LinuxQuestions.org - [SOLVED] No 3D Acceleration on i5-2400 Asus P8Z68-V LE with Slack 13.37 Xfce

- Slackware (http://www.linuxquestions.org/questions/slackware-14/)

- - No 3D Acceleration on i5-2400 Asus P8Z68-V LE with Slack 13.37 Xfce (http://www.linuxquestions.org/questions/slackware-14/no-3d-acceleration-on-i5-2400-asus-p8z68-v-le-with-slack-13-37-xfce-939973/)

No 3D Acceleration on i5-2400 Asus P8Z68-V LE with Slack 13.37 Xfce

When I install the original/unaltered 32bit Slack 13.37 on an Asus P8Z68-V LE motherboard everything works fine when I bring up Xfce....except there's no 3D Acceleration and I could use some guidance. glxgears shows 60 FPS. I'm using the built in graphics, no separate video card.

Over the years I've always used NVIDIA video chips/cards usually with AMD processors. This is my first experience with Intel video hardware so I'm probably just ignorant of the tasks needed to get 3d acceleration with Intel. Please don't look for a bug if I've just not done my installation job properly.

I've spent many hours searching online looking for "instructions" on this without success. I didn't see anything in slackbuilds.org that looked like it applied. I'd appreciate pointers to a link showing how to set the Xorg.conf file or stating what combination of libraries, etc are needed to get 3d acceleration to work on this Z68 Asus board with i5-2400 Sandybridge processor and Slack 13.37.

Thanks.

--------------

Some highlights. Full logs are available if they would be helpful.

I've tried this with and without an X config file (/etc/X11/Xorg.conf). When I used the config file I set the Device driver to "intel".

lsmod shows the i915, drm, drm_kms_helper,intel_agp, intel_gtt, and others loaded.

glxinfo shows "direct rendering: Yes" and an OpenGL render string of "Mesa DRI Intel(R) Sandybridge Desktop x86/MMX/SSE2"

lspci shows "VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09)"

I get the same results whether running Xfce as a user or as root.

Boot up messages showed an MTRR allocation failure due to a type mismatch (write-back vs write-combining) which could result in poor graphics performance. (probably not related)

Ran trials with the motherboard TPU switch set either way (enable/disable). didn't find BIOS settings to affect 3d acceleration. BIOS version 3702. "Z68 Express" motherboard chipset. 16GB RAM.

I have logs of this activity but didn't want to take up space in this message if all I need to do is add a one-liner to Xorg.conf and try again.

How are you measuring 3D acceleration? The glxinfo output suggests that hardware acceleration should be working. If you're just going by glxgears output it should be noted that some graphics card drivers sync the framerate to the display's refresh rate (hence the 60 FPS). glxgears really isn't a benchmarking tool, and treating it as one, *especially* with synchronized rendering, is erroneous. Is there some other way you're measuring poor 3D acceleration?

The Intel drivers in 13.37 have problems with the second generation Intel HD graphics. Intel the newer kernel, Mesa and libdrm from /testing and you should be fine.

Quote:

Originally Posted by T3slider (Post 4653715)

I considered that perhaps I was chasing a problem that doesn't exist. glxgears on my other slack systems with NVIDIA all show the message about syncing the frame rate with the refresh rate but still show numbers like 4000 FPS.

This started out by installing Virtualbox on a 32bit Slack host with Windows 7 running as one of the guest OSes. MS Windows was working well but when I installed a game to see if the 3d acceleration would work, the game would crash (instead of running). The game runs just fine on identical hardware with only Windows 7 installed. Anyway I didn't want to add too many variables and the cause of the guest OS failure could be other things as virtualbox is new to me. I took a step back to see if the host operating system had 3d working which is where I'm at now.

Do you have a suggestion for another (simple) way to test that 3d acceleration is working? I'd like to try it. I'm not interested in the "benchmarking" aspect, just that it works.

Thanks for your response.

