host(vcpu->kvm, vcpu->sregs.cr3);
-	if (!host_ptr_in_ram(vcpu->kvm, pte4))
-		return;
-
-	pte3	= guest_flat_to_host(vcpu->kvm, (*pte4 & ~0xfff));
-	if (!host_ptr_in_ram(vcpu->kvm, pte3))
-		return;
-
-	pte2	= guest_flat_to_host(vcpu->kvm, (*pte3 & ~0xfff));
-	if (!host_ptr_in_ram(vcpu->kvm, pte2))
-		return;
-
-	pte1	= guest_flat_to_host(vcpu->kvm, (*pte2 & ~0xfff));
-	if (!host_ptr_in_ram(vcpu->kvm, pte1))
-		return;
-
-	dprintf(debug_fd, "Page Tables:\n");
-	if (*pte2 & (1 << 7))
-		dprintf(debug_fd, " pte4: %016llx   pte3: %016llx"
-			"   pte2: %016llx\n",
-			*pte4, *pte3, *pte2);
-	else
-		dprintf(debug_fd, " pte4: %016llx  pte3: %016llx   pte2: %016"
-			"llx   pte1: %016llx\n",
-			*pte4, *pte3, *pte2, *pte1);
-}
-
 void kvm_cpu__run(struct kvm_cpu *vcpu)
 {
 	int err;

@@ -454,7 +83,6 @@ int kvm_cpu__start(struct kvm_cpu *cpu)
 	signal(SIGKVMEXIT, kvm_cpu_signal_handler);
 	signal(SIGKVMPAUSE, kvm_cpu_signal_handler);
 
-	kvm_cpu__setup_cpuid(cpu);
 	kvm_cpu__reset_vcpu(cpu);
 
 	if (cpu->kvm->single_step)

tools/kvm/kvm.c +6 -317

index 252bd18..7ce1640 100644
@@ -1,10 +1,5 @@
 #include "kvm/kvm.h"
-
-#include "kvm/boot-protocol.h"
-#include "kvm/cpufeature.h"
 #include "kvm/read-write.h"
-#include "kvm/interrupt.h"
-#include "kvm/mptable.h"
 #include "kvm/util.h"
 #include "kvm/mutex.h"
 #include "kvm/kvm-cpu.h"

@@ -12,14 +7,11 @@
 
 #include <linux/kvm.h>
 
-#include <asm/bootparam.h>
-
 #include <sys/un.h>
 #include <sys/types.h>
 #include <sys/socket.h>
 #include <sys/ioctl.h>
 #include <sys/mman.h>
-#include <sys/stat.h>
 #include <stdbool.h>
 #include <assert.h>
 #include <limits.h>

@@ -58,29 +50,11 @@ const char *kvm_exit_reasons[] = {
 	DEFINE_KVM_EXIT_REASON(KVM_EXIT_INTERNAL_ERROR),
 };
 
-#define DEFINE_KVM_EXT(ext)		\
-	.name = #ext,			\
-	.code = ext
-
-struct {
-	const char *name;
-	int code;
-} kvm_req_ext[] = {
-	{ DEFINE_KVM_EXT(KVM_CAP_COALESCED_MMIO) },
-	{ DEFINE_KVM_EXT(KVM_CAP_SET_TSS_ADDR) },
-	{ DEFINE_KVM_EXT(KVM_CAP_PIT2) },
-	{ DEFINE_KVM_EXT(KVM_CAP_USER_MEMORY) },
-	{ DEFINE_KVM_EXT(KVM_CAP_IRQ_ROUTING) },
-	{ DEFINE_KVM_EXT(KVM_CAP_IRQCHIP) },
-	{ DEFINE_KVM_EXT(KVM_CAP_HLT) },
-	{ DEFINE_KVM_EXT(KVM_CAP_IRQ_INJECT_STATUS) },
-	{ DEFINE_KVM_EXT(KVM_CAP_EXT_CPUID) },
-};
-
 extern struct kvm *kvm;
 extern struct kvm_cpu *kvm_cpus[KVM_NR_CPUS];
 static int pause_event;
 static DEFINE_MUTEX(pause_lock);
+extern struct kvm_ext kvm_req_ext[];
 
 static char kvm_dir[PATH_MAX];
 

@@ -127,7 +101,9 @@ static int kvm__check_extensions(struct kvm *kvm)
 {
 	unsigned int i;
 
-	for (i = 0; i < ARRAY_SIZE(kvm_req_ext); i++) {
+	for (i = 0; ; i++) {
+		if (!kvm_req_ext[i].name)
+			break;
 		if (!kvm__supports_extension(kvm, kvm_req_ext[i].code)) {
 			pr_error("Unsuppored KVM extension detected: %s",
 				kvm_req_ext[i].name);

@@ -261,48 +237,6 @@ void kvm__delete(struct kvm *kvm)
 	free(kvm);
 }
 
-static bool kvm__cpu_supports_vm(void)
-{
-	struct cpuid_regs regs;
-	u32 eax_base;
-	int feature;
-
-	regs	= (struct cpuid_regs) {
-		.eax		= 0x00,
-	};
-	host_cpuid(&regs);
-
-	switch (regs.ebx) {
-	case CPUID_VENDOR_INTEL_1:
-		eax_base	= 0x00;
-		feature		= KVM__X86_FEATURE_VMX;
-		break;
-
-	case CPUID_VENDOR_AMD_1:
-		eax_base	= 0x80000000;
-		feature		= KVM__X86_FEATURE_SVM;
-		break;
-
-	default:
-		return false;
-	}
-
-	regs	= (struct cpuid_regs) {
-		.eax		= eax_base,
-	};
-	host_cpuid(&regs);
-
-	if (regs.eax < eax_base + 0x01)
-		return false;
-
-	regs	= (struct cpuid_regs) {
-		.eax		= eax_base + 0x01
-	};
-	host_cpuid(&regs);
-
-	return regs.ecx & (1 << feature);
-}
-
 /*
  * Note: KVM_SET_USER_MEMORY_REGION assumes that we don't pass overlapping
  * memory regions to it. Therefore, be careful if you use this function for

@@ -325,47 +259,6 @@ void kvm__register_mem(struct kvm *kvm, u64 guest_phys, u64 size, void *userspac
 		die_perror("KVM_SET_USER_MEMORY_REGION ioctl");
 }
 
-/*
- * Allocating RAM size bigger than 4GB requires us to leave a gap
- * in the RAM which is used for PCI MMIO, hotplug, and unconfigured
- * devices (see documentation of e820_setup_gap() for details).
- *
- * If we're required to initialize RAM bigger than 4GB, we will create
- * a gap between 0xe0000000 and 0x100000000 in the guest virtual mem space.
- */
-
-void kvm__init_ram(struct kvm *kvm)
-{
-	u64	phys_start, phys_size;
-	void	*host_mem;
-
-	if (kvm->ram_size < KVM_32BIT_GAP_START) {
-		/* Use a single block of RAM for 32bit RAM */
-
-		phys_start = 0;
-		phys_size  = kvm->ram_size;
-		host_mem   = kvm->ram_start;
-
-		kvm__register_mem(kvm, phys_start, phys_size, host_mem);
-	} else {
-		/* First RAM range from zero to the PCI gap: */
-
-		phys_start = 0;
-		phys_size  = KVM_32BIT_GAP_START;
-		host_mem   = kvm->ram_start;
-
-		kvm__register_mem(kvm, phys_start, phys_size, host_mem);
-
-		/* Second RAM range from 4GB to the end of RAM: */
-
-		phys_start = 0x100000000ULL;
-		phys_size  = kvm->ram_size - phys_size;
-		host_mem   = kvm->ram_start + phys_start;
-
-		kvm__register_mem(kvm, phys_start, phys_size, host_mem);
-	}
-}
-
 int kvm__recommended_cpus(struct kvm *kvm)
 {
 	int ret;

@@ -410,11 +303,10 @@ int kvm__max_cpus(struct kvm *kvm)
 
 struct kvm *kvm__init(const char *kvm_dev, u64 ram_size, const char *name)
 {
-	struct kvm_pit_config pit_config = { .flags = 0, };
 	struct kvm *kvm;
 	int ret;
 
-	if (!kvm__cpu_supports_vm())
+	if (!kvm__arch_cpu_supports_vm())
 		die("Your CPU does not support hardware virtualization");
 
 	kvm = kvm__new();

@@ -442,36 +334,7 @@ struct kvm *kvm__init(const char *kvm_dev, u64 ram_size, const char *name)
 	if (kvm__check_extensions(kvm))
 		die("A required KVM extention is not supported by OS");
 
