2008-11-05 17:29:27 +01:00
|
|
|
/*
|
|
|
|
* QEMU KVM support
|
|
|
|
*
|
|
|
|
* Copyright IBM, Corp. 2008
|
2008-11-24 20:36:26 +01:00
|
|
|
* Red Hat, Inc. 2008
|
2008-11-05 17:29:27 +01:00
|
|
|
*
|
|
|
|
* Authors:
|
|
|
|
* Anthony Liguori <aliguori@us.ibm.com>
|
2008-11-24 20:36:26 +01:00
|
|
|
* Glauber Costa <gcosta@redhat.com>
|
2008-11-05 17:29:27 +01:00
|
|
|
*
|
|
|
|
* This work is licensed under the terms of the GNU GPL, version 2 or later.
|
|
|
|
* See the COPYING file in the top-level directory.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/types.h>
|
|
|
|
#include <sys/ioctl.h>
|
|
|
|
#include <sys/mman.h>
|
2008-11-13 20:21:00 +01:00
|
|
|
#include <stdarg.h>
|
2008-11-05 17:29:27 +01:00
|
|
|
|
|
|
|
#include <linux/kvm.h>
|
|
|
|
|
|
|
|
#include "qemu-common.h"
|
|
|
|
#include "sysemu.h"
|
2009-05-02 00:29:37 +02:00
|
|
|
#include "hw/hw.h"
|
2009-03-12 21:12:48 +01:00
|
|
|
#include "gdbstub.h"
|
2008-11-05 17:29:27 +01:00
|
|
|
#include "kvm.h"
|
|
|
|
|
2008-12-09 21:09:57 +01:00
|
|
|
/* KVM uses PAGE_SIZE in it's definition of COALESCED_MMIO_MAX */
|
|
|
|
#define PAGE_SIZE TARGET_PAGE_SIZE
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
//#define DEBUG_KVM
|
|
|
|
|
|
|
|
#ifdef DEBUG_KVM
|
|
|
|
#define dprintf(fmt, ...) \
|
|
|
|
do { fprintf(stderr, fmt, ## __VA_ARGS__); } while (0)
|
|
|
|
#else
|
|
|
|
#define dprintf(fmt, ...) \
|
|
|
|
do { } while (0)
|
|
|
|
#endif
|
|
|
|
|
2008-11-19 18:41:58 +01:00
|
|
|
typedef struct KVMSlot
|
|
|
|
{
|
2009-10-01 23:12:16 +02:00
|
|
|
target_phys_addr_t start_addr;
|
|
|
|
ram_addr_t memory_size;
|
|
|
|
ram_addr_t phys_offset;
|
2008-11-19 18:41:58 +01:00
|
|
|
int slot;
|
|
|
|
int flags;
|
|
|
|
} KVMSlot;
|
2008-11-05 17:29:27 +01:00
|
|
|
|
2008-11-24 20:36:26 +01:00
|
|
|
typedef struct kvm_dirty_log KVMDirtyLog;
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
int kvm_allowed = 0;
|
|
|
|
|
|
|
|
struct KVMState
|
|
|
|
{
|
|
|
|
KVMSlot slots[32];
|
|
|
|
int fd;
|
|
|
|
int vmfd;
|
2008-12-09 21:09:57 +01:00
|
|
|
int coalesced_mmio;
|
2010-01-26 12:21:16 +01:00
|
|
|
#ifdef KVM_CAP_COALESCED_MMIO
|
|
|
|
struct kvm_coalesced_mmio_ring *coalesced_mmio_ring;
|
|
|
|
#endif
|
2009-05-01 20:42:15 +02:00
|
|
|
int broken_set_mem_region;
|
2009-05-01 20:52:46 +02:00
|
|
|
int migration_log;
|
2009-11-25 00:33:03 +01:00
|
|
|
int vcpu_events;
|
2009-03-12 21:12:48 +01:00
|
|
|
#ifdef KVM_CAP_SET_GUEST_DEBUG
|
|
|
|
struct kvm_sw_breakpoint_head kvm_sw_breakpoints;
|
|
|
|
#endif
|
2009-07-21 17:26:58 +02:00
|
|
|
int irqchip_in_kernel;
|
|
|
|
int pit_in_kernel;
|
2008-11-05 17:29:27 +01:00
|
|
|
};
|
|
|
|
|
|
|
|
static KVMState *kvm_state;
|
|
|
|
|
|
|
|
static KVMSlot *kvm_alloc_slot(KVMState *s)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(s->slots); i++) {
|
2008-11-18 16:41:18 +01:00
|
|
|
/* KVM private memory slots */
|
|
|
|
if (i >= 8 && i < 12)
|
|
|
|
continue;
|
2008-11-05 17:29:27 +01:00
|
|
|
if (s->slots[i].memory_size == 0)
|
|
|
|
return &s->slots[i];
|
|
|
|
}
|
|
|
|
|
2009-04-17 16:26:29 +02:00
|
|
|
fprintf(stderr, "%s: no free slot available\n", __func__);
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
|
|
|
static KVMSlot *kvm_lookup_matching_slot(KVMState *s,
|
2009-10-01 23:12:16 +02:00
|
|
|
target_phys_addr_t start_addr,
|
|
|
|
target_phys_addr_t end_addr)
|
2009-04-17 16:26:29 +02:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(s->slots); i++) {
|
|
|
|
KVMSlot *mem = &s->slots[i];
|
|
|
|
|
|
|
|
if (start_addr == mem->start_addr &&
|
|
|
|
end_addr == mem->start_addr + mem->memory_size) {
|
|
|
|
return mem;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2009-04-17 16:26:33 +02:00
|
|
|
/*
|
|
|
|
* Find overlapping slot with lowest start address
|
|
|
|
*/
|
|
|
|
static KVMSlot *kvm_lookup_overlapping_slot(KVMState *s,
|
2009-10-01 23:12:16 +02:00
|
|
|
target_phys_addr_t start_addr,
|
|
|
|
target_phys_addr_t end_addr)
|
2008-11-05 17:29:27 +01:00
|
|
|
{
|
2009-04-17 16:26:33 +02:00
|
|
|
KVMSlot *found = NULL;
|
2008-11-05 17:29:27 +01:00
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(s->slots); i++) {
|
|
|
|
KVMSlot *mem = &s->slots[i];
|
|
|
|
|
2009-04-17 16:26:33 +02:00
|
|
|
if (mem->memory_size == 0 ||
|
|
|
|
(found && found->start_addr < mem->start_addr)) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (end_addr > mem->start_addr &&
|
|
|
|
start_addr < mem->start_addr + mem->memory_size) {
|
|
|
|
found = mem;
|
|
|
|
}
|
2008-11-05 17:29:27 +01:00
|
|
|
}
|
|
|
|
|
2009-04-17 16:26:33 +02:00
|
|
|
return found;
|
2008-11-05 17:29:27 +01:00
|
|
|
}
|
|
|
|
|
2008-11-24 20:36:26 +01:00
|
|
|
static int kvm_set_user_memory_region(KVMState *s, KVMSlot *slot)
|
|
|
|
{
|
|
|
|
struct kvm_userspace_memory_region mem;
|
|
|
|
|
|
|
|
mem.slot = slot->slot;
|
|
|
|
mem.guest_phys_addr = slot->start_addr;
|
|
|
|
mem.memory_size = slot->memory_size;
|
2009-04-11 16:47:08 +02:00
|
|
|
mem.userspace_addr = (unsigned long)qemu_get_ram_ptr(slot->phys_offset);
|
2008-11-24 20:36:26 +01:00
|
|
|
mem.flags = slot->flags;
|
2009-05-01 20:52:46 +02:00
|
|
|
if (s->migration_log) {
|
|
|
|
mem.flags |= KVM_MEM_LOG_DIRTY_PAGES;
|
|
|
|
}
|
2008-11-24 20:36:26 +01:00
|
|
|
return kvm_vm_ioctl(s, KVM_SET_USER_MEMORY_REGION, &mem);
|
|
|
|
}
|
|
|
|
|
2009-06-27 09:24:58 +02:00
|
|
|