Quote:

Originally Posted by TobiSGD (Post 4653723)

The Intel drivers in 13.37 have problems with the second generation Intel HD graphics. Intel the newer kernel, Mesa and libdrm from /testing and you should be fine.

Thanks for your suggestion. I'll try it with the newer components and post the results.

Quote:

Originally Posted by Tracy Tiger (Post 4653739)

This started out by installing Virtualbox on a 32bit Slack host with Windows 7 running as one of the guest OSes. MS Windows was working well but when I installed a game to see if the 3d acceleration would work, the game would crash (instead of running). The game runs just fine on identical hardware with only Windows 7 installed. Anyway I didn't want to add too many variables and the cause of the guest OS failure could be other things as virtualbox is new to me. I took a step back to see if the host operating system had 3d working which is where I'm at now.

Virtualbox, as the name suggests, is creating a virtual PC, in which you installed your Windows. It is not using your physical hardware (except the CPU) at all. So that games don't run in Virtualbox is to 99% caused by either a lack of functions in the virtual graphics chip or simply a lack of performance. You can't compare it at all to an OS that runs on physical hardware when it comes to 3D functionality and performance.

Quote:

Originally Posted by TobiSGD (Post 4653753)

I was exploring the characteristics and limitations of virtualbox to see if it would be a good fit in three different computer setups I have in mind. I didn't know what video resources (e.g. 3d acceleration) on the hosting OS may be available through the virtual graphics interface to the guest OS or not, thus the experiment with the game.

You're right. A game running on real video hardware probably doesn't say anything about the same game running on the virtual video interface presented to it. My mistake in comparing the two.

I'll first get the 3d running on the Slack host system and then see if that affects the virtual video resources presented to the virtual machines. Thanks for your comments.

I'm having difficulty with my new motherboard(s) booting slackware-current from DVD (older motherboards are fine). Once I've resolved those problems. I'll post results/status of the OP to this thread. I may have to open a new thread to resolve my "current" problems, but first I'll try for a day or two on my own.

One step forward and two steps backwards. The more problems I have the more I learn. :-)

Module Troubleshooting Questions

Howdy !

No time to read but would like to help out? Skip to the bottom and respond to one or more of my 8 basic questions.

------

I've been playing with many variations of kernels (Huge, Generic, 2.6, 3.2, 3.3, Slack 13.37 & current, drivers up to the latest stable versions) and making observations. I didn't yet get all the upstream ones (intellinuxgraphics.org, mesa3d.org, freedesktop.org) to compile/install due to my skillset limits to deal with conflicting OS release streams (2.x vs 3.x).

Right now there have been too many changes on multiple builds on my 3 hardware-identical experimental boxes so I'm starting over to control the environment again and use what I've learned so far and follow the advice I've been given.

Recent Results

Here are some results on these Asus P8Z68-V LE motherboards that may be helpful to other owners of the board having problems. This is all 32 bit stuff, no Slack64 yet. xfce is the desktop environment. glxgears is used to measure the frames per second. I don't believe I've ever gotten the intel graphics hardware acceleration to function.

Added "acpi=off noapic" boot parameters and upgrade to current Asus BIOS 3702 to get things started.
http://www.linuxquestions.org/questi...2-bios-940376/

With NO xorg.conf file ...

60 fps with no added boot parameters
190 fps when "acpi=off noapic" boot parameters are used and lilo has vga=0x37F, 773, and others (Tux at top of console screen)
800 fps when "acpi=off noapic" boot parameters are used and lilo has vga=normal or missing. (No Tux at top of console screen)

I got the same fps results as above with xorg.conf using a vesa driver.

When xorg.conf uses the intel driver X fails to run and gives the following messages on the screen.

