[Qemu-devel] [patch 10/11] qemu: make iothread selectable at compile time

[Marcelo Tosatti]

Turned off by default.

Signed-off-by: Marcelo Tosatti <address@hidden>

Index: trunk/vl.c
===================================================================
--- trunk.orig/vl.c
+++ trunk/vl.c
@@ -277,6 +277,7 @@ static QEMUTimer *nographic_timer;
 
 uint8_t qemu_uuid[16];
 
+#ifdef QEMU_THREADED
 static int io_thread_fd = -1;
 
 QemuMutex qemu_global_mutex;
@@ -294,6 +295,7 @@ QemuCond qemu_cpu_cond;
 QemuCond qemu_system_cond;
 
 QemuCond qemu_pause_cond;
+#endif
 
 static void pause_all_vcpus(void);
 static void resume_all_vcpus(void);
@@ -1367,6 +1369,14 @@ static void host_alarm_handler(int host_
         write(alarm_timer_wfd, &byte, sizeof(byte));
 #endif
         alarm_timer->flags |= ALARM_FLAG_EXPIRED;
+#ifndef QEMU_THREADED
+        {
+            CPUState *env = next_cpu;
+            if (env)
+                cpu_exit(env);
+            event_pending = 1;
+        }
+#endif
     }
 }
 
@@ -3343,6 +3353,7 @@ static int ram_load(QEMUFile *f, void *o
 
 void qemu_service_io(void)
 {
+    qemu_notify_event();
 }
 
 /***********************************************************/
@@ -3560,11 +3571,13 @@ static int qemu_vmstop_requested(void)
     return r;
 }
 
+#ifdef QEMU_THREADED
 static void qemu_system_vmstop_request(int reason)
 {
     vmstop_requested = reason;
     main_loop_break();
 }
+#endif
 
 static void __vm_stop(int reason)
 {
@@ -3578,6 +3591,7 @@ static void __vm_stop(int reason)
 
 void vm_stop(int reason)
 {
+#ifdef QEMU_THREADED
     QemuThread me;
     qemu_thread_self(&me);
 
@@ -3588,6 +3602,7 @@ void vm_stop(int reason)
             cpu_single_env->stop = 1;
         return;
     }
+#endif
     __vm_stop(reason);
 }
 
@@ -3671,6 +3686,7 @@ void qemu_system_powerdown_request(void)
     main_loop_break();
 }
 
+#ifdef QEMU_THREADED
 void qemu_notify_event(void)
 {
     main_loop_break();
@@ -3725,74 +3741,6 @@ static void io_thread_wakeup(void *opaqu
     }
 }
 
-#ifdef _WIN32
-static void host_main_loop_wait(int *timeout)
-{
-    int ret, ret2, i;
-    PollingEntry *pe;
-
-
-    /* XXX: need to suppress polling by better using win32 events */
-    ret = 0;
-    for(pe = first_polling_entry; pe != NULL; pe = pe->next) {
-        ret |= pe->func(pe->opaque);
-    }
-    if (ret == 0) {
-        int err;
-        WaitObjects *w = &wait_objects;
-
-        ret = WaitForMultipleObjects(w->num, w->events, FALSE, *timeout);
-        if (WAIT_OBJECT_0 + 0 <= ret && ret <= WAIT_OBJECT_0 + w->num - 1) {
-            if (w->func[ret - WAIT_OBJECT_0])
-                w->func[ret - WAIT_OBJECT_0](w->opaque[ret - WAIT_OBJECT_0]);
-
-            /* Check for additional signaled events */
-            for(i = (ret - WAIT_OBJECT_0 + 1); i < w->num; i++) {
-
-                /* Check if event is signaled */
-                ret2 = WaitForSingleObject(w->events[i], 0);
-                if(ret2 == WAIT_OBJECT_0) {
-                    if (w->func[i])
-                        w->func[i](w->opaque[i]);
-                } else if (ret2 == WAIT_TIMEOUT) {
-                } else {
-                    err = GetLastError();
-                    fprintf(stderr, "WaitForSingleObject error %d %d\n", i, 
err);
-                }
-            }
-        } else if (ret == WAIT_TIMEOUT) {
-        } else {
-            err = GetLastError();
-            fprintf(stderr, "WaitForMultipleObjects error %d %d\n", ret, err);
-        }
-    }
-
-    *timeout = 0;
-}
-#else
-static void host_main_loop_wait(int *timeout)
-{
-}
-#endif
-
-/* Q: is it allowed to enter cpu_exec */
-static int cpu_can_run(CPUState *env)
-{
-    if (env->stop)
-        return 0;
-    if (env->stopped)
-        return 0;
-    if (shutdown_requested)
-        return 0;
-    if (powerdown_requested)
-        return 0;
-    if (reset_requested)
-        return 0;
-    if (vmstop_requested)
-        return 0;
-    return 1;
-}
-
 /* Q: should break out of the wait loop */
 static int cpu_has_work(CPUState *env)
 {
@@ -3950,6 +3898,187 @@ static void resume_all_vcpus(void)
     }
 }
 