static void kvm_reset_vcpu(void *opaque)
|
|
|
|
{
|
|
|
|
CPUState *env = opaque;
|
|
|
|
|
2009-11-06 19:39:24 +01:00
|
|
|
kvm_arch_reset_vcpu(env);
|
2009-06-27 09:24:58 +02:00
|
|
|
if (kvm_arch_put_registers(env)) {
|
|
|
|
fprintf(stderr, "Fatal: kvm vcpu reset failed\n");
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
}
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-07-21 17:26:58 +02:00
|
|
|
int kvm_irqchip_in_kernel(void)
|
|
|
|
{
|
|
|
|
return kvm_state->irqchip_in_kernel;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_pit_in_kernel(void)
|
|
|
|
{
|
|
|
|
return kvm_state->pit_in_kernel;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
int kvm_init_vcpu(CPUState *env)
|
|
|
|
{
|
|
|
|
KVMState *s = kvm_state;
|
|
|
|
long mmap_size;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
dprintf("kvm_init_vcpu\n");
|
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
ret = kvm_vm_ioctl(s, KVM_CREATE_VCPU, env->cpu_index);
|
2008-11-05 17:29:27 +01:00
|
|
|
if (ret < 0) {
|
|
|
|
dprintf("kvm_create_vcpu failed\n");
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
env->kvm_fd = ret;
|
|
|
|
env->kvm_state = s;
|
|
|
|
|
|
|
|
mmap_size = kvm_ioctl(s, KVM_GET_VCPU_MMAP_SIZE, 0);
|
|
|
|
if (mmap_size < 0) {
|
|
|
|
dprintf("KVM_GET_VCPU_MMAP_SIZE failed\n");
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
env->kvm_run = mmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_SHARED,
|
|
|
|
env->kvm_fd, 0);
|
|
|
|
if (env->kvm_run == MAP_FAILED) {
|
|
|
|
ret = -errno;
|
|
|
|
dprintf("mmap'ing vcpu state failed\n");
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
2010-01-26 12:21:16 +01:00
|
|
|
#ifdef KVM_CAP_COALESCED_MMIO
|
|
|
|
if (s->coalesced_mmio && !s->coalesced_mmio_ring)
|
|
|
|
s->coalesced_mmio_ring = (void *) env->kvm_run +
|
|
|
|
s->coalesced_mmio * PAGE_SIZE;
|
|
|
|
#endif
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
ret = kvm_arch_init_vcpu(env);
|
2009-06-27 09:24:58 +02:00
|
|
|
if (ret == 0) {
|
2009-06-27 09:25:07 +02:00
|
|
|
qemu_register_reset(kvm_reset_vcpu, env);
|
2009-11-06 19:39:24 +01:00
|
|
|
kvm_arch_reset_vcpu(env);
|
2009-06-27 09:24:58 +02:00
|
|
|
ret = kvm_arch_put_registers(env);
|
|
|
|
}
|
2008-11-05 17:29:27 +01:00
|
|
|
err:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-11-24 20:36:26 +01:00
|
|
|
/*
|
|
|
|
* dirty pages logging control
|
|
|
|
*/
|
2009-10-01 23:12:16 +02:00
|
|
|
static int kvm_dirty_pages_log_change(target_phys_addr_t phys_addr,
|
|
|
|
ram_addr_t size, int flags, int mask)
|
2008-11-24 20:36:26 +01:00
|
|
|
{
|
|
|
|
KVMState *s = kvm_state;
|
2009-04-17 16:26:29 +02:00
|
|
|
KVMSlot *mem = kvm_lookup_matching_slot(s, phys_addr, phys_addr + size);
|
2009-05-01 20:52:46 +02:00
|
|
|
int old_flags;
|
|
|
|
|
2008-11-24 20:36:26 +01:00
|
|
|
if (mem == NULL) {
|
2009-04-17 16:26:29 +02:00
|
|
|
fprintf(stderr, "BUG: %s: invalid parameters " TARGET_FMT_plx "-"
|
|
|
|
TARGET_FMT_plx "\n", __func__, phys_addr,
|
2009-10-01 23:12:16 +02:00
|
|
|
(target_phys_addr_t)(phys_addr + size - 1));
|
2008-11-24 20:36:26 +01:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2009-05-01 20:52:46 +02:00
|
|
|
old_flags = mem->flags;
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-05-01 20:52:46 +02:00
|
|
|
flags = (mem->flags & ~mask) | flags;
|
2008-11-24 20:36:26 +01:00
|
|
|
mem->flags = flags;
|
|
|
|
|
2009-05-01 20:52:46 +02:00
|
|
|
/* If nothing changed effectively, no need to issue ioctl */
|
|
|
|
if (s->migration_log) {
|
|
|
|
flags |= KVM_MEM_LOG_DIRTY_PAGES;
|
|
|
|
}
|
|
|
|
if (flags == old_flags) {
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-11-24 20:36:26 +01:00
|
|
|
return kvm_set_user_memory_region(s, mem);
|
|
|
|
}
|
|
|
|
|
2009-10-01 23:12:16 +02:00
|
|
|
int kvm_log_start(target_phys_addr_t phys_addr, ram_addr_t size)
|
2008-11-24 20:36:26 +01:00
|
|
|
{
|
2009-04-17 16:26:29 +02:00
|
|
|
return kvm_dirty_pages_log_change(phys_addr, size,
|
2008-11-24 20:36:26 +01:00
|
|
|
KVM_MEM_LOG_DIRTY_PAGES,
|
|
|
|
KVM_MEM_LOG_DIRTY_PAGES);
|
|
|
|
}
|
|
|
|
|
2009-10-01 23:12:16 +02:00
|
|
|
int kvm_log_stop(target_phys_addr_t phys_addr, ram_addr_t size)
|
2008-11-24 20:36:26 +01:00
|
|
|
{
|
2009-04-17 16:26:29 +02:00
|
|
|
return kvm_dirty_pages_log_change(phys_addr, size,
|
2008-11-24 20:36:26 +01:00
|
|
|
0,
|
|
|
|
KVM_MEM_LOG_DIRTY_PAGES);
|
|
|
|
}
|
|
|
|
|
2010-01-27 21:07:21 +01:00
|
|
|
static int kvm_set_migration_log(int enable)
|
2009-05-01 20:52:46 +02:00
|
|
|
{
|
|
|
|
KVMState *s = kvm_state;
|
|
|
|
KVMSlot *mem;
|
|
|
|
int i, err;
|
|
|
|
|
|
|
|
s->migration_log = enable;
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(s->slots); i++) {
|
|
|
|
mem = &s->slots[i];
|
|
|
|
|
|
|
|
if (!!(mem->flags & KVM_MEM_LOG_DIRTY_PAGES) == enable) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
err = kvm_set_user_memory_region(s, mem);
|
|
|
|
if (err) {
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-07-27 12:49:56 +02:00
|
|
|
static int test_le_bit(unsigned long nr, unsigned char *addr)
|
|
|
|
{
|
|
|
|
return (addr[nr >> 3] >> (nr & 7)) & 1;
|
|
|
|
}
|
|
|
|
|
2008-11-24 20:36:26 +01:00
|
|
|
/**
|
|
|
|
* kvm_physical_sync_dirty_bitmap - Grab dirty bitmap from kernel space
|
|
|
|
* This function updates qemu's dirty bitmap using cpu_physical_memory_set_dirty().