-	ret = ioctl(kvm->vm_fd, KVM_SET_TSS_ADDR, 0xfffbd000);
-	if (ret < 0)
-		die_perror("KVM_SET_TSS_ADDR ioctl");
-
-	ret = ioctl(kvm->vm_fd, KVM_CREATE_PIT2, &pit_config);
-	if (ret < 0)
-		die_perror("KVM_CREATE_PIT2 ioctl");
-
-	kvm->ram_size		= ram_size;
-
-	if (kvm->ram_size < KVM_32BIT_GAP_START) {
-		kvm->ram_start = mmap(NULL, ram_size, PROT_RW, MAP_ANON_NORESERVE, -1, 0);
-	} else {
-		kvm->ram_start = mmap(NULL, ram_size + KVM_32BIT_GAP_SIZE, PROT_RW, MAP_ANON_NORESERVE, -1, 0);
-		if (kvm->ram_start != MAP_FAILED) {
-			/*
-			 * We mprotect the gap (see kvm__init_ram() for details) PROT_NONE so that
-			 * if we accidently write to it, we will know.
-			 */
-			mprotect(kvm->ram_start + KVM_32BIT_GAP_START, KVM_32BIT_GAP_SIZE, PROT_NONE);
-		}
-	}
-	if (kvm->ram_start == MAP_FAILED)
-		die("out of memory");
-
-	madvise(kvm->ram_start, kvm->ram_size, MADV_MERGEABLE);
-
-	ret = ioctl(kvm->vm_fd, KVM_CREATE_IRQCHIP);
-	if (ret < 0)
-		die_perror("KVM_CREATE_IRQCHIP ioctl");
+	kvm__arch_init(kvm, kvm_dev, ram_size, name);
 
 	kvm->name = name;
 

@@ -480,141 +343,6 @@ struct kvm *kvm__init(const char *kvm_dev, u64 ram_size, const char *name)
 	return kvm;
 }
 
-#define BOOT_LOADER_SELECTOR	0x1000
-#define BOOT_LOADER_IP		0x0000
-#define BOOT_LOADER_SP		0x8000
-#define BOOT_CMDLINE_OFFSET	0x20000
-
-#define BOOT_PROTOCOL_REQUIRED	0x206
-#define LOAD_HIGH		0x01
-
-static int load_flat_binary(struct kvm *kvm, int fd)
-{
-	void *p;
-	int nr;
-
-	if (lseek(fd, 0, SEEK_SET) < 0)
-		die_perror("lseek");
-
-	p = guest_real_to_host(kvm, BOOT_LOADER_SELECTOR, BOOT_LOADER_IP);
-
-	while ((nr = read(fd, p, 65536)) > 0)
-		p += nr;
-
-	kvm->boot_selector	= BOOT_LOADER_SELECTOR;
-	kvm->boot_ip		= BOOT_LOADER_IP;
-	kvm->boot_sp		= BOOT_LOADER_SP;
-
-	return true;
-}
-
-static const char *BZIMAGE_MAGIC	= "HdrS";
-
-static bool load_bzimage(struct kvm *kvm, int fd_kernel,
-			int fd_initrd, const char *kernel_cmdline, u16 vidmode)
-{
-	struct boot_params *kern_boot;
-	unsigned long setup_sects;
-	struct boot_params boot;
-	size_t cmdline_size;
-	ssize_t setup_size;
-	void *p;
-	int nr;
-
-	/*
-	 * See Documentation/x86/boot.txt for details no bzImage on-disk and
-	 * memory layout.
-	 */
-
-	if (lseek(fd_kernel, 0, SEEK_SET) < 0)
-		die_perror("lseek");
-
-	if (read(fd_kernel, &boot, sizeof(boot)) != sizeof(boot))
-		return false;
-
-	if (memcmp(&boot.hdr.header, BZIMAGE_MAGIC, strlen(BZIMAGE_MAGIC)))
-		return false;
-
-	if (boot.hdr.version < BOOT_PROTOCOL_REQUIRED)
-		die("Too old kernel");
-
-	if (lseek(fd_kernel, 0, SEEK_SET) < 0)
-		die_perror("lseek");
-
-	if (!boot.hdr.setup_sects)
-		boot.hdr.setup_sects = BZ_DEFAULT_SETUP_SECTS;
-	setup_sects = boot.hdr.setup_sects + 1;
-
-	setup_size = setup_sects << 9;
-	p = guest_real_to_host(kvm, BOOT_LOADER_SELECTOR, BOOT_LOADER_IP);
-
-	/* copy setup.bin to mem*/
-	if (read(fd_kernel, p, setup_size) != setup_size)
-		die_perror("read");
-
-	/* copy vmlinux.bin to BZ_KERNEL_START*/
-	p = guest_flat_to_host(kvm, BZ_KERNEL_START);
-
-	while ((nr = read(fd_kernel, p, 65536)) > 0)
-		p += nr;
-
-	p = guest_flat_to_host(kvm, BOOT_CMDLINE_OFFSET);
-	if (kernel_cmdline) {
-		cmdline_size = strlen(kernel_cmdline) + 1;
-		if (cmdline_size > boot.hdr.cmdline_size)
-			cmdline_size = boot.hdr.cmdline_size;
-
-		memset(p, 0, boot.hdr.cmdline_size);
-		memcpy(p, kernel_cmdline, cmdline_size - 1);
-	}
-
-	kern_boot	= guest_real_to_host(kvm, BOOT_LOADER_SELECTOR, 0x00);
-
-	kern_boot->hdr.cmd_line_ptr	= BOOT_CMDLINE_OFFSET;
-	kern_boot->hdr.type_of_loader	= 0xff;
-	kern_boot->hdr.heap_end_ptr	= 0xfe00;
-	kern_boot->hdr.loadflags	|= CAN_USE_HEAP;
-	kern_boot->hdr.vid_mode		= vidmode;
-
-	/*
-	 * Read initrd image into guest memory
-	 */
-	if (fd_initrd >= 0) {
-		struct stat initrd_stat;
-		unsigned long addr;
-
-		if (fstat(fd_initrd, &initrd_stat))
-			die_perror("fstat");
-
-		addr = boot.hdr.initrd_addr_max & ~0xfffff;
-		for (;;) {
-			if (addr < BZ_KERNEL_START)
-				die("Not enough memory for initrd");
-			else if (addr < (kvm->ram_size - initrd_stat.st_size))
-				break;
-			addr -= 0x100000;
-		}
-
-		p = guest_flat_to_host(kvm, addr);
-		nr = read(fd_initrd, p, initrd_stat.st_size);
-		if (nr != initrd_stat.st_size)
-			die("Failed to read initrd");
-
-		kern_boot->hdr.ramdisk_image	= addr;
-		kern_boot->hdr.ramdisk_size	= initrd_stat.st_size;
-	}
-
-	kvm->boot_selector	= BOOT_LOADER_SELECTOR;
-	/*
-	 * The real-mode setup code starts at offset 0x200 of a bzImage. See
-	 * Documentation/x86/boot.txt for details.
-	 */
-	kvm->boot_ip		= BOOT_LOADER_IP + 0x200;
-	kvm->boot_sp		= BOOT_LOADER_SP;
-
-	return true;
-}
-
 /* RFC 1952 */
 #define GZIP_ID1		0x1f
 #define GZIP_ID2		0x8b

@@ -675,24 +403,6 @@ found_kernel:
 	return ret;
 }
 
-/**
- * kvm__setup_bios - inject BIOS into guest system memory
- * @kvm - guest system descriptor
- *
- * This function is a main routine where we poke guest memory
- * and install BIOS there.
- */
-void kvm__setup_bios(struct kvm *kvm)
-{
-	/* standart minimal configuration */
-	setup_bios(kvm);
-
-	/* FIXME: SMP, ACPI and friends here */
-
-	/* MP table */
-	mptable_setup(kvm, kvm->nrcpus);
-}
-
 #define TIMER_INTERVAL_NS 1000000	/* 1 msec */
 
 /*

@@ -732,27 +442,6 @@ void kvm__stop_timer(struct kvm *kvm)
 	kvm->timerid = 0;
 }
 
-void kvm__irq_line(struct kvm *kvm, int irq, int level)
-{
-	struct kvm_irq_level irq_level;
-
-	irq_level	= (struct kvm_irq_level) {
-		{
-			.irq		= irq,
-		},
-		.level		= level,
-	};
-
-	if (ioctl(kvm->vm_fd, KVM_IRQ_LINE, &irq_level) < 0)
-		die_perror("KVM_IRQ_LINE failed");
-}
-
-void kvm__irq_trigger(struct kvm *kvm, int irq)
-{
-	kvm__irq_line(kvm, irq, 1);
-	kvm__irq_line(kvm, irq, 0);
-}
-
 void kvm__dump_mem(struct kvm *kvm, unsigned long addr, unsigned long size)
 {
 	unsigned char *p;

tools/kvm/x86/bios.c renamed from tools/kvm/bios.c +0 -0

similarity index 100%
rename from tools/kvm/bios.c
rename to tools/kvm/x86/bios.c

tools/kvm/x86/bios/.gitignore renamed from tools/kvm/bios/.gitignore +0 -0

similarity index 100%
rename from tools/kvm/bios/.gitignore
rename to tools/kvm/x86/bios/.gitignore

tools/kvm/x86/bios/bios-rom.S renamed from tools/kvm/bios/bios-rom.S +1 -1

similarity index 80%
rename from tools/kvm/bios/bios-rom.S
rename to tools/kvm/x86/bios/bios-rom.S
index dc52b1e..3269ce9 100644
@@ -8,5 +8,5 @@
 #endif
 