+static void *cpu_thread(void *arg);
+
+void qemu_init_vcpu(void *_env)
+{
+    CPUState *env = _env;
+    /* share a single thread for all cpus with TCG */
+    if (!tcg_cpu_thread) {
+        env->thread = qemu_mallocz(sizeof(QemuThread));
+        env->halt_cond = qemu_mallocz(sizeof(QemuCond));
+        qemu_cond_init(env->halt_cond);
+        qemu_thread_create(env->thread, cpu_thread, env);
+        while (env->created == 0)
+            qemu_cond_timedwait(&qemu_cpu_cond, &qemu_global_mutex, 100);
+        tcg_cpu_thread = env->thread;
+        tcg_halt_cond = env->halt_cond;
+    } else {
+        env->thread = tcg_cpu_thread;
+        env->halt_cond = tcg_halt_cond;
+    }
+}
+
+static void qemu_init_state(void)
+{
+    qemu_cond_init(&qemu_pause_cond);
+    qemu_mutex_init(&qemu_fair_mutex);
+    qemu_mutex_init(&qemu_global_mutex);
+    qemu_mutex_lock(&qemu_global_mutex);
+}
+
+static void setup_iothread_fd(void);
+static void pre_main_loop(void)
+{
+    qemu_thread_self(&io_thread);
+    setup_iothread_fd();
+
+    unblock_io_signals();
+
+    qemu_system_ready = 1;
+    qemu_cond_broadcast(&qemu_system_cond);
+}
+
+static int qemu_cpu_exec(CPUState *env);
+static void *cpu_thread(void *arg)
+{
+    int r;
+    CPUState *env = arg;
+
+    block_io_signals();
+    qemu_thread_self(env->thread);
+
+    /* signal CPU creation */
+    qemu_mutex_lock(&qemu_global_mutex);
+    env->created = 1;
+    qemu_cond_signal(&qemu_cpu_cond);
+
+    /* and wait for machine initialization */
+    while (!qemu_system_ready)
+        qemu_cond_timedwait(&qemu_system_cond, &qemu_global_mutex, 100);
+
+    while (1) {
+        r = qemu_cpu_exec(env);
+        qemu_wait_io_event(env, r);
+    }
+
+    return NULL;
+}
+
+static void setup_iothread_fd(void)
+{
+    int fds[2];
+
+    if (pipe(fds) == -1) {
+        fprintf(stderr, "failed to create iothread pipe");
+        exit(0);
+    }
+
+    qemu_set_fd_handler2(fds[0], NULL, io_thread_wakeup, NULL,
+                         (void *)(unsigned long)fds[0]);
+    io_thread_fd = fds[1];
+    fcntl(io_thread_fd, F_SETFL, O_NONBLOCK);
+}
+
+#else /* QEMU_THREADED */
+#define qemu_signal_lock(a)         do { } while(0)
+#define qemu_mutex_unlock(a)         do { } while(0)
+#define pre_main_loop()             do { } while(0)
+#define qemu_init_state()           do { } while(0)
+
+void pause_all_vcpus(void)
+{
+    return;
+}
+
+void resume_all_vcpus(void)
+{
+    return;
+}
+
+void qemu_cpu_kick(void *env)
+{
+    return;
+}
+
+void qemu_init_vcpu(void *_env)
+{
+    CPUState *env = _env;
+
+    if (kvm_enabled())
+        kvm_init_vcpu(env);
+    return;
+}
+
+void qemu_notify_event(void)
+{
+    CPUState *env = cpu_single_env;
+
+    if (env)
+        cpu_exit(env);
+}
+
+void main_loop_break(void)
+{
+    qemu_notify_event();
+}
+
+int qemu_cpu_self(void *env)
+{
+    return 1;
+}
+#endif
+
+#ifdef _WIN32
+static void host_main_loop_wait(int *timeout)
+{
+    int ret, ret2, i;
+    PollingEntry *pe;
+
+
+    /* XXX: need to suppress polling by better using win32 events */
+    ret = 0;
+    for(pe = first_polling_entry; pe != NULL; pe = pe->next) {
+        ret |= pe->func(pe->opaque);
+    }
+    if (ret == 0) {
+        int err;
+        WaitObjects *w = &wait_objects;
+
+        ret = WaitForMultipleObjects(w->num, w->events, FALSE, *timeout);
+        if (WAIT_OBJECT_0 + 0 <= ret && ret <= WAIT_OBJECT_0 + w->num - 1) {
+            if (w->func[ret - WAIT_OBJECT_0])
+                w->func[ret - WAIT_OBJECT_0](w->opaque[ret - WAIT_OBJECT_0]);
+
+            /* Check for additional signaled events */
+            for(i = (ret - WAIT_OBJECT_0 + 1); i < w->num; i++) {
+
+                /* Check if event is signaled */
+                ret2 = WaitForSingleObject(w->events[i], 0);
+                if(ret2 == WAIT_OBJECT_0) {
+                    if (w->func[i])
+                        w->func[i](w->opaque[i]);
+                } else if (ret2 == WAIT_TIMEOUT) {
+                } else {
+                    err = GetLastError();
+                    fprintf(stderr, "WaitForSingleObject error %d %d\n", i, 
err);
+                }
+            }
+        } else if (ret == WAIT_TIMEOUT) {
+        } else {
+            err = GetLastError();
+            fprintf(stderr, "WaitForMultipleObjects error %d %d\n", ret, err);
+        }
+    }
+
+    *timeout = 0;
+}
+#else
+static void host_main_loop_wait(int *timeout)
+{
+}
+#endif
+
 void main_loop_wait(int timeout)
 {
     IOHandlerRecord *ioh;
@@ -4051,203 +4180,155 @@ void main_loop_wait(int timeout)
 
