diff --git a/arch_init.c b/arch_init.c
index 0471cd5a6bf41c31bcb0461bc5d7e177fd4ad87a..e47e1399bb5fcf961c3c5d95a3bc4a7469b81842 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -1196,15 +1196,14 @@ static void mig_sleep_cpu(void *opq)
    much time in the VM. The migration thread will try to catchup.
    Workload will experience a performance drop.
 */
-static void mig_throttle_cpu_down(CPUState *cpu, void *data)
-{
-    async_run_on_cpu(cpu, mig_sleep_cpu, NULL);
-}
-
 static void mig_throttle_guest_down(void)
 {
+    CPUState *cpu;
+
     qemu_mutex_lock_iothread();
-    qemu_for_each_cpu(mig_throttle_cpu_down, NULL);
+    CPU_FOREACH(cpu) {
+        async_run_on_cpu(cpu, mig_sleep_cpu, NULL);
+    }
     qemu_mutex_unlock_iothread();
 }
 
diff --git a/cpus.c b/cpus.c
index d74cc117b38b947961db8ea5611de09230621d54..e566297bd30c12d81724221b0cccb83ba66c86ac 100644
--- a/cpus.c
+++ b/cpus.c
@@ -86,7 +86,7 @@ static bool all_cpu_threads_idle(void)
 {
     CPUState *cpu;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         if (!cpu_thread_is_idle(cpu)) {
             return false;
         }
@@ -416,7 +416,7 @@ void hw_error(const char *fmt, ...)
     fprintf(stderr, "qemu: hardware error: ");
     vfprintf(stderr, fmt, ap);
     fprintf(stderr, "\n");
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         fprintf(stderr, "CPU #%d:\n", cpu->cpu_index);
         cpu_dump_state(cpu, stderr, fprintf, CPU_DUMP_FPU);
     }
@@ -428,7 +428,7 @@ void cpu_synchronize_all_states(void)
 {
     CPUState *cpu;
 
-    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         cpu_synchronize_state(cpu);
     }
 }
@@ -437,7 +437,7 @@ void cpu_synchronize_all_post_reset(void)
 {
     CPUState *cpu;
 
-    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         cpu_synchronize_post_reset(cpu);
     }
 }
@@ -446,7 +446,7 @@ void cpu_synchronize_all_post_init(void)
 {
     CPUState *cpu;
 
-    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         cpu_synchronize_post_init(cpu);
     }
 }
@@ -760,7 +760,7 @@ static void qemu_tcg_wait_io_event(void)
         qemu_cond_wait(&qemu_io_proceeded_cond, &qemu_global_mutex);
     }
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         qemu_wait_io_event_common(cpu);
     }
 }
@@ -854,12 +854,6 @@ static void *qemu_dummy_cpu_thread_fn(void *arg)
 
 static void tcg_exec_all(void);
 
-static void tcg_signal_cpu_creation(CPUState *cpu, void *data)
-{
-    cpu->thread_id = qemu_get_thread_id();
-    cpu->created = true;
-}
-
 static void *qemu_tcg_cpu_thread_fn(void *arg)
 {
     CPUState *cpu = arg;
@@ -868,15 +862,18 @@ static void *qemu_tcg_cpu_thread_fn(void *arg)
     qemu_thread_get_self(cpu->thread);
 
     qemu_mutex_lock(&qemu_global_mutex);
-    qemu_for_each_cpu(tcg_signal_cpu_creation, NULL);
+    CPU_FOREACH(cpu) {
+        cpu->thread_id = qemu_get_thread_id();
+        cpu->created = true;
+    }
     qemu_cond_signal(&qemu_cpu_cond);
 
     /* wait for initial kick-off after machine start */
-    while (first_cpu->stopped) {
+    while (QTAILQ_FIRST(&cpus)->stopped) {
         qemu_cond_wait(tcg_halt_cond, &qemu_global_mutex);
 
         /* process any pending work */
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             qemu_wait_io_event_common(cpu);
         }
     }
@@ -991,13 +988,12 @@ void qemu_mutex_unlock_iothread(void)
 
 static int all_vcpus_paused(void)
 {
-    CPUState *cpu = first_cpu;
+    CPUState *cpu;
 
-    while (cpu) {
+    CPU_FOREACH(cpu) {
         if (!cpu->stopped) {
             return 0;
         }
-        cpu = cpu->next_cpu;
     }
 
     return 1;
@@ -1005,23 +1001,20 @@ static int all_vcpus_paused(void)
 
 void pause_all_vcpus(void)
 {
-    CPUState *cpu = first_cpu;
+    CPUState *cpu;
 
     qemu_clock_enable(QEMU_CLOCK_VIRTUAL, false);
-    while (cpu) {
+    CPU_FOREACH(cpu) {
         cpu->stop = true;
         qemu_cpu_kick(cpu);
-        cpu = cpu->next_cpu;
     }
 
     if (qemu_in_vcpu_thread()) {
         cpu_stop_current();
         if (!kvm_enabled()) {
-            cpu = first_cpu;
-            while (cpu) {
+            CPU_FOREACH(cpu) {
                 cpu->stop = false;
                 cpu->stopped = true;
-                cpu = cpu->next_cpu;
             }
             return;
         }
@@ -1029,10 +1022,8 @@ void pause_all_vcpus(void)
 
     while (!all_vcpus_paused()) {
         qemu_cond_wait(&qemu_pause_cond, &qemu_global_mutex);
-        cpu = first_cpu;
-        while (cpu) {
+        CPU_FOREACH(cpu) {
             qemu_cpu_kick(cpu);
-            cpu = cpu->next_cpu;
         }
     }
 }
@@ -1046,12 +1037,11 @@ void cpu_resume(CPUState *cpu)
 
 void resume_all_vcpus(void)
 {
-    CPUState *cpu = first_cpu;
+    CPUState *cpu;
 
     qemu_clock_enable(QEMU_CLOCK_VIRTUAL, true);
-    while (cpu) {
+    CPU_FOREACH(cpu) {
         cpu_resume(cpu);
-        cpu = cpu->next_cpu;
     }
 }
 
@@ -1215,7 +1205,7 @@ static void tcg_exec_all(void)
     if (next_cpu == NULL) {
         next_cpu = first_cpu;
     }
-    for (; next_cpu != NULL && !exit_request; next_cpu = next_cpu->next_cpu) {
+    for (; next_cpu != NULL && !exit_request; next_cpu = CPU_NEXT(next_cpu)) {
         CPUState *cpu = next_cpu;
         CPUArchState *env = cpu->env_ptr;
 
