Archipelago » qemu

/*

 * QEMU KVM support

 *

 * Copyright IBM, Corp. 2008

 *           Red Hat, Inc. 2008

 *

 * Authors:

 *  Anthony Liguori   <aliguori@us.ibm.com>

 *  Glauber Costa     <gcosta@redhat.com>

 *

 * 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>

#include <stdarg.h>

#include <linux/kvm.h>

#include "qemu-common.h"

#include "qemu/atomic.h"

#include "qemu/option.h"

#include "qemu/config-file.h"

#include "sysemu/sysemu.h"

#include "hw/hw.h"

#include "hw/pci/msi.h"

#include "exec/gdbstub.h"

#include "sysemu/kvm.h"

#include "qemu/bswap.h"

#include "exec/memory.h"

#include "exec/address-spaces.h"

#include "qemu/event_notifier.h"

#include "trace.h"

/* This check must be after config-host.h is included */

#ifdef CONFIG_EVENTFD

#include <sys/eventfd.h>

#endif

#ifdef CONFIG_VALGRIND_H

#include <valgrind/memcheck.h>

#endif

/* KVM uses PAGE_SIZE in its definition of COALESCED_MMIO_MAX */

#define PAGE_SIZE TARGET_PAGE_SIZE

//#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

#define KVM_MSI_HASHTAB_SIZE    256

typedef struct KVMSlot

{

    hwaddr start_addr;

    ram_addr_t memory_size;

    void *ram;

    int slot;

    int flags;

} KVMSlot;

typedef struct kvm_dirty_log KVMDirtyLog;

struct KVMState

{

    KVMSlot slots[32];

    int fd;

    int vmfd;

    int coalesced_mmio;

    struct kvm_coalesced_mmio_ring *coalesced_mmio_ring;

    bool coalesced_flush_in_progress;

    int broken_set_mem_region;

    int migration_log;

    int vcpu_events;

    int robust_singlestep;

    int debugregs;

#ifdef KVM_CAP_SET_GUEST_DEBUG

    struct kvm_sw_breakpoint_head kvm_sw_breakpoints;

#endif

    int pit_state2;

    int xsave, xcrs;

    int many_ioeventfds;

    int intx_set_mask;

    /* The man page (and posix) say ioctl numbers are signed int, but

     * they're not.  Linux, glibc and *BSD all treat ioctl numbers as

     * unsigned, and treating them as signed here can break things */

    unsigned irq_set_ioctl;

#ifdef KVM_CAP_IRQ_ROUTING

    struct kvm_irq_routing *irq_routes;

    int nr_allocated_irq_routes;

    uint32_t *used_gsi_bitmap;

    unsigned int gsi_count;

    QTAILQ_HEAD(msi_hashtab, KVMMSIRoute) msi_hashtab[KVM_MSI_HASHTAB_SIZE];

    bool direct_msi;

#endif

};

KVMState *kvm_state;

bool kvm_kernel_irqchip;

bool kvm_async_interrupts_allowed;

bool kvm_irqfds_allowed;

bool kvm_msi_via_irqfd_allowed;

bool kvm_gsi_routing_allowed;

bool kvm_allowed;

bool kvm_readonly_mem_allowed;

static const KVMCapabilityInfo kvm_required_capabilites[] = {

    KVM_CAP_INFO(USER_MEMORY),

    KVM_CAP_INFO(DESTROY_MEMORY_REGION_WORKS),

    KVM_CAP_LAST_INFO

};

static KVMSlot *kvm_alloc_slot(KVMState *s)

{

    int i;

    for (i = 0; i < ARRAY_SIZE(s->slots); i++) {

        if (s->slots[i].memory_size == 0) {

            return &s->slots[i];

        }

    }

    fprintf(stderr, "%s: no free slot available\n", __func__);

    abort();

}

static KVMSlot *kvm_lookup_matching_slot(KVMState *s,

                                         hwaddr start_addr,

                                         hwaddr end_addr)

{

    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;

        }

    }

    return NULL;

}

/*

 * Find overlapping slot with lowest start address

 */

static KVMSlot *kvm_lookup_overlapping_slot(KVMState *s,

                                            hwaddr start_addr,

                                            hwaddr end_addr)

{

    KVMSlot *found = NULL;

    int i;

    for (i = 0; i < ARRAY_SIZE(s->slots); i++) {

        KVMSlot *mem = &s->slots[i];

        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;

        }

    }

    return found;

}

int kvm_physical_memory_addr_from_host(KVMState *s, void *ram,

                                       hwaddr *phys_addr)

{

    int i;

    for (i = 0; i < ARRAY_SIZE(s->slots); i++) {

        KVMSlot *mem = &s->slots[i];

        if (ram >= mem->ram && ram < mem->ram + mem->memory_size) {

            *phys_addr = mem->start_addr + (ram - mem->ram);

            return 1;