|
|
|
|
* This means all bits are set to dirty.
|
|
|
|
*
|
2009-04-17 16:26:29 +02:00
|
|
|
* @start_add: start of logged region.
|
2008-11-24 20:36:26 +01:00
|
|
|
* @end_addr: end of logged region.
|
|
|
|
*/
|
2010-01-27 21:07:21 +01:00
|
|
|
static int kvm_physical_sync_dirty_bitmap(target_phys_addr_t start_addr,
|
|
|
|
target_phys_addr_t end_addr)
|
2008-11-24 20:36:26 +01:00
|
|
|
{
|
|
|
|
KVMState *s = kvm_state;
|
2009-05-01 20:52:47 +02:00
|
|
|
unsigned long size, allocated_size = 0;
|
2009-10-01 23:12:16 +02:00
|
|
|
target_phys_addr_t phys_addr;
|
|
|
|
ram_addr_t addr;
|
2009-05-01 20:52:47 +02:00
|
|
|
KVMDirtyLog d;
|
|
|
|
KVMSlot *mem;
|
|
|
|
int ret = 0;
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-05-01 20:52:47 +02:00
|
|
|
d.dirty_bitmap = NULL;
|
|
|
|
while (start_addr < end_addr) {
|
|
|
|
mem = kvm_lookup_overlapping_slot(s, start_addr, end_addr);
|
|
|
|
if (mem == NULL) {
|
|
|
|
break;
|
|
|
|
}
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-05-01 20:52:47 +02:00
|
|
|
size = ((mem->memory_size >> TARGET_PAGE_BITS) + 7) / 8;
|
|
|
|
if (!d.dirty_bitmap) {
|
|
|
|
d.dirty_bitmap = qemu_malloc(size);
|
|
|
|
} else if (size > allocated_size) {
|
|
|
|
d.dirty_bitmap = qemu_realloc(d.dirty_bitmap, size);
|
|
|
|
}
|
|
|
|
allocated_size = size;
|
|
|
|
memset(d.dirty_bitmap, 0, allocated_size);
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-05-01 20:52:47 +02:00
|
|
|
d.slot = mem->slot;
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-07-27 22:23:59 +02:00
|
|
|
if (kvm_vm_ioctl(s, KVM_GET_DIRTY_LOG, &d) == -1) {
|
2009-05-01 20:52:47 +02:00
|
|
|
dprintf("ioctl failed %d\n", errno);
|
|
|
|
ret = -1;
|
|
|
|
break;
|
|
|
|
}
|
2008-11-24 20:36:26 +01:00
|
|
|
|
2009-05-01 20:52:47 +02:00
|
|
|
for (phys_addr = mem->start_addr, addr = mem->phys_offset;
|
|
|
|
phys_addr < mem->start_addr + mem->memory_size;
|
|
|
|
phys_addr += TARGET_PAGE_SIZE, addr += TARGET_PAGE_SIZE) {
|
2009-07-27 12:49:56 +02:00
|
|
|
unsigned char *bitmap = (unsigned char *)d.dirty_bitmap;
|
2009-05-01 20:52:47 +02:00
|
|
|
unsigned nr = (phys_addr - mem->start_addr) >> TARGET_PAGE_BITS;
|
|
|
|
|
2009-07-27 12:49:56 +02:00
|
|
|
if (test_le_bit(nr, bitmap)) {
|
2009-05-01 20:52:47 +02:00
|
|
|
cpu_physical_memory_set_dirty(addr);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
start_addr = phys_addr;
|
2008-11-24 20:36:26 +01:00
|
|
|
}
|
|
|
|
qemu_free(d.dirty_bitmap);
|
2009-05-01 20:52:47 +02:00
|
|
|
|
|
|
|
return ret;
|
2008-11-24 20:36:26 +01:00
|
|
|
}
|
|
|
|
|
2009-10-01 23:12:16 +02:00
|
|
|
int kvm_coalesce_mmio_region(target_phys_addr_t start, ram_addr_t size)
|
2008-12-09 21:09:57 +01:00
|
|
|
{
|
|
|
|
int ret = -ENOSYS;
|
|
|
|
#ifdef KVM_CAP_COALESCED_MMIO
|
|
|
|
KVMState *s = kvm_state;
|
|
|
|
|
|
|
|
if (s->coalesced_mmio) {
|
|
|
|
struct kvm_coalesced_mmio_zone zone;
|
|
|
|
|
|
|
|
zone.addr = start;
|
|
|
|
zone.size = size;
|
|
|
|
|
|
|
|
ret = kvm_vm_ioctl(s, KVM_REGISTER_COALESCED_MMIO, &zone);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2009-10-01 23:12:16 +02:00
|
|
|
int kvm_uncoalesce_mmio_region(target_phys_addr_t start, ram_addr_t size)
|
2008-12-09 21:09:57 +01:00
|
|
|
{
|
|
|
|
int ret = -ENOSYS;
|
|
|
|
#ifdef KVM_CAP_COALESCED_MMIO
|
|
|
|
KVMState *s = kvm_state;
|
|
|
|
|
|
|
|
if (s->coalesced_mmio) {
|
|
|
|
struct kvm_coalesced_mmio_zone zone;
|
|
|
|
|
|
|
|
zone.addr = start;
|
|
|
|
zone.size = size;
|
|
|
|
|
|
|
|
ret = kvm_vm_ioctl(s, KVM_UNREGISTER_COALESCED_MMIO, &zone);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2009-05-08 22:33:24 +02:00
|
|
|
int kvm_check_extension(KVMState *s, unsigned int extension)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = kvm_ioctl(s, KVM_CHECK_EXTENSION, extension);
|
|
|
|
if (ret < 0) {
|
|
|
|
ret = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2010-01-27 21:07:21 +01:00
|
|
|
static void kvm_set_phys_mem(target_phys_addr_t start_addr,
|
|
|
|
ram_addr_t size,
|
|
|
|
ram_addr_t phys_offset)
|
2010-01-27 21:07:08 +01:00
|
|
|
{
|
|
|
|
KVMState *s = kvm_state;
|
|
|
|
ram_addr_t flags = phys_offset & ~TARGET_PAGE_MASK;
|
|
|
|
KVMSlot *mem, old;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (start_addr & ~TARGET_PAGE_MASK) {
|
|
|
|
if (flags >= IO_MEM_UNASSIGNED) {
|
|
|
|
if (!kvm_lookup_overlapping_slot(s, start_addr,
|
|
|
|
start_addr + size)) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
fprintf(stderr, "Unaligned split of a KVM memory slot\n");
|
|
|
|
} else {
|
|
|
|
fprintf(stderr, "Only page-aligned memory slots supported\n");
|
|
|
|
}
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
|
|
|
/* KVM does not support read-only slots */
|
|
|
|
phys_offset &= ~IO_MEM_ROM;
|
|
|
|
|
|
|
|
while (1) {
|
|
|
|
mem = kvm_lookup_overlapping_slot(s, start_addr, start_addr + size);
|
|
|
|
if (!mem) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (flags < IO_MEM_UNASSIGNED && start_addr >= mem->start_addr &&
