On Fri, 2 Feb 2018 12:49:25 +0000 Daniel P. Berrangé <berrange@xxxxxxxxxx> wrote: > On Thu, Feb 01, 2018 at 12:54:41PM -0500, Luiz Capitulino wrote: > > > > Libvirt needs to know when a vCPU is halted. To get this information, > > libvirt has started using the query-cpus command from QEMU. However, > > if in kernel irqchip is in use, query-cpus will force all vCPUs > > to user-space since they have to issue the KVM_GET_MP_STATE ioctl. > > This has catastrophic implications to low-latency workloads like > > KVM-RT and zero packet loss with DPDK. To make matters worse, there's > > an OpenStack service called ceilometer that causes libvirt to > > issue query-cpus every few minutes. > > > > The solution proposed in this patch is to export the vCPU > > halted state in the already existing vcpu directory in sysfs. > > This way, libvirt can read the vCPU halted state from sysfs and avoid > > using the query-cpus command. This solution seems to be sufficient > > for libvirt needs, but it has the following cons: > > > > * vcpu information in sysfs lives in a debug directory, so > > libvirt would be basing its API on debug info > > Is this part of regular sysfs mount point, or does it require a > debug fs to be mounted separately at /sys/fs/debug ? Yeah, it depends on debugfs being separately mounted at /sys/fs/debug. > > * Currently, only x86 supports the vcpu dir in sysfs, so > > we'd have to expand this to other archs (should be doable) > > Yep, that would be fairly important That's doable. The code that exports the vcpu information to sysfs is pretty arch-independent. Only x86 does it today because it's the only arch that actually exports anything per vcpu (which is the tsc-offset).