        }

    }

    return 0;

}

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.userspace_addr = (unsigned long)slot->ram;

    mem.flags = slot->flags;

    if (s->migration_log) {

        mem.flags |= KVM_MEM_LOG_DIRTY_PAGES;

    }

    if (slot->memory_size && mem.flags & KVM_MEM_READONLY) {

        /* Set the slot size to 0 before setting the slot to the desired

         * value. This is needed based on KVM commit 75d61fbc. */

        mem.memory_size = 0;

        kvm_vm_ioctl(s, KVM_SET_USER_MEMORY_REGION, &mem);

    }

    mem.memory_size = slot->memory_size;

    return kvm_vm_ioctl(s, KVM_SET_USER_MEMORY_REGION, &mem);

}

static void kvm_reset_vcpu(void *opaque)

{

    CPUState *cpu = opaque;

    kvm_arch_reset_vcpu(cpu);

}

int kvm_init_vcpu(CPUState *cpu)

{

    KVMState *s = kvm_state;

    long mmap_size;

    int ret;

    DPRINTF("kvm_init_vcpu\n");

    ret = kvm_vm_ioctl(s, KVM_CREATE_VCPU, (void *)kvm_arch_vcpu_id(cpu));

    if (ret < 0) {

        DPRINTF("kvm_create_vcpu failed\n");

        goto err;

    }

    cpu->kvm_fd = ret;

    cpu->kvm_state = s;

    cpu->kvm_vcpu_dirty = true;

    mmap_size = kvm_ioctl(s, KVM_GET_VCPU_MMAP_SIZE, 0);

    if (mmap_size < 0) {

        ret = mmap_size;

        DPRINTF("KVM_GET_VCPU_MMAP_SIZE failed\n");

        goto err;

    }

    cpu->kvm_run = mmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_SHARED,

                        cpu->kvm_fd, 0);

    if (cpu->kvm_run == MAP_FAILED) {

        ret = -errno;

        DPRINTF("mmap'ing vcpu state failed\n");

        goto err;

    }

    if (s->coalesced_mmio && !s->coalesced_mmio_ring) {

        s->coalesced_mmio_ring =

            (void *)cpu->kvm_run + s->coalesced_mmio * PAGE_SIZE;

    }

    ret = kvm_arch_init_vcpu(cpu);

    if (ret == 0) {

        qemu_register_reset(kvm_reset_vcpu, cpu);

        kvm_arch_reset_vcpu(cpu);

    }

err:

    return ret;

}

/*

 * dirty pages logging control

 */

static int kvm_mem_flags(KVMState *s, bool log_dirty, bool readonly)

{

    int flags = 0;

    flags = log_dirty ? KVM_MEM_LOG_DIRTY_PAGES : 0;

    if (readonly && kvm_readonly_mem_allowed) {

        flags |= KVM_MEM_READONLY;

    }

    return flags;

}

static int kvm_slot_dirty_pages_log_change(KVMSlot *mem, bool log_dirty)

{

    KVMState *s = kvm_state;

    int flags, mask = KVM_MEM_LOG_DIRTY_PAGES;

    int old_flags;

    old_flags = mem->flags;

    flags = (mem->flags & ~mask) | kvm_mem_flags(s, log_dirty, false);

    mem->flags = flags;

    /* 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;

    }

    return kvm_set_user_memory_region(s, mem);

}

static int kvm_dirty_pages_log_change(hwaddr phys_addr,

                                      ram_addr_t size, bool log_dirty)

{

    KVMState *s = kvm_state;

    KVMSlot *mem = kvm_lookup_matching_slot(s, phys_addr, phys_addr + size);

    if (mem == NULL)  {

        fprintf(stderr, "BUG: %s: invalid parameters " TARGET_FMT_plx "-"

                TARGET_FMT_plx "\n", __func__, phys_addr,

                (hwaddr)(phys_addr + size - 1));

        return -EINVAL;

    }

    return kvm_slot_dirty_pages_log_change(mem, log_dirty);

}

static void kvm_log_start(MemoryListener *listener,

                          MemoryRegionSection *section)

{

    int r;

    r = kvm_dirty_pages_log_change(section->offset_within_address_space,

                                   int128_get64(section->size), true);

    if (r < 0) {

        abort();

    }

}

static void kvm_log_stop(MemoryListener *listener,

                          MemoryRegionSection *section)

{

    int r;

    r = kvm_dirty_pages_log_change(section->offset_within_address_space,

                                   int128_get64(section->size), false);

    if (r < 0) {

        abort();

    }

}

static int kvm_set_migration_log(int enable)

{

    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->memory_size) {

            continue;

        }

        if (!!(mem->flags & KVM_MEM_LOG_DIRTY_PAGES) == enable) {

            continue;

