X with kernel 2.6.7 produces black screen then freezes computer

[Debian Bug Importer]

Automatically imported from Debian bug report #257465 http://bugs.debian.org/257465

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #257465 http://bugs.debian.org/257465

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sat, 03 Jul 2004 18:21:31 +0100
From: Helen Faulkner <email address hidden>
To: <email address hidden>
Subject: X with kernel 2.6.7 produces black screen then freezes computer

Package: xserver-xfree86
Version: 4.3.0.dfsg.1-5
Severity: grave

This problem appears to be similar to Bug #256356, though not I'm not
getting exactly the same symptoms. I just installed the new kernel
image 2.6.7-1-686, and also a new nvidia-kernel-source package. I guess
the problem is with X not with nvidia, because I haven't gotten the
usual sort of "I can't deal with this nvidia thing" errors that X has
given me on occasion before, and because lsmod showed the nvidia module
had loaded correctly, even though X wasn't working at all.

What happens is this:
  - boot the computer with the 2.6.7 kernel

  - everything looks normal until we are about to get the first X screen

  - then the monitor goes completely black, and seems totally
unresponsive. ctrl-alt-backspace to try to restart x doesn't seem to do
anything, but ctrl-alt-1 gives me a text login prompt as usual.

  - logging in at the text console allows me to run ps aux, producing
the output at the bottom of this email (maybe it's useful), There
seems to be some X stuff there, so maybe some things are working.

  - trying to run startx from the text console, to see if it gives error
messages, produces the same black screen, and this time the computer
appears to be totally frozen with no response to the keyboard at all.

  - after reading the bug report for #256356, I tried adding
 Option "NoAccel" "True"
to my /etc/X11/XF86Config-4, but that didn't make any difference. My
symptoms seem slightly different to that bug report, so maybe that's
irrelevant anyway. My /etc/X11/XF86Config-4 is attached below incase it
helps.

Thanks,

Helen

__________________________________________________________________________

Output of "ps aux" after logging into text console:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.2 0.1 1528 528 ? S 17:27 0:00 init [2]
root 2 0.0 0.0 0 0 ? SN 17:27 0:00 [ksoftirqd/0]
root 3 0.0 0.0 0 0 ? S< 17:27 0:00 [events/0]
root 4 0.0 0.0 0 0 ? S< 17:27 0:00 [khelper]
root 5 0.0 0.0 0 0 ? S< 17:27 0:00 [kacpid]
root 39 0.0 0.0 0 0 ? S< 17:27 0:00 [kblockd/0]
root 49 0.0 0.0 0 0 ? S 17:27 0:00 [pdflush]
root 50 0.0 0.0 0 0 ? S 17:27 0:00 [pdflush]
root 52 0.0 0.0 0 0 ? S< 17:27 0:00 [aio/0]
root 51 0.0 0.0 0 0 ? S 17:27 0:00 [kswapd0]
root 53 0.0 0.0 0 0 ? S 17:27 0:00 [kpnpbiosd]
root 188 0.0 0.0 0 0 ? S 17:27 0:00 [kseriod]
root 304 0.0 0.0 0 0 ? S 17:27 0:00 [kjournald]
root 538 0.0 0.0 0 0 ? S 17:27 0:00 [scsi_eh_0]
root 769 0.0 0.0 0 0 ? S 17:27 0:00 [khubd]
root 865 0.0...

[Fabio Massimo Di Nitto (fabbione)]

tag 257465 moreinfo
stop

Hi,
 does X work if you revert the kernel to 2.6.6-2 ? does X work with
2.6.7 if you disable the nvidia drivers?

Fabio

--
<user> fajita: step one
<fajita> Whatever the problem, step one is always to look in the error log.
<user> fajita: step two
<fajita> When in danger or in doubt, step two is to scream and shout.

[Helen Faulkner (helen-ml-faulkner)]

Hello,

> does X work if you revert the kernel to 2.6.6-2 ?

Yes. Everything is working fine then.

> does X work with
> 2.6.7 if you disable the nvidia drivers?

I'm not sure how to do this? Do you mean to change "nvidia" in
XF86Config-4 to something else? In which case I don't see how it would
work properly, because nvidia is the right (presumably the only) driver
for my graphics card.

