Re: [Xen-devel] [xen-unstable-smoke test] 113438: regressions - FAIL

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Thu, Sep 14, 2017 at 04:29:14AM -0600, Jan Beulich wrote:
> >>> On 14.09.17 at 12:19, <wei.liu2@xxxxxxxxxx> wrote:
> > On Thu, Sep 14, 2017 at 11:17:23AM +0100, Wei Liu wrote:
> >> On Thu, Sep 14, 2017 at 04:14:29AM -0600, Jan Beulich wrote:
> >> > >>> On 14.09.17 at 10:49, <osstest-admin@xxxxxxxxxxxxxx> wrote:
> >> > > flight 113438 xen-unstable-smoke real [real]
> >> > > http://logs.test-lab.xenproject.org/osstest/logs/113438/ 
> >> > > 
> >> > > Regressions :-(
> >> > > 
> >> > > Tests which did not succeed and are blocking,
> >> > > including tests which could not be run:
> >> > >  test-amd64-amd64-xl-qemuu-debianhvm-i386 15 guest-saverestore.2 fail 
> >> > > REGR. vs. 
> > 113384
> >> > 
> >> > The first instance of this regression occurred with just the
> >> > "libxl: add PV display device driver interface" series applied, so it's
> >> > pretty certain there's something wrong in there. Are you in the
> >> > process of looking into this?
> >> > 
> >> 
> >> I looked into this this morning.
> >> 
> >> libxl: debug: libxl_qmp.c:296:qmp_handle_response: Domain 4:message type: 
> >> return
> >> xc: error: Dom 4 not suspended: (shutdown 0, reason 255): Internal error
> >> libxl: error: libxl_dom_suspend.c:474:libxl__domain_resume: Domain 
> >> 4:xc_domain_resume failed: Invalid argument
> >> libxl: debug: libxl_event.c:1869:libxl__ao_complete: ao 0x9e6700: 
> >> complete, rc=-3
> >> libxl: debug: libxl_domain.c:183:libxl_domain_resume: Domain 4:ao 
> >> 0x9e6700: inprogress: poller=0x9e4f80, flags=ic
> >> libxl: debug: libxl_event.c:1838:libxl__ao__destroy: ao 0x9e6700: destroy
> >> 
> >> It is a libxc call that failed. It didn't look to be related to the device
> >> framework changes.
> >> 
> >> I will have a closer look in the afternoon.
> > 
> > Also the failure happened in the second attempt of save/restore while
> > the first attempt succeeded. IIRC we've seen error like this before so I
> > was a bit hesitant to draw the conclusion whether it was related to
> > libxl changes at the moment.
> 
> Indeed, that's why I didn't reply to the first failure, but to the
> 8th in a row.

Hmm... 8th in a row. There is definitely something wrong. I will have a
closer look in the afternoon. Worst case we can revert that series
tomorrow to see how osstest fairs over the weekend.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel