Re: [ovmf test] 58956: regressions - FAIL

[Ian Campbell <ian.campbell@citrix.com>]

On Tue, 2015-06-30 at 17:53 +0100, Anthony PERARD wrote:
> On Mon, Jun 29, 2015 at 10:32:43AM +0100, Ian Campbell wrote:
> > On Sun, 2015-06-28 at 22:32 +0000, osstest service user wrote:
> > > flight 58956 ovmf real [real]
> > > http://logs.test-lab.xenproject.org/osstest/logs/58956/
> > > 
> > > Regressions :-(
> > > 
> > > Tests which did not succeed and are blocking,
> > > including tests which could not be run:
> > >  test-amd64-i386-xl-qemuu-win7-amd64  9 windows-install    fail REGR. vs. 58919
> > 
> > http://logs.test-lab.xenproject.org/osstest/results/bisect/ovmf/test-amd64-i386-xl-qemuu-win7-amd64.windows-install.html smells like something in the range:
> > $ git log --oneline 495ee9b85..79d274b8b
> > 79d274b OvmfPkg: PlatformPei: invert MTRR setup in QemuInitializeRam()
> > cfc80e2 OvmfPkg: PlatformPei: beautify memory HOB order in QemuInitializeRam()
> > 86a14b0 OvmfPkg: PlatformPei: create the CPU HOB with dynamic memory space width
> > bc89fe4 OvmfPkg: PlatformPei: enable larger permanent PEI RAM
> > 
> > has made windows installs less reliable. From
> > http://logs.test-lab.xenproject.org/osstest/results/history/test-amd64-i386-xl-qemuu-win7-amd64/ovmf.html they mostly seem to be okay, or else we've gotten unluckY 3 in a row.
> 
> Wait we are testing win7 install with OVMF in osstest? Did you made a new
> iso that is capable of EFI booting? Also I don't think I've ever found a
> win7 iso within Citrix that was capables of booting with OVMF. But I
> haven't look very hard either.
> 
> Looking at the guest config, I don't see any mention of using OVMF instead
> of SeaBIOS.
> http://logs.test-lab.xenproject.org/osstest/logs/58956/test-amd64-i386-xl-qemuu-win7-amd64/fiano1---etc-xen-win.guest.osstest.cfg
> 
> So, I guest OVMF have nothing to do with the failure.

Yes, it looks like you are right from the runvars of 58956:

test-amd64-amd64-xl-qemuu-debianhvm-amd64     bios                   seabios                                                       
test-amd64-amd64-xl-qemuu-debianhvm-amd64-xsm bios                   seabios                                                       
test-amd64-amd64-xl-qemuu-ovmf-amd64          bios                   ovmf                                                          
test-amd64-i386-xl-qemuu-debianhvm-amd64      bios                   seabios                                                       
test-amd64-i386-xl-qemuu-debianhvm-amd64-xsm  bios                   seabios                                                       
test-amd64-i386-xl-qemuu-ovmf-amd64           bios                   ovmf                      

+ there are tests with no bios specified at all (=> gets the default = seabios):

test-amd64-amd64-xl-qemuu-win7-amd64
test-amd64-amd64-xl-qemuu-winxpsp3
test-amd64-i386-qemuu-rhel6hvm-amd
test-amd64-i386-qemuu-rhel6hvm-intel
test-amd64-i386-xl-qemuu-win7-amd64
test-amd64-i386-xl-qemuu-winxpsp3
test-amd64-i386-xl-qemuu-winxpsp3-vcpus1

Wei, is that expected? (sorry if I've misremembered who added these
tests...)

Is the fix to stop wasting time with some of those tests on the ovmf
branch or to switch them to ovmf variants (or some combination of the
two)?

Ian.