@@ -1240,7 +1230,7 @@ void set_numa_modes(void)
     CPUState *cpu;
     int i;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         for (i = 0; i < nb_numa_nodes; i++) {
             if (test_bit(cpu->cpu_index, node_cpumask[i])) {
                 cpu->numa_node = i;
@@ -1262,7 +1252,7 @@ CpuInfoList *qmp_query_cpus(Error **errp)
     CpuInfoList *head = NULL, *cur_item = NULL;
     CPUState *cpu;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         CpuInfoList *info;
 #if defined(TARGET_I386)
         X86CPU *x86_cpu = X86_CPU(cpu);
@@ -1391,7 +1381,7 @@ void qmp_inject_nmi(Error **errp)
 #if defined(TARGET_I386)
     CPUState *cs;
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         X86CPU *cpu = X86_CPU(cs);
         CPUX86State *env = &cpu->env;
 
@@ -1405,7 +1395,7 @@ void qmp_inject_nmi(Error **errp)
     CPUState *cs;
     S390CPU *cpu;
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         cpu = S390_CPU(cs);
         if (cpu->env.cpu_num == monitor_get_cpu_index()) {
             if (s390_cpu_restart(S390_CPU(cs)) == -1) {
diff --git a/cputlb.c b/cputlb.c
index 977c0ca59d3c527941084fc255ed10b5185d6b32..19ecf60983ef8b7f1c495c3d2c9823b10af92615 100644
--- a/cputlb.c
+++ b/cputlb.c
@@ -189,7 +189,7 @@ void cpu_tlb_reset_dirty_all(ram_addr_t start1, ram_addr_t length)
     CPUState *cpu;
     CPUArchState *env;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         int mmu_idx;
 
         env = cpu->env_ptr;
diff --git a/dump.c b/dump.c
index c0dae2c3ff934f372e6e2b773465d8a1148d1a99..846155cbc90a970e755e177e4ba9aa16e4dac1c7 100644
--- a/dump.c
+++ b/dump.c
@@ -277,7 +277,7 @@ static int write_elf64_notes(DumpState *s)
     int ret;
     int id;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         id = cpu_index(cpu);
         ret = cpu_write_elf64_note(fd_write_vmcore, cpu, id, s);
         if (ret < 0) {
@@ -286,7 +286,7 @@ static int write_elf64_notes(DumpState *s)
         }
     }
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         ret = cpu_write_elf64_qemunote(fd_write_vmcore, cpu, s);
         if (ret < 0) {
             dump_error(s, "dump: failed to write CPU status.\n");
@@ -327,7 +327,7 @@ static int write_elf32_notes(DumpState *s)
     int ret;
     int id;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         id = cpu_index(cpu);
         ret = cpu_write_elf32_note(fd_write_vmcore, cpu, id, s);
         if (ret < 0) {
@@ -336,7 +336,7 @@ static int write_elf32_notes(DumpState *s)
         }
     }
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         ret = cpu_write_elf32_qemunote(fd_write_vmcore, cpu, s);
         if (ret < 0) {
             dump_error(s, "dump: failed to write CPU status.\n");
@@ -734,7 +734,7 @@ static int dump_init(DumpState *s, int fd, bool paging, bool has_filter,
      */
     cpu_synchronize_all_states();
     nr_cpus = 0;
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         nr_cpus++;
     }
 
diff --git a/exec.c b/exec.c
index 3ca9381214b8002e0ac32eacb23f56a195d74764..87b0b39b904154bc1deb3ff640b9fb56e7ed09c6 100644
--- a/exec.c
+++ b/exec.c
@@ -69,7 +69,7 @@ static MemoryRegion io_mem_unassigned;
 
 #endif
 
-CPUState *first_cpu;
+struct CPUTailQ cpus = QTAILQ_HEAD_INITIALIZER(cpus);
 /* current CPU in the current thread. It is only valid inside
    cpu_exec() */
 DEFINE_TLS(CPUState *, current_cpu);
@@ -351,44 +351,29 @@ const VMStateDescription vmstate_cpu_common = {
 
 CPUState *qemu_get_cpu(int index)
 {
-    CPUState *cpu = first_cpu;
+    CPUState *cpu;
 
-    while (cpu) {
+    CPU_FOREACH(cpu) {
         if (cpu->cpu_index == index) {
-            break;
+            return cpu;
         }
-        cpu = cpu->next_cpu;
     }
 
-    return cpu;
-}
-
-void qemu_for_each_cpu(void (*func)(CPUState *cpu, void *data), void *data)
-{
-    CPUState *cpu;
-
-    cpu = first_cpu;
-    while (cpu) {
-        func(cpu, data);
-        cpu = cpu->next_cpu;
-    }
+    return NULL;
 }
 
 void cpu_exec_init(CPUArchState *env)
 {
     CPUState *cpu = ENV_GET_CPU(env);
     CPUClass *cc = CPU_GET_CLASS(cpu);
-    CPUState **pcpu;
+    CPUState *some_cpu;
     int cpu_index;
 
 #if defined(CONFIG_USER_ONLY)
     cpu_list_lock();
 #endif
-    cpu->next_cpu = NULL;
-    pcpu = &first_cpu;
     cpu_index = 0;
-    while (*pcpu != NULL) {
-        pcpu = &(*pcpu)->next_cpu;
+    CPU_FOREACH(some_cpu) {
         cpu_index++;
     }
     cpu->cpu_index = cpu_index;
@@ -398,7 +383,7 @@ void cpu_exec_init(CPUArchState *env)
 #ifndef CONFIG_USER_ONLY
     cpu->thread_id = qemu_get_thread_id();
 #endif
-    *pcpu = cpu;
+    QTAILQ_INSERT_TAIL(&cpus, cpu, node);
 #if defined(CONFIG_USER_ONLY)
     cpu_list_unlock();
 #endif
@@ -1762,7 +1747,7 @@ static void tcg_commit(MemoryListener *listener)
     /* since each CPU stores ram addresses in its TLB cache, we must
        reset the modified entries */
     /* XXX: slow ! */
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         CPUArchState *env = cpu->env_ptr;
 