 GLOBAL(bios_rom)
-	.incbin "bios/bios.bin"
+	.incbin "x86/bios/bios.bin"
 END(bios_rom)

tools/kvm/x86/bios/e820.c renamed from tools/kvm/bios/e820.c +0 -0

similarity index 100%
rename from tools/kvm/bios/e820.c
rename to tools/kvm/x86/bios/e820.c

tools/kvm/x86/bios/entry.S renamed from tools/kvm/bios/entry.S +0 -0

similarity index 100%
rename from tools/kvm/bios/entry.S
rename to tools/kvm/x86/bios/entry.S

tools/kvm/x86/bios/gen-offsets.sh renamed from tools/kvm/bios/gen-offsets.sh +0 -0

similarity index 100%
rename from tools/kvm/bios/gen-offsets.sh
rename to tools/kvm/x86/bios/gen-offsets.sh

tools/kvm/x86/bios/int10.c renamed from tools/kvm/bios/int10.c +0 -0

similarity index 100%
rename from tools/kvm/bios/int10.c
rename to tools/kvm/x86/bios/int10.c

tools/kvm/x86/bios/int15.c renamed from tools/kvm/bios/int15.c +0 -0

similarity index 100%
rename from tools/kvm/bios/int15.c
rename to tools/kvm/x86/bios/int15.c

tools/kvm/x86/bios/local.S renamed from tools/kvm/bios/local.S +0 -0

similarity index 100%
rename from tools/kvm/bios/local.S
rename to tools/kvm/x86/bios/local.S

tools/kvm/x86/bios/macro.S renamed from tools/kvm/bios/macro.S +0 -0

similarity index 100%
rename from tools/kvm/bios/macro.S
rename to tools/kvm/x86/bios/macro.S

tools/kvm/x86/bios/memcpy.c renamed from tools/kvm/bios/memcpy.c +0 -0

similarity index 100%
rename from tools/kvm/bios/memcpy.c
rename to tools/kvm/x86/bios/memcpy.c

tools/kvm/x86/bios/rom.ld.S renamed from tools/kvm/bios/rom.ld.S +0 -0

similarity index 100%
rename from tools/kvm/bios/rom.ld.S
rename to tools/kvm/x86/bios/rom.ld.S

tools/kvm/x86/cpuid.c renamed from tools/kvm/cpuid.c +0 -0

similarity index 100%
rename from tools/kvm/cpuid.c
rename to tools/kvm/x86/cpuid.c

tools/kvm/x86/include/kvm/assembly.h renamed from tools/kvm/include/kvm/assembly.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/assembly.h
rename to tools/kvm/x86/include/kvm/assembly.h

tools/kvm/x86/include/kvm/barrier.h renamed from tools/kvm/include/kvm/barrier.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/barrier.h
rename to tools/kvm/x86/include/kvm/barrier.h

tools/kvm/x86/include/kvm/bios-export.h renamed from tools/kvm/include/kvm/bios-export.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/bios-export.h
rename to tools/kvm/x86/include/kvm/bios-export.h

tools/kvm/x86/include/kvm/bios.h renamed from tools/kvm/include/kvm/bios.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/bios.h
rename to tools/kvm/x86/include/kvm/bios.h

tools/kvm/x86/include/kvm/boot-protocol.h renamed from tools/kvm/include/kvm/boot-protocol.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/boot-protocol.h
rename to tools/kvm/x86/include/kvm/boot-protocol.h

tools/kvm/x86/include/kvm/cpufeature.h renamed from tools/kvm/include/kvm/cpufeature.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/cpufeature.h
rename to tools/kvm/x86/include/kvm/cpufeature.h

tools/kvm/x86/include/kvm/interrupt.h renamed from tools/kvm/include/kvm/interrupt.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/interrupt.h
rename to tools/kvm/x86/include/kvm/interrupt.h

tools/kvm/x86/include/kvm/kvm-arch.h new file +59 -0

new file mode 100644
index 0000000..02aa8b9
@@ -0,0 +1,59 @@
+#ifndef KVM__KVM_ARCH_H
+#define KVM__KVM_ARCH_H
+
+#include "kvm/interrupt.h"
+#include "kvm/segment.h"
+
+#include <stdbool.h>
+#include <linux/types.h>
+#include <time.h>
+
+#define KVM_NR_CPUS		(255)
+
+/*
+ * The hole includes VESA framebuffer and PCI memory.
+ */
+#define KVM_32BIT_GAP_SIZE	(768 << 20)
+#define KVM_32BIT_GAP_START	((1ULL << 32) - KVM_32BIT_GAP_SIZE)
+
+#define KVM_MMIO_START		KVM_32BIT_GAP_START
+
+struct kvm {
+	int			sys_fd;		/* For system ioctls(), i.e. /dev/kvm */
+	int			vm_fd;		/* For VM ioctls() */
+	timer_t			timerid;	/* Posix timer for interrupts */
+
+	int			nrcpus;		/* Number of cpus to run */
+
+	u32			mem_slots;	/* for KVM_SET_USER_MEMORY_REGION */
+
+	u64			ram_size;
+	void			*ram_start;
+
+	bool			nmi_disabled;
+
+	bool			single_step;
+
+	u16			boot_selector;
+	u16			boot_ip;
+	u16			boot_sp;
+
+	struct interrupt_table	interrupt_table;
+
+	const char		*vmlinux;
+	struct disk_image       **disks;
+	int                     nr_disks;
+
+	const char		*name;
+};
+
+static inline void *guest_flat_to_host(struct kvm *kvm, unsigned long offset); /* In kvm.h */
+
+static inline void *guest_real_to_host(struct kvm *kvm, u16 selector, u16 offset)
+{
+	unsigned long flat = segment_to_flat(selector, offset);
+
+	return guest_flat_to_host(kvm, flat);
+}
+
+#endif /* KVM__KVM_ARCH_H */

tools/kvm/x86/include/kvm/kvm-cpu-arch.h new file +33 -0

new file mode 100644
index 0000000..ed1c727
@@ -0,0 +1,33 @@
+#ifndef KVM__KVM_CPU_ARCH_H
+#define KVM__KVM_CPU_ARCH_H
+
+/* Architecture-specific kvm_cpu definitions. */
+
+#include <linux/kvm.h>	/* for struct kvm_regs */
+
+#include <pthread.h>
+
+struct kvm;
+
+struct kvm_cpu {
+	pthread_t		thread;		/* VCPU thread */
+
+	unsigned long		cpu_id;
+
+	struct kvm		*kvm;		/* parent KVM */
+	int			vcpu_fd;	/* For VCPU ioctls() */
+	struct kvm_run		*kvm_run;
+
+	struct kvm_regs		regs;
+	struct kvm_sregs	sregs;
+	struct kvm_fpu		fpu;
+
+	struct kvm_msrs		*msrs;		/* dynamically allocated */
+
+	u8			is_running;
+	u8			paused;
+
+	struct kvm_coalesced_mmio_ring	*ring;
+};
+
+#endif /* KVM__KVM_CPU_ARCH_H */

tools/kvm/x86/include/kvm/mptable.h renamed from tools/kvm/include/kvm/mptable.h +0 -0

similarity index 100%
rename from tools/kvm/include/kvm/mptable.h
rename to tools/kvm/x86/include/kvm/mptable.h

tools/kvm/x86/interrupt.c renamed from tools/kvm/interrupt.c +0 -0

similarity index 100%
rename from tools/kvm/interrupt.c
rename to tools/kvm/x86/interrupt.c

tools/kvm/x86/ioport.c new file +59 -0

new file mode 100644
index 0000000..8a91bf2
@@ -0,0 +1,59 @@
+#include "kvm/ioport.h"
+
+#include <stdlib.h>
+
+static bool debug_io_out(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size)
+{
+	exit(EXIT_SUCCESS);
+}
+
+static struct ioport_operations debug_ops = {
+	.io_out		= debug_io_out,
+};
+
+static bool dummy_io_in(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size)
+{
+	return true;
+}
+
+static bool dummy_io_out(struct ioport *ioport, struct kvm *kvm, u16 port, void *data, int size)
+{
+	return true;
+}
+
+static struct ioport_operations dummy_read_write_ioport_ops = {
+	.io_in		= dummy_io_in,
+	.io_out		= dummy_io_out,
+};
+
+static struct ioport_operations dummy_write_only_ioport_ops = {
+	.io_out		= dummy_io_out,
+};
+
+void ioport__setup_arch(void)
+{
+	/* Legacy ioport setup */
+
+	/* 0x0020 - 0x003F - 8259A PIC 1 */
+	ioport__register(0x0020, &dummy_read_write_ioport_ops, 2, NULL);
+
+	/* PORT 0040-005F - PIT - PROGRAMMABLE INTERVAL TIMER (8253, 8254) */
+	ioport__register(0x0040, &dummy_read_write_ioport_ops, 4, NULL);
+
+	/* 0x00A0 - 0x00AF - 8259A PIC 2 */
+	ioport__register(0x00A0, &dummy_read_write_ioport_ops, 2, NULL);
+
+	/* PORT 00E0-00EF are 'motherboard specific' so we use them for our
+	   internal debugging purposes.  */
+	ioport__register(IOPORT_DBG, &debug_ops, 1, NULL);
+
+	/* PORT 00ED - DUMMY PORT FOR DELAY??? */
+	ioport__register(0x00ED, &dummy_write_only_ioport_ops, 1, NULL);
+
+	/* 0x00F0 - 0x00FF - Math co-processor */
+	ioport__register(0x00F0, &dummy_write_only_ioport_ops, 2, NULL);
+
+	/* PORT 03D4-03D5 - COLOR VIDEO - CRT CONTROL REGISTERS */
+	ioport__register(0x03D4, &dummy_read_write_ioport_ops, 1, NULL);
+	ioport__register(0x03D5, &dummy_write_only_ioport_ops, 1, NULL);
+}