        }

        err = kvm_set_user_memory_region(s, mem);

        if (err) {

            return err;

        }

    }

    return 0;

}

/* get kvm's dirty pages bitmap and update qemu's */

static int kvm_get_dirty_pages_log_range(MemoryRegionSection *section,

                                         unsigned long *bitmap)

{

    unsigned int i, j;

    unsigned long page_number, c;

    hwaddr addr, addr1;

    unsigned int pages = int128_get64(section->size) / getpagesize();

    unsigned int len = (pages + HOST_LONG_BITS - 1) / HOST_LONG_BITS;

    unsigned long hpratio = getpagesize() / TARGET_PAGE_SIZE;

    /*

     * bitmap-traveling is faster than memory-traveling (for addr...)

     * especially when most of the memory is not dirty.

     */

    for (i = 0; i < len; i++) {

        if (bitmap[i] != 0) {

            c = leul_to_cpu(bitmap[i]);

            do {

                j = ffsl(c) - 1;

                c &= ~(1ul << j);

                page_number = (i * HOST_LONG_BITS + j) * hpratio;

                addr1 = page_number * TARGET_PAGE_SIZE;

                addr = section->offset_within_region + addr1;

                memory_region_set_dirty(section->mr, addr,

                                        TARGET_PAGE_SIZE * hpratio);

            } while (c != 0);

        }

    }

    return 0;

}

#define ALIGN(x, y)  (((x)+(y)-1) & ~((y)-1))

/**

 * kvm_physical_sync_dirty_bitmap - Grab dirty bitmap from kernel space

 * This function updates qemu's dirty bitmap using

 * memory_region_set_dirty().  This means all bits are set

 * to dirty.

 *

 * @start_add: start of logged region.

 * @end_addr: end of logged region.

 */

static int kvm_physical_sync_dirty_bitmap(MemoryRegionSection *section)

{

    KVMState *s = kvm_state;

    unsigned long size, allocated_size = 0;

    KVMDirtyLog d;

    KVMSlot *mem;

    int ret = 0;

    hwaddr start_addr = section->offset_within_address_space;

    hwaddr end_addr = start_addr + int128_get64(section->size);

    d.dirty_bitmap = NULL;

    while (start_addr < end_addr) {

        mem = kvm_lookup_overlapping_slot(s, start_addr, end_addr);

        if (mem == NULL) {

            break;

        }

        /* XXX bad kernel interface alert

         * For dirty bitmap, kernel allocates array of size aligned to

         * bits-per-long.  But for case when the kernel is 64bits and

         * the userspace is 32bits, userspace can't align to the same

         * bits-per-long, since sizeof(long) is different between kernel

         * and user space.  This way, userspace will provide buffer which

         * may be 4 bytes less than the kernel will use, resulting in

         * userspace memory corruption (which is not detectable by valgrind

         * too, in most cases).

         * So for now, let's align to 64 instead of HOST_LONG_BITS here, in

         * a hope that sizeof(long) wont become >8 any time soon.

         */

        size = ALIGN(((mem->memory_size) >> TARGET_PAGE_BITS),

                     /*HOST_LONG_BITS*/ 64) / 8;

        if (!d.dirty_bitmap) {

            d.dirty_bitmap = g_malloc(size);

        } else if (size > allocated_size) {

            d.dirty_bitmap = g_realloc(d.dirty_bitmap, size);

        }

        allocated_size = size;

        memset(d.dirty_bitmap, 0, allocated_size);

        d.slot = mem->slot;

        if (kvm_vm_ioctl(s, KVM_GET_DIRTY_LOG, &d) == -1) {

            DPRINTF("ioctl failed %d\n", errno);

            ret = -1;

            break;

        }

        kvm_get_dirty_pages_log_range(section, d.dirty_bitmap);

        start_addr = mem->start_addr + mem->memory_size;

    }

    g_free(d.dirty_bitmap);

    return ret;

}

static void kvm_coalesce_mmio_region(MemoryListener *listener,

                                     MemoryRegionSection *secion,

                                     hwaddr start, hwaddr size)

{

    KVMState *s = kvm_state;

    if (s->coalesced_mmio) {

        struct kvm_coalesced_mmio_zone zone;

        zone.addr = start;

        zone.size = size;

        zone.pad = 0;

        (void)kvm_vm_ioctl(s, KVM_REGISTER_COALESCED_MMIO, &zone);

    }

}

static void kvm_uncoalesce_mmio_region(MemoryListener *listener,

                                       MemoryRegionSection *secion,

                                       hwaddr start, hwaddr size)

{

    KVMState *s = kvm_state;

    if (s->coalesced_mmio) {

        struct kvm_coalesced_mmio_zone zone;

        zone.addr = start;

        zone.size = size;

        zone.pad = 0;