         tlb_flush(env, 1);
diff --git a/gdbstub.c b/gdbstub.c
index 9d067d6b807e7d2d81023295743cc09ed0536361..2b7f22b2d2b8c70af89954294fa069ebf23a5c54 100644
--- a/gdbstub.c
+++ b/gdbstub.c
@@ -648,7 +648,7 @@ static int gdb_breakpoint_insert(target_ulong addr, target_ulong len, int type)
     switch (type) {
     case GDB_BREAKPOINT_SW:
     case GDB_BREAKPOINT_HW:
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             env = cpu->env_ptr;
             err = cpu_breakpoint_insert(env, addr, BP_GDB, NULL);
             if (err)
@@ -659,7 +659,7 @@ static int gdb_breakpoint_insert(target_ulong addr, target_ulong len, int type)
     case GDB_WATCHPOINT_WRITE:
     case GDB_WATCHPOINT_READ:
     case GDB_WATCHPOINT_ACCESS:
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             env = cpu->env_ptr;
             err = cpu_watchpoint_insert(env, addr, len, xlat_gdb_type[type],
                                         NULL);
@@ -686,7 +686,7 @@ static int gdb_breakpoint_remove(target_ulong addr, target_ulong len, int type)
     switch (type) {
     case GDB_BREAKPOINT_SW:
     case GDB_BREAKPOINT_HW:
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             env = cpu->env_ptr;
             err = cpu_breakpoint_remove(env, addr, BP_GDB);
             if (err)
@@ -697,7 +697,7 @@ static int gdb_breakpoint_remove(target_ulong addr, target_ulong len, int type)
     case GDB_WATCHPOINT_WRITE:
     case GDB_WATCHPOINT_READ:
     case GDB_WATCHPOINT_ACCESS:
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             env = cpu->env_ptr;
             err = cpu_watchpoint_remove(env, addr, len, xlat_gdb_type[type]);
             if (err)
@@ -720,7 +720,7 @@ static void gdb_breakpoint_remove_all(void)
         return;
     }
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         env = cpu->env_ptr;
         cpu_breakpoint_remove_all(env, BP_GDB);
 #ifndef CONFIG_USER_ONLY
@@ -744,7 +744,7 @@ static CPUState *find_cpu(uint32_t thread_id)
 {
     CPUState *cpu;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         if (cpu_index(cpu) == thread_id) {
             return cpu;
         }
@@ -1070,7 +1070,7 @@ static int gdb_handle_packet(GDBState *s, const char *line_buf)
             if (s->query_cpu) {
                 snprintf(buf, sizeof(buf), "m%x", cpu_index(s->query_cpu));
                 put_packet(s, buf);
-                s->query_cpu = s->query_cpu->next_cpu;
+                s->query_cpu = CPU_NEXT(s->query_cpu);
             } else
                 put_packet(s, "l");
             break;
diff --git a/hw/acpi/piix4.c b/hw/acpi/piix4.c
index 613d98736a12f793161b4a7c90e1eee4abc84a44..0b8d1d9da9c10b4559bf5d06a492a875bc3b018a 100644
--- a/hw/acpi/piix4.c
+++ b/hw/acpi/piix4.c
@@ -667,22 +667,14 @@ static void piix4_cpu_added_req(Notifier *n, void *opaque)
     piix4_cpu_hotplug_req(s, CPU(opaque), PLUG);
 }
 
-static void piix4_init_cpu_status(CPUState *cpu, void *data)
-{
-    CPUStatus *g = (CPUStatus *)data;
-    CPUClass *k = CPU_GET_CLASS(cpu);
-    int64_t id = k->get_arch_id(cpu);
-
-    g_assert((id / 8) < PIIX4_PROC_LEN);
-    g->sts[id / 8] |= (1 << (id % 8));
-}
-
 static int piix4_device_hotplug(DeviceState *qdev, PCIDevice *dev,
                                 PCIHotplugState state);
 
 static void piix4_acpi_system_hot_add_init(MemoryRegion *parent,
                                            PCIBus *bus, PIIX4PMState *s)
 {
+    CPUState *cpu;
+
     memory_region_init_io(&s->io_gpe, OBJECT(s), &piix4_gpe_ops, s,
                           "acpi-gpe0", GPE_LEN);
     memory_region_add_subregion(parent, GPE_BASE, &s->io_gpe);
@@ -693,7 +685,13 @@ static void piix4_acpi_system_hot_add_init(MemoryRegion *parent,
                                 &s->io_pci);
     pci_bus_hotplug(bus, piix4_device_hotplug, DEVICE(s));
 
-    qemu_for_each_cpu(piix4_init_cpu_status, &s->gpe_cpu);
+    CPU_FOREACH(cpu) {
+        CPUClass *cc = CPU_GET_CLASS(cpu);
+        int64_t id = cc->get_arch_id(cpu);
+
+        g_assert((id / 8) < PIIX4_PROC_LEN);
+        s->gpe_cpu.sts[id / 8] |= (1 << (id % 8));
+    }
     memory_region_init_io(&s->io_cpu, OBJECT(s), &cpu_hotplug_ops, s,
                           "acpi-cpu-hotplug", PIIX4_PROC_LEN);
     memory_region_add_subregion(parent, PIIX4_PROC_BASE, &s->io_cpu);
diff --git a/hw/arm/boot.c b/hw/arm/boot.c
index 2cbeefdcba6c2446ed5a85a0de14d4dbdf3da32f..1e313afe8d43c5629103c2853c65ca969df7a1b2 100644
--- a/hw/arm/boot.c
+++ b/hw/arm/boot.c
@@ -468,7 +468,7 @@ void arm_load_kernel(ARMCPU *cpu, struct arm_boot_info *info)
     }
     info->is_linux = is_linux;
 
-    for (; cs; cs = cs->next_cpu) {
+    for (; cs; cs = CPU_NEXT(cs)) {
         cpu = ARM_CPU(cs);
         cpu->env.boot_info = info;
         qemu_register_reset(do_cpu_reset, cpu);
diff --git a/hw/cpu/a15mpcore.c b/hw/cpu/a15mpcore.c
index af182da4eef400adcf4c57c75dcbf092d5f49dc6..9abba67632be83d086f4533508e4dc6e3dbd6a53 100644
--- a/hw/cpu/a15mpcore.c
+++ b/hw/cpu/a15mpcore.c
@@ -50,7 +50,6 @@ static int a15mp_priv_init(SysBusDevice *dev)
     SysBusDevice *busdev;
     const char *gictype = "arm_gic";
     int i;
-    CPUState *cpu;
 