 }
 
-static void setup_iothread_fd(void)
+/* Q: is it allowed to enter cpu_exec */
+static int cpu_can_run(CPUState *env)
 {
-    int fds[2];
-
-    if (pipe(fds) == -1) {
-        fprintf(stderr, "failed to create iothread pipe");
-        exit(0);
-    }
-
-    qemu_set_fd_handler2(fds[0], NULL, io_thread_wakeup, NULL,
-                         (void *)(unsigned long)fds[0]);
-    io_thread_fd = fds[1];
-    fcntl(io_thread_fd, F_SETFL, O_NONBLOCK);
+    if (env->stop)
+        return 0;
+    if (env->stopped)
+        return 0;
+    if (shutdown_requested)
+        return 0;
+    if (powerdown_requested)
+        return 0;
+    if (reset_requested)
+        return 0;
+    if (vmstop_requested)
+        return 0;
+    return 1;
 }
 
-static void *cpu_main_loop(void *arg)
+static int qemu_cpu_exec(CPUState *env)
 {
     int ret, timeout;
 #ifdef CONFIG_PROFILER
     int64_t ti;
 #endif
-    CPUState *env = arg;
-
-    block_io_signals();
-    qemu_thread_self(env->thread);
-
-    /* signal CPU creation */
-    qemu_mutex_lock(&qemu_global_mutex);
-    env->created = 1;
-    qemu_cond_signal(&qemu_cpu_cond);
-
-    /* and wait for machine initialization */
-    while (!qemu_system_ready)
-        qemu_cond_timedwait(&qemu_system_cond, &qemu_global_mutex, 100);
-
-    cur_cpu = env = first_cpu;
-    next_cpu = cur_cpu->next_cpu ?: first_cpu;
     for(;;) {
-        if (vm_running) {
-
-            for(;;) {
-                /* get next cpu */
-                env = next_cpu;
+        /* get next cpu */
+        env = next_cpu;
 #ifdef CONFIG_PROFILER
-                ti = profile_getclock();
+        ti = profile_getclock();
 #endif
-                if (use_icount) {
-                    int64_t count;
-                    int decr;
-                    qemu_icount -= (env->icount_decr.u16.low + 
env->icount_extra);
-                    env->icount_decr.u16.low = 0;
-                    env->icount_extra = 0;
-                    count = qemu_next_deadline();
-                    count = (count + (1 << icount_time_shift) - 1)
-                            >> icount_time_shift;
-                    qemu_icount += count;
-                    decr = (count > 0xffff) ? 0xffff : count;
-                    count -= decr;
-                    env->icount_decr.u16.low = decr;
-                    env->icount_extra = count;
-                }
-                ret = EXCP_HALTED;
-                if (cpu_can_run(env))
-                    ret = cpu_exec(env);
+        if (use_icount) {
+            int64_t count;
+            int decr;
+            qemu_icount -= (env->icount_decr.u16.low + env->icount_extra);
+            env->icount_decr.u16.low = 0;
+            env->icount_extra = 0;
+            count = qemu_next_deadline();
+            count = (count + (1 << icount_time_shift) - 1)
+                    >> icount_time_shift;
+            qemu_icount += count;
+            decr = (count > 0xffff) ? 0xffff : count;
+            count -= decr;
+            env->icount_decr.u16.low = decr;
+            env->icount_extra = count;
+        }
+        ret = EXCP_HALTED;
+        if (cpu_can_run(env))
+            ret = cpu_exec(env);
 #ifdef CONFIG_PROFILER
-                qemu_time += profile_getclock() - ti;
+        qemu_time += profile_getclock() - ti;
 #endif
-                if (use_icount) {
-                    /* Fold pending instructions back into the
-                       instruction counter, and clear the interrupt flag.  */
-                    qemu_icount -= (env->icount_decr.u16.low
-                                    + env->icount_extra);
-                    env->icount_decr.u32 = 0;
-                    env->icount_extra = 0;