tools/kvm/x86/irq.c renamed from tools/kvm/irq.c +0 -0

similarity index 100%
rename from tools/kvm/irq.c
rename to tools/kvm/x86/irq.c

tools/kvm/x86/kvm-cpu.c new file +383 -0

new file mode 100644
index 0000000..b26b208
@@ -0,0 +1,383 @@
+#include "kvm/kvm-cpu.h"
+
+#include "kvm/symbol.h"
+#include "kvm/util.h"
+#include "kvm/kvm.h"
+
+#include <asm/msr-index.h>
+
+#include <sys/ioctl.h>
+#include <sys/mman.h>
+#include <signal.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+#include <stdio.h>
+
+static int debug_fd;
+
+void kvm_cpu__set_debug_fd(int fd)
+{
+	debug_fd = fd;
+}
+
+int kvm_cpu__get_debug_fd(void)
+{
+	return debug_fd;
+}
+
+static inline bool is_in_protected_mode(struct kvm_cpu *vcpu)
+{
+	return vcpu->sregs.cr0 & 0x01;
+}
+
+static inline u64 ip_to_flat(struct kvm_cpu *vcpu, u64 ip)
+{
+	u64 cs;
+
+	/*
+	 * NOTE! We should take code segment base address into account here.
+	 * Luckily it's usually zero because Linux uses flat memory model.
+	 */
+	if (is_in_protected_mode(vcpu))
+		return ip;
+
+	cs = vcpu->sregs.cs.selector;
+
+	return ip + (cs << 4);
+}
+
+static inline u32 selector_to_base(u16 selector)
+{
+	/*
+	 * KVM on Intel requires 'base' to be 'selector * 16' in real mode.
+	 */
+	return (u32)selector * 16;
+}
+
+static struct kvm_cpu *kvm_cpu__new(struct kvm *kvm)
+{
+	struct kvm_cpu *vcpu;
+
+	vcpu		= calloc(1, sizeof *vcpu);
+	if (!vcpu)
+		return NULL;
+
+	vcpu->kvm	= kvm;
+
+	return vcpu;
+}
+
+void kvm_cpu__delete(struct kvm_cpu *vcpu)
+{
+	if (vcpu->msrs)
+		free(vcpu->msrs);
+
+	free(vcpu);
+}
+
+struct kvm_cpu *kvm_cpu__init(struct kvm *kvm, unsigned long cpu_id)
+{
+	struct kvm_cpu *vcpu;
+	int mmap_size;
+	int coalesced_offset;
+
+	vcpu		= kvm_cpu__new(kvm);
+	if (!vcpu)
+		return NULL;
+
+	vcpu->cpu_id	= cpu_id;
+
+	vcpu->vcpu_fd = ioctl(vcpu->kvm->vm_fd, KVM_CREATE_VCPU, cpu_id);
+	if (vcpu->vcpu_fd < 0)
+		die_perror("KVM_CREATE_VCPU ioctl");
+
+	mmap_size = ioctl(vcpu->kvm->sys_fd, KVM_GET_VCPU_MMAP_SIZE, 0);
+	if (mmap_size < 0)
+		die_perror("KVM_GET_VCPU_MMAP_SIZE ioctl");
+
+	vcpu->kvm_run = mmap(NULL, mmap_size, PROT_RW, MAP_SHARED, vcpu->vcpu_fd, 0);
+	if (vcpu->kvm_run == MAP_FAILED)
+		die("unable to mmap vcpu fd");
+
+	coalesced_offset = ioctl(kvm->sys_fd, KVM_CHECK_EXTENSION, KVM_CAP_COALESCED_MMIO);
+	if (coalesced_offset)
+		vcpu->ring = (void *)vcpu->kvm_run + (coalesced_offset * PAGE_SIZE);
+
+	vcpu->is_running = true;
+
+	return vcpu;
+}
+
+static struct kvm_msrs *kvm_msrs__new(size_t nmsrs)
+{
+	struct kvm_msrs *vcpu = calloc(1, sizeof(*vcpu) + (sizeof(struct kvm_msr_entry) * nmsrs));
+
+	if (!vcpu)
+		die("out of memory");
+
+	return vcpu;
+}
+
+#define KVM_MSR_ENTRY(_index, _data)	\
+	(struct kvm_msr_entry) { .index = _index, .data = _data }
+
+static void kvm_cpu__setup_msrs(struct kvm_cpu *vcpu)
+{
+	unsigned long ndx = 0;
+
+	vcpu->msrs = kvm_msrs__new(100);
+
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_SYSENTER_CS,	0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_SYSENTER_ESP,	0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_SYSENTER_EIP,	0x0);
+#ifdef CONFIG_X86_64
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_STAR,			0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_CSTAR,			0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_KERNEL_GS_BASE,		0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_SYSCALL_MASK,		0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_LSTAR,			0x0);
+#endif
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_TSC,		0x0);
+	vcpu->msrs->entries[ndx++] = KVM_MSR_ENTRY(MSR_IA32_MISC_ENABLE,
+						MSR_IA32_MISC_ENABLE_FAST_STRING);
+
+	vcpu->msrs->nmsrs	= ndx;
+
+	if (ioctl(vcpu->vcpu_fd, KVM_SET_MSRS, vcpu->msrs) < 0)
+		die_perror("KVM_SET_MSRS failed");
+}
+
+static void kvm_cpu__setup_fpu(struct kvm_cpu *vcpu)
+{
+	vcpu->fpu = (struct kvm_fpu) {
+		.fcw		= 0x37f,
+		.mxcsr		= 0x1f80,
+	};
+
+	if (ioctl(vcpu->vcpu_fd, KVM_SET_FPU, &vcpu->fpu) < 0)
+		die_perror("KVM_SET_FPU failed");
+}
+
+static void kvm_cpu__setup_regs(struct kvm_cpu *vcpu)
+{
+	vcpu->regs = (struct kvm_regs) {
+		/* We start the guest in 16-bit real mode  */
+		.rflags		= 0x0000000000000002ULL,
+
+		.rip		= vcpu->kvm->boot_ip,
+		.rsp		= vcpu->kvm->boot_sp,
+		.rbp		= vcpu->kvm->boot_sp,
+	};
+
+	if (vcpu->regs.rip > USHRT_MAX)
+		die("ip 0x%llx is too high for real mode", (u64) vcpu->regs.rip);
+
+	if (ioctl(vcpu->vcpu_fd, KVM_SET_REGS, &vcpu->regs) < 0)
+		die_perror("KVM_SET_REGS failed");
+}
+
+static void kvm_cpu__setup_sregs(struct kvm_cpu *vcpu)
+{
+
+	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &vcpu->sregs) < 0)
+		die_perror("KVM_GET_SREGS failed");
+
+	vcpu->sregs.cs.selector	= vcpu->kvm->boot_selector;
+	vcpu->sregs.cs.base	= selector_to_base(vcpu->kvm->boot_selector);
+	vcpu->sregs.ss.selector	= vcpu->kvm->boot_selector;
+	vcpu->sregs.ss.base	= selector_to_base(vcpu->kvm->boot_selector);
+	vcpu->sregs.ds.selector	= vcpu->kvm->boot_selector;
+	vcpu->sregs.ds.base	= selector_to_base(vcpu->kvm->boot_selector);
+	vcpu->sregs.es.selector	= vcpu->kvm->boot_selector;
+	vcpu->sregs.es.base	= selector_to_base(vcpu->kvm->boot_selector);
+	vcpu->sregs.fs.selector	= vcpu->kvm->boot_selector;
+	vcpu->sregs.fs.base	= selector_to_base(vcpu->kvm->boot_selector);
+	vcpu->sregs.gs.selector	= vcpu->kvm->boot_selector;
+	vcpu->sregs.gs.base	= selector_to_base(vcpu->kvm->boot_selector);
+
+	if (ioctl(vcpu->vcpu_fd, KVM_SET_SREGS, &vcpu->sregs) < 0)
+		die_perror("KVM_SET_SREGS failed");
+}
+
+/**
+ * kvm_cpu__reset_vcpu - reset virtual CPU to a known state
+ */
+void kvm_cpu__reset_vcpu(struct kvm_cpu *vcpu)