Or do you mean to uninstall the nvidia kernel drivers that I made from
the nvidia-kernel-source? Usually if I don't build the nvidia drivers
when a new kernel is installed, X doesn't work, and it gives me error
messages about the monitor until I build and install a new kernel
package from nvidia-kernel-source.

Sorry if I'm stating the obvious or being stupid. I'm not very clear
about how this stuff works.

Helen.

[Crimsun-fungus (crimsun-fungus)]

On Sun, Jul 04, 2004 at 10:27:00AM +0100, Helen Faulkner wrote:
> >does X work with
> >2.6.7 if you disable the nvidia drivers?
>
> I'm not sure how to do this? Do you mean to change "nvidia" in
> XF86Config-4 to something else? In which case I don't see how it would
> work properly, because nvidia is the right (presumably the only) driver
> for my graphics card.

You should be able to change just
Driver "nvidia" to
Driver "nv"
in Section "Device" of /etc/X11/XF86Config-4. You'll also want to
comment out
Load "glx"
in Section "Module".

--
Daniel T. Chen <email address hidden>
GPG key: www.sh.nu/~crimsun/pubkey.gpg.asc

[Helen Faulkner (helen-ml-faulkner)]

<email address hidden> wrote:
> On Sun, Jul 04, 2004 at 10:27:00AM +0100, Helen Faulkner wrote:
>
>>>does X work with
>>>2.6.7 if you disable the nvidia drivers?

> You should be able to change just
> Driver "nvidia" to
> Driver "nv"
> in Section "Device" of /etc/X11/XF86Config-4. You'll also want to
> comment out
> Load "glx"
> in Section "Module".

OK, I made those changes and rebooted with the 2.6.7 kernel. X didn't
work, but it gave error messages in the console rather than just hanging
with a black screen. This is probably similar to what I've seen before
with the nv module, which is why I was using the nvidia modules.

The end of the messages in /var/log/XFree86.0.log was like this:

(WW) open /dev/fb0: No such device
(WW) open /dev/fb1: No such device
(WW) open /dev/fb2: No such device
(WW) open /dev/fb3: No such device
(WW) open /dev/fb4: No such device
(WW) open /dev/fb5: No such device
(WW) open /dev/fb6: No such device
(WW) open /dev/fb7: No such device
(EE) Unable to find a valid framebuffer device
(EE) NV(0): Failed to open framebuffer device, consult warnings and/or
errors above for possible reasons
 (you may have to look at the server log to see warnings)
(II) UnloadModule: "nv"
(II) UnloadModule: "fbdevhw"
(II) Unloading /usr/X11R6/lib/modules/linux/libfbdevhw.a
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(II) UnloadModule: "int10"
(EE) Screen(s) found, but none have a usable configuration.

But, (here's where it got strange): I then carefully changed my
/etc/X11/XF86Config-4 file back to how it was to begin with. But then X
wouldn't boot with the older (2.6.6-2) kernel. I tried several times
and it doesn't work. The end of the messages in /var/log/XFree86.0.log
now go like this:

(--) NVIDIA(0): Linear framebuffer at 0xD0000000
(--) NVIDIA(0): MMIO registers at 0xFC000000
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
(EE) NVIDIA(0): *** Aborting ***
(II) UnloadModule: "nvidia"
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(EE) Screen(s) found, but none have a usable configuration.

However (the good news, I think...) X now does work with the 2.6.7
kernel, which is definitely wasn't doing last night. So I don't seem to
be able to reproduce the bug I reported (which is good if it's gone and
bad if it's merely hiding), but am now getting a different wierdness
(that X no longer works with the old kernel).

I did dselect-update-install today, to update anything that might fix
the bug. I didn't notice anything new that related to X or to nvidia,
but I guess something may have changed so the nvidia driver now works
with the 2.6.7 kernel but not with the 2.6.6 kernel, if that's possible.
  Or maybe trying to use the nv module has changed something.

Sorry if this too confusing to be helpful. I am confused too.

Helen.

[Fabio Massimo Di Nitto (fabbione)]

On Sun, 4 Jul 2004, Helen Faulkner wrote:

> > You should be able to change just
> > Driver "nvidia" to
> > Driver "nv"
> > in Section "Device" of /etc/X11/XF86Config-4. You'll also want to
> > comment out
> > Load "glx"
> > in Section "Module".
>
> OK, I made those changes and rebooted with the 2.6.7 kernel. X didn't
> work, but it gave error messages in the console rather than just hanging
> with a black screen. This is probably similar to what I've seen before
> with the nv module, which is why I was using the nvidia modules.