|
|
|
|
(start_addr + size <= mem->start_addr + mem->memory_size) &&
|
|
|
|
(phys_offset - start_addr == mem->phys_offset - mem->start_addr)) {
|
|
|
|
/* The new slot fits into the existing one and comes with
|
|
|
|
* identical parameters - nothing to be done. */
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
old = *mem;
|
|
|
|
|
|
|
|
/* unregister the overlapping slot */
|
|
|
|
mem->memory_size = 0;
|
|
|
|
err = kvm_set_user_memory_region(s, mem);
|
|
|
|
if (err) {
|
|
|
|
fprintf(stderr, "%s: error unregistering overlapping slot: %s\n",
|
|
|
|
__func__, strerror(-err));
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Workaround for older KVM versions: we can't join slots, even not by
|
|
|
|
* unregistering the previous ones and then registering the larger
|
|
|
|
* slot. We have to maintain the existing fragmentation. Sigh.
|
|
|
|
*
|
|
|
|
* This workaround assumes that the new slot starts at the same
|
|
|
|
* address as the first existing one. If not or if some overlapping
|
|
|
|
* slot comes around later, we will fail (not seen in practice so far)
|
|
|
|
* - and actually require a recent KVM version. */
|
|
|
|
if (s->broken_set_mem_region &&
|
|
|
|
old.start_addr == start_addr && old.memory_size < size &&
|
|
|
|
flags < IO_MEM_UNASSIGNED) {
|
|
|
|
mem = kvm_alloc_slot(s);
|
|
|
|
mem->memory_size = old.memory_size;
|
|
|
|
mem->start_addr = old.start_addr;
|
|
|
|
mem->phys_offset = old.phys_offset;
|
|
|
|
mem->flags = 0;
|
|
|
|
|
|
|
|
err = kvm_set_user_memory_region(s, mem);
|
|
|
|
if (err) {
|
|
|
|
fprintf(stderr, "%s: error updating slot: %s\n", __func__,
|
|
|
|
strerror(-err));
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
|
|
|
start_addr += old.memory_size;
|
|
|
|
phys_offset += old.memory_size;
|
|
|
|
size -= old.memory_size;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* register prefix slot */
|
|
|
|
if (old.start_addr < start_addr) {
|
|
|
|
mem = kvm_alloc_slot(s);
|
|
|
|
mem->memory_size = start_addr - old.start_addr;
|
|
|
|
mem->start_addr = old.start_addr;
|
|
|
|
mem->phys_offset = old.phys_offset;
|
|
|
|
mem->flags = 0;
|
|
|
|
|
|
|
|
err = kvm_set_user_memory_region(s, mem);
|
|
|
|
if (err) {
|
|
|
|
fprintf(stderr, "%s: error registering prefix slot: %s\n",
|
|
|
|
__func__, strerror(-err));
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* register suffix slot */
|
|
|
|
if (old.start_addr + old.memory_size > start_addr + size) {
|
|
|
|
ram_addr_t size_delta;
|
|
|
|
|
|
|
|
mem = kvm_alloc_slot(s);
|
|
|
|
mem->start_addr = start_addr + size;
|
|
|
|
size_delta = mem->start_addr - old.start_addr;
|
|
|
|
mem->memory_size = old.memory_size - size_delta;
|
|
|
|
mem->phys_offset = old.phys_offset + size_delta;
|
|
|
|
mem->flags = 0;
|
|
|
|
|
|
|
|
err = kvm_set_user_memory_region(s, mem);
|
|
|
|
if (err) {
|
|
|
|
fprintf(stderr, "%s: error registering suffix slot: %s\n",
|
|
|
|
__func__, strerror(-err));
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* in case the KVM bug workaround already "consumed" the new slot */
|
|
|
|
if (!size)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/* KVM does not need to know about this memory */
|
|
|
|
if (flags >= IO_MEM_UNASSIGNED)
|
|
|
|
return;
|
|
|
|
|
|
|
|
mem = kvm_alloc_slot(s);
|
|
|
|
mem->memory_size = size;
|
|
|
|
mem->start_addr = start_addr;
|
|
|
|
mem->phys_offset = phys_offset;
|
|
|
|
mem->flags = 0;
|
|
|
|
|
|
|
|
err = kvm_set_user_memory_region(s, mem);
|
|
|
|
if (err) {
|
|
|
|
fprintf(stderr, "%s: error registering slot: %s\n", __func__,
|
|
|
|
strerror(-err));
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-01-27 21:07:21 +01:00
|
|
|
static void kvm_client_set_memory(struct CPUPhysMemoryClient *client,
|
|
|
|
target_phys_addr_t start_addr,
|
|
|
|
ram_addr_t size,
|
|
|
|
ram_addr_t phys_offset)
|
|
|
|
{
|
|
|
|
kvm_set_phys_mem(start_addr, size, phys_offset);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_client_sync_dirty_bitmap(struct CPUPhysMemoryClient *client,
|
|
|
|
target_phys_addr_t start_addr,
|
|
|
|
target_phys_addr_t end_addr)
|
|
|
|
{
|
|
|
|
return kvm_physical_sync_dirty_bitmap(start_addr, end_addr);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_client_migration_log(struct CPUPhysMemoryClient *client,
|
|
|
|
int enable)
|
|
|
|
{
|
|
|
|
return kvm_set_migration_log(enable);
|
|
|
|
}
|
|
|
|
|
|
|
|
static CPUPhysMemoryClient kvm_cpu_phys_memory_client = {
|
|
|
|
.set_memory = kvm_client_set_memory,
|
|
|
|
.sync_dirty_bitmap = kvm_client_sync_dirty_bitmap,
|
|
|
|
.migration_log = kvm_client_migration_log,
|
|
|
|
};
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
int kvm_init(int smp_cpus)
|
|
|
|
{
|
2009-06-07 11:30:25 +02:00
|
|
|
static const char upgrade_note[] =
|
|
|
|
"Please upgrade to at least kernel 2.6.29 or recent kvm-kmod\n"
|
|
|
|
"(see http://sourceforge.net/projects/kvm).\n";
|
2008-11-05 17:29:27 +01:00
|
|
|