     if (kvm_irqchip_in_kernel()) {
         gictype = "kvm-arm-gic";
@@ -72,8 +71,8 @@ static int a15mp_priv_init(SysBusDevice *dev)
     /* Wire the outputs from each CPU's generic timer to the
      * appropriate GIC PPI inputs
      */
-    for (i = 0, cpu = first_cpu; i < s->num_cpu; i++, cpu = cpu->next_cpu) {
-        DeviceState *cpudev = DEVICE(cpu);
+    for (i = 0; i < s->num_cpu; i++) {
+        DeviceState *cpudev = DEVICE(qemu_get_cpu(i));
         int ppibase = s->num_irq - 32 + i * 32;
         /* physical timer; we wire it up to the non-secure timer's ID,
          * since a real A15 always has TrustZone but QEMU doesn't.
diff --git a/hw/i386/kvm/clock.c b/hw/i386/kvm/clock.c
index e89e2f768e1fa3cabaaa9e869e23242bee11fa9e..92aabb83b5693243460608e9a1349352510b36b2 100644
--- a/hw/i386/kvm/clock.c
+++ b/hw/i386/kvm/clock.c
@@ -59,7 +59,7 @@ static void kvmclock_vm_state_change(void *opaque, int running,
         if (!cap_clock_ctrl) {
             return;
         }
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             ret = kvm_vcpu_ioctl(cpu, KVM_KVMCLOCK_CTRL, 0);
             if (ret) {
                 if (ret != -EINVAL) {
diff --git a/hw/i386/kvmvapic.c b/hw/i386/kvmvapic.c
index 15beb8044e736582281207c2160e2f0ba2ebbfaf..d3a6fbe1f95c11995692146b8951fbfb2ac717b4 100644
--- a/hw/i386/kvmvapic.c
+++ b/hw/i386/kvmvapic.c
@@ -498,7 +498,7 @@ static void vapic_enable_tpr_reporting(bool enable)
     X86CPU *cpu;
     CPUX86State *env;
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         cpu = X86_CPU(cs);
         env = &cpu->env;
         info.apic = env->apic_state;
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index 3a620a185687872c3eedca970da6015c6ba0fb5b..0c313feb0b297709a5c554051cf1a517e2786340 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -191,13 +191,12 @@ static void pic_irq_request(void *opaque, int irq, int level)
 
     DPRINTF("pic_irqs: %s irq %d\n", level? "raise" : "lower", irq);
     if (env->apic_state) {
-        while (cs) {
+        CPU_FOREACH(cs) {
             cpu = X86_CPU(cs);
             env = &cpu->env;
             if (apic_accept_pic_intr(env->apic_state)) {
                 apic_deliver_pic_intr(env->apic_state, level);
             }
-            cs = cs->next_cpu;
         }
     } else {
         if (level) {
diff --git a/hw/ppc/e500.c b/hw/ppc/e500.c
index 9059ff9bc7ef44b8323cd7991d611ca76671f979..cfdd84b969afd5537eb28f5753de1c9fafac50b0 100644
--- a/hw/ppc/e500.c
+++ b/hw/ppc/e500.c
@@ -540,7 +540,7 @@ static DeviceState *ppce500_init_mpic_kvm(PPCE500Params *params,
         return NULL;
     }
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         if (kvm_openpic_connect_vcpu(dev, cs)) {
             fprintf(stderr, "%s: failed to connect vcpu to irqchip\n",
                     __func__);
diff --git a/hw/ppc/ppc.c b/hw/ppc/ppc.c
index 59b41cbc6f4dc4cf54ee4a058b59c2870cb90c56..bf2d3d4b350430b98bb358cd648d9d0cf61f42f8 100644
--- a/hw/ppc/ppc.c
+++ b/hw/ppc/ppc.c
@@ -443,7 +443,7 @@ void ppce500_set_mpic_proxy(bool enabled)
 {
     CPUState *cs;
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         PowerPCCPU *cpu = POWERPC_CPU(cs);
 
         cpu->env.mpic_proxy = enabled;
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 279b88af97046db03869b13a536a0daf53d0d72a..004184d84174a0878407cfc55d73001d944e17d0 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -187,7 +187,7 @@ static int spapr_fixup_cpu_dt(void *fdt, sPAPREnvironment *spapr)
 
     assert(spapr->cpu_model);
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         uint32_t associativity[] = {cpu_to_be32(0x5),
                                     cpu_to_be32(0x0),
                                     cpu_to_be32(0x0),
@@ -351,7 +351,7 @@ static void *spapr_create_fdt_skel(const char *cpu_model,
     /* This is needed during FDT finalization */
     spapr->cpu_model = g_strdup(modelname);
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         PowerPCCPU *cpu = POWERPC_CPU(cs);
         CPUPPCState *env = &cpu->env;
         PowerPCCPUClass *pcc = POWERPC_CPU_GET_CLASS(cs);
diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
index 89e6a00dd935a23a79f2a43fa5123a1e7cf70d1e..f10ba8a932b0350c9da93e1c1bb920030aae907e 100644
--- a/hw/ppc/spapr_hcall.c
+++ b/hw/ppc/spapr_hcall.c
@@ -679,7 +679,7 @@ static target_ulong h_set_mode(PowerPCCPU *cpu, sPAPREnvironment *spapr,
 
         switch (mflags) {
         case H_SET_MODE_ENDIAN_BIG:
-            for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+            CPU_FOREACH(cs) {
                 PowerPCCPU *cp = POWERPC_CPU(cs);
                 CPUPPCState *env = &cp->env;
                 env->spr[SPR_LPCR] &= ~LPCR_ILE;
@@ -688,7 +688,7 @@ static target_ulong h_set_mode(PowerPCCPU *cpu, sPAPREnvironment *spapr,
             break;
 
         case H_SET_MODE_ENDIAN_LITTLE:
-            for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+            CPU_FOREACH(cs) {
                 PowerPCCPU *cp = POWERPC_CPU(cs);
                 CPUPPCState *env = &cp->env;
                 env->spr[SPR_LPCR] |= LPCR_ILE;
diff --git a/include/qom/cpu.h b/include/qom/cpu.h
index 3e4993661af0b161fe73d5350dc2a911c1a96fcd..7739e000679f85dd0f7f2a15f4c3d4dace9fabb9 100644
--- a/include/qom/cpu.h
+++ b/include/qom/cpu.h
@@ -23,6 +23,7 @@
 #include <signal.h>
 #include "hw/qdev-core.h"
 #include "exec/hwaddr.h"
+#include "qemu/queue.h"
 #include "qemu/thread.h"
 #include "qemu/tls.h"
 #include "qemu/typedefs.h"
@@ -190,7 +191,7 @@ struct CPUState {
     struct GDBRegisterState *gdb_regs;
     int gdb_num_regs;
     int gdb_num_g_regs;
-    CPUState *next_cpu;
+    QTAILQ_ENTRY(CPUState) node;
 
     int kvm_fd;
     bool kvm_vcpu_dirty;
@@ -202,7 +203,13 @@ struct CPUState {
     uint32_t halted; /* used by alpha, cris, ppc TCG */
 };
 
