On 10/12/2012 03:52 PM, Osier Yang wrote:
Setting pinning policy for vcpu which exceeds current vcpus number
just makes no sense, however, it could cause various problems, E.g.
<vcpu current='1'>4</vcpu>
<cputune>
<vcpupin vcpuid='3' cpuset='4'/>
</cputune>
% virsh start linux
error: Failed to start domain linux
error: cannot set CPU affinity on process 32534: No such process
Well pointed out.
We must have some odd codes underlying which produces the
"on process 32534", but the point is why we not to prevent
earlier when parsing? Note that this is only one of the
problem it could cause.
We should definitely fix it there as well. However I *think* that it is
caused just by us trying to move the processor more times, which would
be fixed by this. But that's just an idea.
This patch is to ignore the <vcpupin> for not onlined vcpus.
---
src/conf/domain_conf.c | 10 +++++++++-
1 files changed, 9 insertions(+), 1 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 3c3d0ae..5141be6 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -8806,7 +8806,15 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps,
goto error;
}
- def->cputune.vcpupin[def->cputune.nvcpupin++] = vcpupin;
+ if (vcpupin->vcpuid >= def->vcpus)
+ /* To avoid the regression when daemon loading
+ * domain confs, we can't simply error out if
+ * <vcpupin> nodes greater than current vcpus,
+ * ignoring them instead.
+ */
+ VIR_WARN("Ignore vcpupin for not onlined vcpus");
+ else
+ def->cputune.vcpupin[def->cputune.nvcpupin++] = vcpupin;
}
VIR_FREE(nodes);
ACK,
Martin