KVMState *s;
|
|
|
|
int ret;
|
|
|
|
int i;
|
|
|
|
|
2009-05-12 13:43:19 +02:00
|
|
|
if (smp_cpus > 1) {
|
|
|
|
fprintf(stderr, "No SMP KVM support, use '-smp 1'\n");
|
2008-11-05 17:29:27 +01:00
|
|
|
return -EINVAL;
|
2009-05-12 13:43:19 +02:00
|
|
|
}
|
2008-11-05 17:29:27 +01:00
|
|
|
|
|
|
|
s = qemu_mallocz(sizeof(KVMState));
|
|
|
|
|
2009-03-12 21:12:48 +01:00
|
|
|
#ifdef KVM_CAP_SET_GUEST_DEBUG
|
2009-09-12 09:36:22 +02:00
|
|
|
QTAILQ_INIT(&s->kvm_sw_breakpoints);
|
2009-03-12 21:12:48 +01:00
|
|
|
#endif
|
2008-11-05 17:29:27 +01:00
|
|
|
for (i = 0; i < ARRAY_SIZE(s->slots); i++)
|
|
|
|
s->slots[i].slot = i;
|
|
|
|
|
|
|
|
s->vmfd = -1;
|
2009-12-02 12:24:42 +01:00
|
|
|
s->fd = qemu_open("/dev/kvm", O_RDWR);
|
2008-11-05 17:29:27 +01:00
|
|
|
if (s->fd == -1) {
|
|
|
|
fprintf(stderr, "Could not access KVM kernel module: %m\n");
|
|
|
|
ret = -errno;
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = kvm_ioctl(s, KVM_GET_API_VERSION, 0);
|
|
|
|
if (ret < KVM_API_VERSION) {
|
|
|
|
if (ret > 0)
|
|
|
|
ret = -EINVAL;
|
|
|
|
fprintf(stderr, "kvm version too old\n");
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ret > KVM_API_VERSION) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
fprintf(stderr, "kvm version not supported\n");
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
|
|
|
s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);
|
|
|
|
if (s->vmfd < 0)
|
|
|
|
goto err;
|
|
|
|
|
|
|
|
/* initially, KVM allocated its own memory and we had to jump through
|
|
|
|
* hooks to make phys_ram_base point to this. Modern versions of KVM
|
2009-04-11 16:47:08 +02:00
|
|
|
* just use a user allocated buffer so we can use regular pages
|
2008-11-05 17:29:27 +01:00
|
|
|
* unmodified. Make sure we have a sufficiently modern version of KVM.
|
|
|
|
*/
|
2009-05-08 22:33:24 +02:00
|
|
|
if (!kvm_check_extension(s, KVM_CAP_USER_MEMORY)) {
|
|
|
|
ret = -EINVAL;
|
2009-06-07 11:30:25 +02:00
|
|
|
fprintf(stderr, "kvm does not support KVM_CAP_USER_MEMORY\n%s",
|
|
|
|
upgrade_note);
|
2008-11-05 17:29:27 +01:00
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
2008-12-09 20:59:09 +01:00
|
|
|
/* There was a nasty bug in < kvm-80 that prevents memory slots from being
|
|
|
|
* destroyed properly. Since we rely on this capability, refuse to work
|
|
|
|
* with any kernel without this capability. */
|
2009-05-08 22:33:24 +02:00
|
|
|
if (!kvm_check_extension(s, KVM_CAP_DESTROY_MEMORY_REGION_WORKS)) {
|
|
|
|
ret = -EINVAL;
|
2008-12-09 20:59:09 +01:00
|
|
|
|
|
|
|
fprintf(stderr,
|
2009-06-07 11:30:25 +02:00
|
|
|
"KVM kernel module broken (DESTROY_MEMORY_REGION).\n%s",
|
|
|
|
upgrade_note);
|
2008-12-09 20:59:09 +01:00
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
2010-01-26 12:21:16 +01:00
|
|
|
s->coalesced_mmio = 0;
|
2008-12-09 21:09:57 +01:00
|
|
|
#ifdef KVM_CAP_COALESCED_MMIO
|
2009-05-08 22:33:24 +02:00
|
|
|
s->coalesced_mmio = kvm_check_extension(s, KVM_CAP_COALESCED_MMIO);
|
2010-01-26 12:21:16 +01:00
|
|
|
s->coalesced_mmio_ring = NULL;
|
2008-12-09 21:09:57 +01:00
|
|
|
#endif
|
|
|
|
|
2009-05-01 20:42:15 +02:00
|
|
|
s->broken_set_mem_region = 1;
|
|
|
|
#ifdef KVM_CAP_JOIN_MEMORY_REGIONS_WORKS
|
|
|
|
ret = kvm_ioctl(s, KVM_CHECK_EXTENSION, KVM_CAP_JOIN_MEMORY_REGIONS_WORKS);
|
|
|
|
if (ret > 0) {
|
|
|
|
s->broken_set_mem_region = 0;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2009-11-25 00:33:03 +01:00
|
|
|
s->vcpu_events = 0;
|
|
|
|
#ifdef KVM_CAP_VCPU_EVENTS
|
|
|
|
s->vcpu_events = kvm_check_extension(s, KVM_CAP_VCPU_EVENTS);
|
|
|
|
#endif
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
ret = kvm_arch_init(s, smp_cpus);
|
|
|
|
if (ret < 0)
|
|
|
|
goto err;
|
|
|
|
|
|
|
|
kvm_state = s;
|
2010-01-27 21:07:21 +01:00
|
|
|
cpu_register_phys_memory_client(&kvm_cpu_phys_memory_client);
|
2008-11-05 17:29:27 +01:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err:
|
|
|
|
if (s) {
|
|
|
|
if (s->vmfd != -1)
|
|
|
|
close(s->vmfd);
|
|
|
|
if (s->fd != -1)
|
|
|
|
close(s->fd);
|
|
|
|
}
|
|
|
|
qemu_free(s);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2009-09-20 18:05:47 +02:00
|
|
|
static int kvm_handle_io(uint16_t port, void *data, int direction, int size,
|
|
|
|
uint32_t count)
|
2008-11-05 17:29:27 +01:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
uint8_t *ptr = data;
|
|
|
|
|
|
|
|
for (i = 0; i < count; i++) {
|
|
|
|
if (direction == KVM_EXIT_IO_IN) {
|
|
|
|
switch (size) {
|
|
|
|
case 1:
|
2009-09-20 18:05:47 +02:00
|
|
|
stb_p(ptr, cpu_inb(port));
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
case 2:
|
2009-09-20 18:05:47 +02:00
|
|
|
stw_p(ptr, cpu_inw(port));
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
case 4:
|
2009-09-20 18:05:47 +02:00
|
|
|
stl_p(ptr, cpu_inl(port));
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
switch (size) {
|
|
|
|
case 1:
|
2009-09-20 18:05:47 +02:00
|
|
|
cpu_outb(port, ldub_p(ptr));