+{
+	kvm_cpu__setup_cpuid(vcpu);
+	kvm_cpu__setup_sregs(vcpu);
+	kvm_cpu__setup_regs(vcpu);
+	kvm_cpu__setup_fpu(vcpu);
+	kvm_cpu__setup_msrs(vcpu);
+}
+
+static void print_dtable(const char *name, struct kvm_dtable *dtable)
+{
+	dprintf(debug_fd, " %s                 %016llx  %08hx\n",
+		name, (u64) dtable->base, (u16) dtable->limit);
+}
+
+static void print_segment(const char *name, struct kvm_segment *seg)
+{
+	dprintf(debug_fd, " %s       %04hx      %016llx  %08x  %02hhx    %x %x   %x  %x %x %x %x\n",
+		name, (u16) seg->selector, (u64) seg->base, (u32) seg->limit,
+		(u8) seg->type, seg->present, seg->dpl, seg->db, seg->s, seg->l, seg->g, seg->avl);
+}
+
+void kvm_cpu__show_registers(struct kvm_cpu *vcpu)
+{
+	unsigned long cr0, cr2, cr3;
+	unsigned long cr4, cr8;
+	unsigned long rax, rbx, rcx;
+	unsigned long rdx, rsi, rdi;
+	unsigned long rbp,  r8,  r9;
+	unsigned long r10, r11, r12;
+	unsigned long r13, r14, r15;
+	unsigned long rip, rsp;
+	struct kvm_sregs sregs;
+	unsigned long rflags;
+	struct kvm_regs regs;
+	int i;
+
+	if (ioctl(vcpu->vcpu_fd, KVM_GET_REGS, &regs) < 0)
+		die("KVM_GET_REGS failed");
+
+	rflags = regs.rflags;
+
+	rip = regs.rip; rsp = regs.rsp;
+	rax = regs.rax; rbx = regs.rbx; rcx = regs.rcx;
+	rdx = regs.rdx; rsi = regs.rsi; rdi = regs.rdi;
+	rbp = regs.rbp; r8  = regs.r8;  r9  = regs.r9;
+	r10 = regs.r10; r11 = regs.r11; r12 = regs.r12;
+	r13 = regs.r13; r14 = regs.r14; r15 = regs.r15;
+
+	dprintf(debug_fd, "\n Registers:\n");
+	dprintf(debug_fd,   " ----------\n");
+	dprintf(debug_fd, " rip: %016lx   rsp: %016lx flags: %016lx\n", rip, rsp, rflags);
+	dprintf(debug_fd, " rax: %016lx   rbx: %016lx   rcx: %016lx\n", rax, rbx, rcx);
+	dprintf(debug_fd, " rdx: %016lx   rsi: %016lx   rdi: %016lx\n", rdx, rsi, rdi);
+	dprintf(debug_fd, " rbp: %016lx    r8: %016lx    r9: %016lx\n", rbp, r8,  r9);
+	dprintf(debug_fd, " r10: %016lx   r11: %016lx   r12: %016lx\n", r10, r11, r12);
+	dprintf(debug_fd, " r13: %016lx   r14: %016lx   r15: %016lx\n", r13, r14, r15);
+
+	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &sregs) < 0)
+		die("KVM_GET_REGS failed");
+
+	cr0 = sregs.cr0; cr2 = sregs.cr2; cr3 = sregs.cr3;
+	cr4 = sregs.cr4; cr8 = sregs.cr8;
+
+	dprintf(debug_fd, " cr0: %016lx   cr2: %016lx   cr3: %016lx\n", cr0, cr2, cr3);
+	dprintf(debug_fd, " cr4: %016lx   cr8: %016lx\n", cr4, cr8);
+	dprintf(debug_fd, "\n Segment registers:\n");
+	dprintf(debug_fd,   " ------------------\n");
+	dprintf(debug_fd, " register  selector  base              limit     type  p dpl db s l g avl\n");
+	print_segment("cs ", &sregs.cs);
+	print_segment("ss ", &sregs.ss);
+	print_segment("ds ", &sregs.ds);
+	print_segment("es ", &sregs.es);
+	print_segment("fs ", &sregs.fs);
+	print_segment("gs ", &sregs.gs);
+	print_segment("tr ", &sregs.tr);
+	print_segment("ldt", &sregs.ldt);
+	print_dtable("gdt", &sregs.gdt);
+	print_dtable("idt", &sregs.idt);
+
+	dprintf(debug_fd, "\n APIC:\n");
+	dprintf(debug_fd,   " -----\n");
+	dprintf(debug_fd, " efer: %016llx  apic base: %016llx  nmi: %s\n",
+		(u64) sregs.efer, (u64) sregs.apic_base,
+		(vcpu->kvm->nmi_disabled ? "disabled" : "enabled"));
+
+	dprintf(debug_fd, "\n Interrupt bitmap:\n");
+	dprintf(debug_fd,   " -----------------\n");
+	for (i = 0; i < (KVM_NR_INTERRUPTS + 63) / 64; i++)
+		dprintf(debug_fd, " %016llx", (u64) sregs.interrupt_bitmap[i]);
+	dprintf(debug_fd, "\n");
+}
+
+#define MAX_SYM_LEN		128
+
+void kvm_cpu__show_code(struct kvm_cpu *vcpu)
+{
+	unsigned int code_bytes = 64;
+	unsigned int code_prologue = code_bytes * 43 / 64;
+	unsigned int code_len = code_bytes;
+	char sym[MAX_SYM_LEN];
+	unsigned char c;
+	unsigned int i;
+	u8 *ip;
+
+	if (ioctl(vcpu->vcpu_fd, KVM_GET_REGS, &vcpu->regs) < 0)
+		die("KVM_GET_REGS failed");
+
+	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &vcpu->sregs) < 0)
+		die("KVM_GET_SREGS failed");
+
+	ip = guest_flat_to_host(vcpu->kvm, ip_to_flat(vcpu, vcpu->regs.rip) - code_prologue);
+
+	dprintf(debug_fd, "\n Code:\n");
+	dprintf(debug_fd,   " -----\n");
+
+	symbol__lookup(vcpu->kvm, vcpu->regs.rip, sym, MAX_SYM_LEN);
+
+	dprintf(debug_fd, " rip: [<%016lx>] %s\n\n", (unsigned long) vcpu->regs.rip, sym);
+
+	for (i = 0; i < code_len; i++, ip++) {
+		if (!host_ptr_in_ram(vcpu->kvm, ip))
+			break;
+
+		c = *ip;
+
+		if (ip == guest_flat_to_host(vcpu->kvm, ip_to_flat(vcpu, vcpu->regs.rip)))
+			dprintf(debug_fd, " <%02x>", c);
+		else
+			dprintf(debug_fd, " %02x", c);
+	}
+
+	dprintf(debug_fd, "\n");
+
+	dprintf(debug_fd, "\n Stack:\n");
+	dprintf(debug_fd,   " ------\n");
+	kvm__dump_mem(vcpu->kvm, vcpu->regs.rsp, 32);
+}
+
+void kvm_cpu__show_page_tables(struct kvm_cpu *vcpu)
+{
+	u64 *pte1;
+	u64 *pte2;
+	u64 *pte3;
+	u64 *pte4;
+
+	if (!is_in_protected_mode(vcpu))
+		return;
+
+	if (ioctl(vcpu->vcpu_fd, KVM_GET_SREGS, &vcpu->sregs) < 0)
+		die("KVM_GET_SREGS failed");
+
+	pte4	= guest_flat_to_host(vcpu->kvm, vcpu->sregs.cr3);
+	if (!host_ptr_in_ram(vcpu->kvm, pte4))
+		return;
+
+	pte3	= guest_flat_to_host(vcpu->kvm, (*pte4 & ~0xfff));
+	if (!host_ptr_in_ram(vcpu->kvm, pte3))
+		return;
+
+	pte2	= guest_flat_to_host(vcpu->kvm, (*pte3 & ~0xfff));
+	if (!host_ptr_in_ram(vcpu->kvm, pte2))
+		return;
+
+	pte1	= guest_flat_to_host(vcpu->kvm, (*pte2 & ~0xfff));
+	if (!host_ptr_in_ram(vcpu->kvm, pte1))
+		return;
+
+	dprintf(debug_fd, "Page Tables:\n");
+	if (*pte2 & (1 << 7))
+		dprintf(debug_fd, " pte4: %016llx   pte3: %016llx"
+			"   pte2: %016llx\n",
+			*pte4, *pte3, *pte2);
+	else
+		dprintf(debug_fd, " pte4: %016llx  pte3: %016llx   pte2: %016"
+			"llx   pte1: %016llx\n",
+			*pte4, *pte3, *pte2, *pte1);
+}

