Trouble configuring xorg on a 64-bit system

Dan Nicholson dbn.lists at gmail.com
Mon Aug 31 12:00:56 PDT 2009


On Mon, Aug 31, 2009 at 11:39 AM, Bruce Dubbs<bruce.dubbs at gmail.com> wrote:
> Dan Nicholson wrote:
>
>> Yeah, that's the one. If I recall, this FATAL is not really the one
>> bringing down X. Ah, now I remember. The "FATAL: Module fbcon not
>> found" is actually coming from modprobe since you have fbcon built in
>> and it can't find a loadable fbcon module. It's not actually fatal to
>> X since the module is loaded (and X wouldn't need it anyway). There's
>> gotta be a different error there, so I think we'd need to look at the
>> X log.
>
> I can't get the i915 log unless I remove th nvidia card.  I'm not sure this is
> worth trouble shooting.  It isn't to me, but if you want me to, I'll do it.

Nah. You may be able to toggle a BIOS switch that says "yes, I really
want to run from the integrated graphics" without removing the
discrete card, though. All I'd really be able to tell you is "that
might be fixed in a newer kernel".

>> This is an Intel G45 or something? What type of monitor and connection
>> are you driving?
>
> (--) NVIDIA(0): --- EDID for DELL IN1910N (CRT-1) ---
> (--) NVIDIA(0): EDID Version                 : 1.3
> (--) NVIDIA(0): Manufacturer                 : DEL
> (--) NVIDIA(0): Monitor Name                 : DELL IN1910N
> (--) NVIDIA(0): Product ID                   : 41036
> (--) NVIDIA(0): 32-bit Serial Number         : 911625043
> (--) NVIDIA(0): Serial Number String         : J349N96N6VKS
> (--) NVIDIA(0): Manufacture Date             : 2009, week 26
> (--) NVIDIA(0): DPMS Capabilities            : Standby Suspend Active Off
> (--) NVIDIA(0): Prefer first detailed timing : Yes
> (--) NVIDIA(0): Supports GTF                 : No
> (--) NVIDIA(0): Maximum Image Size           : 410mm x 230mm
> (--) NVIDIA(0): Valid HSync Range            : 30.0 kHz - 83.0 kHz
> (--) NVIDIA(0): Valid VRefresh Range         : 56 Hz - 75 Hz
> (--) NVIDIA(0): EDID maximum pixel clock     : 90.0 MHz
>
> (--) NVIDIA(0): Detailed Timings:
> (--) NVIDIA(0):   1366 x 768  @ 60 Hz
> (--) NVIDIA(0):     Pixel Clock      : 85.50 MHz
> (--) NVIDIA(0):     HRes, HSyncStart : 1366, 1436
> (--) NVIDIA(0):     HSyncEnd, HTotal : 1579, 1792
> (--) NVIDIA(0):     VRes, VSyncStart : 768, 771
> (--) NVIDIA(0):     VSyncEnd, VTotal : 774, 798
> (--) NVIDIA(0):     H/V Polarity     : +/+
>
> It's a flat panel, but it uses a standard 15-pin VGA connection.
>
>> If I had to take a guess, I would say that KMS is
>> failing for you. They're still working some of the bugs out to get KMS
>> up to the standard of the userspace modesetting from xf86-video-intel.
>> In that case, the options are to boot with i915.modeset=0 or try newer
>> kernels to see if things are fixed. Of course, whether you want to
>> pursue or not is up to you since you've got working nvidia graphics.
>
> I don't know what KMS is.  The system is a stock LFS-6.5 with a 2.6.30.2 kernel.
>  I don't know how fast they develop.  If you mean Keyboard-Mouse-System,  what
> I have now is:

Sorry. KMS is kernel modesetting. This has been a recent change.
Previously, the X drivers did all the modesetting from userspace,
which required a lot of banging on the card as root. Moving all the
privileged code for modesetting into the kernel allows lots of
improvements. Here's a far better explanation than I could ever give:

http://keithp.com/blogs/kernel-mode-drivers/

The rub here is that the intel driver (and now radeon) have very
mature modesetting code for userspace that has to be ported over to
the kernel. This has mostly been done for intel, but people are still
reporting regressions (black screen on intel was one that was reported
often).

--
Dan



More information about the blfs-dev mailing list