Enable cpuset cgroup and synchronous vcpupin info to cgroup.

vcpu threads pin are implemented using sched_setaffinity(), but
not controlled by cgroup. This patch does the following things:

    1) enable cpuset cgroup
    2) reflect all the vcpu threads pin info to cgroup

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>

src/libvirt_private.syms

@@ -82,6 +82,7 @@ virCgroupGetCpuShares;
 virCgroupGetCpuacctPercpuUsage;
 virCgroupGetCpuacctStat;
 virCgroupGetCpuacctUsage;
+virCgroupGetCpusetCpus;
 virCgroupGetCpusetMems;
 virCgroupGetFreezerState;
 virCgroupGetMemSwapHardLimit;
@@ -100,6 +101,7 @@ virCgroupSetBlkioWeight;
 virCgroupSetCpuCfsPeriod;
 virCgroupSetCpuCfsQuota;
 virCgroupSetCpuShares;
+virCgroupSetCpusetCpus;
 virCgroupSetCpusetMems;
 virCgroupSetFreezerState;
 virCgroupSetMemSwapHardLimit;

src/qemu/qemu_cgroup.c

@@ -491,11 +491,45 @@ cleanup:
     return -1;
 }
 
+int qemuSetupCgroupVcpuPin(virCgroupPtr cgroup,
+                           virDomainVcpuPinDefPtr *vcpupin,
+                           int nvcpupin,
+                           int vcpuid)
+{
+    int i, rc = 0;
+    char *new_cpus = NULL;
+
+    for (i = 0; i < nvcpupin; i++) {
+        if (vcpuid == vcpupin[i]->vcpuid) {
+            new_cpus = virDomainCpuSetFormat(vcpupin[i]->cpumask,
+                                             VIR_DOMAIN_CPUMASK_LEN);
+            if (!new_cpus) {
+                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                               _("failed to convert cpu mask"));
+                rc = -1;
+                goto cleanup;
+            }
+            rc = virCgroupSetCpusetCpus(cgroup, new_cpus);
+            if (rc != 0) {
+                virReportSystemError(-rc,
+                                     "%s",
+                                     _("Unable to set cpuset.cpus"));
+                goto cleanup;
+            }
+        }
+    }
+
+cleanup:
+    VIR_FREE(new_cpus);
+    return rc;
+}
+
 int qemuSetupCgroupForVcpu(struct qemud_driver *driver, virDomainObjPtr vm)
 {
     virCgroupPtr cgroup = NULL;
     virCgroupPtr cgroup_vcpu = NULL;
     qemuDomainObjPrivatePtr priv = vm->privateData;
+    virDomainDefPtr def = vm->def;
     int rc;
     unsigned int i;
     unsigned long long period = vm->def->cputune.period;
@@ -567,6 +601,15 @@ int qemuSetupCgroupForVcpu(struct qemud_driver *driver, virDomainObjPtr vm)
             }
         }
 
+        /* Set vcpupin in cgroup if vcpupin xml is provided */
+        if (def->cputune.nvcpupin &&
+            qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_CPUSET) &&
+            qemuSetupCgroupVcpuPin(cgroup_vcpu,
+                                   def->cputune.vcpupin,
+                                   def->cputune.nvcpupin,
+                                   i) < 0)
+            goto cleanup;
+
         virCgroupFree(&cgroup_vcpu);
     }
 

src/qemu/qemu_cgroup.h

@@ -53,6 +53,10 @@ int qemuSetupCgroup(struct qemud_driver *driver,
 int qemuSetupCgroupVcpuBW(virCgroupPtr cgroup,
                           unsigned long long period,
                           long long quota);
+int qemuSetupCgroupVcpuPin(virCgroupPtr cgroup,
+                           virDomainVcpuPinDefPtr *vcpupin,
+                           int nvcpupin,
+                           int vcpuid);
 int qemuSetupCgroupForVcpu(struct qemud_driver *driver, virDomainObjPtr vm);
 int qemuSetupCgroupForEmulator(struct qemud_driver *driver,
                                virDomainObjPtr vm);

src/qemu/qemu_driver.c

@@ -3716,11 +3716,15 @@ qemudDomainPinVcpuFlags(virDomainPtr dom,
     struct qemud_driver *driver = dom->conn->privateData;
     virDomainObjPtr vm;
     virDomainDefPtr persistentDef = NULL;
+    virCgroupPtr cgroup_dom = NULL;
+    virCgroupPtr cgroup_vcpu = NULL;
     int maxcpu, hostcpus;
     virNodeInfo nodeinfo;
     int ret = -1;
     qemuDomainObjPrivatePtr priv;
     bool canResetting = true;
+    int newVcpuPinNum = 0;
+    virDomainVcpuPinDefPtr *newVcpuPin = NULL;
     int pcpu;
 
     virCheckFlags(VIR_DOMAIN_AFFECT_LIVE |
@@ -3769,16 +3773,54 @@ qemudDomainPinVcpuFlags(virDomainPtr dom,
 
     if (flags & VIR_DOMAIN_AFFECT_LIVE) {
 
-        if (priv->vcpupids != NULL) {
+        if (priv->vcpupids == NULL) {
-            if (virProcessInfoSetAffinity(priv->vcpupids[vcpu],
-                                          cpumap, maplen, maxcpu) < 0)
-                goto cleanup;
-        } else {
             virReportError(VIR_ERR_OPERATION_INVALID,
                            "%s", _("cpu affinity is not supported"));
             goto cleanup;
         }
 