tools/kvm/x86/kvm.c new file +330 -0

new file mode 100644
index 0000000..ac6c91e
@@ -0,0 +1,330 @@
+#include "kvm/kvm.h"
+#include "kvm/boot-protocol.h"
+#include "kvm/cpufeature.h"
+#include "kvm/interrupt.h"
+#include "kvm/mptable.h"
+#include "kvm/util.h"
+
+#include <asm/bootparam.h>
+#include <linux/kvm.h>
+
+#include <sys/types.h>
+#include <sys/ioctl.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <stdbool.h>
+#include <assert.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <fcntl.h>
+#include <asm/unistd.h>
+
+struct kvm_ext kvm_req_ext[] = {
+	{ DEFINE_KVM_EXT(KVM_CAP_COALESCED_MMIO) },
+	{ DEFINE_KVM_EXT(KVM_CAP_SET_TSS_ADDR) },
+	{ DEFINE_KVM_EXT(KVM_CAP_PIT2) },
+	{ DEFINE_KVM_EXT(KVM_CAP_USER_MEMORY) },
+	{ DEFINE_KVM_EXT(KVM_CAP_IRQ_ROUTING) },
+	{ DEFINE_KVM_EXT(KVM_CAP_IRQCHIP) },
+	{ DEFINE_KVM_EXT(KVM_CAP_HLT) },
+	{ DEFINE_KVM_EXT(KVM_CAP_IRQ_INJECT_STATUS) },
+	{ DEFINE_KVM_EXT(KVM_CAP_EXT_CPUID) },
+	{ 0, 0 }
+};
+
+bool kvm__arch_cpu_supports_vm(void)
+{
+	struct cpuid_regs regs;
+	u32 eax_base;
+	int feature;
+
+	regs	= (struct cpuid_regs) {
+		.eax		= 0x00,
+	};
+	host_cpuid(&regs);
+
+	switch (regs.ebx) {
+	case CPUID_VENDOR_INTEL_1:
+		eax_base	= 0x00;
+		feature		= KVM__X86_FEATURE_VMX;
+		break;
+
+	case CPUID_VENDOR_AMD_1:
+		eax_base	= 0x80000000;
+		feature		= KVM__X86_FEATURE_SVM;
+		break;
+
+	default:
+		return false;
+	}
+
+	regs	= (struct cpuid_regs) {
+		.eax		= eax_base,
+	};
+	host_cpuid(&regs);
+
+	if (regs.eax < eax_base + 0x01)
+		return false;
+
+	regs	= (struct cpuid_regs) {
+		.eax		= eax_base + 0x01
+	};
+	host_cpuid(&regs);
+
+	return regs.ecx & (1 << feature);
+}
+
+/*
+ * Allocating RAM size bigger than 4GB requires us to leave a gap
+ * in the RAM which is used for PCI MMIO, hotplug, and unconfigured
+ * devices (see documentation of e820_setup_gap() for details).
+ *
+ * If we're required to initialize RAM bigger than 4GB, we will create
+ * a gap between 0xe0000000 and 0x100000000 in the guest virtual mem space.
+ */
+
+void kvm__init_ram(struct kvm *kvm)
+{
+	u64	phys_start, phys_size;
+	void	*host_mem;
+
+	if (kvm->ram_size < KVM_32BIT_GAP_START) {
+		/* Use a single block of RAM for 32bit RAM */
+
+		phys_start = 0;
+		phys_size  = kvm->ram_size;
+		host_mem   = kvm->ram_start;
+
+		kvm__register_mem(kvm, phys_start, phys_size, host_mem);
+	} else {
+		/* First RAM range from zero to the PCI gap: */
+
+		phys_start = 0;
+		phys_size  = KVM_32BIT_GAP_START;
+		host_mem   = kvm->ram_start;
+
+		kvm__register_mem(kvm, phys_start, phys_size, host_mem);
+
+		/* Second RAM range from 4GB to the end of RAM: */
+
+		phys_start = 0x100000000ULL;
+		phys_size  = kvm->ram_size - phys_size;
+		host_mem   = kvm->ram_start + phys_start;
+
+		kvm__register_mem(kvm, phys_start, phys_size, host_mem);
+	}
+}
+
+/* Architecture-specific KVM init */
+void kvm__arch_init(struct kvm *kvm, const char *kvm_dev, u64 ram_size, const char *name)
+{
+	struct kvm_pit_config pit_config = { .flags = 0, };
+	int ret;
+
+	ret = ioctl(kvm->vm_fd, KVM_SET_TSS_ADDR, 0xfffbd000);
+	if (ret < 0)
+		die_perror("KVM_SET_TSS_ADDR ioctl");
+
+	ret = ioctl(kvm->vm_fd, KVM_CREATE_PIT2, &pit_config);
+	if (ret < 0)
+		die_perror("KVM_CREATE_PIT2 ioctl");
+
+	kvm->ram_size		= ram_size;
+
+	if (kvm->ram_size < KVM_32BIT_GAP_START) {
+		kvm->ram_start = mmap(NULL, ram_size, PROT_RW, MAP_ANON_NORESERVE, -1, 0);
+	} else {
+		kvm->ram_start = mmap(NULL, ram_size + KVM_32BIT_GAP_SIZE, PROT_RW, MAP_ANON_NORESERVE, -1, 0);
+		if (kvm->ram_start != MAP_FAILED) {
+			/*
+			 * We mprotect the gap (see kvm__init_ram() for details) PROT_NONE so that
+			 * if we accidently write to it, we will know.
+			 */
+			mprotect(kvm->ram_start + KVM_32BIT_GAP_START, KVM_32BIT_GAP_SIZE, PROT_NONE);
+		}
+	}
+	if (kvm->ram_start == MAP_FAILED)
+		die("out of memory");
+
+	madvise(kvm->ram_start, kvm->ram_size, MADV_MERGEABLE);
+
+	ret = ioctl(kvm->vm_fd, KVM_CREATE_IRQCHIP);
+	if (ret < 0)
+		die_perror("KVM_CREATE_IRQCHIP ioctl");
+}
+
+void kvm__irq_line(struct kvm *kvm, int irq, int level)
+{
+	struct kvm_irq_level irq_level;
+
+	irq_level	= (struct kvm_irq_level) {
+		{
+			.irq		= irq,
+		},
+		.level		= level,
+	};
+
+	if (ioctl(kvm->vm_fd, KVM_IRQ_LINE, &irq_level) < 0)
+		die_perror("KVM_IRQ_LINE failed");
+}
+
+void kvm__irq_trigger(struct kvm *kvm, int irq)
+{
+	kvm__irq_line(kvm, irq, 1);
+	kvm__irq_line(kvm, irq, 0);
+}
+
+#define BOOT_LOADER_SELECTOR	0x1000
+#define BOOT_LOADER_IP		0x0000
+#define BOOT_LOADER_SP		0x8000
+#define BOOT_CMDLINE_OFFSET	0x20000
+
+#define BOOT_PROTOCOL_REQUIRED	0x206
+#define LOAD_HIGH		0x01
+
+int load_flat_binary(struct kvm *kvm, int fd)
+{
+	void *p;
+	int nr;
+
+	if (lseek(fd, 0, SEEK_SET) < 0)
+		die_perror("lseek");
+
+	p = guest_real_to_host(kvm, BOOT_LOADER_SELECTOR, BOOT_LOADER_IP);
+
+	while ((nr = read(fd, p, 65536)) > 0)
+		p += nr;
+
+	kvm->boot_selector	= BOOT_LOADER_SELECTOR;
+	kvm->boot_ip		= BOOT_LOADER_IP;
+	kvm->boot_sp		= BOOT_LOADER_SP;
+
+	return true;
+}
+
+static const char *BZIMAGE_MAGIC	= "HdrS";
+
+bool load_bzimage(struct kvm *kvm, int fd_kernel,
+		  int fd_initrd, const char *kernel_cmdline, u16 vidmode)
+{
+	struct boot_params *kern_boot;
+	unsigned long setup_sects;
+	struct boot_params boot;
+	size_t cmdline_size;
+	ssize_t setup_size;
+	void *p;
+	int nr;
+
+	/*
+	 * See Documentation/x86/boot.txt for details no bzImage on-disk and
+	 * memory layout.
+	 */
+
+	if (lseek(fd_kernel, 0, SEEK_SET) < 0)
+		die_perror("lseek");
+
+	if (read(fd_kernel, &boot, sizeof(boot)) != sizeof(boot))
+		return false;
+
+	if (memcmp(&boot.hdr.header, BZIMAGE_MAGIC, strlen(BZIMAGE_MAGIC)))
+		return false;
+
+	if (boot.hdr.version < BOOT_PROTOCOL_REQUIRED)
+		die("Too old kernel");
+
+	if (lseek(fd_kernel, 0, SEEK_SET) < 0)
+		die_perror("lseek");
+
+	if (!boot.hdr.setup_sects)
+		boot.hdr.setup_sects = BZ_DEFAULT_SETUP_SECTS;
+	setup_sects = boot.hdr.setup_sects + 1;
+
+	setup_size = setup_sects << 9;
+	p = guest_real_to_host(kvm, BOOT_LOADER_SELECTOR, BOOT_LOADER_IP);
+
+	/* copy setup.bin to mem*/
+	if (read(fd_kernel, p, setup_size) != setup_size)
+		die_perror("read");
+
+	/* copy vmlinux.bin to BZ_KERNEL_START*/
+	p = guest_flat_to_host(kvm, BZ_KERNEL_START);
+
+	while ((nr = read(fd_kernel, p, 65536)) > 0)
+		p += nr;
+
+	p = guest_flat_to_host(kvm, BOOT_CMDLINE_OFFSET);
+	if (kernel_cmdline) {
+		cmdline_size = strlen(kernel_cmdline) + 1;
+		if (cmdline_size > boot.hdr.cmdline_size)
+			cmdline_size = boot.hdr.cmdline_size;
+
+		memset(p, 0, boot.hdr.cmdline_size);
+		memcpy(p, kernel_cmdline, cmdline_size - 1);
+	}
+
+	kern_boot	= guest_real_to_host(kvm, BOOT_LOADER_SELECTOR, 0x00);
+
+	kern_boot->hdr.cmd_line_ptr	= BOOT_CMDLINE_OFFSET;
+	kern_boot->hdr.type_of_loader	= 0xff;
+	kern_boot->hdr.heap_end_ptr	= 0xfe00;
+	kern_boot->hdr.loadflags	|= CAN_USE_HEAP;
+	kern_boot->hdr.vid_mode		= vidmode;
+
+	/*
+	 * Read initrd image into guest memory
+	 */
+	if (fd_initrd >= 0) {
+		struct stat initrd_stat;
+		unsigned long addr;
+
+		if (fstat(fd_initrd, &initrd_stat))
+			die_perror("fstat");
+
+		addr = boot.hdr.initrd_addr_max & ~0xfffff;
+		for (;;) {
+			if (addr < BZ_KERNEL_START)
+				die("Not enough memory for initrd");
+			else if (addr < (kvm->ram_size - initrd_stat.st_size))
+				break;
+			addr -= 0x100000;
+		}
+
+		p = guest_flat_to_host(kvm, addr);
+		nr = read(fd_initrd, p, initrd_stat.st_size);
+		if (nr != initrd_stat.st_size)
+			die("Failed to read initrd");
+
+		kern_boot->hdr.ramdisk_image	= addr;
+		kern_boot->hdr.ramdisk_size	= initrd_stat.st_size;
+	}
+
+	kvm->boot_selector	= BOOT_LOADER_SELECTOR;
+	/*
+	 * The real-mode setup code starts at offset 0x200 of a bzImage. See
+	 * Documentation/x86/boot.txt for details.
+	 */
+	kvm->boot_ip		= BOOT_LOADER_IP + 0x200;
+	kvm->boot_sp		= BOOT_LOADER_SP;
+
+	return true;
+}
+
+/**
+ * kvm__arch_setup_firmware - inject BIOS into guest system memory
+ * @kvm - guest system descriptor
+ *
+ * This function is a main routine where we poke guest memory
+ * and install BIOS there.
+ */
+void kvm__arch_setup_firmware(struct kvm *kvm)
+{
+	/* standart minimal configuration */
+	setup_bios(kvm);
+
+	/* FIXME: SMP, ACPI and friends here */
+
+	/* MP table */
+	mptable_setup(kvm, kvm->nrcpus);
+}

