Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.hardware > #1254 > unrolled thread

Intel gcg 830 on Debian Squeeze

Back to article view | Back to comp.os.linux.hardware

Contents

  Intel gcg 830 on Debian Squeeze Thomas Richter <thor@math.tu-berlin.de> - 2012-05-24 11:05 +0200
    Re: Intel gcg 830 on Debian Squeeze Aragorn <stryder@telenet.be.invalid> - 2012-05-24 14:10 +0200
      Re: Intel gcg 830 on Debian Squeeze Thomas Richter <thor@math.tu-berlin.de> - 2012-05-24 14:52 +0200
        Re: Intel gcg 830 on Debian Squeeze Aragorn <stryder@telenet.be.invalid> - 2012-05-24 16:45 +0200
          Re: Intel gcg 830 on Debian Squeeze Thomas Richter <thor@math.tu-berlin.de> - 2012-05-25 14:00 +0200
            Re: Intel gcg 830 on Debian Squeeze Thomas Richter <thor@math.tu-berlin.de> - 2012-05-25 16:15 +0200
        Re: Intel gcg 830 on Debian Squeeze buck <buck@private.mil> - 2012-05-24 18:30 +0000
          Re: Intel gcg 830 on Debian Squeeze Thomas Richter <thor@math.tu-berlin.de> - 2012-06-10 18:14 +0200

#1254 — Intel gcg 830 on Debian Squeeze

Hi folks,

trying to install an old laptop with Debian, I run into a couple of 
problems with its intel chipset graphics. As stated, it identifies 
itself as an intel 82830 CGC rev 04. When booting from Debian Squeeze,
I get no screen at all. Ok, so I deactivated modeset in grub by adding 
the boot parameter i915.modeset=0. Then it booted up to the desktop, 
where the kernel froze for a reason it was never able to tell me. 
Probably related to the network or the wireless chip?

I thus installed kernel 3.2.0 from Squeeze-backports. This then gave me 
no screen at all, probably because the intel driver requires KMS. I 
tried upgrading xorg and xserver-xorg-video-intel, but again, no cigar. 
I then removed the intel driver altogether, leaving only the vesa driver 
in the system, which finally gave me a desktop (and no kernel freeze).

However, what I find rather bewildering is that

*) without kms, I cannot use the intel driver, get only a console - and -
*) with kms, I get no screen, not even a console.

Kind of a deadlock situation, the system doesn't work either way. Is 
there any chance of getting the system to work with anything *but* the 
vesa xorg driver?

Greetings,
	Thomas

[toc] | [next] | [standalone]

#1255

On Thursday 24 May 2012 11:05, Thomas Richter conveyed the following to 
comp.os.linux.hardware...

> Hi folks,
> 
> trying to install an old laptop with Debian, I run into a couple of
> problems with its intel chipset graphics. As stated, it identifies
> itself as an intel 82830 CGC rev 04. When booting from Debian Squeeze,
> I get no screen at all. Ok, so I deactivated modeset in grub by adding
> the boot parameter i915.modeset=0. Then it booted up to the desktop,
> where the kernel froze for a reason it was never able to tell me.

How do you know for certain whether it was the kernel that was frozen?  
Did you try the magic SysRq keys?

> Probably related to the network or the wireless chip?

Your guess is as good as ours, with the information you provide.

> I thus installed kernel 3.2.0 from Squeeze-backports. This then gave
> me no screen at all, probably because the intel driver requires KMS.

It does, yes.

> I tried upgrading xorg and xserver-xorg-video-intel, but again, no
> cigar. I then removed the intel driver altogether, leaving only the
> vesa driver in the system, which finally gave me a desktop (and no
> kernel freeze).
> 
> However, what I find rather bewildering is that
> 
> *) without kms, I cannot use the intel driver, get only a console -
> and - *) with kms, I get no screen, not even a console.
> 
> Kind of a deadlock situation, the system doesn't work either way. Is
> there any chance of getting the system to work with anything *but* the
> vesa xorg driver?

You could try and install kernel 3.4, which has significant improvements 
with regard to the Open Source "intel" driver.  But chances are that 
you're going to have to build it from sources, as I don't think any 
distribution has already included it in its repositories in binary form.

-- 
= Aragorn =
(registered GNU/Linux user #223157)

[toc] | [prev] | [next] | [standalone]

#1256

Am 24.05.2012 14:10, schrieb Aragorn:
>
> How do you know for certain whether it was the kernel that was frozen?
> Did you try the magic SysRq keys?

Not compiled into the default kernel, however the machine was 
inresponsive in every way. I *suspect* it might be the network because a 
similar type of network chip (a prism2 intersil) gave me problems in a 
different machine. However, this is rather unimportant for the problem 
at hand - it was only the motivation for why I had to exchange the 
kernel. The new kernel fixed *this* specific problem, but created 
another. )-:

>> I thus installed kernel 3.2.0 from Squeeze-backports. This then gave
>> me no screen at all, probably because the intel driver requires KMS.
>
> It does, yes.
>
>> I tried upgrading xorg and xserver-xorg-video-intel, but again, no
>> cigar. I then removed the intel driver altogether, leaving only the
>> vesa driver in the system, which finally gave me a desktop (and no
>> kernel freeze).
>>
>> However, what I find rather bewildering is that
>>
>> *) without kms, I cannot use the intel driver, get only a console -
>> and - *) with kms, I get no screen, not even a console.
>>
>> Kind of a deadlock situation, the system doesn't work either way. Is
>> there any chance of getting the system to work with anything *but* the
>> vesa xorg driver?
>
> You could try and install kernel 3.4, which has significant improvements
> with regard to the Open Source "intel" driver.  But chances are that
> you're going to have to build it from sources, as I don't think any
> distribution has already included it in its repositories in binary form.

