[PC-BSD Testing] Recreated crash OpenGL Info screen on "working"9.0 x86 Beta 2.0, WAS: Re: instant unexpected logout..

Julian Elischer julian at freebsd.org
Fri Sep 23 08:38:19 PDT 2011

On 9/23/11 8:32 AM, Arthur Koziol wrote:
> On 09/21/2011 9:32 AM, Arthur Koziol wrote:
>> On 09/20/2011 1:39 PM, Ken Moore wrote:
>>> On 09/20/11 14:20, Arthur Koziol wrote:
>>>> On 09/20/2011 1:18 PM, Julian Elischer wrote:
>>>>> On 9/20/11 11:07 AM, Arthur Koziol wrote:
>>>>>> On 09/19/2011 2:03 PM, Julian Elischer wrote:
>>>>>>> On 9/19/11 10:39 AM, Arthur wrote:
>>>>>>>> I did report it but I think it was working again in Beta1.5
>>>>>>>> x86. It
>>>>>>>> was originally borked in Beta 1. I did not retry in Beta 2 yet on
>>>>>>>> x86.
>>>>>>> fails on beta2
>>>>>> Well, after my last message I again left the screen sitting at
>>>>>> OpenGL
>>>>>> Information and this time around when I moved the mouse, the
>>>>>> monitor
>>>>>> did wake from sleep but I was back at the login screen. So some
>>>>>> time
>>>>>> between the monitor going to sleep and just a few moments ago it
>>>>>> took
>>>>>> a dive. There's got to be a log or crash info for this event.
>>>>>> Anyone
>>>>>> able to point me to where it would be so I can up the info?
>>>>> no idea but for me it's *Instant*.. as soon as I click on it I get
>>>>> logged out.
>>>> Julian,
>>>> What VGA adapter do you have? Also, if this is integrated, do you
>>>> have
>>>> the latest BIOS for your motherboard as sometimes HW makers (like
>>>> Dell) update the VGA hardware BIOS as part of the motherboard BIOS
>>>> update.
>>>> Arthur
>>>> ______________________________________________________________________
>>>> This email has been scanned by the MessageLabs Email Security System.
>>>> For more information please visit http://www.messagelabs.com/email
>>>> ______________________________________________________________________
>>>> _______________________________________________
>>>> Testing mailing list
>>>> Testing at lists.pcbsd.org
>>>> http://lists.pcbsd.org/mailman/listinfo/testing
>>> If you could, run the following command and see if it works.
>>>> glxgears
>>> It is just a quick test of the 3D acceleration for the video card. The
>>> reason is that I have noticed that NVIDIA Settings -> OpenGL/GLX
>>> information tab crashes if the 3D acceleration is not working. This
>>> might be part of the issue, because if you upgraded from Beta1.5 to
>>> Beta2 (as opposed to a fresh install of Beta2) we have noticed that
>>> there is a problem with the nvidia driver enabling 3D acceleration
>>> (this
>>> has already been fixed for the next release BTW).
>>> If this is the case, you might want to switch to the "nv" driver
>>> instead
>>> of the "nvidia" driver. It doesn't run nearly as fast and there are
>>> occasional graphical issues, but at least the 3D acceleration works
>>> and
>>> prevents crashes like this.
>>> Also, restoring from sleep/suspend is new for FreeBSD 9.0 so there
>>> might
>>> still be some bugs in it, but I think that is unrelated to the first
>>> issue here.
>> Ok, I have figured out the problem on my end with this issue. When I
>> have a screen saver set (in my case, OpenGL> Solar Winds) and it runs
>> and then the monitor goes to sleep, after it wakes, it must crash and
>> kick the user out to the login screen. Last night I turned off the
>> screen saver and just let the PC sit at desktop. I came in this morning
>> and woke it via mouse and it was still at desktop. Now I am going to
>> try
>> the same being on the "OpenGL Information" screen to see what happens.
>> Incidentally, Julian has an ATI card, not NVidia.

and using the VESA driver

> Well, my experiment has proven that it must be related to OpenGL /
> Mesa or *maybe* the Intel driver. I left the machine running overnight
> on the OpenGL Information screen with no screen set active. Came in
> morning and shook the mouse to wake the monitor and it was still at
> the OpenGL Information screen. I am satisfied with my conclusions.

well, that much is likely here too.. entering the OpenGL info page 
makes the session leader process or the Xserver crash.. though I can't 
say which.. probably the first.

More information about the Testing mailing list