[libvirt] [PATCHv7 18/18] qemu: Setting up vcpu and adding pids to resctrl monitor groups during reconnection

[Wang, Huaqiang]

> -----Original Message-----
> From: John Ferlan [mailto:jferlan at redhat.com]
> Sent: Tuesday, November 6, 2018 4:09 AM
> To: Wang, Huaqiang <huaqiang.wang at intel.com>; libvir-list at redhat.com
> Cc: Feng, Shaohe <shaohe.feng at intel.com>; Niu, Bing <bing.niu at intel.com>;
> Ding, Jian-feng <jian-feng.ding at intel.com>; Zang, Rui <rui.zang at intel.com>
> Subject: Re: [PATCHv7 18/18] qemu: Setting up vcpu and adding pids to resctrl
> monitor groups during reconnection
> 
> 
> 
> On 10/22/18 4:01 AM, Wang Huaqiang wrote:
> > Invoking qemuProcessSetupVcpus in process of VM reconnection.
> >
> > The vcpu pid information need to be refilled to resctrl monitor after
> > a VM reconnection./
> >
> > Signed-off-by: Wang Huaqiang <huaqiang.wang at intel.com>
> > ---
> >  src/qemu/qemu_process.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index
> > fba4fb4..ed0330b 100644
> > --- a/src/qemu/qemu_process.c
> > +++ b/src/qemu/qemu_process.c
> > @@ -8008,6 +8008,9 @@ qemuProcessReconnect(void *opaque)
> >          }
> >      }
> >
> > +    if (qemuProcessSetupVcpus(obj) < 0)
> > +        goto error;
> > +
> 
> IIRC the reason for not needing this is that the assumption is that the initial
> startup (ProcessLaunch) and any hotplug thereafter would end up calling
> qemuProcessSetupVcpu.
> 
> Not sure this is necessary for everything except perhaps the resctrl monitor
> *pids list which I'm not even sure is necessary.
> 

This time I think this patch is not necessary. This patch is only 
required if *tasks file content is changed in process of re-connection.
Now I don't think the *tasks file will be changed because the vcpu 
processes are still there
in process of reconnection and no add no removal.

Patch will be removed.

> John
> 

Thanks for review.
Huaqiang

> >      /* update domain state XML with possibly updated state in virDomainObj */
> >      if (virDomainSaveStatus(driver->xmlopt, cfg->stateDir, obj, driver->caps) < 0)
> >          goto error;
> >