        (void)kvm_vm_ioctl(s, KVM_UNREGISTER_COALESCED_MMIO, &zone);

    }

}

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;

}

static int kvm_set_ioeventfd_mmio(int fd, uint32_t addr, uint32_t val,

                                  bool assign, uint32_t size, bool datamatch)

{

    int ret;

    struct kvm_ioeventfd iofd;

    iofd.datamatch = datamatch ? val : 0;

    iofd.addr = addr;

    iofd.len = size;

    iofd.flags = 0;

    iofd.fd = fd;

    if (!kvm_enabled()) {

        return -ENOSYS;

    }

    if (datamatch) {

        iofd.flags |= KVM_IOEVENTFD_FLAG_DATAMATCH;

    }

    if (!assign) {

        iofd.flags |= KVM_IOEVENTFD_FLAG_DEASSIGN;

    }

    ret = kvm_vm_ioctl(kvm_state, KVM_IOEVENTFD, &iofd);

    if (ret < 0) {

        return -errno;

    }

    return 0;

}

static int kvm_set_ioeventfd_pio(int fd, uint16_t addr, uint16_t val,

                                 bool assign, uint32_t size, bool datamatch)

{

    struct kvm_ioeventfd kick = {

        .datamatch = datamatch ? val : 0,

        .addr = addr,

        .flags = KVM_IOEVENTFD_FLAG_PIO,

        .len = size,

        .fd = fd,

    };

    int r;

    if (!kvm_enabled()) {

        return -ENOSYS;

    }

    if (datamatch) {

        kick.flags |= KVM_IOEVENTFD_FLAG_DATAMATCH;

    }

    if (!assign) {

        kick.flags |= KVM_IOEVENTFD_FLAG_DEASSIGN;

    }

    r = kvm_vm_ioctl(kvm_state, KVM_IOEVENTFD, &kick);

    if (r < 0) {

        return r;

    }

    return 0;

}

static int kvm_check_many_ioeventfds(void)

{

    /* Userspace can use ioeventfd for io notification.  This requires a host

     * that supports eventfd(2) and an I/O thread; since eventfd does not

     * support SIGIO it cannot interrupt the vcpu.

     *

     * Older kernels have a 6 device limit on the KVM io bus.  Find out so we

     * can avoid creating too many ioeventfds.

     */

#if defined(CONFIG_EVENTFD)

    int ioeventfds[7];

    int i, ret = 0;

    for (i = 0; i < ARRAY_SIZE(ioeventfds); i++) {

        ioeventfds[i] = eventfd(0, EFD_CLOEXEC);

        if (ioeventfds[i] < 0) {

            break;

        }

        ret = kvm_set_ioeventfd_pio(ioeventfds[i], 0, i, true, 2, true);

        if (ret < 0) {

            close(ioeventfds[i]);

            break;

        }

    }

    /* Decide whether many devices are supported or not */

    ret = i == ARRAY_SIZE(ioeventfds);

    while (i-- > 0) {

        kvm_set_ioeventfd_pio(ioeventfds[i], 0, i, false, 2, true);

        close(ioeventfds[i]);

    }

    return ret;

#else

    return 0;

#endif

}

static const KVMCapabilityInfo *

kvm_check_extension_list(KVMState *s, const KVMCapabilityInfo *list)

{

    while (list->name) {

        if (!kvm_check_extension(s, list->value)) {

            return list;

        }

        list++;

    }

    return NULL;

}

static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)

{

    KVMState *s = kvm_state;

    KVMSlot *mem, old;

    int err;

    MemoryRegion *mr = section->mr;

    bool log_dirty = memory_region_is_logging(mr);

    bool writeable = !mr->readonly && !mr->rom_device;

    bool readonly_flag = mr->readonly || memory_region_is_romd(mr);

    hwaddr start_addr = section->offset_within_address_space;

    ram_addr_t size = int128_get64(section->size);

    void *ram = NULL;

    unsigned delta;

    /* kvm works in page size chunks, but the function may be called

       with sub-page size and unaligned start address. */

    delta = TARGET_PAGE_ALIGN(size) - size;

    if (delta > size) {

        return;

    }

    start_addr += delta;

    size -= delta;

    size &= TARGET_PAGE_MASK;

    if (!size || (start_addr & ~TARGET_PAGE_MASK)) {

        return;

    }

    if (!memory_region_is_ram(mr)) {

        if (writeable || !kvm_readonly_mem_allowed) {

            return;

        } else if (!mr->romd_mode) {

            /* If the memory device is not in romd_mode, then we actually want

             * to remove the kvm memory slot so all accesses will trap. */

            add = false;

        }

    }

    ram = memory_region_get_ram_ptr(mr) + section->offset_within_region + delta;

    while (1) {

        mem = kvm_lookup_overlapping_slot(s, start_addr, start_addr + size);

        if (!mem) {

            break;

        }

        if (add && start_addr >= mem->start_addr &&

            (start_addr + size <= mem->start_addr + mem->memory_size) &&

            (ram - start_addr == mem->ram - mem->start_addr)) {

            /* The new slot fits into the existing one and comes with

             * identical parameters - update flags and done. */

            kvm_slot_dirty_pages_log_change(mem, log_dirty);

            return;

        }

        old = *mem;

        if (mem->flags & KVM_MEM_LOG_DIRTY_PAGES) {

            kvm_physical_sync_dirty_bitmap(section);

        }

        /* 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 && add) {

            mem = kvm_alloc_slot(s);

            mem->memory_size = old.memory_size;

            mem->start_addr = old.start_addr;

            mem->ram = old.ram;

            mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);

            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;

            ram += 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->ram = old.ram;

            mem->flags =  kvm_mem_flags(s, log_dirty, readonly_flag);

            err = kvm_set_user_memory_region(s, mem);

            if (err) {

                fprintf(stderr, "%s: error registering prefix slot: %s\n",

                        __func__, strerror(-err));

#ifdef TARGET_PPC

                fprintf(stderr, "%s: This is probably because your kernel's " \

                                "PAGE_SIZE is too big. Please try to use 4k " \

                                "PAGE_SIZE!\n", __func__);

#endif

                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->ram = old.ram + size_delta;

            mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);

            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;

    }

    if (!add) {

        return;

    }

    mem = kvm_alloc_slot(s);

    mem->memory_size = size;

    mem->start_addr = start_addr;

    mem->ram = ram;

    mem->flags = kvm_mem_flags(s, log_dirty, readonly_flag);

    err = kvm_set_user_memory_region(s, mem);

    if (err) {

        fprintf(stderr, "%s: error registering slot: %s\n", __func__,

                strerror(-err));

        abort();

    }

}

static void kvm_region_add(MemoryListener *listener,

                           MemoryRegionSection *section)

{

    kvm_set_phys_mem(section, true);

}

static void kvm_region_del(MemoryListener *listener,

                           MemoryRegionSection *section)

{

    kvm_set_phys_mem(section, false);

}

static void kvm_log_sync(MemoryListener *listener,

                         MemoryRegionSection *section)

{

    int r;

    r = kvm_physical_sync_dirty_bitmap(section);

    if (r < 0) {

        abort();

    }

}

static void kvm_log_global_start(struct MemoryListener *listener)

{

    int r;

    r = kvm_set_migration_log(1);

    assert(r >= 0);

}

static void kvm_log_global_stop(struct MemoryListener *listener)

{

    int r;

    r = kvm_set_migration_log(0);

    assert(r >= 0);

}

static void kvm_mem_ioeventfd_add(MemoryListener *listener,

                                  MemoryRegionSection *section,

                                  bool match_data, uint64_t data,

                                  EventNotifier *e)

{

    int fd = event_notifier_get_fd(e);

    int r;

    r = kvm_set_ioeventfd_mmio(fd, section->offset_within_address_space,

                               data, true, int128_get64(section->size),

                               match_data);

    if (r < 0) {

        abort();

    }

}

static void kvm_mem_ioeventfd_del(MemoryListener *listener,

                                  MemoryRegionSection *section,

                                  bool match_data, uint64_t data,

                                  EventNotifier *e)

{

    int fd = event_notifier_get_fd(e);

    int r;

    r = kvm_set_ioeventfd_mmio(fd, section->offset_within_address_space,

                               data, false, int128_get64(section->size),

                               match_data);

    if (r < 0) {

        abort();

    }

}

static void kvm_io_ioeventfd_add(MemoryListener *listener,

                                 MemoryRegionSection *section,

                                 bool match_data, uint64_t data,

                                 EventNotifier *e)

{

    int fd = event_notifier_get_fd(e);

    int r;

    r = kvm_set_ioeventfd_pio(fd, section->offset_within_address_space,

                              data, true, int128_get64(section->size),

                              match_data);

    if (r < 0) {

        abort();

    }

}

static void kvm_io_ioeventfd_del(MemoryListener *listener,

                                 MemoryRegionSection *section,

                                 bool match_data, uint64_t data,

                                 EventNotifier *e)

{

    int fd = event_notifier_get_fd(e);

    int r;

    r = kvm_set_ioeventfd_pio(fd, section->offset_within_address_space,

                              data, false, int128_get64(section->size),

                              match_data);

    if (r < 0) {

        abort();