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
case 2:
|
2009-09-20 18:05:47 +02:00
|
|
|
cpu_outw(port, lduw_p(ptr));
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
case 4:
|
2009-09-20 18:05:47 +02:00
|
|
|
cpu_outl(port, ldl_p(ptr));
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
ptr += size;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2010-01-26 12:21:16 +01:00
|
|
|
void kvm_flush_coalesced_mmio_buffer(void)
|
2008-12-09 21:09:57 +01:00
|
|
|
{
|
|
|
|
#ifdef KVM_CAP_COALESCED_MMIO
|
|
|
|
KVMState *s = kvm_state;
|
2010-01-26 12:21:16 +01:00
|
|
|
if (s->coalesced_mmio_ring) {
|
|
|
|
struct kvm_coalesced_mmio_ring *ring = s->coalesced_mmio_ring;
|
2008-12-09 21:09:57 +01:00
|
|
|
while (ring->first != ring->last) {
|
|
|
|
struct kvm_coalesced_mmio *ent;
|
|
|
|
|
|
|
|
ent = &ring->coalesced_mmio[ring->first];
|
|
|
|
|
|
|
|
cpu_physical_memory_write(ent->phys_addr, ent->data, ent->len);
|
|
|
|
/* FIXME smp_wmb() */
|
|
|
|
ring->first = (ring->first + 1) % KVM_COALESCED_MMIO_MAX;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2009-08-17 22:19:53 +02:00
|
|
|
void kvm_cpu_synchronize_state(CPUState *env)
|
|
|
|
{
|
2010-02-03 21:17:05 +01:00
|
|
|
if (!env->kvm_vcpu_dirty) {
|
2009-08-17 22:19:53 +02:00
|
|
|
kvm_arch_get_registers(env);
|
2010-02-03 21:17:05 +01:00
|
|
|
env->kvm_vcpu_dirty = 1;
|
2009-08-17 22:19:53 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
int kvm_cpu_exec(CPUState *env)
|
|
|
|
{
|
|
|
|
struct kvm_run *run = env->kvm_run;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
dprintf("kvm_cpu_exec()\n");
|
|
|
|
|
|
|
|
do {
|
2009-03-06 22:48:00 +01:00
|
|
|
if (env->exit_request) {
|
2008-11-05 17:29:27 +01:00
|
|
|
dprintf("interrupt exit requested\n");
|
|
|
|
ret = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2010-02-03 21:17:05 +01:00
|
|
|
if (env->kvm_vcpu_dirty) {
|
2009-08-17 22:19:53 +02:00
|
|
|
kvm_arch_put_registers(env);
|
2010-02-03 21:17:05 +01:00
|
|
|
env->kvm_vcpu_dirty = 0;
|
2009-08-17 22:19:53 +02:00
|
|
|
}
|
|
|
|
|
2009-05-30 10:01:45 +02:00
|
|
|
kvm_arch_pre_run(env, run);
|
2009-10-07 21:38:03 +02:00
|
|
|
qemu_mutex_unlock_iothread();
|
2008-11-05 17:29:27 +01:00
|
|
|
ret = kvm_vcpu_ioctl(env, KVM_RUN, 0);
|
2009-10-07 21:38:03 +02:00
|
|
|
qemu_mutex_lock_iothread();
|
2008-11-05 17:29:27 +01:00
|
|
|
kvm_arch_post_run(env, run);
|
|
|
|
|
|
|
|
if (ret == -EINTR || ret == -EAGAIN) {
|
|
|
|
dprintf("io window exit\n");
|
|
|
|
ret = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ret < 0) {
|
|
|
|
dprintf("kvm run failed %s\n", strerror(-ret));
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
2010-01-26 12:21:16 +01:00
|
|
|
kvm_flush_coalesced_mmio_buffer();
|
2008-12-09 21:09:57 +01:00
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
ret = 0; /* exit loop */
|
|
|
|
switch (run->exit_reason) {
|
|
|
|
case KVM_EXIT_IO:
|
|
|
|
dprintf("handle_io\n");
|
2009-09-20 18:05:47 +02:00
|
|
|
ret = kvm_handle_io(run->io.port,
|
2008-11-05 17:29:27 +01:00
|
|
|
(uint8_t *)run + run->io.data_offset,
|
|
|
|
run->io.direction,
|
|
|
|
run->io.size,
|
|
|
|
run->io.count);
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_MMIO:
|
|
|
|
dprintf("handle_mmio\n");
|
|
|
|
cpu_physical_memory_rw(run->mmio.phys_addr,
|
|
|
|
run->mmio.data,
|
|
|
|
run->mmio.len,
|
|
|
|
run->mmio.is_write);
|
|
|
|
ret = 1;
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_IRQ_WINDOW_OPEN:
|
|
|
|
dprintf("irq_window_open\n");
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_SHUTDOWN:
|
|
|
|
dprintf("shutdown\n");
|
|
|
|
qemu_system_reset_request();
|
|
|
|
ret = 1;
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_UNKNOWN:
|
|
|
|
dprintf("kvm_exit_unknown\n");
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_FAIL_ENTRY:
|
|
|
|
dprintf("kvm_exit_fail_entry\n");
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_EXCEPTION:
|
|
|
|
dprintf("kvm_exit_exception\n");
|
|
|
|
break;
|
|
|
|
case KVM_EXIT_DEBUG:
|
|
|
|
dprintf("kvm_exit_debug\n");
|
2009-03-12 21:12:48 +01:00
|
|
|
#ifdef KVM_CAP_SET_GUEST_DEBUG
|
|
|
|
if (kvm_arch_debug(&run->debug.arch)) {
|
|
|
|
gdb_set_stop_cpu(env);
|
|
|
|
vm_stop(EXCP_DEBUG);
|
|
|
|
env->exception_index = EXCP_DEBUG;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
/* re-enter, this exception was guest-internal */
|
|
|
|
ret = 1;
|
|
|
|
#endif /* KVM_CAP_SET_GUEST_DEBUG */
|
2008-11-05 17:29:27 +01:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
dprintf("kvm_arch_handle_exit\n");
|
|
|
|
ret = kvm_arch_handle_exit(env, run);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
} while (ret > 0);
|
|
|
|
|
2009-03-06 22:48:00 +01:00
|
|
|
if (env->exit_request) {
|
|
|
|
env->exit_request = 0;
|
2008-11-10 16:55:14 +01:00
|
|
|
env->exception_index = EXCP_INTERRUPT;
|
|
|
|
}
|
|
|
|
|
2008-11-05 17:29:27 +01:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
int kvm_ioctl(KVMState *s, int type, ...)