tools/kvm/x86/mptable.c renamed from tools/kvm/mptable.c +0 -1

similarity index 100%
rename from tools/kvm/mptable.c
rename to tools/kvm/x86/mptable.c
--

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

The code doesn't build after this patch due to missing header issues which you fixed in patches #10 & #11. Could you please move those two to the beginning of the series for the sake of bisectablilty?

Not every architecture has an i8042 kbd controller, so only use this when building for x86.

Signed-off-by Matt Evans <[email protected]>

---
 tools/kvm/Makefile      |    2 +-
 tools/kvm/builtin-run.c |    2 ++
 2 files changed, 3 insertions(+), 1 deletions(-)

tools/kvm/Makefile +1 -1

index 243886e..f58a1d8 100644
@@ -77,7 +77,6 @@ OBJS	+= util/strbuf.o
 OBJS	+= virtio/9p.o
 OBJS	+= virtio/9p-pdu.o
 OBJS	+= hw/vesa.o
-OBJS	+= hw/i8042.o
 OBJS	+= hw/pci-shmem.o
 OBJS	+= kvm-ipc.o
 

@@ -153,6 +152,7 @@ ifeq ($(ARCH),x86)
 	OBJS	+= x86/kvm.o
 	OBJS	+= x86/kvm-cpu.o
 	OBJS	+= x86/mptable.o
+	OBJS	+= hw/i8042.o
 # Exclude BIOS object files from header dependencies.
 	OTHEROBJS	+= x86/bios.o
 	OTHEROBJS	+= x86/bios/bios-rom.o

tools/kvm/builtin-run.c +2 -1

index 9148d83..e4aa87e 100644
@@ -941,7 +941,9 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 
 	kvm__init_ram(kvm);
 
+#ifdef CONFIG_X86
 	kbd__init(kvm);
+#endif
 
 	pci_shmem__init(kvm);
 
--

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

Not every architecture has an i8042 kbd controller, so only use this when building for x86.

Quoted 2 lines · Matt Evans

Not every architecture has an i8042 kbd controller, so only use this when building for x86.

There are non-x86 machines that have one, though -- does KVM tool have any sort of target configuration mechanism?

On LP64 systems our u64s are just longs; remove the %llx'es in favour of PRIx64 etc.

This patch also adds CFLAGS to the final link, so that any -m64 is obeyed when linking, too.

Signed-off-by Matt Evans <[email protected]>

---
 tools/kvm/Makefile       |    2 +-
 tools/kvm/builtin-run.c  |   14 ++++++++------
 tools/kvm/builtin-stat.c |    4 +++-
 tools/kvm/disk/core.c    |    4 +++-
 tools/kvm/mmio.c         |    4 +++-
 5 files changed, 18 insertions(+), 10 deletions(-)

tools/kvm/Makefile +1 -1

index 009a6ba..57dc521 100644
@@ -218,7 +218,7 @@ KVMTOOLS-VERSION-FILE:
 
 $(PROGRAM): $(DEPS) $(OBJS)
 	$(E) "  LINK    " $@
-	$(Q) $(CC) $(OBJS) $(LIBS) -o $@
+	$(Q) $(CC) $(CFLAGS) $(OBJS) $(LIBS) -o $@
 
 $(GUEST_INIT): guest/init.c
 	$(E) "  LINK    " $@

tools/kvm/builtin-run.c +8 -6

index e4aa87e..7cf208d 100644
@@ -42,6 +42,8 @@
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
+#define __STDC_FORMAT_MACROS
+#include <inttypes.h>
 #include <ctype.h>
 #include <stdio.h>
 

@@ -383,8 +385,8 @@ static int shmem_parser(const struct option *opt, const char *arg, int unset)
 		strcpy(handle, default_handle);
 	}
 	if (verbose) {
-		pr_info("shmem: phys_addr = %llx", phys_addr);
-		pr_info("shmem: size      = %llx", size);
+		pr_info("shmem: phys_addr = %"PRIx64, phys_addr);
+		pr_info("shmem: size      = %"PRIx64, size);
 		pr_info("shmem: handle    = %s", handle);
 		pr_info("shmem: create    = %d", create);
 	}

@@ -545,7 +547,7 @@ panic_kvm:
 		current_kvm_cpu->kvm_run->exit_reason,
 		kvm_exit_reasons[current_kvm_cpu->kvm_run->exit_reason]);
 	if (current_kvm_cpu->kvm_run->exit_reason = KVM_EXIT_UNKNOWN)
-		fprintf(stderr, "KVM exit code: 0x%Lu\n",
+		fprintf(stderr, "KVM exit code: 0x%"PRIx64"\n",
 			current_kvm_cpu->kvm_run->hw.hardware_exit_reason);
 
 	kvm_cpu__set_debug_fd(STDOUT_FILENO);

@@ -760,10 +762,10 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 		ram_size	= get_ram_size(nrcpus);
 
 	if (ram_size < MIN_RAM_SIZE_MB)
-		die("Not enough memory specified: %lluMB (min %lluMB)", ram_size, MIN_RAM_SIZE_MB);
+		die("Not enough memory specified: %"PRIu64"MB (min %lluMB)", ram_size, MIN_RAM_SIZE_MB);
 
 	if (ram_size > host_ram_size())
-		pr_warning("Guest memory size %lluMB exceeds host physical RAM size %lluMB", ram_size, host_ram_size());
+		pr_warning("Guest memory size %"PRIu64"MB exceeds host physical RAM size %"PRIu64"MB", ram_size, host_ram_size());
 
 	ram_size <<= MB_SHIFT;
 

@@ -878,7 +880,7 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 		virtio_blk__init_all(kvm);
 	}
 
-	printf("  # kvm run -k %s -m %Lu -c %d --name %s\n", kernel_filename, ram_size / 1024 / 1024, nrcpus, guest_name);
+	printf("  # kvm run -k %s -m %"PRId64" -c %d --name %s\n", kernel_filename, ram_size / 1024 / 1024, nrcpus, guest_name);
 
 	if (!kvm__load_kernel(kvm, kernel_filename, initrd_filename,
 				real_cmdline, vidmode))

tools/kvm/builtin-stat.c +3 -1

index e28eb5b..c1f2605 100644
@@ -9,6 +9,8 @@
 #include <stdio.h>
 #include <string.h>
 #include <signal.h>
+#define __STDC_FORMAT_MACROS
+#include <inttypes.h>
 
 #include <linux/virtio_balloon.h>
 

@@ -97,7 +99,7 @@ static int do_memstat(const char *name, int sock)
 			printf("The total amount of memory available (in bytes):");
 			break;
 		}
-		printf("%llu\n", stats[i].val);
+		printf("%"PRId64"\n", stats[i].val);
 	}
 	printf("\n");
 

tools/kvm/disk/core.c +3 -1

index 4915efd..a135851 100644
@@ -4,6 +4,8 @@
 
 #include <sys/eventfd.h>
 #include <sys/poll.h>
+#define __STDC_FORMAT_MACROS
+#include <inttypes.h>
 