-                }
-                next_cpu = env->next_cpu ?: first_cpu;
-                if (event_pending && likely(ret != EXCP_DEBUG)) {
-                    ret = EXCP_INTERRUPT;
-                    event_pending = 0;
-                    break;
-                }
-                if (ret == EXCP_HLT) {
-                    /* Give the next CPU a chance to run.  */
-                    cur_cpu = env;
-                    continue;
-                }
-                if (ret != EXCP_HALTED)
-                    break;
-                /* all CPUs are halted ? */
-                if (env == cur_cpu)
-                    break;
-            }
+        if (use_icount) {
+            /* Fold pending instructions back into the
+            instruction counter, and clear the interrupt flag.  */
+            qemu_icount -= (env->icount_decr.u16.low
+                            + env->icount_extra);
+            env->icount_decr.u32 = 0;
+            env->icount_extra = 0;
+        }
+        next_cpu = env->next_cpu ?: first_cpu;
+        if (event_pending && likely(ret != EXCP_DEBUG)) {
+            ret = EXCP_INTERRUPT;
+            event_pending = 0;
+            break;
+        }
+        if (ret == EXCP_HLT) {
+            /* Give the next CPU a chance to run.  */
             cur_cpu = env;
-
-            if (unlikely(ret == EXCP_DEBUG)) {
-                gdb_set_stop_cpu(cur_cpu);
-                vm_stop(EXCP_DEBUG);
-            }
-            /* If all cpus are halted then wait until the next IRQ */
-            /* XXX: use timeout computed from timers */
-            if (ret == EXCP_HALTED) {
-                if (use_icount) {
-                    int64_t add;
-                    int64_t delta;
-                    /* Advance virtual time to the next event.  */
-                    if (use_icount == 1) {
-                        /* When not using an adaptive execution frequency
-                           we tend to get badly out of sync with real time,
-                           so just delay for a reasonable amount of time.  */
-                        delta = 0;
-                    } else {
-                        delta = cpu_get_icount() - cpu_get_clock();
-                    }
-                    if (delta > 0) {
-                        /* If virtual time is ahead of real time then just
-                           wait for IO.  */
-                        timeout = (delta / 1000000) + 1;
-                    } else {
-                        /* Wait for either IO to occur or the next
-                           timer event.  */
-                        add = qemu_next_deadline();
-                        /* We advance the timer before checking for IO.
-                           Limit the amount we advance so that early IO
-                           activity won't get the guest too far ahead.  */
-                        if (add > 10000000)
-                            add = 10000000;
-                        delta += add;
-                        add = (add + (1 << icount_time_shift) - 1)
-                              >> icount_time_shift;
-                        qemu_icount += add;
-                        timeout = delta / 1000000;
-                        if (timeout < 0)
-                            timeout = 0;
-                    }
-                } else {
-                    timeout = 5000;
-                }
-            } else {
-                timeout = 0;
-            }
-        } else {
-            env = env->next_cpu ?: first_cpu;
-            timeout = 5000;
+            continue;
         }
+        if (ret != EXCP_HALTED)
+            break;
+        /* all CPUs are halted ? */
