qemu: Keep the affinity when creating cgroup for emulator thread

When the cpu placement model is "auto", it sets the affinity for
domain process with the advisory nodeset from numad, however,
creating cgroup for the domain process (called emulator thread
in some contexts) later overrides that with pinning it to all
available pCPUs.

How to reproduce:

  * Configure the domain with "auto" placement for <vcpu>, e.g.
    <vcpu placement='auto'>4</vcpu>
  * % virsh start dom
  * % cat /proc/$dompid/status

Though the emulator cgroup cause conflicts, but we can't simply
prohibit creating it, as other tunables are still useful, such
as "emulator_period", which is used by API
virDomainSetSchedulerParameter. So this patch doesn't prohibit
creating the emulator cgroup, but inherit the nodeset from numad,
and reset the affinity for domain process.

* src/qemu/qemu_cgroup.h: Modify definition of qemuSetupCgroupForEmulator
                          to accept the passed nodenet
* src/qemu/qemu_cgroup.c: Set the affinity with the passed nodeset

diff --git a/src/qemu/qemu_cgroup.c b/src/qemu/qemu_cgroup.c
index db371a07731d749c4a9e47cbb5e5dc53412b0275..5ce748a2302dfa1152759b7b9acecf6452773f53 100644 (file)
--- a/src/qemu/qemu_cgroup.c
+++ b/src/qemu/qemu_cgroup.c
@@ -25,6 +25,7 @@
 
 #include "qemu_cgroup.h"
 #include "qemu_domain.h"
+#include "qemu_process.h"
 #include "cgroup.h"
 #include "logging.h"
 #include "memory.h"
@@ -637,9 +638,11 @@ cleanup:
 }
 
 int qemuSetupCgroupForEmulator(struct qemud_driver *driver,
-                               virDomainObjPtr vm)
+                               virDomainObjPtr vm,
+                               virBitmapPtr nodemask)
 {
     virBitmapPtr cpumask = NULL;
+    virBitmapPtr cpumap = NULL;
     virCgroupPtr cgroup = NULL;
     virCgroupPtr cgroup_emulator = NULL;
     virDomainDefPtr def = vm->def;
@@ -687,10 +690,15 @@ int qemuSetupCgroupForEmulator(struct qemud_driver *driver,
         }
     }
 
-    if (def->cputune.emulatorpin)
+    if (def->placement_mode == VIR_DOMAIN_CPU_PLACEMENT_MODE_AUTO) {
+        if (!(cpumap = qemuPrepareCpumap(driver, nodemask)))
+            goto cleanup;
+        cpumask = cpumap;
+    } else if (def->cputune.emulatorpin) {
         cpumask = def->cputune.emulatorpin->cpumask;
-    else if (def->cpumask)
+    } else if (def->cpumask) {
         cpumask = def->cpumask;
+    }
 
     if (cpumask) {
         if (qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_CPUSET)) {
@@ -711,9 +719,12 @@ int qemuSetupCgroupForEmulator(struct qemud_driver *driver,
 
     virCgroupFree(&cgroup_emulator);
     virCgroupFree(&cgroup);
+    virBitmapFree(cpumap);
     return 0;
 
 cleanup:
+    virBitmapFree(cpumap);
+
     if (cgroup_emulator) {
         virCgroupRemove(cgroup_emulator);
         virCgroupFree(&cgroup_emulator);

diff --git a/src/qemu/qemu_cgroup.h b/src/qemu/qemu_cgroup.h
index c55216238d01a98da6c544d1b44de7272978c8b3..50ee09248b0831e3c657da094482503227da6b90 100644 (file)
--- a/src/qemu/qemu_cgroup.h
+++ b/src/qemu/qemu_cgroup.h
@@ -58,7 +58,8 @@ int qemuSetupCgroupVcpuPin(virCgroupPtr cgroup,
 int qemuSetupCgroupEmulatorPin(virCgroupPtr cgroup, virBitmapPtr cpumask);
 int qemuSetupCgroupForVcpu(struct qemud_driver *driver, virDomainObjPtr vm);
 int qemuSetupCgroupForEmulator(struct qemud_driver *driver,
-                               virDomainObjPtr vm);
+                               virDomainObjPtr vm,
+                               virBitmapPtr nodemask);
 int qemuRemoveCgroup(struct qemud_driver *driver,
                      virDomainObjPtr vm,
                      int quiet);

diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index 26be35ae83c34fa8cd2370adde88c13edafbfb53..5004e9ba30b557a3a1ae745f444808d9f3418a8a 100644 (file)
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -3812,7 +3812,7 @@ int qemuProcessStart(virConnectPtr conn,
         goto cleanup;
 
     VIR_DEBUG("Setting cgroup for emulator (if required)");
-    if (qemuSetupCgroupForEmulator(driver, vm) < 0)
+    if (qemuSetupCgroupForEmulator(driver, vm, nodemask) < 0)
         goto cleanup;
 
     VIR_DEBUG("Setting VCPU affinities");