[Qemu-devel] [RFC V2 10/10] cpus: reclaim allocated vCPU objects

[Gu Zheng <guz.fnst@cn.fujitsu.com>]

After ACPI get a signal to eject a vCPU, the vCPU must be
removed from CPU list，before the vCPU really removed,  then
release the all related vCPU objects.
But we do not close KVM vcpu fd, just record it into a list, in
order to reuse it.

Signed-off-by: Chen Fan <chen.fan.fnst@cn.fujitsu.com>
Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
---
 cpus.c               |   37 ++++++++++++++++++++++++++++++++
 include/sysemu/kvm.h |    1 +
 kvm-all.c            |   57 +++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 94 insertions(+), 1 deletions(-)

diff --git a/cpus.c b/cpus.c
index eee693b..0608b41 100644
--- a/cpus.c
+++ b/cpus.c
@@ -851,6 +851,24 @@ void async_run_on_cpu(CPUState *cpu, void (*func)(void *data), void *data)
     qemu_cpu_kick(cpu);
 }
 
+static void qemu_kvm_destroy_vcpu(CPUState *cpu)
+{
+    CPU_REMOVE(cpu);
+
+    if (kvm_destroy_vcpu(cpu) < 0) {
+        fprintf(stderr, "kvm_destroy_vcpu failed.\n");
+        exit(1);
+    }
+
+    object_unparent(OBJECT(cpu));
+}
+
+static void qemu_tcg_destroy_vcpu(CPUState *cpu)
+{
+    CPU_REMOVE(cpu);
+    object_unparent(OBJECT(cpu));
+}
+
 static void flush_queued_work(CPUState *cpu)
 {
     struct qemu_work_item *wi;
@@ -942,6 +960,11 @@ static void *qemu_kvm_cpu_thread_fn(void *arg)
             }
         }
         qemu_kvm_wait_io_event(cpu);
+        if (cpu->exit && !cpu_can_run(cpu)) {
+            qemu_kvm_destroy_vcpu(cpu);
+            qemu_mutex_unlock(&qemu_global_mutex);
+            return NULL;
+        }
     }
 
     return NULL;
@@ -994,6 +1017,7 @@ static void tcg_exec_all(void);
 static void *qemu_tcg_cpu_thread_fn(void *arg)
 {
     CPUState *cpu = arg;
+    CPUState *remove_cpu = NULL;
 
     qemu_tcg_init_cpu_signals();
     qemu_thread_get_self(cpu->thread);
@@ -1026,6 +1050,16 @@ static void *qemu_tcg_cpu_thread_fn(void *arg)
             }
         }
         qemu_tcg_wait_io_event();
+        CPU_FOREACH(cpu) {
+            if (cpu->exit && !cpu_can_run(cpu)) {
+                remove_cpu = cpu;
+                break;
+            }
+        }
+        if (remove_cpu) {
+            qemu_tcg_destroy_vcpu(remove_cpu);
+            remove_cpu = NULL;
+        }
     }
 
     return NULL;
@@ -1383,6 +1417,9 @@ static void tcg_exec_all(void)
                 break;
             }
         } else if (cpu->stop || cpu->stopped) {
+            if (cpu->exit) {
+                next_cpu = CPU_NEXT(cpu);
+            }
             break;
         }
     }
diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h
index 174ea36..88e2403 100644
--- a/include/sysemu/kvm.h
+++ b/include/sysemu/kvm.h
@@ -178,6 +178,7 @@ int kvm_has_intx_set_mask(void);
 
 int kvm_init_vcpu(CPUState *cpu);
 int kvm_cpu_exec(CPUState *cpu);
+int kvm_destroy_vcpu(CPUState *cpu);
 
 #ifdef NEED_CPU_H
 
diff --git a/kvm-all.c b/kvm-all.c
index 1402f4f..d0caeff 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -74,6 +74,12 @@ typedef struct KVMSlot
 
 typedef struct kvm_dirty_log KVMDirtyLog;
 
+struct KVMParkedVcpu {
+    unsigned long vcpu_id;
+    int kvm_fd;
+    QLIST_ENTRY(KVMParkedVcpu) node;
+};
+
 struct KVMState
 {
     KVMSlot *slots;
@@ -108,6 +114,7 @@ struct KVMState
     QTAILQ_HEAD(msi_hashtab, KVMMSIRoute) msi_hashtab[KVM_MSI_HASHTAB_SIZE];
     bool direct_msi;
 #endif
+    QLIST_HEAD(, KVMParkedVcpu) kvm_parked_vcpus;
 };
 
 KVMState *kvm_state;
@@ -226,6 +233,53 @@ static int kvm_set_user_memory_region(KVMState *s, KVMSlot *slot)
     return kvm_vm_ioctl(s, KVM_SET_USER_MEMORY_REGION, &mem);
 }
 
+int kvm_destroy_vcpu(CPUState *cpu)
+{
+    KVMState *s = kvm_state;
+    long mmap_size;
+    struct KVMParkedVcpu *vcpu = NULL;
+    int ret = 0;
+
+    DPRINTF("kvm_destroy_vcpu\n");
+
+    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;
+    }
+
+    ret = munmap(cpu->kvm_run, mmap_size);
+    if (ret < 0) {
+        goto err;
+    }
+
+    vcpu = g_malloc0(sizeof(*vcpu));
+    vcpu->vcpu_id = kvm_arch_vcpu_id(cpu);
+    vcpu->kvm_fd = cpu->kvm_fd;
+    QLIST_INSERT_HEAD(&kvm_state->kvm_parked_vcpus, vcpu, node);
+err:
+    return ret;
+}
+
+static int kvm_get_vcpu(KVMState *s, unsigned long vcpu_id)
+{
+    struct KVMParkedVcpu *cpu;
+
+    QLIST_FOREACH(cpu, &s->kvm_parked_vcpus, node) {
+        if (cpu->vcpu_id == vcpu_id) {
+            int kvm_fd;
+
+            QLIST_REMOVE(cpu, node);
+            kvm_fd = cpu->kvm_fd;
+            g_free(cpu);
+            return kvm_fd;
+        }
+    }
+
+    return kvm_vm_ioctl(s, KVM_CREATE_VCPU, (void *)vcpu_id);
+}
+
 int kvm_init_vcpu(CPUState *cpu)
 {
     KVMState *s = kvm_state;
@@ -234,7 +288,7 @@ int kvm_init_vcpu(CPUState *cpu)
 
     DPRINTF("kvm_init_vcpu\n");
 
-    ret = kvm_vm_ioctl(s, KVM_CREATE_VCPU, (void *)kvm_arch_vcpu_id(cpu));
+    ret = kvm_get_vcpu(s, kvm_arch_vcpu_id(cpu));
     if (ret < 0) {
         DPRINTF("kvm_create_vcpu failed\n");
         goto err;
@@ -1404,6 +1458,7 @@ int kvm_init(MachineClass *mc)
 #ifdef KVM_CAP_SET_GUEST_DEBUG
     QTAILQ_INIT(&s->kvm_sw_breakpoints);
 #endif
+    QLIST_INIT(&s->kvm_parked_vcpus);
     s->vmfd = -1;
     s->fd = qemu_open("/dev/kvm", O_RDWR);
     if (s->fd == -1) {
-- 
1.7.7