Code:

(==) Log file: "/var/log/Xorg.0.log", Time: Sat Apr 21 00:09:46 2012

(==) Using config file: "/etc/X11/xorg.conf"

(==) Using system config directory "/usr/share/X11/xorg.conf.d"

WARNING: Error inserting i2c_algo_bit (/lib/modules/3.2.13-smp/kernel/drivers/i2c/algos/i2c-algo-bit.ko): No such device

WARNING: Error inserting intel_agp (/lib/modules/3.2.13-smp/kernel/drivers/char/agp/intel-agp.ko): No such device

WARNING: Error inserting drm (/lib/modules/3.2.13-smp/kernel/drivers/gpu/drm/drm.ko): No such device

WARNING: Error inserting drm_kms_helper (/lib/modules/3.2.13-smp/kernel/drivers/gpu/drm/drm_kms_helper.ko): No such device

FATAL: Error inserting i915 (/lib/modules/3.2.13-smp/kernel/drivers/gpu/drm/i915/i915.ko): No such device

(EE) No devices detected.



Fatal server error:

no screens found

Xorg.0.log shows

Code:

...

...

[ 18990.982] (II) LoadModule: "intel"

[ 18990.982] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so

[ 18990.982] (II) Module intel: vendor="X.Org Foundation"

[ 18990.982]         compiled for 1.9.5, module version = 2.15.0

[ 18990.982]         Module class: X.Org Video Driver

[ 18990.982]         ABI class: X.Org Video Driver, version 8.0

[ 18990.982] (II) intel: Driver for Intel Integrated Graphics Chipsets: i810,

        i810-dc100, i810e, i815, i830M, 845G, 854, 852GM/855GM, 865G, 915G,

        E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM, Pineview G,

        965G, G35, 965Q, 946GZ, 965GM, 965GME/GLE, G33, Q35, Q33, GM45,

        4 Series, G45/G43, Q45/Q43, G41, B43, B43, Clarkdale, Arrandale,

        Sandybridge, Sandybridge, Sandybridge, Sandybridge, Sandybridge,

        Sandybridge, Sandybridge

[ 18990.982] (--) using VT number 7



[ 18991.046] (EE) No devices detected.

[ 18991.046] 

Fatal server error:

This is probably just the result of a misbehaving intel driver but I wanted to post it for others that may have the same issue.

These Xorg errors went away with the 3.3.2 kernel (and perhaps earlier kernels) and X ran, but the intel graphics driver never worked.

lsmod in an early trial showed "i915" loaded but never in later trials. (I probably got it to successfully load manually at the time perhaps with the 3506 BIOS.)

That same early "i915" trial showed a glxinfo lines of:

Code:

direct rendering: Yes

OpenGL render string: "Mesa DRI Intel(R) Sandybridge Desktop x86/MMX/SSE2"

Without the i915 loaded in all recent trials glxinfo shows:

Code:

direct rendering: Yes

OpenGL renderer string: Software Rasterizer

I assume no hardware available to use so it's done in sofware?

I'm guessing I'm having basic problems and at this point need not worry about more refined details such as using the Gallium3D option when building mesa.

I noticed these other recent related threads on this site which may be helpful to someone (me).
http://www.linuxquestions.org/questi...s-card-935776/
http://www.linuxquestions.org/questi...driver-938020/

Questions

Right now I'm NOT looking for help with my specific set of software as I've changed too many variables. Others have already given me advice regarding software versions. I'm going to start over and once again carefully control the changes made for each trial.

With a vanilla build I will follow the suggestions of TobiSGD, dugan, ReaperX7, and others. I've probably blown right past the correct combination of software to make things work correctly. For example I ran huge-smp instead of generic-smp for about half of the trials. I become curious and get carried away exploring (that's what make computers fun). Though I do carefully control the trial variables (for a while).

