Re: [BUG REPORT] My Xen 4.3-unstable Changeset 26666 Installation is Broken!

[Gordan Bobic <gordan@bobich.net>]

On 07/27/2013 10:52 AM, Singapore Citizen Mr. Teo En Ming (Zhang Enming) 
wrote:
> On 27/07/2013 17:49, Singapore Citizen Mr. Teo En Ming (Zhang Enming)
> wrote:
>> On 26/07/2013 16:52, Ren, Yongjie wrote:
>>>> -----Original Message-----
>>>> From: Singapore Citizen Mr. Teo En Ming (Zhang Enming)
>>>> [mailto:singapore.mr.teo.en.ming@gmail.com]
>>>> Sent: Friday, July 26, 2013 4:39 PM
>>>> To: xen-devel@lists.xen.org; Teo En Ming (Zhang Enming); Ian Campbell;
>>>> Konrad Rzeszutek Wilk; Ren, Yongjie
>>>> Subject: [BUG REPORT] My Xen 4.3-unstable Changeset 26666 Installation
>>>> is Broken!
>>>>
>>>> Hi,
>>>>
>>>> My Xen 4.3-unstable changeset 26666 installation is broken. The
>>>> following is my bug report.
>>>>
>>>> teo-en-ming@intel-core-i5-4430-desktop:~$ sudo xl create
>>>> /etc/xen/windows7
>>>> Parsing config from /etc/xen/windows7
>>>> failed to free memory for the domain
>>>>
>>>> teo-en-ming@intel-core-i5-4430-desktop:~$ sudo xl list
>>>> libxl: error: libxl.c:557:libxl_list_domain: geting domain info list:
>>>> Permission denied
>>>> libxl_domain_infolist failed.
>>>>
>>>> teo-en-ming@intel-core-i5-4430-desktop:~$ sudo xl dmesg
>>>> libxl: error: libxl.c:4759:libxl_xen_console_read_line: reading console
>>>> ring buffer: Permission denied
>>>>
>>>> teo-en-ming@intel-core-i5-4430-desktop:~$ sudo xl info
>>>> host                   : intel-core-i5-4430-desktop
>>>> release                : 3.10.0-xen-teo.en.ming-sgp
>>>> version                : #1 SMP Sat Jul 13 00:07:55 SGT 2013
>>>> machine                : x86_64
>>>> libxl: error: libxl.c:3938:libxl_get_physinfo: getting physinfo:
>>>> Permission denied
>>>> libxl_physinfo failed.
>>>> libxl: error: libxl.c:4320:libxl_get_scheduler: getting domain info
>>>> list: Permission denied
>>>> xen_major              : 4
>>>> xen_minor              : 3
>>>> xen_extra              : -unstable
>>>> xen_version            : 4.3-unstable
>>>> xen_caps               : xen-3.0-x86_64 xen-3.0-x86_32p
>>>> hvm-3.0-x86_32
>>>> hvm-3.0-x86_32p hvm-3.0-x86_64
>>>> xen_scheduler          : (null)
>>>> xen_pagesize           : 4096
>>>> platform_params        : virt_start=0xffff800000000000
>>>> xen_changeset          : Thu Feb 28 09:22:41 2013 +0100
>>>> 26666:94ece33caae2
>>>> xen_commandline        :
>>>> cc_compiler            : gcc (Ubuntu/Linaro 4.7.3-1ubuntu1) 4.7.3
>>>> cc_compile_by          : root
>>>> cc_compile_domain      :
>>>> cc_compile_date        : Thu Jul 25 23:20:43 SGT 2013
>>>> xend_config_format     : 4
>>>>
>>>> Please let me know how to solve this bug.
>>>>
>>> I don't know how to resolve it.
>>> But if you want to do IGD passthrough, you can try with Xen 4.3.0
>>> release with latest traditional qemu.
>>> It can also successful for me with Xen 4.3.0 release with latest
>>> traditional qemu.
>>>
>>> Best Regards,
>>>       Yongjie (Jay)
>>>
>>>
>>>> Thank you very much.
>>>>
>>>> --
>>>> Yours sincerely,
>>>>
>>>> Singapore Citizen Mr. Teo En Ming (Zhang Enming)
>>>
>> Hi Yongjie,
>>
>> I tried to use Xen 4.3.0 official release and Linux kernel 3.8.1 but
>> IGD passthrough is still unsuccessful. Yellow triangle with
>> exclamation mark and error code 43 still show up in Device Manager for
>> Intel HD Graphics 4600.
>>
>> I have attached my configuration files. What else can I do?
>>
>> I have also filed a bug for Xen 4.3-unstable changeset 26666 at
>> http://bugzilla-archived.xenproject.org/bugzilla/show_bug.cgi?id=1869
>>
> Hi All,
>
> Please check through my configuration files to see if there is any error
> or mistake.
>
> Thank you very much.

Have you tried a vBAR=pBAR patch?

See here for more details:

http://www.davidgis.fr/blog/index.php?2011/12/07/860-xen-42unstable-patches-for-vga-pass-through

This is mainly aimed at Nvidia GeForce cards (GeForce drivers don't 
handle the case where vBAR != pBAR, but Quadro drivers do - I modified 
my GeForce cards to equivalent Quadros to work around the issue). But - 
it may also help with your situation. It is entirely plausible that your 
GPU is having problems finding some of it's BARs, which might also 
explain why it is stuck in a low resolution / 4-bit colour mode.

I am not sure how much you might need to adapt this and which patches 
you can skip or will have to modify for IGD instead of GeForce, but it's 
probably worth a shot.

Gordan