    }

}

static MemoryListener kvm_memory_listener = {

    .region_add = kvm_region_add,

    .region_del = kvm_region_del,

    .log_start = kvm_log_start,

    .log_stop = kvm_log_stop,

    .log_sync = kvm_log_sync,

    .log_global_start = kvm_log_global_start,

    .log_global_stop = kvm_log_global_stop,

    .eventfd_add = kvm_mem_ioeventfd_add,

    .eventfd_del = kvm_mem_ioeventfd_del,

    .coalesced_mmio_add = kvm_coalesce_mmio_region,

    .coalesced_mmio_del = kvm_uncoalesce_mmio_region,

    .priority = 10,

};

static MemoryListener kvm_io_listener = {

    .eventfd_add = kvm_io_ioeventfd_add,

    .eventfd_del = kvm_io_ioeventfd_del,

    .priority = 10,

};

static void kvm_handle_interrupt(CPUState *cpu, int mask)

{

    cpu->interrupt_request |= mask;

    if (!qemu_cpu_is_self(cpu)) {

        qemu_cpu_kick(cpu);

    }

}

int kvm_set_irq(KVMState *s, int irq, int level)

{

    struct kvm_irq_level event;

    int ret;

    assert(kvm_async_interrupts_enabled());

    event.level = level;

    event.irq = irq;

    ret = kvm_vm_ioctl(s, s->irq_set_ioctl, &event);

    if (ret < 0) {

        perror("kvm_set_irq");

        abort();

    }

    return (s->irq_set_ioctl == KVM_IRQ_LINE) ? 1 : event.status;

}

#ifdef KVM_CAP_IRQ_ROUTING

typedef struct KVMMSIRoute {

    struct kvm_irq_routing_entry kroute;

    QTAILQ_ENTRY(KVMMSIRoute) entry;

} KVMMSIRoute;

static void set_gsi(KVMState *s, unsigned int gsi)

{

    s->used_gsi_bitmap[gsi / 32] |= 1U << (gsi % 32);

}

static void clear_gsi(KVMState *s, unsigned int gsi)

{

    s->used_gsi_bitmap[gsi / 32] &= ~(1U << (gsi % 32));

}

static void kvm_init_irq_routing(KVMState *s)

{

    int gsi_count, i;

    gsi_count = kvm_check_extension(s, KVM_CAP_IRQ_ROUTING);

    if (gsi_count > 0) {

        unsigned int gsi_bits, i;

        /* Round up so we can search ints using ffs */

        gsi_bits = ALIGN(gsi_count, 32);

        s->used_gsi_bitmap = g_malloc0(gsi_bits / 8);

        s->gsi_count = gsi_count;

        /* Mark any over-allocated bits as already in use */

        for (i = gsi_count; i < gsi_bits; i++) {

            set_gsi(s, i);

        }

    }

    s->irq_routes = g_malloc0(sizeof(*s->irq_routes));

    s->nr_allocated_irq_routes = 0;

    if (!s->direct_msi) {

        for (i = 0; i < KVM_MSI_HASHTAB_SIZE; i++) {

            QTAILQ_INIT(&s->msi_hashtab[i]);

        }

    }

    kvm_arch_init_irq_routing(s);

}

static void kvm_irqchip_commit_routes(KVMState *s)

{

    int ret;

    s->irq_routes->flags = 0;

    ret = kvm_vm_ioctl(s, KVM_SET_GSI_ROUTING, s->irq_routes);

    assert(ret == 0);

}

static void kvm_add_routing_entry(KVMState *s,

                                  struct kvm_irq_routing_entry *entry)

{

    struct kvm_irq_routing_entry *new;

    int n, size;

    if (s->irq_routes->nr == s->nr_allocated_irq_routes) {

        n = s->nr_allocated_irq_routes * 2;

        if (n < 64) {

            n = 64;

        }

        size = sizeof(struct kvm_irq_routing);

        size += n * sizeof(*new);

        s->irq_routes = g_realloc(s->irq_routes, size);

        s->nr_allocated_irq_routes = n;

    }

    n = s->irq_routes->nr++;

    new = &s->irq_routes->entries[n];

    memset(new, 0, sizeof(*new));

    new->gsi = entry->gsi;

    new->type = entry->type;

    new->flags = entry->flags;

    new->u = entry->u;

    set_gsi(s, entry->gsi);

    kvm_irqchip_commit_routes(s);

}

static int kvm_update_routing_entry(KVMState *s,

                                    struct kvm_irq_routing_entry *new_entry)

{

    struct kvm_irq_routing_entry *entry;

    int n;

    for (n = 0; n < s->irq_routes->nr; n++) {

        entry = &s->irq_routes->entries[n];

        if (entry->gsi != new_entry->gsi) {

            continue;

        }

        entry->type = new_entry->type;

        entry->flags = new_entry->flags;

        entry->u = new_entry->u;

        kvm_irqchip_commit_routes(s);

        return 0;

    }

    return -ESRCH;

}

void kvm_irqchip_add_irq_route(KVMState *s, int irq, int irqchip, int pin)

