X problems

[Robin Laing]

Jeffrey Ross wrote:
>> Jeffrey Ross wrote:
>>> having problems with X on Fedora 7, these are problems I did not have
>>> with FC6.
>>>
>>> The big problem is X is locking up, the system continues to run and it
>>> accessable via the network, but a CTRL-ALT-BKspace does not restore the
>>> X session.
>> Does X die and leave whatever was previously displayed on the screen? I
>> had a problem with an Intel 865G card a long time back. X would actually
>> abort and the screen would not reset until the Graphics card is reset by
>> a system reboot.
> 
> As X crashed/hung while the screensaver was active, or probably more
> correctly coming active, the screen would be blank with just the mouse
> pointer on the screen.  I *think* but not 100% positive that X was trying
> to restart itself.  The only way I could recover was to reboot the system,
> I could not get to any of the virtual screens but ssh'ing into the system
> worked fine.
>>
>>> I found in the /var/log/Xorg.0.log.old the following:
>>>
>>> (II) AIGLX: Suspending AIGLX clients for VT switch
>>> Error in I830WaitLpRing(), timeout for 2 seconds
>>> pgetbl_ctl: 0x7ef80001 pgetbl_err: 0x0
>>> ipeir: 0 iphdr: bf089b31
>>> LP ring tail: 700 head: 0 len: 1f001 start 0
>>> Err ID (eir): 0 Err Status (esr): 1 Err Mask (emr): ffffffdf
>>> instdone: ffe5fafd instdone_1: fffff
>>> instpm: 0
>>> memmode: 0 instps: 8001e022
>>> HW Status mask (hwstam): fffedffe
>>> IRQ enable (ier): 2 imr: fffe0000 iir: 80
>>> acthd: 2dd8024 dma_fadd_p: 2dd8100
>>> ecoskpd: 307 excc: 0
>>> cache_mode: 6800/180
>>> mi_arb_state: 44
>>> IA_VERTICES_COUNT_QW 0/0
>>> IA_PRIMITIVES_COUNT_QW 0/0
>>> VS_INVOCATION_COUNT_QW 0/0
>>> GS_INVOCATION_COUNT_QW 0/0
>>> GS_PRIMITIVES_COUNT_QW 0/0
>>> CL_INVOCATION_COUNT_QW 0/0
>>> CL_PRIMITIVES_COUNT_QW 0/0
>>> PS_INVOCATION_COUNT_QW 0/0
>>> PS_DEPTH_COUNT_QW 0/0
>>> WIZ_CTL 0
>>> TS_CTL 0  TS_DEBUG_DATA fef7fdf6
>>> TD_CTL 0 / 0
>>> space: 129272 wanted 131064
>> The space xxx wanted yyy error sounds familiar with one system that I
>> tried running two different video cards on. The error could have been
>> from an Intel graphics using one graphics card. Regardless this sounds
>> like a problem introduced into the newer version of the driver.
>>
>>
>>> FatalError re-entered, aborting
>>> lockup
>> Did you add lockup or was this part of the excerpt from the log.
> 
> That is exactly how it appeared in the log, it appeared to end at the line
> "lockup".  Actually the log had at least 2 complete sets of the error
> sequence.
>>>
>>> The motherboard is an Intel DG965RY with a built in video card lspci
>>> reports the VGA controller as:
>>> 00:02.0 VGA compatible controller: Intel Corporation 82G965 Integrated
>>> Graphics Controller (rev 02)
>>
>> Which driver are you using? (i810 or intel)
> at the time I was using the intel driver, I have now switched to the i810
> and things seem to be better.  Not quite 100% but much better and livable.
>  One of the problems I am having now that I did not have with FC6 is when
> using mplayer in a full screen mode I am missing a small area near the
> bottom the screen.
>>> I believe one of the screen savers is triggering the bug as it just
>>> started happening after I loaded a bunch of screensavers and set it to
>>> randomly choose one, plus it only happens when the system brings up a
>>> screensaver.
>> noof was a good one for my testing on an Intel driver where X crashed
>> but left a residual display on the monitor.
> 
> I'll have to give that one a try.
>>
>>> In addition I have noticed that I can not play  video's with mplayer
>>> which by default uses xv (?) unless I specify "-vo x11"  in FC6 I didn't
>>> have this issue.
>> I think that you might get away with downloading the fc6 version and
>> reverting using rpm and the --oldpackage argument to see if it is
>> specifically caused by  changes in the driver.
> 
> Not 100% sure which driver/package I need to download but I'll see if I
> can identify it.
>>> I'm guessing that F7 is using a modified or new version of the Intel
>>> graphics driver.  Any way to either load the version from FC6 or specify
>>> the equivalent driver in Fedora 7?
>> Hopefully you only need to revert the driver. File a bug report for it
>> if the reverting to the older driver helps. Otherwise you might have
>> problems with some other X package which brings out the bug with the
>> Intel driver.
>>
>> Jim
>>
> 
> 
> 

I am having the same issues with a desktop with nvidia driver.  I was 
looking at blaming the nvidia driver but if the same issues with xorg 
taking up all the processing power is happening with the Intel driver, 
maybe there is an issue with the Fedora 7 xorg files and how it works 
with closed source drivers.

I see at least on lockup a day if not more.  ssh into the computer and 
kill the xorg process and the login screen returns.

I don't have enough info to submit a bug report yet.

-- 
Due to the move to Exchange Server,
    anything that is a priority, please phone.
Robin Laing