I would like assistance to increase my knowledge about the Linux modules environment/tools so I can improve my basic Linux driver/module troubleshooting technique and eventually solve my current intel driver issue. I hope someone can respond to the following LinuxQuestions.org I have.

lspci shows:

Code:

00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])

        Subsystem: ASUSTeK Computer Inc. Device 844d

        Flags: bus master, fast devsel, latency 0, IRQ 11

        Memory at f7800000 (64-bit, non-prefetchable) [size=4M]

        Memory at e0000000 (64-bit, prefetchable) [size=256M]

        I/O ports at f000 [size=64]

        Expansion ROM at <unassigned> [disabled]

        Capabilities: [90] MSI: Enable- Count=1/1 Maskable- 64bit-

        Capabilities: [d0] Power Management version 2

        Capabilities: [a4] PCI Advanced Features

        Kernel modules: i915

Other hardware (usb, pci, etc) displays the Kernel modules AND the "Kernel driver in use:"

Is this lspci command with the missing kernel driver in use value a reliable indication that the driver is NOT loaded and functioning on a running system? If this value was present is it a reliable indication that the driver IS loaded and (perhaps) functioning?
Is "i915" module name wrong or does a correctly functioning system with Intel Sanydbridge HD (i5-2400) hardware always show i915? Might one see this value change when a proper functioning driver is later loaded and running or is this static information from system startup?
Are there /proc files, that when viewed, do a better job than the lspci interpretation of determining relevant information about the hardware?
If the intel HD graphics driver was built as a module and was functioning properly on my motherboards, what values would I see in an lsmod listing, lspci listing (last two lines)? What other indications should I look for to verify either it's loaded or it's loaded and functioning properly?
When a driver is compiled into the kernel instead of built as a module, is there a straight forward way to know some basic info (name,date,version) of that driver that's in the currently running kernel. ... as opposed to the driver that was intended to be in the kernel? (configured vs built vs loaded/running, Intentions don't always equal fact.) For example are there some /proc files that can be inspected to see some basic info about a specific driver in the currently running kernel?
For troubleshooting purposes could I just load a basic bare bones intel intel driver without any 3D acceleration or most other options? Perhaps scale down the driver with compile-time or run-time options? Then gradually add features with their required libraries.
Would someone remind me where modules get recorded (after a manual modprobe) so they will automatically load at next reboot. I'm also looking for result of doing a modprobe command when no error is returned and the module doesn't show up in lsmod.
What other tools/commands should I be using for this basic level of troubleshooting?

I believe my (huge) knowledge gaps are resulting in my poor troubleshooting technique of even basic Linux stuff. That's what happened when I stopped experimenting with Linux 10 years ago and just became a Linux user to get work done.

-------

At first, glxgears is in no way a reliable benchmark, the numbers you got say exactly nothing.

To your questions:
1. lspci is reliable: If a kernel module is mentioned it is loaded and used.
2. The driver name is i915 and is not related to the actual hardware.
3. lspci and /proc give information about different parts of hardware. As the name states, lspci gives only information about PCI(e) devices.
4. If the driver is loaded you should see the i915 module with lsmod/lspci, since this is the driver for that device.
5. ???
6. You realize that the kernel module is not the only part of the driver? You also need a Mesa/libdrm version that supports 3D acceleration for your device.
7. You don'z need a "record" for that, if the module dependencies are correct (man depmod) the driver will be loaded automatically.
8. Except /var/log/Xorg.0.log I can't think of any, except glxinfo.

Having said that, since you have used -current, what are your results in -current without xorg.conf (meaning the output of the renderer string of glxinfo and the framerate of a real 3D application)? In current your hardware should run out of the box with working 3D acceleration.

I didn't realize that the same driver name i915 was used even with the newer intel HD hardware. I was incorrectly thinking that out of date software was incorrectly identifying the hardware and wanted to use the wrong driver.

Quote:

Originally Posted by TobiSGD (Post 4660246)

Having said that, since you have used -current, what are your results in -current without xorg.conf (meaning the output of the renderer string of glxinfo and the framerate of a real 3D application)? In current your hardware should run out of the box with working 3D acceleration.