{

    struct kvm_irq_routing_entry e;

    assert(pin < s->gsi_count);

    e.gsi = irq;

    e.type = KVM_IRQ_ROUTING_IRQCHIP;

    e.flags = 0;

    e.u.irqchip.irqchip = irqchip;

    e.u.irqchip.pin = pin;

    kvm_add_routing_entry(s, &e);

}

void kvm_irqchip_release_virq(KVMState *s, int virq)

{

    struct kvm_irq_routing_entry *e;

    int i;

    for (i = 0; i < s->irq_routes->nr; i++) {

        e = &s->irq_routes->entries[i];

        if (e->gsi == virq) {

            s->irq_routes->nr--;

            *e = s->irq_routes->entries[s->irq_routes->nr];

        }

    }

    clear_gsi(s, virq);

}

static unsigned int kvm_hash_msi(uint32_t data)

{

    /* This is optimized for IA32 MSI layout. However, no other arch shall

     * repeat the mistake of not providing a direct MSI injection API. */

    return data & 0xff;

}

static void kvm_flush_dynamic_msi_routes(KVMState *s)

{

    KVMMSIRoute *route, *next;

    unsigned int hash;

    for (hash = 0; hash < KVM_MSI_HASHTAB_SIZE; hash++) {

        QTAILQ_FOREACH_SAFE(route, &s->msi_hashtab[hash], entry, next) {

            kvm_irqchip_release_virq(s, route->kroute.gsi);

            QTAILQ_REMOVE(&s->msi_hashtab[hash], route, entry);

            g_free(route);

        }

    }

}

static int kvm_irqchip_get_virq(KVMState *s)

{

    uint32_t *word = s->used_gsi_bitmap;

    int max_words = ALIGN(s->gsi_count, 32) / 32;

    int i, bit;

    bool retry = true;

again:

    /* Return the lowest unused GSI in the bitmap */

    for (i = 0; i < max_words; i++) {

        bit = ffs(~word[i]);

        if (!bit) {

            continue;

        }

        return bit - 1 + i * 32;

    }

    if (!s->direct_msi && retry) {

        retry = false;

        kvm_flush_dynamic_msi_routes(s);

        goto again;

    }

    return -ENOSPC;

}

static KVMMSIRoute *kvm_lookup_msi_route(KVMState *s, MSIMessage msg)

{

    unsigned int hash = kvm_hash_msi(msg.data);

    KVMMSIRoute *route;

    QTAILQ_FOREACH(route, &s->msi_hashtab[hash], entry) {

        if (route->kroute.u.msi.address_lo == (uint32_t)msg.address &&

            route->kroute.u.msi.address_hi == (msg.address >> 32) &&

            route->kroute.u.msi.data == msg.data) {

            return route;

        }

    }

    return NULL;

}

int kvm_irqchip_send_msi(KVMState *s, MSIMessage msg)

{

    struct kvm_msi msi;

    KVMMSIRoute *route;

    if (s->direct_msi) {

        msi.address_lo = (uint32_t)msg.address;

        msi.address_hi = msg.address >> 32;

        msi.data = msg.data;

        msi.flags = 0;

        memset(msi.pad, 0, sizeof(msi.pad));

        return kvm_vm_ioctl(s, KVM_SIGNAL_MSI, &msi);

    }

    route = kvm_lookup_msi_route(s, msg);

    if (!route) {

        int virq;

        virq = kvm_irqchip_get_virq(s);

        if (virq < 0) {

            return virq;

        }

        route = g_malloc(sizeof(KVMMSIRoute));

        route->kroute.gsi = virq;

        route->kroute.type = KVM_IRQ_ROUTING_MSI;

        route->kroute.flags = 0;

        route->kroute.u.msi.address_lo = (uint32_t)msg.address;

        route->kroute.u.msi.address_hi = msg.address >> 32;

        route->kroute.u.msi.data = msg.data;

        kvm_add_routing_entry(s, &route->kroute);

        QTAILQ_INSERT_TAIL(&s->msi_hashtab[kvm_hash_msi(msg.data)], route,

                           entry);

    }

    assert(route->kroute.type == KVM_IRQ_ROUTING_MSI);

    return kvm_set_irq(s, route->kroute.gsi, 1);

}

int kvm_irqchip_add_msi_route(KVMState *s, MSIMessage msg)

{

    struct kvm_irq_routing_entry kroute;

    int virq;

    if (!kvm_gsi_routing_enabled()) {

        return -ENOSYS;

    }

    virq = kvm_irqchip_get_virq(s);

    if (virq < 0) {

        return virq;

    }

    kroute.gsi = virq;

    kroute.type = KVM_IRQ_ROUTING_MSI;

    kroute.flags = 0;

    kroute.u.msi.address_lo = (uint32_t)msg.address;

    kroute.u.msi.address_hi = msg.address >> 32;

    kroute.u.msi.data = msg.data;

    kvm_add_routing_entry(s, &kroute);

    return virq;