>
> But, (here's where it got strange): I then carefully changed my
> /etc/X11/XF86Config-4 file back to how it was to begin with. But then X
> wouldn't boot with the older (2.6.6-2) kernel. I tried several times
> and it doesn't work. The end of the messages in /var/log/XFree86.0.log
> now go like this:
>
> (--) NVIDIA(0): Linear framebuffer at 0xD0000000
> (--) NVIDIA(0): MMIO registers at 0xFC000000
> (EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
> (EE) NVIDIA(0): *** Aborting ***
> (II) UnloadModule: "nvidia"
> (II) UnloadModule: "vgahw"
> (II) Unloading /usr/X11R6/lib/modules/libvgahw.a
> (EE) Screen(s) found, but none have a usable configuration.
>
> However (the good news, I think...) X now does work with the 2.6.7
> kernel, which is definitely wasn't doing last night. So I don't seem to
> be able to reproduce the bug I reported (which is good if it's gone and
> bad if it's merely hiding), but am now getting a different wierdness
> (that X no longer works with the old kernel).

I think there was a simple mistake your side compiling the NVIDIA driver.
You always have to build it with the running kernel you want to use later.

For example what i do is:

install new kernel
stop any graphical login
reboot into the new kernel
install nvidia drivers
start graphical login

and this always worked for me 100% of the times.

> Sorry if this too confusing to be helpful. I am confused too.

No problem. I am closing this bug but if it happens again please inform us
immediatly.

Thanks
Fabio

--
<user> fajita: step one
<fajita> Whatever the problem, step one is always to look in the error log.
<user> fajita: step two
<fajita> When in danger or in doubt, step two is to scream and shout.

[Fabio Massimo Di Nitto (fabbione)]

User error compiling nvidia drivers

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 4 Jul 2004 09:05:48 +0200 (CEST)
From: Fabio Massimo Di Nitto <email address hidden>
To: <email address hidden>
Cc: <email address hidden>
Subject: more info

tag 257465 moreinfo
stop

Hi,
 does X work if you revert the kernel to 2.6.6-2 ? does X work with
2.6.7 if you disable the nvidia drivers?

Fabio

--
<user> fajita: step one
<fajita> Whatever the problem, step one is always to look in the error log.
<user> fajita: step two
<fajita> When in danger or in doubt, step two is to scream and shout.

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 04 Jul 2004 10:27:00 +0100
From: Helen Faulkner <email address hidden>
To: Fabio Massimo Di Nitto <email address hidden>,
 <email address hidden>
Subject: Re: Bug#257465: more info

Hello,

> does X work if you revert the kernel to 2.6.6-2 ?

Yes. Everything is working fine then.

> does X work with
> 2.6.7 if you disable the nvidia drivers?

I'm not sure how to do this? Do you mean to change "nvidia" in
XF86Config-4 to something else? In which case I don't see how it would
work properly, because nvidia is the right (presumably the only) driver
for my graphics card.

Or do you mean to uninstall the nvidia kernel drivers that I made from
the nvidia-kernel-source? Usually if I don't build the nvidia drivers
when a new kernel is installed, X doesn't work, and it gives me error
messages about the monitor until I build and install a new kernel
package from nvidia-kernel-source.

Sorry if I'm stating the obvious or being stupid. I'm not very clear
about how this stuff works.

Helen.

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 4 Jul 2004 14:51:35 -0700
From: <email address hidden>
To: Helen Faulkner <email address hidden>,
 <email address hidden>
Subject: Re: Bug#257465: more info

--nFreZHaLTZJo0R7j
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Jul 04, 2004 at 10:27:00AM +0100, Helen Faulkner wrote:
> >does X work with
> >2.6.7 if you disable the nvidia drivers?
>=20
> I'm not sure how to do this? Do you mean to change "nvidia" in=20
> XF86Config-4 to something else? In which case I don't see how it would=
=20
> work properly, because nvidia is the right (presumably the only) driver=
=20
> for my graphics card.

You should be able to change just
Driver "nvidia" to
Driver "nv"
in Section "Device" of /etc/X11/XF86Config-4. You'll also want to
comment out
Load "glx"
in Section "Module".

