[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH 2/5] Support CPU placement in LXC driver



On Thu, Nov 24, 2011 at 09:55:26PM +0800, Daniel Veillard wrote:
> On Thu, Nov 24, 2011 at 11:38:13AM +0000, Daniel P. Berrange wrote:
> > From: "Daniel P. Berrange" <berrange redhat com>
> > 
> > While LXC does not have the concept of VCPUS, so we cann't do
> 
>  s/nn/n/
> 
> > per-VCPU pCPU placement, we can support the VM level CPU
> > placement. Todo this simply set the CPU affinity of the LXC
> > controller at startup. All child processes will inherit this
> > affinity.
> > 
> > * src/lxc/lxc_controller.c: Set process affinity
> > ---
> >  src/lxc/lxc_controller.c |   63 ++++++++++++++++++++++++++++++++++++++++++++++
> >  1 files changed, 63 insertions(+), 0 deletions(-)
> > 
> > diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c
> > index 4718fa8..1db25fb 100644
> > --- a/src/lxc/lxc_controller.c
> > +++ b/src/lxc/lxc_controller.c
> > @@ -66,6 +66,8 @@
> >  #include "virfile.h"
> >  #include "virpidfile.h"
> >  #include "command.h"
> > +#include "processinfo.h"
> > +#include "nodeinfo.h"
> >  
> >  #define VIR_FROM_THIS VIR_FROM_LXC
> >  
> > @@ -322,6 +324,64 @@ static int lxcSetContainerNUMAPolicy(virDomainDefPtr def)
> >  }
> >  #endif
> >  
> > +
> > +/*
> > + * To be run while still single threaded
> > + */
> > +static int lxcSetContainerCpuAffinity(virDomainDefPtr def)
> > +{
> > +    int i, hostcpus, maxcpu = CPU_SETSIZE;
> > +    virNodeInfo nodeinfo;
> > +    unsigned char *cpumap;
> > +    int cpumaplen;
> > +
> > +    VIR_DEBUG("Setting CPU affinity");
> > +
> > +    if (nodeGetInfo(NULL, &nodeinfo) < 0)
> > +        return -1;
> > +
> > +    /* setaffinity fails if you set bits for CPUs which
> > +     * aren't present, so we have to limit ourselves */
> > +    hostcpus = VIR_NODEINFO_MAXCPUS(nodeinfo);
> > +    if (maxcpu > hostcpus)
> > +        maxcpu = hostcpus;
> > +
> > +    cpumaplen = VIR_CPU_MAPLEN(maxcpu);
> > +    if (VIR_ALLOC_N(cpumap, cpumaplen) < 0) {
> > +        virReportOOMError();
> > +        return -1;
> > +    }
> > +
> > +    if (def->cpumask) {
> > +        /* XXX why don't we keep 'cpumask' in the libvirt cpumap
> > +         * format to start with ?!?! */
> > +        for (i = 0 ; i < maxcpu && i < def->cpumasklen ; i++)
> > +            if (def->cpumask[i])
> > +                VIR_USE_CPU(cpumap, i);
> > +    } else {
> > +        /* You may think this is redundant, but we can't assume libvirtd
> > +         * itself is running on all pCPUs, so we need to explicitly set
> > +         * the spawned QEMU instance to all pCPUs if no map is given in
> 
>   Hum you mean LXC container instead of QEmu, right ?
> 
> > +         * its config file */
> > +        for (i = 0 ; i < maxcpu ; i++)
> > +            VIR_USE_CPU(cpumap, i);
> > +    }
> > +
> > +    /* We are pressuming we are running between fork/exec of QEMU
> 
>   idem

Yes, these should obviously have been  s/QEMU/LXC/  :-)


Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]