Fwd: [Bug 469470] Re: proprietary nvidia driver does not work under karmic

Tue Nov 17 07:31:29 UTC 2009

NoOp wrote:
> On 11/16/2009 07:17 PM, Leonard wrote:
>> NoOp wrote:
>>> On 11/16/2009 02:02 PM, Leonard Chatagnier wrote:
>>>>
>>>>
>>>> --- On Mon, 11/16/09, NoOp...
>>>>> OK.
>>>>>
>>>>> When you login&   use gnome instead of kde do you still have the
>>>>> same issues?
>>>>>
>>>>>
>>>> When the kdm login screen does come up, however i get it to come up,
>>>> I can login to anything listed in the drop down menu and the desktop
>>>> runs fine without issue.  I just logged in to gnome and strangely
>>>> this time I had no error messages that the bug report lists.  It
>>>> happens that way some times but for the most part I get those error
>>>> messages about not recognizing my monitor, keyboard or graphics card
>>>> and I have to go and start logging into the tty and suddenly the kdm
>>>> login screen pops up before even starting to login most of the time.
>>>> Very strange isn't it.  Now, if you want me to switch to gdm login
>>>> and see what happens, let me know.
>>>
>>> It sounds like a race condition somewhere. Either that or you've got
>>> some hardware going boarderline. Run the usual memtest, check video card
>>> &   CPU temps etc., etc.
>>>
>> It's a race alright; a race to see if I can fix it before the devs ever
>> start to work on the bug.
>
> I am talking about a race condition where the kernel options are perhaps
> loading something before the nvidia drivers&  vice versa. My point is
> that the situation doesn't seem to be consistent on your system, so
> _something_ is conflicting on sytstem start.
>
I knew what you meant.  Should have put a smiley at the end.
Is there something we can put in rc.local to make the driver start early 
or first. no smiley!
>
> IMHO, memtest is a waste of time.  Have run
>> it many times without finding an error. I doubt hardware is going
>> borderling as whenever X starts normally or when I make it start by
>> dropping to a tty and waiting a little, kdm generally starts on its own
>> without me even logging on if I wait long enough.
>
> And that tends to confirm my point above&  also seems to indicate that
> kdm *will* start eventually. So, something is getting hung up in the
> boot process. Have a look at your logs&  see if you can spot anything
> obvious.
>
Well, I don't think you want to look at all related things in my 
logs-:o) but I'll list a few that may confirm your race condition 
theory. BTW, I think so too.  Anyway:
xorg.0.log shows
(==) ModulePath set to "/usr/lib/xorg/modules"
(WW) AllowEmptyInput is on, devices using drivers 'kbd', 'mouse' or 
'vmmouse' will be disabled.
(WW) Disabling Keyboard0
(WW) Disabling Mouse0
But later on toward the end they are initialized.

after an entry saying configuring hal and a few lines later it shows
(II) AT Translated Set 2 keyboard: Device reopened after 1 attempts.
(II) Power Button: Device reopened after 1 attempts.
(II) Power Button: Device reopened after 1 attempts.
(II) Macintosh mouse button emulation: Device reopened after 1 attempts.
(II) USB Optical Mouse: Device reopened after 1 attempts.
These are the last 5 lines in the log.
kdm.log shows
first line
error setting MTRR (base = 0xe0000000, size = 0x10000000, type = 1) 
Invalid argument (22)
  ddxSigGiveUp: Closing log
This shows up and repeated multiple times:
(EE) open /dev/fb0: No such file or directory
(EE) Failed to initialize GLX extension (Compatible NVIDIA X driver not 
found)
error setting MTRR (base = 0xe0000000, size = 0x10000000, type = 1) 
Invalid argument (22)
  ddxSigGiveUp: Closing log

Fatal server error:
Server is already active for display 0
         If this server is no longer running, remove /tmp/.X0-lock
         and start again.

Please consult the The X.Org Foundation support
          at http://wiki.x.org
  for help.

  ddxSigGiveUp: Closing log
Warning:  Failsafe mode was already attempted within 30 seconds.
Warning:  Falling back to gdm to report the issue.( I have no gdm log 
only a gdm dir with some log files(many). Let me know if you want to see 
them and I'll send them)

Fatal server error:
Server is already active for display 0
         If this server is no longer running, remove /tmp/.X0-lock
         and start again.

Please consult the The X.Org Foundation support
          at http://wiki.x.org
  for help.

  ddxSigGiveUp: Closing log
However, I don't think it is current as I have seen it many times 
earlier before my latest setup here.  The last entry which looks like an 
xorg.log log entry has no entries as below:
  ddxSigGiveUp: Closing log
Warning:  Failsafe mode was already attempted within 30 seconds.
Warning:  Falling back to gdm to report the issue.

X.Org X Server 1.6.4
Release Date: 2009-9-27
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.24-23-server x86_64 Ubuntu
Current Operating System: Linux karmic-desktop 2.6.31-15-generic 
#50-Ubuntu SMP Tue Nov 10 14:53:52 UTC 2009 x86_64
Kernel command line: BOOT_IMAGE=/boot/vmlinuz-2.6.31-15-generic 
root=UUID=5045eac2-7204-4029-90b5-bfa2285681b2 ro quiet splash
Build Date: 26 October 2009  05:19:56PM
xorg-server 2:1.6.4-2ubuntu4 (buildd@)
         Before reporting problems, check http://wiki.x.org
         to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
         (++) from command line, (!!) notice, (II) informational,
         (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Mon Nov 16 22:43:56 2009
(==) Using config file: "/etc/X11/xorg.conf"
(END)
syslog had:
Nov 16 22:43:41 karmic-desktop kdm[1653]: X server died during startup
Nov 16 22:43:41 karmic-desktop kdm[1653]: Failed to start X server. 
Starting failsafe X server.
Repeated at least once but didn't scroll through the whole file.
Messaged had:
Nov 16 22:43:26 karmic-desktop kernel: [   13.252180] usplash:361 
freeing invalid memtype fffffffff9000000-fffffffff9e00000
Nov 16 22:43:26 karmic-desktop kernel: [   13.949931] mtrr: 
base(0xf9000000) is not aligned on a size(0xe00000) boundary
Nov 16 22:44:05 karmic-desktop kernel: [   53.002523] Clocksource tsc 
unstable (delta = -111214015 ns)
Nov 16 22:44:33 karmic-desktop pulseaudio[2017]: sink-input.c: Failed to 
create sink input: sink is suspended.
Nov 16 23:21:13 karmic-desktop pulseaudio[2017]: ratelimit.c: 68 events 
suppressed