--=20
Daniel T. Chen <email address hidden>
GPG key: www.sh.nu/~crimsun/pubkey.gpg.asc

--nFreZHaLTZJo0R7j
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQFA6Hvme9GwFciKvaMRAt16AJ9UayNc1sQeqeV94XjJM+gtEn6/WQCfSl5D
xdePz/ODa0IU87qZFr9k0H8=
=n4HD
-----END PGP SIGNATURE-----

--nFreZHaLTZJo0R7j--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 04 Jul 2004 23:48:13 +0100
From: Helen Faulkner <email address hidden>
To: <email address hidden>, <email address hidden>
Subject: Re: Bug#257465: more info

<email address hidden> wrote:
> On Sun, Jul 04, 2004 at 10:27:00AM +0100, Helen Faulkner wrote:
>
>>>does X work with
>>>2.6.7 if you disable the nvidia drivers?

> You should be able to change just
> Driver "nvidia" to
> Driver "nv"
> in Section "Device" of /etc/X11/XF86Config-4. You'll also want to
> comment out
> Load "glx"
> in Section "Module".

OK, I made those changes and rebooted with the 2.6.7 kernel. X didn't
work, but it gave error messages in the console rather than just hanging
with a black screen. This is probably similar to what I've seen before
with the nv module, which is why I was using the nvidia modules.

The end of the messages in /var/log/XFree86.0.log was like this:

(WW) open /dev/fb0: No such device
(WW) open /dev/fb1: No such device
(WW) open /dev/fb2: No such device
(WW) open /dev/fb3: No such device
(WW) open /dev/fb4: No such device
(WW) open /dev/fb5: No such device
(WW) open /dev/fb6: No such device
(WW) open /dev/fb7: No such device
(EE) Unable to find a valid framebuffer device
(EE) NV(0): Failed to open framebuffer device, consult warnings and/or
errors above for possible reasons
 (you may have to look at the server log to see warnings)
(II) UnloadModule: "nv"
(II) UnloadModule: "fbdevhw"
(II) Unloading /usr/X11R6/lib/modules/linux/libfbdevhw.a
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(II) UnloadModule: "int10"
(EE) Screen(s) found, but none have a usable configuration.

But, (here's where it got strange): I then carefully changed my
/etc/X11/XF86Config-4 file back to how it was to begin with. But then X
wouldn't boot with the older (2.6.6-2) kernel. I tried several times
and it doesn't work. The end of the messages in /var/log/XFree86.0.log
now go like this:

(--) NVIDIA(0): Linear framebuffer at 0xD0000000
(--) NVIDIA(0): MMIO registers at 0xFC000000
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
(EE) NVIDIA(0): *** Aborting ***
(II) UnloadModule: "nvidia"
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(EE) Screen(s) found, but none have a usable configuration.

However (the good news, I think...) X now does work with the 2.6.7
kernel, which is definitely wasn't doing last night. So I don't seem to
be able to reproduce the bug I reported (which is good if it's gone and
bad if it's merely hiding), but am now getting a different wierdness
(that X no longer works with the old kernel).

I did dselect-update-install today, to update anything that might fix
the bug. I didn't notice anything new that related to X or to nvidia,
but I guess something may have changed so the nvidia driver now works
with the 2.6.7 kernel but not with the 2.6.6 kernel, if that's possible.
  Or maybe trying to use the nv module has changed something.

Sorry if this too confusing to be helpful. I am confused too.

Helen.

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 5 Jul 2004 07:29:35 +0200 (CEST)
From: Fabio Massimo Di Nitto <email address hidden>
To: Helen Faulkner <email address hidden>,
 <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#257465: more info

On Sun, 4 Jul 2004, Helen Faulkner wrote:

> > You should be able to change just
> > Driver "nvidia" to
> > Driver "nv"
> > in Section "Device" of /etc/X11/XF86Config-4. You'll also want to
> > comment out
> > Load "glx"
> > in Section "Module".
>
> OK, I made those changes and rebooted with the 2.6.7 kernel. X didn't
> work, but it gave error messages in the console rather than just hanging
> with a black screen. This is probably similar to what I've seen before
> with the nv module, which is why I was using the nvidia modules.