-extern CPUState *first_cpu;
+QTAILQ_HEAD(CPUTailQ, CPUState);
+extern struct CPUTailQ cpus;
+#define CPU_NEXT(cpu) QTAILQ_NEXT(cpu, node)
+#define CPU_FOREACH(cpu) QTAILQ_FOREACH(cpu, &cpus, node)
+#define CPU_FOREACH_SAFE(cpu, next_cpu) \
+    QTAILQ_FOREACH_SAFE(cpu, &cpus, node, next_cpu)
+#define first_cpu QTAILQ_FIRST(&cpus)
 
 DECLARE_TLS(CPUState *, current_cpu);
 #define current_cpu tls_var(current_cpu)
@@ -395,15 +402,6 @@ void run_on_cpu(CPUState *cpu, void (*func)(void *data), void *data);
  */
 void async_run_on_cpu(CPUState *cpu, void (*func)(void *data), void *data);
 
-/**
- * qemu_for_each_cpu:
- * @func: The function to be executed.
- * @data: Data to pass to the function.
- *
- * Executes @func for each CPU.
- */
-void qemu_for_each_cpu(void (*func)(CPUState *cpu, void *data), void *data);
-
 /**
  * qemu_get_cpu:
  * @index: The CPUState@cpu_index value of the CPU to obtain.
diff --git a/kvm-all.c b/kvm-all.c
index 875e32ec875b6242338e3b89e74e8bb8d928058a..c29a015ccaae9dffcdf935e9cc806f507cbad50c 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -1925,7 +1925,7 @@ int kvm_insert_breakpoint(CPUState *cpu, target_ulong addr,
         }
     }
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         err = kvm_update_guest_debug(cpu, 0);
         if (err) {
             return err;
@@ -1965,7 +1965,7 @@ int kvm_remove_breakpoint(CPUState *cpu, target_ulong addr,
         }
     }
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         err = kvm_update_guest_debug(cpu, 0);
         if (err) {
             return err;
@@ -1982,7 +1982,7 @@ void kvm_remove_all_breakpoints(CPUState *cpu)
     QTAILQ_FOREACH_SAFE(bp, &s->kvm_sw_breakpoints, entry, next) {
         if (kvm_arch_remove_sw_breakpoint(cpu, bp) != 0) {
             /* Try harder to find a CPU that currently sees the breakpoint. */
-            for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+            CPU_FOREACH(cpu) {
                 if (kvm_arch_remove_sw_breakpoint(cpu, bp) == 0) {
                     break;
                 }
@@ -1993,7 +1993,7 @@ void kvm_remove_all_breakpoints(CPUState *cpu)
     }
     kvm_arch_remove_all_hw_breakpoints();
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         kvm_update_guest_debug(cpu, 0);
     }
 }
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index 7ce2eab1bbe7bf2281af59bb644630ddcc6318fe..72d92707c61bd6d3113670948235ceda0b26fc82 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -2668,7 +2668,7 @@ static int fill_note_info(struct elf_note_info *info,
 
     /* read and fill status of all threads */
     cpu_list_lock();
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         if (cpu == thread_cpu) {
             continue;
         }
diff --git a/linux-user/main.c b/linux-user/main.c
index 03859bcc23adb407e897e442275fd6aafb6fd7cf..5c2f7b26b40191973785038ee9f843b6930a90bf 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -117,10 +117,14 @@ void fork_end(int child)
 {
     mmap_fork_end(child);
     if (child) {
+        CPUState *cpu, *next_cpu;
         /* Child processes created by fork() only have a single thread.
            Discard information about the parent threads.  */
-        first_cpu = thread_cpu;
-        first_cpu->next_cpu = NULL;
+        CPU_FOREACH_SAFE(cpu, next_cpu) {
+            if (cpu != thread_cpu) {
+                QTAILQ_REMOVE(&cpus, thread_cpu, node);
+            }
+        }
         pending_cpus = 0;
         pthread_mutex_init(&exclusive_lock, NULL);
         pthread_mutex_init(&cpu_list_mutex, NULL);
@@ -154,7 +158,7 @@ static inline void start_exclusive(void)
 
     pending_cpus = 1;
     /* Make all other cpus stop executing.  */
-    for (other_cpu = first_cpu; other_cpu; other_cpu = other_cpu->next_cpu) {
+    CPU_FOREACH(other_cpu) {
         if (other_cpu->running) {
             pending_cpus++;
             cpu_exit(other_cpu);
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index f986548c2dc62154a862ec90e33eaafa12d5e78b..ecead512a0269ef6bb2cddf09fd0ad4483df5291 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -5113,25 +5113,12 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
            Do thread termination if we have more then one thread.  */
         /* FIXME: This probably breaks if a signal arrives.  We should probably
            be disabling signals.  */
-        if (first_cpu->next_cpu) {
+        if (CPU_NEXT(first_cpu)) {
             TaskState *ts;
-            CPUState **lastp;
-            CPUState *p;
 
             cpu_list_lock();
-            lastp = &first_cpu;
-            p = first_cpu;
-            while (p && p != cpu) {
-                lastp = &p->next_cpu;
-                p = p->next_cpu;
-            }
-            /* If we didn't find the CPU for this thread then something is
-               horribly wrong.  */
-            if (!p) {
-                abort();
-            }
             /* Remove the CPU from the list.  */
-            *lastp = p->next_cpu;
+            QTAILQ_REMOVE(&cpus, cpu, node);
             cpu_list_unlock();
             ts = ((CPUArchState *)cpu_env)->opaque;
             if (ts->child_tidptr) {
diff --git a/memory_mapping.c b/memory_mapping.c
index eeeeb440266b6aa76fb2d3cc6bbf3bb7945fa34d..87a6ed5c8e198c2055dcfbdb15adbd2330735e8f 100644
--- a/memory_mapping.c
+++ b/memory_mapping.c
@@ -270,7 +270,7 @@ static CPUState *find_paging_enabled_cpu(CPUState *start_cpu)
 {
     CPUState *cpu;
 