Is that enough.  Want more? Say which one to review.

> JMO. I can run kde all
>> day after logging in and never an issue comes up.  Nvidia settings shows
>> the card core temp in thermal as 57-61 C with a slowdown threshold of
>> 255 C. Looks good to me.
> ...
>
>> Well, that didn't help; got any idea how to run
>> computertemp or can suggest a better program(kde preferred) to check cpu
>> temperatures.
>
> Gnome :-)
>
No!  Well maybe but so far haven't had time to switch to gnome and look 
for it.
>>
>>> You might also try connecting via an analog cable if you have one to see
>>> if the dvi is an issue... or, set to login automatically, boot up
>>> _without_ the monitor powered on/connected&   when you see the machine
>>> settle, power up the monitor. You're screen will most likely be set to
>>> 800x600 at that point (reset it using System|Display&   don't forget to
>>> set the proper refresh rate 60hz, but try a few others just to test).
>>>     It will be interesting to see if: 1) the autologin makes it through
>>> the process, and 2) if resetting via System|Display works properly. Then
>>> try with the nvidia-settings utility.
>>>
>> First, let me find an analog usb cable if I can.  The HDMI cable I
>> bought wasn't a cheap one and I will be upset if it is bad.  I doubt it
>> is.  I'll try the auto login tonight but here is some new info I added
>> the the bug report:
>
> Turn off the monitor w/autologin and test. Takes about 5 minutes to do.
> My point here is that some other bugs, similar to yours, indicated
> issues with digital cable connections. Doesn't take a genious to set
> autologin&  reboot with the monitor off :-)
>
Got auto login set up and rebooted with monitor on which caused startup 
X errors for first boot in about 6 times.  Will now try without monitor 
being on.  Wait, wait; must save draft, must save draft. With monitor 
off first reboot gave a highpitched rapid beeping noise; a warning 
right?  On second reboot with monitor off the beeping didn't stop and in 
both cases I had a black screen with a blinking cursor and only the 
magic blue button held down for several long seconds would shut it down.
Rebooting with monitor on went through auto login to kde without issue.
Seems like my new xorg.conf is still working-:o)

>> I rmed xorg.conf and got the message shown on the bug report but can't
>> remember whether it was on the first reboot or not.  Anyway,
>> nvidia-setting wouldn't work
>
> So reinstall it.
>
It was installed and had an output but no xorg.conf to read; thus the error.

> $ sudo apt-get purge nvidia-settings
> $ sudo apt-get install nvidia-settings
>
Yeah, I know.
>
> so I ran nvidia-xconfig as it said to do
>
> Who is 'it'?
>
It is error message from nvidia-settings.

>> which put a new but old style xorg.conf file in place and so far I've
>> rebooted 3 times with any error messages.  The data is on the bug report
>> if you want to see.
>
> Nah. nvidia-xconfig died after intrepid:
> http://packages.ubuntu.com/search?searchon=names&keywords=nvidia-xconfig
>
Yes, I know and it's not phantomly installed but I kid you not and I 
don't fib and I ran nvidia-xconfig just as I said and presto the new old 
fasioned xorg.conf fell in place in /etc/X11 as reported and shown in 
the bug report.  Strange but true or do you think I just imagined it. No 
smiley here!

> Purge&  reinstall nvidia-settings
>
Don't see the need. It's installed and funchioning correctly.  All the 
data show is correct for my system and hasn't changed from what it was.
But if you insist, I'll reinstall it.

> <http://packages.ubuntu.com/search?suite=default&section=all&arch=any&searchon=names&keywords=nvidia-settings>
> ...
>>> My nVidia cards are old NV25GL [Quadro4 900 XGL] cards&   use the -96
>>> drivers. Maybe you should downgrade to one of those&   a 32bit CPU :-)
>>>
>> A better option IMHO, is for you to install nvidia-glx-185 on your old
>> NV25gl and run it for a while rebooting ofter and see if you have any X
>> issues.
>
> Why would I do that? 185 doesn't support my card.
>
So right.  With all the card listings I didn't check to see if yours was 
listed.  I couldn't find it.  You'll have to get a supported card so you 
can help me-:O))  Smiley intended.

> You can post your findings to the bug report. So far there is
>> me and the OP and I have already put more on the report than he has and
>> never a followup from him. I fear the bug will never be addressed if
>> there are only two people having issues.  Help please.
>
> He probably bought a $10 NV25GL [Quadro4 900 XGL] and is happily
> computing away :-)
>
Yeah, He, He.  I hope you get something useful out of this as it took a 
long time to do.  Note: not counting the monitor off reboots, I'm still 
booting up without X issues with the new xorg.conf generated with the 
dead nvidia-xconfig.  Spirits are powerful beings-:o) Smiley intended.
>
>
>

-- 
Leonard
lenc5570 at sbcglobal.net