+        if (env == cur_cpu)
+            break;
+     }
+     cur_cpu = env;
+
+     if (unlikely(ret == EXCP_DEBUG)) {
+         gdb_set_stop_cpu(cur_cpu);
+         vm_stop(EXCP_DEBUG);
+     }
+     /* If all cpus are halted then wait until the next IRQ */
+     /* XXX: use timeout computed from timers */
+     if (ret == EXCP_HALTED) {
+         if (use_icount) {
+             int64_t add;
+             int64_t delta;
+             /* Advance virtual time to the next event.  */
+             if (use_icount == 1) {
+                 /* When not using an adaptive execution frequency
+                    we tend to get badly out of sync with real time,
+                    so just delay for a reasonable amount of time.  */
+                 delta = 0;
+             } else {
+                 delta = cpu_get_icount() - cpu_get_clock();
+             }
+             if (delta > 0) {
+                 /* If virtual time is ahead of real time then just
+                    wait for IO.  */
+                 timeout = (delta / 1000000) + 1;
+             } else {
+                 /* Wait for either IO to occur or the next
+                    timer event.  */
+                 add = qemu_next_deadline();
+                 /* We advance the timer before checking for IO.
+                    Limit the amount we advance so that early IO
+                    activity won't get the guest too far ahead.  */
+                 if (add > 10000000)
+                     add = 10000000;
+                 delta += add;
+                 add = (add + (1 << icount_time_shift) - 1)
+                       >> icount_time_shift;
+                 qemu_icount += add;
+                 timeout = delta / 1000000;
+                 if (timeout < 0)
+                     timeout = 0;
+             }
+         } else {
+             timeout = 5000;
+         }
+     } else {
+         timeout = 0;
+     }
 #ifdef CONFIG_PROFILER
-        ti = profile_getclock();
+     ti = profile_getclock();
 #endif
-        qemu_wait_io_event(env, timeout);
 #ifdef CONFIG_PROFILER
-        dev_time += profile_getclock() - ti;
+     dev_time += profile_getclock() - ti;
 #endif
-    }
-    cpu_disable_ticks();
-    return NULL;
-}
-
-void qemu_init_vcpu(void *_env)
-{
-    CPUState *env = _env;
-    /* share a single thread for all cpus with TCG */
-    if (!tcg_cpu_thread) {
-        env->thread = qemu_mallocz(sizeof(QemuThread));
-        env->halt_cond = qemu_mallocz(sizeof(QemuCond));
-        qemu_cond_init(env->halt_cond);
-        qemu_thread_create(env->thread, cpu_main_loop, env);
-        while (env->created == 0)
-            qemu_cond_timedwait(&qemu_cpu_cond, &qemu_global_mutex, 100);
-        tcg_cpu_thread = env->thread;
-        tcg_halt_cond = env->halt_cond;
-    } else {
-        env->thread = tcg_cpu_thread;
-        env->halt_cond = tcg_halt_cond;
-    }
-}
-
-static void qemu_init_state(void)
-{
-    qemu_cond_init(&qemu_pause_cond);
-    qemu_mutex_init(&qemu_fair_mutex);
-    qemu_mutex_init(&qemu_global_mutex);
-    qemu_mutex_lock(&qemu_global_mutex);
+    return timeout;
 }
 
 static void main_loop(void)
 {
     int r;
 
-    qemu_thread_self(&io_thread);
-    setup_iothread_fd();
-
-    unblock_io_signals();
+    pre_main_loop();
 
-    qemu_system_ready = 1;
-    qemu_cond_broadcast(&qemu_system_cond);
+    cur_cpu = first_cpu;
+    next_cpu = cur_cpu->next_cpu ?: first_cpu;
 
     while (1) {
+#ifdef QEMU_THREADED
         main_loop_wait(1000);
+#else
+        r = qemu_cpu_exec(next_cpu);
+        main_loop_wait(r);
+#endif
         if (qemu_shutdown_requested()) {
             pause_all_vcpus();
             if (no_shutdown)

--