I'll install a game that shows frame rate and see what it shows. If there was a simple to use program that's commonly used for 3D testing purposes I would like to have a pointer to it.

As I mentioned my systems have had a lot of changes/experiments. I'll be starting again from scratch and will build one with a vanilla slack-current release. I'll report the results and will only make changes to it that you suggest.

I don't want to waste your time by my giving you possibly inaccurate information about my experience with -current. I'll report after the -current build.

Thanks for your multiple responses.

Quote:

Originally Posted by Tracy Tiger (Post 4660273)

If there was a simple to use program that's commonly used for 3D testing purposes I would like to have a pointer to it.

Try this benchmark: http://unigine.com/products/heaven/download/

Quote:

Originally Posted by TobiSGD (Post 4660246)

I've reinstalled -current (Apr 23rd version) using a 13.37 DVD to boot/run setup and the -current files from osuosl.org. I've used all default values except I selected Xfce instead of KDE and use the boot parameter "acpi=off" so the 3.2.13-smp kernel will run. I'm using the default huge-smp kernel. No xorg.conf file used.

I don't believe the driver is being loaded and am looking for further suggestions.

Details below.

--------------------

lspci shows in part:

Code:

00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])

        Subsystem: ASUSTeK Computer Inc. Device 844d

        Flags: bus master, fast devsel, latency 0, IRQ 11

        Memory at f7800000 (64-bit, non-prefetchable) [size=4M]

        Memory at e0000000 (64-bit, prefetchable) [size=256M]

        I/O ports at f000 [size=64]

        Expansion ROM at <unassigned> [disabled]

        Capabilities: [90] MSI: Enable- Count=1/1 Maskable- 64bit-

        Capabilities: [d0] Power Management version 2

        Capabilities: [a4] PCI Advanced Features

        Kernel modules: i915

Note that there is no "Kernel driver in use: " row displayed.

lsmod shows:

Code:

Module                  Size  Used by

snd_seq_dummy           1119  0

snd_seq_oss            24956  0

snd_seq_midi_event      4568  1 snd_seq_oss

snd_seq                41599  5 snd_seq_dummy,snd_seq_oss,snd_seq_midi_event

snd_seq_device          4276  3 snd_seq_dummy,snd_seq

_oss,snd_seq

snd_pcm_oss            33254  0

snd_mixer_oss          12984  1 snd_pcm_oss

ipv6                  238149  18

lp                      7119  0

ppdev                   4710  0

parport_pc             16755  0

parport                26159  3 lp,ppdev,parport_pc

fuse                   58282  1

snd_hda_codec_hdmi     21753  1

snd_hda_codec_realtek   150144  1

snd_hda_intel          19471  0

snd_hda_codec          70505  3 snd_hda_codec_hdmi,snd_hda_codec_realtek,snd_hda_intel

i2c_i801                6952  0

snd_hwdep               4802  1 snd_hda_codec

snd_pcm                59045  4 snd_pcm_oss,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec

i2c_core               16454  1 i2c_i801

snd_timer              14798  2 snd_seq,snd_pcm

r8169                  42084  0

usbhid                 31118  0

sg                     20942  0

snd                    43657  12 snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_hda_codec_hdmi,snd_hda_codec_realtek,snd_hda_intel,snd_hda_co

dec,snd_hwdep,snd_pcm,snd_timer

psmouse                54344  0

soundcore               4570  1 snd

snd_page_alloc          5925  2 snd_hda_intel,snd_pcm

intel_agp               8680  0

hid                    65632  1 usb

intel_gtt              11571  2 intel_agp

agpgart                22087  2 intel_agp,intel_gtt

video                   9938  0

thermal_sys            12122  1 video

hwmon                   1033  1 thermal_sys

mii                     3335  1 r8169

mei                    27885  0

serio_raw               3349  0

evdev                   7314  6

loop                   14650  0

Note that there is no i915 module loaded. (built into huge kernel or missing?)

glxinfo shows:

