Re: [ovmf test] 87014: regressions - FAIL

[Wei Liu <wei.liu2@citrix.com>]

On Wed, Mar 23, 2016 at 06:21:58PM -0600, Jim Fehlig wrote:
> On 03/23/2016 09:27 AM, osstest service owner wrote:
> > flight 87014 ovmf real [real]
> > http://logs.test-lab.xenproject.org/osstest/logs/87014/
> >
> > Regressions :-(
> >
> > Tests which did not succeed and are blocking,
> > including tests which could not be run:
> >  test-amd64-amd64-xl-qemuu-ovmf-amd64 9 debian-hvm-install fail REGR. vs. 65543
> >  test-amd64-i386-xl-qemuu-ovmf-amd64  9 debian-hvm-install fail REGR. vs. 65543
> 
> I've been testing Anthony's series to 'Load BIOS via toolstack instead of been
> embedded in hvmloader' [0] in an attempt to use the distro ovmf instead of
> "ovmf-xen". openSUSE Factory's ovmf package [1] is updated regularly with an
> upstream git snapshot and I noticed VMs did not boot when using it. I bisected
> and found ovmf commit 7b0a1ead the culprit
> 
> commit 7b0a1ead7d2efa7f9eae4c2b254ff154d9c5f74f
> Author: Ruiyu Ni <ruiyu.ni@intel.com>
> Date:   Wed Feb 17 18:06:36 2016 +0800
> 
>     MdeModuelPkg/PciBus: Return AddrTranslationOffset in GetBarAttributes
>    
>     Some platform doesn't use CPU(HOST)/Device 1:1 mapping for PCI Bus.
>     But PCI IO doesn't have interface to tell caller (device driver)
>     whether the address returned by GetBarAttributes() is HOST address
>     or device address.
>     UEFI Spec 2.6 addresses this issue by clarifying the address returned
>     is HOST address and caller can use AddrTranslationOffset to calculate
>     the device address.
>  
> I'm sure it's biting this test too.

Yes, and OSSTest already fingered that commit. According to Anthony OVMF
is even more broken now because more bugs crept in.

Anthony is working on fixing OVMF.

> 
> BTW, any pointers on getting debug info from
> hvmloader+ovmf+qemu+other-components-involved? With the broken ovmf, I got
> nothing on the VM serial port, an empty, black vfb, nothing beyond "Invoking
> OVMF..." in xl dmesg, and no errors in the qemu log file. Beyond bisecting, I
> wasn't sure how to debug :-). I tried adding
> 
> device_model_args=[ "-debugcon file:debug.log", "-global isa-debugcon.iobase=0x402"]
> 

Can you try

  device_model_args = ["-debugcon", "file:debug.log", "-global", "isa-debugcon.iobase=0x402"]

?

Wei.

> to the VM config, but qemu failed to start with "-debugcon file:debug.log:
> invalid option"
> 
> Regards,
> Jim
> 
> [0] http://lists.xenproject.org/archives/html/xen-devel/2016-03/msg01626.html
> [1] https://build.opensuse.org/package/show/Virtualization/ovmf
> 

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