+        if (vm->def->cputune.vcpupin) {
+            newVcpuPin = virDomainVcpuPinDefCopy(vm->def->cputune.vcpupin,
+                                                 vm->def->cputune.nvcpupin);
+            if (!newVcpuPin)
+                goto cleanup;
+
+            newVcpuPinNum = vm->def->cputune.nvcpupin;
+        } else {
+            if (VIR_ALLOC(newVcpuPin) < 0) {
+                virReportOOMError();
+                goto cleanup;
+            }
+            newVcpuPinNum = 0;
+        }
+
+        if (virDomainVcpuPinAdd(newVcpuPin, &newVcpuPinNum, cpumap, maplen, vcpu) < 0) {
+            virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                           _("failed to update vcpupin"));
+            virDomainVcpuPinDefFree(newVcpuPin, newVcpuPinNum);
+            goto cleanup;
+        }
+
+        /* Configure the corresponding cpuset cgroup before set affinity. */
+        if (qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_CPUSET)) {
+            if (virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup_dom, 0) == 0 &&
+                virCgroupForVcpu(cgroup_dom, vcpu, &cgroup_vcpu, 0) == 0 &&
+                qemuSetupCgroupVcpuPin(cgroup_vcpu, newVcpuPin, newVcpuPinNum, vcpu) < 0) {
+                virReportError(VIR_ERR_OPERATION_INVALID,
+                               _("failed to set cpuset.cpus in cgroup"
+                                 " for vcpu %d"), vcpu);
+                goto cleanup;
+            }
+        } else {
+            if (virProcessInfoSetAffinity(priv->vcpupids[vcpu],
+                                          cpumap, maplen, maxcpu) < 0) {
+                virReportError(VIR_ERR_SYSTEM_ERROR,
+                               _("failed to set cpu affinity for vcpu %d"),
+                               vcpu);
+                goto cleanup;
+            }
+        }
+
         if (canResetting) {
             if (virDomainVcpuPinDel(vm->def, vcpu) < 0) {
                 virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
@@ -3787,25 +3829,17 @@ qemudDomainPinVcpuFlags(virDomainPtr dom,
                 goto cleanup;
             }
         } else {
-            if (!vm->def->cputune.vcpupin) {
+            if (vm->def->cputune.vcpupin)
-                if (VIR_ALLOC(vm->def->cputune.vcpupin) < 0) {
+                virDomainVcpuPinDefFree(vm->def->cputune.vcpupin, vm->def->cputune.nvcpupin);
-                    virReportOOMError();
+
-                    goto cleanup;
+            vm->def->cputune.vcpupin = newVcpuPin;
-                }
+            vm->def->cputune.nvcpupin = newVcpuPinNum;
-                vm->def->cputune.nvcpupin = 0;
+            newVcpuPin = NULL;
-            }
-            if (virDomainVcpuPinAdd(vm->def->cputune.vcpupin,
-                                    &vm->def->cputune.nvcpupin,
-                                    cpumap,
-                                    maplen,
-                                    vcpu) < 0) {
-                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
-                               _("failed to update or add vcpupin xml of "
-                                 "a running domain"));
-                goto cleanup;
-            }
         }
 
+        if (newVcpuPin)
+            virDomainVcpuPinDefFree(newVcpuPin, newVcpuPinNum);
+
         if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
             goto cleanup;
     }
@@ -3846,6 +3880,10 @@ qemudDomainPinVcpuFlags(virDomainPtr dom,
     ret = 0;
 
 cleanup:
+    if (cgroup_vcpu)
+        virCgroupFree(&cgroup_vcpu);
+    if (cgroup_dom)
+        virCgroupFree(&cgroup_dom);
     if (vm)
         virDomainObjUnlock(vm);
     return ret;

src/util/cgroup.c

@@ -543,7 +543,8 @@ static int virCgroupMakeGroup(virCgroupPtr parent, virCgroupPtr group,
         /* We need to control cpu bandwidth for each vcpu now */
         if ((flags & VIR_CGROUP_VCPU) &&
             (i != VIR_CGROUP_CONTROLLER_CPU &&
-             i != VIR_CGROUP_CONTROLLER_CPUACCT)) {
+             i != VIR_CGROUP_CONTROLLER_CPUACCT &&
+             i != VIR_CGROUP_CONTROLLER_CPUSET)) {
             /* treat it as unmounted and we can use virCgroupAddTask */
             VIR_FREE(group->controllers[i].mountPoint);
             continue;
@@ -1401,6 +1402,38 @@ int virCgroupGetCpusetMems(virCgroupPtr group, char **mems)
                                 mems);
 }
 
+/**
+ * virCgroupSetCpusetCpus:
+ *
+ * @group: The cgroup to set cpuset.cpus for
+ * @cpus: the cpus to set
+ *
+ * Retuens: 0 on success
+ */
+int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus)
+{
+    return virCgroupSetValueStr(group,
+                                VIR_CGROUP_CONTROLLER_CPUSET,
+                                "cpuset.cpus",
+                                cpus);
+}
+
+/**
+ * virCgroupGetCpusetCpus:
+ *
+ * @group: The cgroup to get cpuset.cpus for
+ * @cpus: the cpus to get
+ *
+ * Retuens: 0 on success
+ */
+int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus)
+{
+    return virCgroupGetValueStr(group,
+                                VIR_CGROUP_CONTROLLER_CPUSET,
+                                "cpuset.cpus",
+                                cpus);
+}
+
 /**
  * virCgroupDenyAllDevices:
  *

src/util/cgroup.h

@@ -151,6 +151,9 @@ int virCgroupGetFreezerState(virCgroupPtr group, char **state);
 int virCgroupSetCpusetMems(virCgroupPtr group, const char *mems);
 int virCgroupGetCpusetMems(virCgroupPtr group, char **mems);
 
+int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus);
+int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus);
+
 int virCgroupRemove(virCgroupPtr group);
 
 void virCgroupFree(virCgroupPtr *group);