|
2008-11-05 17:29:27 +01:00
|
|
|
{
|
|
|
|
int ret;
|
2008-11-13 20:21:00 +01:00
|
|
|
void *arg;
|
|
|
|
va_list ap;
|
2008-11-05 17:29:27 +01:00
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
va_start(ap, type);
|
|
|
|
arg = va_arg(ap, void *);
|
|
|
|
va_end(ap);
|
|
|
|
|
|
|
|
ret = ioctl(s->fd, type, arg);
|
2008-11-05 17:29:27 +01:00
|
|
|
if (ret == -1)
|
|
|
|
ret = -errno;
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
int kvm_vm_ioctl(KVMState *s, int type, ...)
|
2008-11-05 17:29:27 +01:00
|
|
|
{
|
|
|
|
int ret;
|
2008-11-13 20:21:00 +01:00
|
|
|
void *arg;
|
|
|
|
va_list ap;
|
|
|
|
|
|
|
|
va_start(ap, type);
|
|
|
|
arg = va_arg(ap, void *);
|
|
|
|
va_end(ap);
|
2008-11-05 17:29:27 +01:00
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
ret = ioctl(s->vmfd, type, arg);
|
2008-11-05 17:29:27 +01:00
|
|
|
if (ret == -1)
|
|
|
|
ret = -errno;
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
int kvm_vcpu_ioctl(CPUState *env, int type, ...)
|
2008-11-05 17:29:27 +01:00
|
|
|
{
|
|
|
|
int ret;
|
2008-11-13 20:21:00 +01:00
|
|
|
void *arg;
|
|
|
|
va_list ap;
|
|
|
|
|
|
|
|
va_start(ap, type);
|
|
|
|
arg = va_arg(ap, void *);
|
|
|
|
va_end(ap);
|
2008-11-05 17:29:27 +01:00
|
|
|
|
2008-11-13 20:21:00 +01:00
|
|
|
ret = ioctl(env->kvm_fd, type, arg);
|
2008-11-05 17:29:27 +01:00
|
|
|
if (ret == -1)
|
|
|
|
ret = -errno;
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
2008-12-04 21:33:06 +01:00
|
|
|
|
|
|
|
int kvm_has_sync_mmu(void)
|
|
|
|
{
|
2008-12-18 23:42:51 +01:00
|
|
|
#ifdef KVM_CAP_SYNC_MMU
|
2008-12-04 21:33:06 +01:00
|
|
|
KVMState *s = kvm_state;
|
|
|
|
|
2009-05-08 22:33:24 +02:00
|
|
|
return kvm_check_extension(s, KVM_CAP_SYNC_MMU);
|
|
|
|
#else
|
2008-12-04 21:33:06 +01:00
|
|
|
return 0;
|
2009-05-08 22:33:24 +02:00
|
|
|
#endif
|
2008-12-04 21:33:06 +01:00
|
|
|
}
|
2009-03-12 21:12:48 +01:00
|
|
|
|
2009-11-25 00:33:03 +01:00
|
|
|
int kvm_has_vcpu_events(void)
|
|
|
|
{
|
|
|
|
return kvm_state->vcpu_events;
|
|
|
|
}
|
|
|
|
|
2009-04-26 18:03:40 +02:00
|
|
|
void kvm_setup_guest_memory(void *start, size_t size)
|
|
|
|
{
|
|
|
|
if (!kvm_has_sync_mmu()) {
|
|
|
|
#ifdef MADV_DONTFORK
|
|
|
|
int ret = madvise(start, size, MADV_DONTFORK);
|
|
|
|
|
|
|
|
if (ret) {
|
|
|
|
perror("madvice");
|
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
fprintf(stderr,
|
|
|
|
"Need MADV_DONTFORK in absence of synchronous KVM MMU\n");
|
|
|
|
exit(1);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-03-12 21:12:48 +01:00
|
|
|
#ifdef KVM_CAP_SET_GUEST_DEBUG
|
2009-07-22 20:32:52 +02:00
|
|
|
static void on_vcpu(CPUState *env, void (*func)(void *data), void *data)
|
|
|
|
{
|
2009-09-17 20:10:06 +02:00
|
|
|
#ifdef CONFIG_IOTHREAD
|
2010-02-04 15:46:22 +01:00
|
|
|
if (env != cpu_single_env) {
|
|
|
|
abort();
|
2009-07-22 20:32:52 +02:00
|
|
|
}
|
2009-09-17 20:10:06 +02:00
|
|
|
#endif
|
2010-02-04 15:46:22 +01:00
|
|
|
func(data);
|
2009-07-22 20:32:52 +02:00
|
|
|
}
|
|
|
|
|
2009-03-12 21:12:48 +01:00
|
|
|
struct kvm_sw_breakpoint *kvm_find_sw_breakpoint(CPUState *env,
|
|
|
|
target_ulong pc)
|
|
|
|
{
|
|
|
|
struct kvm_sw_breakpoint *bp;
|
|
|
|
|
2009-09-12 09:36:22 +02:00
|
|
|
QTAILQ_FOREACH(bp, &env->kvm_state->kvm_sw_breakpoints, entry) {
|
2009-03-12 21:12:48 +01:00
|
|
|
if (bp->pc == pc)
|
|
|
|
return bp;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_sw_breakpoints_active(CPUState *env)
|
|
|
|
{
|
2009-09-12 09:36:22 +02:00
|
|
|
return !QTAILQ_EMPTY(&env->kvm_state->kvm_sw_breakpoints);
|
2009-03-12 21:12:48 +01:00
|
|
|
}
|
|
|
|
|
2009-07-16 23:55:28 +02:00
|
|
|
struct kvm_set_guest_debug_data {
|
|
|
|
struct kvm_guest_debug dbg;
|
|
|
|
CPUState *env;
|
|
|
|
int err;
|
|
|
|
};
|
|
|
|
|
|
|
|
static void kvm_invoke_set_guest_debug(void *data)
|
|
|
|
{
|
|
|
|
struct kvm_set_guest_debug_data *dbg_data = data;
|
2009-09-17 20:05:58 +02:00
|
|
|
CPUState *env = dbg_data->env;
|
|
|
|
|
2010-02-03 21:17:05 +01:00
|
|
|
if (env->kvm_vcpu_dirty) {
|
2009-09-17 20:05:58 +02:00
|
|
|
kvm_arch_put_registers(env);
|
2010-02-03 21:17:05 +01:00
|
|
|
env->kvm_vcpu_dirty = 0;
|
2009-09-17 20:05:58 +02:00
|
|
|
}
|
|
|
|
dbg_data->err = kvm_vcpu_ioctl(env, KVM_SET_GUEST_DEBUG, &dbg_data->dbg);
|
2009-07-16 23:55:28 +02:00
|
|
|
}
|
|
|
|
|
2009-03-12 21:12:48 +01:00
|
|
|
int kvm_update_guest_debug(CPUState *env, unsigned long reinject_trap)