}

int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg)

{

    struct kvm_irq_routing_entry kroute;

    if (!kvm_irqchip_in_kernel()) {

        return -ENOSYS;

    }

    kroute.gsi = virq;

    kroute.type = KVM_IRQ_ROUTING_MSI;

    kroute.flags = 0;

    kroute.u.msi.address_lo = (uint32_t)msg.address;

    kroute.u.msi.address_hi = msg.address >> 32;

    kroute.u.msi.data = msg.data;

    return kvm_update_routing_entry(s, &kroute);

}

static int kvm_irqchip_assign_irqfd(KVMState *s, int fd, int virq, bool assign)

{

    struct kvm_irqfd irqfd = {

        .fd = fd,

        .gsi = virq,

        .flags = assign ? 0 : KVM_IRQFD_FLAG_DEASSIGN,

    };

    if (!kvm_irqfds_enabled()) {

        return -ENOSYS;

    }

    return kvm_vm_ioctl(s, KVM_IRQFD, &irqfd);

}

#else /* !KVM_CAP_IRQ_ROUTING */

static void kvm_init_irq_routing(KVMState *s)

{

}

void kvm_irqchip_release_virq(KVMState *s, int virq)

{

}

int kvm_irqchip_send_msi(KVMState *s, MSIMessage msg)

{

    abort();

}

int kvm_irqchip_add_msi_route(KVMState *s, MSIMessage msg)

{

    return -ENOSYS;

}

static int kvm_irqchip_assign_irqfd(KVMState *s, int fd, int virq, bool assign)

{

    abort();

}

int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg)

{

    return -ENOSYS;

}

#endif /* !KVM_CAP_IRQ_ROUTING */

int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, int virq)

{

    return kvm_irqchip_assign_irqfd(s, event_notifier_get_fd(n), virq, true);

}

int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq)

{

    return kvm_irqchip_assign_irqfd(s, event_notifier_get_fd(n), virq, false);

}

static int kvm_irqchip_create(KVMState *s)

{

    QemuOptsList *list = qemu_find_opts("machine");

    int ret;

    if (QTAILQ_EMPTY(&list->head) ||

        !qemu_opt_get_bool(QTAILQ_FIRST(&list->head),

                           "kernel_irqchip", true) ||

        !kvm_check_extension(s, KVM_CAP_IRQCHIP)) {

        return 0;

    }

    ret = kvm_vm_ioctl(s, KVM_CREATE_IRQCHIP);

    if (ret < 0) {

        fprintf(stderr, "Create kernel irqchip failed\n");

        return ret;

    }

    kvm_kernel_irqchip = true;

    /* If we have an in-kernel IRQ chip then we must have asynchronous

     * interrupt delivery (though the reverse is not necessarily true)

     */

    kvm_async_interrupts_allowed = true;

    kvm_init_irq_routing(s);

    return 0;

}

static int kvm_max_vcpus(KVMState *s)

{

    int ret;

    /* Find number of supported CPUs using the recommended

     * procedure from the kernel API documentation to cope with

     * older kernels that may be missing capabilities.

     */

    ret = kvm_check_extension(s, KVM_CAP_MAX_VCPUS);

    if (ret) {

        return ret;

    }

    ret = kvm_check_extension(s, KVM_CAP_NR_VCPUS);

    if (ret) {

        return ret;

    }

    return 4;

}

int kvm_init(void)

{

    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";

    KVMState *s;

    const KVMCapabilityInfo *missing_cap;

    int ret;

    int i;

    int max_vcpus;

    s = g_malloc0(sizeof(KVMState));

    /*

     * On systems where the kernel can support different base page

     * sizes, host page size may be different from TARGET_PAGE_SIZE,

     * even with KVM.  TARGET_PAGE_SIZE is assumed to be the minimum

     * page size for the system though.

     */

    assert(TARGET_PAGE_SIZE <= getpagesize());

#ifdef KVM_CAP_SET_GUEST_DEBUG

    QTAILQ_INIT(&s->kvm_sw_breakpoints);

#endif

    for (i = 0; i < ARRAY_SIZE(s->slots); i++) {

        s->slots[i].slot = i;

    }

    s->vmfd = -1;

    s->fd = qemu_open("/dev/kvm", O_RDWR);

    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;

    }

    max_vcpus = kvm_max_vcpus(s);

    if (smp_cpus > max_vcpus) {

        ret = -EINVAL;

        fprintf(stderr, "Number of SMP cpus requested (%d) exceeds max cpus "

                "supported by KVM (%d)\n", smp_cpus, max_vcpus);

        goto err;

    }

    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);

    if (s->vmfd < 0) {

#ifdef TARGET_S390X