>
> But, (here's where it got strange): I then carefully changed my
> /etc/X11/XF86Config-4 file back to how it was to begin with. But then X
> wouldn't boot with the older (2.6.6-2) kernel. I tried several times
> and it doesn't work. The end of the messages in /var/log/XFree86.0.log
> now go like this:
>
> (--) NVIDIA(0): Linear framebuffer at 0xD0000000
> (--) NVIDIA(0): MMIO registers at 0xFC000000
> (EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
> (EE) NVIDIA(0): *** Aborting ***
> (II) UnloadModule: "nvidia"
> (II) UnloadModule: "vgahw"
> (II) Unloading /usr/X11R6/lib/modules/libvgahw.a
> (EE) Screen(s) found, but none have a usable configuration.
>
> However (the good news, I think...) X now does work with the 2.6.7
> kernel, which is definitely wasn't doing last night. So I don't seem to
> be able to reproduce the bug I reported (which is good if it's gone and
> bad if it's merely hiding), but am now getting a different wierdness
> (that X no longer works with the old kernel).

I think there was a simple mistake your side compiling the NVIDIA driver.
You always have to build it with the running kernel you want to use later.

For example what i do is:

install new kernel
stop any graphical login
reboot into the new kernel
install nvidia drivers
start graphical login

and this always worked for me 100% of the times.

> Sorry if this too confusing to be helpful. I am confused too.

No problem. I am closing this bug but if it happens again please inform us
immediatly.

Thanks
Fabio

--
<user> fajita: step one
<fajita> Whatever the problem, step one is always to look in the error log.
<user> fajita: step two
<fajita> When in danger or in doubt, step two is to scream and shout.

[Branden Robinson (branden)]

> > Package: xserver-xfree86
> > Version: 4.3.0.dfsg.1-5
> > Severity: grave
[...]
> I think there was a simple mistake your side compiling the NVIDIA driver.
> You always have to build it with the running kernel you want to use later.
[...]
> No problem. I am closing this bug but if it happens again please inform us
> immediatly.

Also, in the future please remember that the property severity for this
sort of problem is "important", not "grave"...particularly when non-Debian
software (NVIDIA's proprietary driver) is in the mix.

Grave bugs affect or potentially affect every user of the package. But not
everyone has NVIDIA hardware, and not all of those folks use the non-free
kernel and XFree86 drivers for that hardware.

Before filing a release-critical bug, you should ask yourself, "should the
entire Debian release be held up due to this bug?"

It would be wise not to encourage people to form the impression that NVIDIA
video card users care about no one but themselves. :)

--
G. Branden Robinson | Why should I allow that same God
Debian GNU/Linux | to tell me how to raise my kids,
<email address hidden> | who had to drown His own?
http://people.debian.org/~branden/ | -- Robert Green Ingersoll

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sat, 10 Jul 2004 12:42:22 -0500
From: Branden Robinson <email address hidden>
To: <email address hidden>
Subject: Re: Bug#257465: marked as done (X with kernel 2.6.7 produces black screen then freezes
 computer)

--PEWkPmBCIbC5Tcj0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

> > Package: xserver-xfree86
> > Version: 4.3.0.dfsg.1-5
> > Severity: grave
[...]
> I think there was a simple mistake your side compiling the NVIDIA driver.
> You always have to build it with the running kernel you want to use later.
[...]
> No problem. I am closing this bug but if it happens again please inform us
> immediatly.

Also, in the future please remember that the property severity for this
sort of problem is "important", not "grave"...particularly when non-Debian
software (NVIDIA's proprietary driver) is in the mix.

Grave bugs affect or potentially affect every user of the package. But not
everyone has NVIDIA hardware, and not all of those folks use the non-free
kernel and XFree86 drivers for that hardware.

Before filing a release-critical bug, you should ask yourself, "should the
entire Debian release be held up due to this bug?"

It would be wise not to encourage people to form the impression that NVIDIA
video card users care about no one but themselves. :)

--=20
G. Branden Robinson | Why should I allow that same God
Debian GNU/Linux | to tell me how to raise my kids,
<email address hidden> | who had to drown His own?
http://people.debian.org/~branden/ | -- Robert Green Ingersoll

--PEWkPmBCIbC5Tcj0
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iEYEARECAAYFAkDwKn4ACgkQ6kxmHytGonxJ8gCgi6g+Id6jvG6kJbfjK98t4maB
074AmgK+wGlvQaCE+eK9nP7yBJE3nbqs
=7cOr
-----END PGP SIGNATURE-----

--PEWkPmBCIbC5Tcj0--