-    for (cpu = start_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         if (cpu_paging_enabled(cpu)) {
             return cpu;
         }
@@ -289,7 +289,8 @@ void qemu_get_guest_memory_mapping(MemoryMappingList *list,
 
     first_paging_enabled_cpu = find_paging_enabled_cpu(first_cpu);
     if (first_paging_enabled_cpu) {
-        for (cpu = first_paging_enabled_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        for (cpu = first_paging_enabled_cpu; cpu != NULL;
+             cpu = CPU_NEXT(cpu)) {
             Error *err = NULL;
             cpu_get_memory_mapping(cpu, list, &err);
             if (err) {
diff --git a/monitor.c b/monitor.c
index ec31b6e6a527677cd157f9476a42017679be6042..74f3f1ba603c60390f767ab0720389d1d2eed10a 100644
--- a/monitor.c
+++ b/monitor.c
@@ -2003,7 +2003,7 @@ static void do_info_numa(Monitor *mon, const QDict *qdict)
     monitor_printf(mon, "%d nodes\n", nb_numa_nodes);
     for (i = 0; i < nb_numa_nodes; i++) {
         monitor_printf(mon, "node %d cpus:", i);
-        for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+        CPU_FOREACH(cpu) {
             if (cpu->numa_node == i) {
                 monitor_printf(mon, " %d", cpu->cpu_index);
             }
diff --git a/qom/cpu.c b/qom/cpu.c
index e71e57bd6b2634a19c8dd4c05ee5511b47dee3d4..fa7ec6b199c4204006d321fdda22af8f3217d812 100644
--- a/qom/cpu.c
+++ b/qom/cpu.c
@@ -25,30 +25,18 @@
 #include "qemu/log.h"
 #include "sysemu/sysemu.h"
 
-typedef struct CPUExistsArgs {
-    int64_t id;
-    bool found;
-} CPUExistsArgs;
-
-static void cpu_exist_cb(CPUState *cpu, void *data)
-{
-    CPUClass *klass = CPU_GET_CLASS(cpu);
-    CPUExistsArgs *arg = data;
-
-    if (klass->get_arch_id(cpu) == arg->id) {
-        arg->found = true;
-    }
-}
-
 bool cpu_exists(int64_t id)
 {
-    CPUExistsArgs data = {
-        .id = id,
-        .found = false,
-    };
+    CPUState *cpu;
+
+    CPU_FOREACH(cpu) {
+        CPUClass *cc = CPU_GET_CLASS(cpu);
 
-    qemu_for_each_cpu(cpu_exist_cb, &data);
-    return data.found;
+        if (cc->get_arch_id(cpu) == id) {
+            return true;
+        }
+    }
+    return false;
 }
 
 bool cpu_paging_enabled(const CPUState *cpu)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index 42c5de034ecc84cddbc53a25e29a0721b900f67f..c36345e426d70930e3998d31e69da014f67fd877 100644
--- a/target-i386/cpu.c
+++ b/target-i386/cpu.c
@@ -48,6 +48,118 @@
 #include "hw/i386/apic_internal.h"
 #endif
 
+
+/* Cache topology CPUID constants: */
+
+/* CPUID Leaf 2 Descriptors */
+
+#define CPUID_2_L1D_32KB_8WAY_64B 0x2c
+#define CPUID_2_L1I_32KB_8WAY_64B 0x30
+#define CPUID_2_L2_2MB_8WAY_64B   0x7d
+
+
+/* CPUID Leaf 4 constants: */
+
+/* EAX: */
+#define CPUID_4_TYPE_DCACHE  1
+#define CPUID_4_TYPE_ICACHE  2
+#define CPUID_4_TYPE_UNIFIED 3
+
+#define CPUID_4_LEVEL(l)          ((l) << 5)
+
+#define CPUID_4_SELF_INIT_LEVEL (1 << 8)
+#define CPUID_4_FULLY_ASSOC     (1 << 9)
+
+/* EDX: */
+#define CPUID_4_NO_INVD_SHARING (1 << 0)
+#define CPUID_4_INCLUSIVE       (1 << 1)
+#define CPUID_4_COMPLEX_IDX     (1 << 2)
+
+#define ASSOC_FULL 0xFF
+
+/* AMD associativity encoding used on CPUID Leaf 0x80000006: */
+#define AMD_ENC_ASSOC(a) (a <=   1 ? a   : \
+                          a ==   2 ? 0x2 : \
+                          a ==   4 ? 0x4 : \
+                          a ==   8 ? 0x6 : \
+                          a ==  16 ? 0x8 : \
+                          a ==  32 ? 0xA : \
+                          a ==  48 ? 0xB : \
+                          a ==  64 ? 0xC : \
+                          a ==  96 ? 0xD : \
+                          a == 128 ? 0xE : \
+                          a == ASSOC_FULL ? 0xF : \
+                          0 /* invalid value */)
+
+
+/* Definitions of the hardcoded cache entries we expose: */
+
+/* L1 data cache: */
+#define L1D_LINE_SIZE         64
+#define L1D_ASSOCIATIVITY      8
+#define L1D_SETS              64
+#define L1D_PARTITIONS         1
+/* Size = LINE_SIZE*ASSOCIATIVITY*SETS*PARTITIONS = 32KiB */
+#define L1D_DESCRIPTOR CPUID_2_L1D_32KB_8WAY_64B
+/*FIXME: CPUID leaf 0x80000005 is inconsistent with leaves 2 & 4 */
+#define L1D_LINES_PER_TAG      1
+#define L1D_SIZE_KB_AMD       64
+#define L1D_ASSOCIATIVITY_AMD  2
+
+/* L1 instruction cache: */
+#define L1I_LINE_SIZE         64
+#define L1I_ASSOCIATIVITY      8
+#define L1I_SETS              64
+#define L1I_PARTITIONS         1
+/* Size = LINE_SIZE*ASSOCIATIVITY*SETS*PARTITIONS = 32KiB */
+#define L1I_DESCRIPTOR CPUID_2_L1I_32KB_8WAY_64B
+/*FIXME: CPUID leaf 0x80000005 is inconsistent with leaves 2 & 4 */
+#define L1I_LINES_PER_TAG      1
+#define L1I_SIZE_KB_AMD       64
+#define L1I_ASSOCIATIVITY_AMD  2
+
+/* Level 2 unified cache: */
+#define L2_LINE_SIZE          64
+#define L2_ASSOCIATIVITY      16
+#define L2_SETS             4096
+#define L2_PARTITIONS          1
+/* Size = LINE_SIZE*ASSOCIATIVITY*SETS*PARTITIONS = 4MiB */
+/*FIXME: CPUID leaf 2 descriptor is inconsistent with CPUID leaf 4 */
+#define L2_DESCRIPTOR CPUID_2_L2_2MB_8WAY_64B
+/*FIXME: CPUID leaf 0x80000006 is inconsistent with leaves 2 & 4 */
+#define L2_LINES_PER_TAG       1
+#define L2_SIZE_KB_AMD       512
+
+/* No L3 cache: */
+#define L3_SIZE_KB             0 /* disabled */
+#define L3_ASSOCIATIVITY       0 /* disabled */
+#define L3_LINES_PER_TAG       0 /* disabled */
+#define L3_LINE_SIZE           0 /* disabled */
+
+/* TLB definitions: */
+
+#define L1_DTLB_2M_ASSOC       1
+#define L1_DTLB_2M_ENTRIES   255
+#define L1_DTLB_4K_ASSOC       1
+#define L1_DTLB_4K_ENTRIES   255
+
+#define L1_ITLB_2M_ASSOC       1
+#define L1_ITLB_2M_ENTRIES   255
+#define L1_ITLB_4K_ASSOC       1
+#define L1_ITLB_4K_ENTRIES   255
+
+#define L2_DTLB_2M_ASSOC       0 /* disabled */
+#define L2_DTLB_2M_ENTRIES     0 /* disabled */
+#define L2_DTLB_4K_ASSOC       4
+#define L2_DTLB_4K_ENTRIES   512
+
+#define L2_ITLB_2M_ASSOC       0 /* disabled */
+#define L2_ITLB_2M_ENTRIES     0 /* disabled */
+#define L2_ITLB_4K_ASSOC       4
+#define L2_ITLB_4K_ENTRIES   512
+
+
+
 static void x86_cpu_vendor_words2str(char *dst, uint32_t vendor1,
                                      uint32_t vendor2, uint32_t vendor3)
 {
@@ -1950,10 +2062,12 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
         break;
     case 2:
         /* cache info: needed for Pentium Pro compatibility */
-        *eax = 1;
+        *eax = 1; /* Number of CPUID[EAX=2] calls required */
         *ebx = 0;
         *ecx = 0;
-        *edx = 0x2c307d;
+        *edx = (L1D_DESCRIPTOR << 16) | \
+               (L1I_DESCRIPTOR <<  8) | \
+               (L2_DESCRIPTOR);
         break;
     case 4:
         /* cache info: needed for Core compatibility */
@@ -1964,25 +2078,37 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
         }
         switch (count) {
             case 0: /* L1 dcache info */
-                *eax |= 0x0000121;
-                *ebx = 0x1c0003f;
-                *ecx = 0x000003f;
-                *edx = 0x0000001;
+                *eax |= CPUID_4_TYPE_DCACHE | \
+                        CPUID_4_LEVEL(1) | \
+                        CPUID_4_SELF_INIT_LEVEL;
+                *ebx = (L1D_LINE_SIZE - 1) | \
+                       ((L1D_PARTITIONS - 1) << 12) | \
+                       ((L1D_ASSOCIATIVITY - 1) << 22);
+                *ecx = L1D_SETS - 1;
+                *edx = CPUID_4_NO_INVD_SHARING;
                 break;
             case 1: /* L1 icache info */
-                *eax |= 0x0000122;
-                *ebx = 0x1c0003f;
-                *ecx = 0x000003f;
-                *edx = 0x0000001;
+                *eax |= CPUID_4_TYPE_ICACHE | \
+                        CPUID_4_LEVEL(1) | \
+                        CPUID_4_SELF_INIT_LEVEL;
+                *ebx = (L1I_LINE_SIZE - 1) | \
+                       ((L1I_PARTITIONS - 1) << 12) | \
+                       ((L1I_ASSOCIATIVITY - 1) << 22);
+                *ecx = L1I_SETS - 1;
+                *edx = CPUID_4_NO_INVD_SHARING;
                 break;
             case 2: /* L2 cache info */
-                *eax |= 0x0000143;
+                *eax |= CPUID_4_TYPE_UNIFIED | \
+                        CPUID_4_LEVEL(2) | \
+                        CPUID_4_SELF_INIT_LEVEL;
                 if (cs->nr_threads > 1) {
                     *eax |= (cs->nr_threads - 1) << 14;
                 }
-                *ebx = 0x3c0003f;
-                *ecx = 0x0000fff;
-                *edx = 0x0000001;
+                *ebx = (L2_LINE_SIZE - 1) | \
+                       ((L2_PARTITIONS - 1) << 12) | \
+                       ((L2_ASSOCIATIVITY - 1) << 22);
+                *ecx = L2_SETS - 1;
+                *edx = CPUID_4_NO_INVD_SHARING;
                 break;
             default: /* end of info */
                 *eax = 0;
@@ -2102,17 +2228,31 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
         break;
     case 0x80000005:
         /* cache info (L1 cache) */
-        *eax = 0x01ff01ff;
-        *ebx = 0x01ff01ff;
-        *ecx = 0x40020140;
-        *edx = 0x40020140;
+        *eax = (L1_DTLB_2M_ASSOC << 24) | (L1_DTLB_2M_ENTRIES << 16) | \
+               (L1_ITLB_2M_ASSOC <<  8) | (L1_ITLB_2M_ENTRIES);
+        *ebx = (L1_DTLB_4K_ASSOC << 24) | (L1_DTLB_4K_ENTRIES << 16) | \
+               (L1_ITLB_4K_ASSOC <<  8) | (L1_ITLB_4K_ENTRIES);
+        *ecx = (L1D_SIZE_KB_AMD << 24) | (L1D_ASSOCIATIVITY_AMD << 16) | \
+               (L1D_LINES_PER_TAG << 8) | (L1D_LINE_SIZE);
+        *edx = (L1I_SIZE_KB_AMD << 24) | (L1I_ASSOCIATIVITY_AMD << 16) | \
+               (L1I_LINES_PER_TAG << 8) | (L1I_LINE_SIZE);
         break;
     case 0x80000006:
         /* cache info (L2 cache) */
-        *eax = 0;
-        *ebx = 0x42004200;
-        *ecx = 0x02008140;
-        *edx = 0;
+        *eax = (AMD_ENC_ASSOC(L2_DTLB_2M_ASSOC) << 28) | \
+               (L2_DTLB_2M_ENTRIES << 16) | \
+               (AMD_ENC_ASSOC(L2_ITLB_2M_ASSOC) << 12) | \
+               (L2_ITLB_2M_ENTRIES);
+        *ebx = (AMD_ENC_ASSOC(L2_DTLB_4K_ASSOC) << 28) | \
+               (L2_DTLB_4K_ENTRIES << 16) | \
+               (AMD_ENC_ASSOC(L2_ITLB_4K_ASSOC) << 12) | \
+               (L2_ITLB_4K_ENTRIES);
+        *ecx = (L2_SIZE_KB_AMD << 16) | \
+               (AMD_ENC_ASSOC(L2_ASSOCIATIVITY) << 12) | \
+               (L2_LINES_PER_TAG << 8) | (L2_LINE_SIZE);
+        *edx = ((L3_SIZE_KB/512) << 18) | \
+               (AMD_ENC_ASSOC(L3_ASSOCIATIVITY) << 12) | \
+               (L3_LINES_PER_TAG << 8) | (L3_LINE_SIZE);
         break;
     case 0x80000008:
         /* virtual & phys address size in low 2 bytes. */
diff --git a/target-i386/helper.c b/target-i386/helper.c
index bf3e2ac73dec3380e192877f7357c87253ad4983..7c58e274d928d291dc6cf8f52ce942060aeaa996 100644
--- a/target-i386/helper.c
+++ b/target-i386/helper.c
@@ -1231,8 +1231,7 @@ void cpu_x86_inject_mce(Monitor *mon, X86CPU *cpu, int bank,
         params.mcg_status = MCG_STATUS_MCIP | MCG_STATUS_RIPV;
         params.addr = 0;
         params.misc = 0;
-        for (other_cs = first_cpu; other_cs != NULL;
-             other_cs = other_cs->next_cpu) {
+        CPU_FOREACH(other_cs) {
             if (other_cs == cs) {
                 continue;
             }
diff --git a/target-i386/misc_helper.c b/target-i386/misc_helper.c
index 957926ced7aff8702899b75804a6eb9b05c08968..93933fd1628d2048b7ddf437409282f3072b738a 100644
--- a/target-i386/misc_helper.c
+++ b/target-i386/misc_helper.c
@@ -610,7 +610,7 @@ void helper_mwait(CPUX86State *env, int next_eip_addend)
     cpu = x86_env_get_cpu(env);
     cs = CPU(cpu);
     /* XXX: not complete but not completely erroneous */
-    if (cs->cpu_index != 0 || cs->next_cpu != NULL) {
+    if (cs->cpu_index != 0 || CPU_NEXT(cs) != NULL) {
         /* more than one CPU: do not sleep because another CPU may
            wake this one */
     } else {
diff --git a/target-mips/op_helper.c b/target-mips/op_helper.c
index b828375714e0f4a8ca02b43eb95131c87937985d..8e3a6d7da672fe89cfdde8bc6579e378f1c3c66a 100644
--- a/target-mips/op_helper.c
+++ b/target-mips/op_helper.c
@@ -1699,15 +1699,14 @@ target_ulong helper_dvpe(CPUMIPSState *env)
     CPUState *other_cs = first_cpu;
     target_ulong prev = env->mvp->CP0_MVPControl;
 