Code:

name of display: :0.0

display: :0  screen: 0

direct rendering: Yes

server glx vendor string: SGI

server glx version string: 1.4

server glx extensions:

    GLX_ARB_multisample, GLX_EXT_visual_info, GLX_EXT_visual_rating,

    GLX_EXT_import_context, GLX_EXT_texture_from_pixmap, GLX_OML_swap_method,

    GLX_SGI_make_current_read, GLX_SGIS_multisample, GLX_SGIX_hyperpipe,

    GLX_SGIX_swap_barrier, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,

    GLX_MESA_copy_sub_buffer, GLX_INTEL_swap_event

client glx vendor string: Mesa Project and SGI

client glx version string: 1.4

client glx extensions:

name of display: :0.0

display: :0  screen: 0

direct rendering: Yes

server glx vendor string: SGI

server glx version string: 1.4

server glx extensions:

    GLX_ARB_multisample, GLX_EXT_visual_info, GLX_EXT_visual_rating,

    GLX_EXT_import_context, GLX_EXT_texture_from_pixmap, GLX_OML_swap_method,

    GLX_SGI_make_current_read, GLX_SGIS_multisample, GLX_SGIX_hyperpipe,

    GLX_SGIX_swap_barrier, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,

    GLX_MESA_copy_sub_buffer, GLX_INTEL_swap_event

client glx vendor string: Mesa Project and SGI

client glx version string: 1.4

client glx extensions:

    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,

    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_copy_sub_buffer,

    GLX_MESA_swap_control, GLX_OML_swap_method, GLX_OML_sync_control,

    GLX_SGI_make_current_read, GLX_SGI_swap_control, GLX_SGI_video_sync,

    GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,

    GLX_SGIX_visual_select_group, GLX_EXT_texture_from_pixmap,

    GLX_INTEL_swap_event

GLX version: 1.4

GLX extensions:

    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,

    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_OML_swap_method,

    GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,

    GLX_INTEL_swap_event

OpenGL vendor string: Mesa Project

OpenGL renderer string: Software Rasterizer

OpenGL version string: 2.1 Mesa 7.10.2

OpenGL shading language version string: 1.20

OpenGL extensions:



 ... lines removed ...

Note the "OpenGL renderer string: Software Rasterizer" line.

HEAVEN

I installed the Unigine Heaven benchmark. It installed and came up with the initial screen. When "RUN" is selected the terminal it was started from displays in part:

Code:

 .... lines removed ...



---- System ----

System: Linux 3.2.13-smp i686

CPU: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz 3193MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT x4

GPU: Software Rasterizer 2.1 Mesa 7.10.2 x1

System memory: 15935 Mb

Video memory:  256 Mb

Sync threads:  3

Async threads: 4



---- MathLib ----

Set SSE2 simd processor



---- Sound ----

NULL



---- Render ----

GLRender::GLRender(): Unknown GPU

OpenGL vendor:   Mesa Project

OpenGL renderer: Software Rasterizer

OpenGL version:  2.1 Mesa 7.10.2

Found required GL_ARB_vertex_array_object



 ... lines removed ...



Loading "core/materials/default/unigine_decal.mat" 1 material 99 shaders 1ms

Loading "core/properties/unigine.prop" 2 properties 0ms

OpenGL error: invalid enum

Note the "GLRender::GLRender(): Unknown GPU" Line.

No Sound because the OpenAL software wasn't installed for this trial. I got the same results days earlier with OpenAL loaded and using a generic-smp kernel.

The Unigine web site shows ATI and NVIDIA graphic card requirements but doesn't mention Intel. Perhaps this software doesn't support (the newer) Intel graphic chips?

EXTREME TUX RACER

I installed this game from Slackbuilds.org and got around 4 frames per second.

------------------

I ran Xfce with an xorg.conf file using the supplied vesa config file. When I changed the video driver to "intel" X failed as it had in the past.

