libvirt/libvirt-qemuDomainSetNumaParamsLive-set-nodeset-for-root-cgr.patch
Jin Yan 676ca2f0f6 bugfix: cherry-pick some bugfix patches from open source community
bugfix patches list:
d677de  remote: fix driver name check for libxl driver
075641  systemd: start libvirtd after firewalld iptables services
06fc99  qemuDomainCleanupRun: Actually run cleanup callbacks in reverse order
01626c  virDevMapperGetTargetsImpl: quit early if device is not a devmapper target
006782  qemu: only stop external devices after the domain
42a415  qemuDomainStorageSourcePrivateDispose: Free httpcookie
cc8c29  Don't require secdrivers to implement .domainMoveImageMetadata
0230e3  qemuProcessStop: Use XATTRs to restore seclabels on disks a domain is mirroring into
8fd274  qemuProcessStop: Reattach NVMe disks a domain is mirroring into
55029d  security: don't fail if built without attr support
a5a297  qemu: Skip pre creation of NVMe disks
2a372  Fix some wrong usage of ATTRIBUTE_NONNULL
e728ff  conf: Increase cpuset length limit for CPU pinning
8f58a4  virQEMUDriverConfigNew: Add slash to cfg defaultTLSx509certdir for non-embeded driver
1b22dd  qemuDomainSetNumaParamsLive: set nodeset for root cgroup
bdb8f2  qemu: do not add model when actual iface type is hostdev

Signed-off-by: Jin Yan <jinyan12@huawei.com>
2020-09-29 16:39:38 +08:00

50 lines
1.7 KiB
Diff

From 222d16af976dc08ee3ea2008d1dc68b00a1ccc9f Mon Sep 17 00:00:00 2001
From: Daniel Henrique Barboza <danielhb413@gmail.com>
Date: Thu, 11 Jun 2020 15:54:57 -0300
Subject: [PATCH] qemuDomainSetNumaParamsLive: set nodeset for root cgroup
This function handles the change of NUMA nodeset for a given
guest, setting CpusetMems for the emulator, vcpus and IOThread
sub-groups. It doesn't set the same nodeset to the root cgroup
though. This means that cpuset.mems of the root cgroup ends up
holding the new nodeset and the old nodeset as well. For
a guest with placement=strict, nodeset='0', doing
virsh numatune <vm> 0 8 --live
Will make cpuset.mems of emulator, vcpus and iothread to be
"8", but cpuset.mems of the root cgroup will be "0,8".
This means that any new tasks that ends up landing in the
root cgroup, aside from the emulator/vcpus/iothread sub-groups,
will be split between the old nodeset and the new nodeset,
which is not what we want.
cherry-pick from commit: 1b22dd6dd44202094e0f78f887cbe790c00e9ebc
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Jin Yan <jinyan12@huawei.com>
---
src/qemu/qemu_driver.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 8bc5368..cb049fe 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -9728,6 +9728,10 @@ qemuDomainSetNumaParamsLive(virDomainObjPtr vm,
virCgroupFree(&cgroup_temp);
}
+ /* set nodeset for root cgroup */
+ if (virCgroupSetCpusetMems(priv->cgroup, nodeset_str) < 0)
+ goto cleanup;
+
ret = 0;
cleanup:
virCgroupFree(&cgroup_temp);
--
1.8.3.1