-    do {
+    CPU_FOREACH(other_cs) {
         MIPSCPU *other_cpu = MIPS_CPU(other_cs);
         /* Turn off all VPEs except the one executing the dvpe.  */
         if (&other_cpu->env != env) {
             other_cpu->env.mvp->CP0_MVPControl &= ~(1 << CP0MVPCo_EVP);
             mips_vpe_sleep(other_cpu);
         }
-        other_cs = other_cs->next_cpu;
-    } while (other_cs);
+    }
     return prev;
 }
 
@@ -1716,7 +1715,7 @@ target_ulong helper_evpe(CPUMIPSState *env)
     CPUState *other_cs = first_cpu;
     target_ulong prev = env->mvp->CP0_MVPControl;
 
-    do {
+    CPU_FOREACH(other_cs) {
         MIPSCPU *other_cpu = MIPS_CPU(other_cs);
 
         if (&other_cpu->env != env
@@ -1726,8 +1725,7 @@ target_ulong helper_evpe(CPUMIPSState *env)
             other_cpu->env.mvp->CP0_MVPControl |= (1 << CP0MVPCo_EVP);
             mips_vpe_wake(other_cpu); /* And wake it up.  */
         }
-        other_cs = other_cs->next_cpu;
-    } while (other_cs);
+    }
     return prev;
 }
 #endif /* !CONFIG_USER_ONLY */
diff --git a/target-ppc/excp_helper.c b/target-ppc/excp_helper.c
index e9577611094a5cce7fffffde1a68066acc9ae99a..c959460f7087d9c9f9c9a3ed2e8dafd0bc645183 100644
--- a/target-ppc/excp_helper.c
+++ b/target-ppc/excp_helper.c
@@ -1002,7 +1002,7 @@ void helper_msgsnd(target_ulong rb)
         return;
     }
 
-    for (cs = first_cpu; cs != NULL; cs = cs->next_cpu) {
+    CPU_FOREACH(cs) {
         PowerPCCPU *cpu = POWERPC_CPU(cs);
         CPUPPCState *cenv = &cpu->env;
 
diff --git a/target-s390x/misc_helper.c b/target-s390x/misc_helper.c
index 4afd7dab1c2a81f31f67ffecddb5fd0888ce5b47..16909071696bda612436b6ad3cdbeb7618337081 100644
--- a/target-s390x/misc_helper.c
+++ b/target-s390x/misc_helper.c
@@ -183,12 +183,12 @@ uint32_t HELPER(servc)(CPUS390XState *env, uint64_t r1, uint64_t r2)
 #ifndef CONFIG_USER_ONLY
 static void cpu_reset_all(void)
 {
-    CPUState *cpu;
+    CPUState *cs;
     S390CPUClass *scc;
 
-    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
-        scc = S390_CPU_GET_CLASS(CPU(cpu));
-        scc->cpu_reset(CPU(cpu));
+    CPU_FOREACH(cs) {
+        scc = S390_CPU_GET_CLASS(cs);
+        scc->cpu_reset(cs);
     }
 }
 
diff --git a/translate-all.c b/translate-all.c
index 3b5fc7c901db015fa1dd22b20da3c7c035dd223f..2c923c644b023180e2043a080a53e5a56aca671d 100644
--- a/translate-all.c
+++ b/translate-all.c
@@ -696,7 +696,7 @@ void tb_flush(CPUArchState *env1)
     }
     tcg_ctx.tb_ctx.nb_tbs = 0;
 
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         CPUArchState *env = cpu->env_ptr;
 
         memset(env->tb_jmp_cache, 0, TB_JMP_CACHE_SIZE * sizeof(void *));
@@ -850,7 +850,7 @@ void tb_phys_invalidate(TranslationBlock *tb, tb_page_addr_t page_addr)
 
     /* remove the TB from the hash list */
     h = tb_jmp_cache_hash_func(tb->pc);
-    for (cpu = first_cpu; cpu != NULL; cpu = cpu->next_cpu) {
+    CPU_FOREACH(cpu) {
         CPUArchState *env = cpu->env_ptr;
 
         if (env->tb_jmp_cache[h] == tb) {