Xorg.0.log Excerpt:

Code:

[ 38035.952] (II) LoadModule: "intel"

[ 38035.952] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so

[ 38035.952] (II) Module intel: vendor="X.Org Foundation"

[ 38035.952]         compiled for 1.9.5, module version = 2.15.0

[ 38035.952]         Module class: X.Org Video Driver

[ 38035.952]         ABI class: X.Org Video Driver, version 8.0

[ 38035.952] (II) intel: Driver for Intel Integrated Graphics Chipsets: i810,

        i810-dc100, i810e, i815, i830M, 845G, 854, 852GM/855GM, 865G, 915G,

        E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM, Pineview G,

        965G, G35, 965Q, 946GZ, 965GM, 965GME/GLE, G33, Q35, Q33, GM45,

        4 Series, G45/G43, Q45/Q43, G41, B43, B43, Clarkdale, Arrandale,

        Sandybridge, Sandybridge, Sandybridge, Sandybridge, Sandybridge,

        Sandybridge, Sandybridge

[ 38035.952] (--) using VT number 7



[ 38035.960] (EE) No devices detected.

[ 38035.960] 

Fatal server error:

[ 38035.960] no screens found

Earlier, on another machine with the same motherboard version, I installed the Apr 23rd version of Slack current with a generic kernel and got the same results. I can try the generic kernel again if that would help identify the driver module that should be used.

Suggestions are welcome.

That is weird, the system should automatically load the driver (what happens when you modprobe it manually before starting X?), also the Unigine website mentions that the HD3000 from Intel is supported.
May be this whole thing is because the acpi=off parameter, is that really needed?

Quote:

Originally Posted by TobiSGD (Post 4669667)

That is weird, the system should automatically load the driver (what happens when you modprobe it manually before starting X?)

I just now ran modprobe i915 and received the following error messages. If "i915" is not the correct name please tell me.

Code:

WARNING: Error inserting i2c_algo_bit (/lib/modules/3.2.13-smp/kernel/drivers/i2c/algos/i2c-algo-bit.ko): No such device

WARNING: Error inserting intel_agp (/lib/modules/3.2.13-smp/kernel/drivers/char/agp/intel-agp.ko): No such device

WARNING: Error inserting drm (/lib/modules/3.2.13-smp/kernel/drivers/gpu/drm/drm.ko): No such device

WARNING: Error inserting drm_kms_helper (/lib/modules/3.2.13-smp/kernel/drivers/gpu/drm/drm_kms_helper.ko): No such device

FATAL: Error inserting i915 (/lib/modules/3.2.13-smp/kernel/drivers/gpu/drm/i915/i915.ko): No such device

I'd seen sequences of errors beginning with this "i2c" error many times from X when attempting to use the intel driver in xorg.conf. Probably for the same reason.

Remember I'm running with the default huge-smp kernel. I can create an initrd and switch to the generic kernel if that helps narrow down the problem. I'm making changes judiciously so as not to complicate your efforts to assist me.

Quote:

Originally Posted by TobiSGD (Post 4669667)

May be this whole thing is because the acpi=off parameter, is that really needed?

I had also considered this. The kernel hangs on boot unless this parameter is used. I'm wondering if it needs to be "off" for the kernel to run and "on" for the driver to work. :(

In one of my early trials I got the i915 module loaded somehow on one of the three machines. If I remember correctly I believe the kernel (2.6? 3.x?) would run with the 3506 BIOS without needing to turn "off" acpi.

When I got the i915 module loaded early in the trials there had not been may changes yet. That configuration has been replaced with a new installation so it no longer exists.

As I noted in my original post ... glxinfo shows "direct rendering: Yes" and an OpenGL render string of "Mesa DRI Intel(R) Sandybridge Desktop x86/MMX/SSE2".

So I probably had at least partial success and didn't know it. That was before you and T3Slider opened my eyes to the fallacy of using the FPS rate in glxgears to determine if hardware acceleration was functioning.