Well, I'm willing to try that, even though it takes several hours to 
compile a kernel on this old machine (a 1GHz Pentium III). The question 
would rather be whether I would also need to recompile Xorg and the 
intel X driver - I suspect so. This would probably go beyond my time limit.

Greetings,
	Thomas

[toc] | [prev] | [next] | [standalone]

#1257

On Thursday 24 May 2012 14:52, Thomas Richter conveyed the following to 
comp.os.linux.hardware...

> Am 24.05.2012 14:10, schrieb Aragorn:
>>
>> You could try and install kernel 3.4, which has significant
>> improvements with regard to the Open Source "intel" driver.  But
>> chances are that you're going to have to build it from sources, as I
>> don't think any distribution has already included it in its
>> repositories in binary form.
> 
> Well, I'm willing to try that, even though it takes several hours to
> compile a kernel on this old machine (a 1GHz Pentium III). The
> question would rather be whether I would also need to recompile Xorg
> and the intel X driver - I suspect so. This would probably go beyond
> my time limit.

For the record, you would not need to recompile X.Org, and the intel 
driver is already part of the kernel - you can opt to compile the driver 
as a module or directly into the kernel.  

That's what the whole modesetting thing is about.  With "intel", 
"radeon" and "nouveau", the console and X11 drivers are unified into a 
single driver, so you no longer need separate drivers for console and 
X11.

-- 
= Aragorn =
(registered GNU/Linux user #223157)

[toc] | [prev] | [next] | [standalone]

#1259

Am 24.05.2012 16:45, schrieb Aragorn:
> On Thursday 24 May 2012 14:52, Thomas Richter conveyed the following to
> comp.os.linux.hardware...
>
>> Am 24.05.2012 14:10, schrieb Aragorn:
>>>
>>> You could try and install kernel 3.4, which has significant
>>> improvements with regard to the Open Source "intel" driver.  But
>>> chances are that you're going to have to build it from sources, as I
>>> don't think any distribution has already included it in its
>>> repositories in binary form.
>>
>> Well, I'm willing to try that, even though it takes several hours to
>> compile a kernel on this old machine (a 1GHz Pentium III). The
>> question would rather be whether I would also need to recompile Xorg
>> and the intel X driver - I suspect so. This would probably go beyond
>> my time limit.
>
> For the record, you would not need to recompile X.Org, and the intel
> driver is already part of the kernel - you can opt to compile the driver
> as a module or directly into the kernel.
>
> That's what the whole modesetting thing is about.  With "intel",
> "radeon" and "nouveau", the console and X11 drivers are unified into a
> single driver, so you no longer need separate drivers for console and
> X11.

I just tried kernel 3.4.0, with the same results. Nothing on the display 
as soon as modeswitching is enabled in the boot.

[toc] | [prev] | [next] | [standalone]

#1260

For the records: I finally got the machine to work. However, I can 
neither use the vesa nor the intel driver. The intel driver just tells 
me that I don't have a LVDS (local laptop flat panel) display connect, 
though it works with external monitors.

The vesa driver tells me that it cannot set the MTRR registers, and that 
the range start is not aligned to the range size. So why doesn't that 
stupid driver clip or round the settings accordingly? Oh well...

Finally, and the following worked:

*) disable the intel KMS modesettings: i915.modeset=0 at the kernel 
parameters,

*) enable the vesa frame buffer on the kernel: video=vesafb:mtrr:3,ywrap

*) set the vesa mode to the resolution of the display: vga=792
This is 1024x768.

*) Use in xorg.conf the fbdev driver.

Of course, the result is not exactly fast. But it works.

Greetings,
	Thomas

[toc] | [prev] | [next] | [standalone]

#1258

Thomas Richter <thor@math.tu-berlin.de> wrote in
news:jplav8$g17$1@news.belwue.de: 

I worked around this by using a 2.6.37.6 kernel.  Kernel 3.2.13 is 
hopelessly broken.

If you find a kernel version 3.X that works, I'd sure like to hear about 
it.
--
buck

[toc] | [prev] | [next] | [standalone]

#1263

On 24.05.2012 20:30, buck wrote:
> Thomas Richter<thor@math.tu-berlin.de>  wrote in
> news:jplav8$g17$1@news.belwue.de:
>
> I worked around this by using a 2.6.37.6 kernel.  Kernel 3.2.13 is
> hopelessly broken.
>
> If you find a kernel version 3.X that works, I'd sure like to hear about
> it.

It turned out that the problem was that the adapter chip that connects 
the i830 chipset to the flat panel was not supported with the current 
kernel. However, after tweaking a already partially existing solution, I 
could come up with a kernel patch that adds support for my specific 
laptop model. You'll find the kernel patch here:

https://bugs.freedesktop.org/show_bug.cgi?id=17902

Scroll right to the end - this patches against 3.4.1 and makes it 
working - for me.

So long,
	Thomas

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.hardware

csiph-web