|
|
|
|
{
|
2009-07-16 23:55:28 +02:00
|
|
|
struct kvm_set_guest_debug_data data;
|
2009-03-12 21:12:48 +01:00
|
|
|
|
2009-07-16 23:55:28 +02:00
|
|
|
data.dbg.control = 0;
|
2009-03-12 21:12:48 +01:00
|
|
|
if (env->singlestep_enabled)
|
2009-07-16 23:55:28 +02:00
|
|
|
data.dbg.control = KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_SINGLESTEP;
|
2009-03-12 21:12:48 +01:00
|
|
|
|
2009-07-16 23:55:28 +02:00
|
|
|
kvm_arch_update_guest_debug(env, &data.dbg);
|
|
|
|
data.dbg.control |= reinject_trap;
|
|
|
|
data.env = env;
|
2009-03-12 21:12:48 +01:00
|
|
|
|
2009-07-16 23:55:28 +02:00
|
|
|
on_vcpu(env, kvm_invoke_set_guest_debug, &data);
|
|
|
|
return data.err;
|
2009-03-12 21:12:48 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_insert_breakpoint(CPUState *current_env, target_ulong addr,
|
|
|
|
target_ulong len, int type)
|
|
|
|
{
|
|
|
|
struct kvm_sw_breakpoint *bp;
|
|
|
|
CPUState *env;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (type == GDB_BREAKPOINT_SW) {
|
|
|
|
bp = kvm_find_sw_breakpoint(current_env, addr);
|
|
|
|
if (bp) {
|
|
|
|
bp->use_count++;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
bp = qemu_malloc(sizeof(struct kvm_sw_breakpoint));
|
|
|
|
if (!bp)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
bp->pc = addr;
|
|
|
|
bp->use_count = 1;
|
|
|
|
err = kvm_arch_insert_sw_breakpoint(current_env, bp);
|
|
|
|
if (err) {
|
|
|
|
free(bp);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2009-09-12 09:36:22 +02:00
|
|
|
QTAILQ_INSERT_HEAD(¤t_env->kvm_state->kvm_sw_breakpoints,
|
2009-03-12 21:12:48 +01:00
|
|
|
bp, entry);
|
|
|
|
} else {
|
|
|
|
err = kvm_arch_insert_hw_breakpoint(addr, len, type);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (env = first_cpu; env != NULL; env = env->next_cpu) {
|
|
|
|
err = kvm_update_guest_debug(env, 0);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_remove_breakpoint(CPUState *current_env, target_ulong addr,
|
|
|
|
target_ulong len, int type)
|
|
|
|
{
|
|
|
|
struct kvm_sw_breakpoint *bp;
|
|
|
|
CPUState *env;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (type == GDB_BREAKPOINT_SW) {
|
|
|
|
bp = kvm_find_sw_breakpoint(current_env, addr);
|
|
|
|
if (!bp)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
if (bp->use_count > 1) {
|
|
|
|
bp->use_count--;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
err = kvm_arch_remove_sw_breakpoint(current_env, bp);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
2009-09-12 09:36:22 +02:00
|
|
|
QTAILQ_REMOVE(¤t_env->kvm_state->kvm_sw_breakpoints, bp, entry);
|
2009-03-12 21:12:48 +01:00
|
|
|
qemu_free(bp);
|
|
|
|
} else {
|
|
|
|
err = kvm_arch_remove_hw_breakpoint(addr, len, type);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (env = first_cpu; env != NULL; env = env->next_cpu) {
|
|
|
|
err = kvm_update_guest_debug(env, 0);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_remove_all_breakpoints(CPUState *current_env)
|
|
|
|
{
|
|
|
|
struct kvm_sw_breakpoint *bp, *next;
|
|
|
|
KVMState *s = current_env->kvm_state;
|
|
|
|
CPUState *env;
|
|
|
|
|
2009-09-12 09:36:22 +02:00
|
|
|
QTAILQ_FOREACH_SAFE(bp, &s->kvm_sw_breakpoints, entry, next) {
|
2009-03-12 21:12:48 +01:00
|
|
|
if (kvm_arch_remove_sw_breakpoint(current_env, bp) != 0) {
|
|
|
|
/* Try harder to find a CPU that currently sees the breakpoint. */
|
|
|
|
for (env = first_cpu; env != NULL; env = env->next_cpu) {
|
|
|
|
if (kvm_arch_remove_sw_breakpoint(env, bp) == 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
kvm_arch_remove_all_hw_breakpoints();
|
|
|
|
|
|
|
|
for (env = first_cpu; env != NULL; env = env->next_cpu)
|
|
|
|
kvm_update_guest_debug(env, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
#else /* !KVM_CAP_SET_GUEST_DEBUG */
|
|
|
|
|
|
|
|
int kvm_update_guest_debug(CPUState *env, unsigned long reinject_trap)
|
|
|
|
{
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_insert_breakpoint(CPUState *current_env, target_ulong addr,
|
|
|
|
target_ulong len, int type)
|
|
|
|
{
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_remove_breakpoint(CPUState *current_env, target_ulong addr,
|
|
|
|
target_ulong len, int type)
|
|
|
|
{
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_remove_all_breakpoints(CPUState *current_env)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
#endif /* !KVM_CAP_SET_GUEST_DEBUG */
|