 #define AIO_MAX 32
 

@@ -232,7 +234,7 @@ ssize_t disk_image__get_serial(struct disk_image *disk, void *buffer, ssize_t *l
 	if (fstat(disk->fd, &st) != 0)
 		return 0;
 
-	*len = snprintf(buffer, *len, "%llu%llu%llu", (u64)st.st_dev, (u64)st.st_rdev, (u64)st.st_ino);
+	*len = snprintf(buffer, *len, "%"PRId64"%"PRId64"%"PRId64, (u64)st.st_dev, (u64)st.st_rdev, (u64)st.st_ino);
 	return *len;
 }
 

tools/kvm/mmio.c +3 -2

index de7320f..1158bff 100644
@@ -9,6 +9,8 @@
 #include <linux/kvm.h>
 #include <linux/types.h>
 #include <linux/rbtree.h>
+#define __STDC_FORMAT_MACROS
+#include <inttypes.h>
 
 #define mmio_node(n) rb_entry(n, struct mmio_mapping, node)
 

@@ -124,7 +126,7 @@ bool kvm__emulate_mmio(struct kvm *kvm, u64 phys_addr, u8 *data, u32 len, u8 is_
 	if (mmio)
 		mmio->mmio_fn(phys_addr, data, len, is_write, mmio->ptr);
 	else
-		fprintf(stderr, "Warning: Ignoring MMIO %s at %016llx (length %u)\n",
+		fprintf(stderr, "Warning: Ignoring MMIO %s at %016"PRIx64" (length %u)\n",
 			to_direction(is_write), phys_addr, len);
 	br_read_unlock();
 
--

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

Ingo actually got us to remove all the PRI* specifiers, but that was back when we only did x86 :)

Ingo, does it make sense to use them now when we support different architectures?

Quoted 3 lines · Sasha Levin

Ingo actually got us to remove all the PRI* specifiers, but that was back when we only did x86 :)

Quoted 2 lines

Ingo, does it make sense to use them now when we support different architectures?

Not at all - ppc should use a sane u64/s64 definition, i.e. int-ll64.h instead of the int-l64.h crap.

The powerpc maintainers indicated that they'd fix that, a couple of years ago, but nothing appears to have come out of that.

Quoted 12 lines · Ingo Molnar

Quoted 5 lines · Sasha Levin

Ingo actually got us to remove all the PRI* specifiers, but that was back when we only did x86 :)

Ingo, does it make sense to use them now when we support different architectures?

Not at all - ppc should use a sane u64/s64 definition, i.e. int-ll64.h instead of the int-l64.h crap.

Quoted 2 lines

The powerpc maintainers indicated that they'd fix that, a couple of years ago, but nothing appears to have come out of that.

We changed it for the kernel, but not for userspace (i.e. glibc) because of concerns about possibly breaking existing userspace programs. [...]

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

Quoted 13 lines · Paul Mackerras

Quoted 12 lines · Ingo Molnar

Quoted 5 lines · Sasha Levin

Ingo actually got us to remove all the PRI* specifiers, but that was back when we only did x86 :)

Ingo, does it make sense to use them now when we support different architectures?

Not at all - ppc should use a sane u64/s64 definition, i.e. int-ll64.h instead of the int-l64.h crap.

Quoted 6 lines

Quoted 2 lines

The powerpc maintainers indicated that they'd fix that, a couple of years ago, but nothing appears to have come out of that.

We changed it for the kernel, but not for userspace (i.e. glibc) because of concerns about possibly breaking existing userspace programs. [...]

Indeed, you do:

#if defined(__powerpc64__) && !defined(__KERNEL__)

 # include <asm-generic/int-l64.h>
 #else
 # include <asm-generic/int-ll64.h>
 #endif

which correctly uses int-ll64.h everywhere except 64-bit userspace inclusions. So i take back my 'nothing appears to have come out of that' comment - it's all nicely fixed!

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

Quoted 8 lines

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

Quoted 5 lines

Quoted 5 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

Quoted 2 lines

Quoted 2 lines

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Sane userspace should not be prevented from using the same sane types the kernel is already using:-)

Quoted 13 lines

Quoted 7 lines

Quoted 7 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Sane userspace should not be prevented from using the same sane types the kernel is already using:-)

I should dig up that patent of mine for "apparatus for conversion of circular motion to rectilinear ('wheel')".

Quoted 9 lines · Matt Evans

Quoted 8 lines

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

Quoted 7 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Sane userspace should not be prevented from using the same sane types the kernel is already using :-)

Quoted 24 lines · Ingo Molnar

Quoted 9 lines · Matt Evans

Quoted 8 lines

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

Quoted 7 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Quoted 2 lines

Sane userspace should not be prevented from using the same sane types the kernel is already using :-)

How does perf handle this? I'm sure it has the exact same issue, doesn't it?

Quoted 26 lines · Pekka Enberg

Quoted 24 lines · Ingo Molnar

Quoted 9 lines · Matt Evans

Quoted 8 lines

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

Quoted 7 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Quoted 5 lines

Quoted 2 lines

Sane userspace should not be prevented from using the same sane types the kernel is already using :-)

How does perf handle this? I'm sure it has the exact same issue, doesn't it?

It does; ironically it uses PRIblah, so I had followed its example.

Quoted 30 lines · Matt Evans

Quoted 29 lines · Pekka Enberg

Quoted 26 lines · Ingo Molnar

Quoted 9 lines · Matt Evans

Quoted 8 lines

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

Quoted 7 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Sane userspace should not be prevented from using the same sane types the kernel is already using :-)

Quoted 5 lines

Quoted 2 lines

How does perf handle this? I'm sure it has the exact same issue, doesn't it?

It does; ironically it uses PRIblah, so I had followed its example.

Sadly it regressed lately in that area - it was certainly PRIblah-less in the early stages :-)

Pekka, how do the headers react if we define __KERNEL__? Alternatively, allowing an override in powerpc/types.h beyond __KERNEL__ looks sensible as well.

Quoted 33 lines · Ingo Molnar

Quoted 30 lines · Matt Evans

Quoted 29 lines · Pekka Enberg

Quoted 26 lines · Ingo Molnar

Quoted 9 lines · Matt Evans

Quoted 8 lines

Quoted 3 lines

[...] I haven't looked closely at Matt's
patches, but it should be possible to use [un]signed long long for the u64/s64 types, I would think.

In tools/kvm/ we are using our own u64/s64 definitions, not glibc's, so i think it should be fine - as long as we don't pick up int-l64.h accidentally via the arch/powerpc/include/asm/types.h exception for user-space.

Quoted 7 lines

That's what's happening here; we're __powerpc64__ and !__KERNEL__, tools/kvm/include/linux/types.h includes asm/types.h so gets the int-l64.h definition of __u64, as above. :/

builtin-run.c:389: error: format %llx' expects type long long unsigned int', but argument 2 has type `u64'

So either define __KERNEL__ or patch a __NEW_USERSPACE__ define into power/asm/types.h and use it - if the PowerPC folks agree with that approach.

Sane userspace should not be prevented from using the same sane types the kernel is already using :-)

Quoted 8 lines

Quoted 5 lines

Quoted 2 lines

How does perf handle this? I'm sure it has the exact same issue, doesn't it?

It does; ironically it uses PRIblah, so I had followed its example.

Sadly it regressed lately in that area - it was certainly PRIblah-less in the early stages :-)

Oh dear :-)

Quoted 3 lines

Pekka, how do the headers react if we define __KERNEL__? Alternatively, allowing an override in powerpc/types.h beyond __KERNEL__ looks sensible as well.

Well, I just tried it and it ended in tears; various things bring in tons more (ppc) stuff from asm/, quite a few conflicts.

I've resorted to defining __KERNEL__ in linux/types.h *only* around #include <asm/types.h> (i.e. undefining it afterwards). This picks up the correct int-ll64.h on PPC64 and doesn't break everything else.

Since tools/kvm/include/linux/types.h only requires __u32, __u64 et al from <asm/types.h>, wouldn't it be most straightforward to just #include <asm-generic/int-ll64.h>? This avoids #define __KERNEL__ breaking other includes brought into userland, avoids changing system headers/distros, and includes the file we're really interested in on both x86 & PPC.

Quoted 7 lines · Matt Evans

Since tools/kvm/include/linux/types.h only requires __u32, __u64 et al from <asm/types.h>, wouldn't it be most straightforward to just #include <asm-generic/int-ll64.h>? This avoids #define __KERNEL__ breaking other includes brought into userland, avoids changing system headers/distros, and includes the file we're really interested in on both x86 & PPC.

No system headers need to be changed: you can just patch powerpc/types.h with the extra __SANE_USERSPACE__ define (or any suitable name) and the KVM tool should work as-is.

The beauty of tools/kvm/ being integrated into the kernel tree.

Really, if that works and if the PowerPC folks agree then we have an immediate, fully adequate, instantly applicable solution.