new Intel driver is crashing X

Darth Vader · 04-20-2011, 05:59 PM

Quote:

Originally Posted by rfernandez

It was a test, not a question of stability. A comparison, should you wish it.

The video card is the one from Intel i5-460M, I never remember the name. As for ~/.xsession-errors, I'll test it again, later.

These tweaks refer to ...

/etc/kde/kdm/kdmrc : X crash like a happy kid in the past. This just launch a new'n'clean X session.

/usr/bin/startkde : In some G31/G33/G41, KWin have right direct rendering support but it don't see right. Brought to you by Mesa Devs!

BUT, this tweak bypass the normal GL detection. Use with care. Is something like: "I known for sure that should work there, but this loser (KWin) don't see it."

AND IS TESTED ON G31/G33/G41 ONLY.

Finally, your ~/.xsession-errors have now the first interesting information, IF you deactivated the tweaks ...

andrew.46 · 04-21-2011, 02:32 AM

Quote:

Originally Posted by volkerdi

After testing nearly every permutation of libdrm/mesa/xf86-video-intel, I have seen crashes with every combination. The very latest stuff seemed like it might have a slight stability edge, but the crashes are somewhat random so it is hard to tell.

Well, I am a little relieved as I was feeling like a bit of an idiot with my own claims of graphics instability

. My own system has been rock solid with the 2.14 driver so that is where I will be staying. BTW thanks for the upcoming release of 13.37, it looks like a great release and I get a sense of the huge amount of work that has gone into it!

audriusk · 04-21-2011, 11:14 AM

Here's my report: GM965/GL960 chipset, currently using 2.6.38.3 kernel with intel-2.15.0, libdrm-2.4.25 and mesa-7.10.2. I also have Lanczos filter blacklisted in ~/.kde/share/config/kwinrc:

Code:

[Blacklist][Lanczos]
Intel=965GM:-:7.10.2

No crashes so far, though I had one with 2.6.38.2 kernel, which felt "not right" (occasional strange flickering) in general. No crashes at all with previous intel, libdrm and mesa combinations, but I was using 2.6.37.x kernel then.

brixtoncalling · 04-22-2011, 04:43 PM

Is 2.14 available somewhere? I need to drop to it down from 2.15 because my system is just too unstable with the new driver -- even after trying all DarthVader's tricks and all the latest kernels.

guanx · 04-22-2011, 04:56 PM

Quote:

Originally Posted by brixtoncalling

Is 2.14 available somewhere? I need to drop to it down from 2.15 because my system is just too unstable with the new driver -- even after trying all DarthVader's tricks and all the latest kernels.

In the pasture/ directory.

andrew.46 · 04-22-2011, 04:57 PM

Quote:

Originally Posted by brixtoncalling

Is 2.14 available somewhere? I need to drop to it down from 2.15 because my system is just too unstable with the new driver -- even after trying all DarthVader's tricks and all the latest kernels.

64bit is here:
http://slackware.osuosl.org/slackwar...rrent/pasture/

32bit is here:
http://slackware.osuosl.org/slackware-current/pasture/

Darth Vader · 04-22-2011, 04:58 PM

Quote:

Originally Posted by brixtoncalling

Is 2.14 available somewhere? I need to drop to it down from 2.15 because my system is just too unstable with the new driver -- even after trying all DarthVader's tricks and all the latest kernels.

It's in /pasture.

But, if you like, I will love to resolve this problem. Don't be proud!

I have to keep alive 17 Intels. Call me paranoid, but I want/need to known how they crash, and what I have to do ...

brixtoncalling · 04-22-2011, 05:10 PM

Huh, the mirror I was looking at (AlienBob's) doesn't include pasture: http://taper.alienbase.nl/mirrors/sl...are64-current/

I thought that maybe Pat didn't include pasture until the release or something...

Thanks guys.

PS Darth Vader, please solve my problem

Darth Vader · 04-22-2011, 05:37 PM

@brixtoncalling This is perfect. But to not be disappointed, I will try to find the most stable solution, not the fastest.

First, I need to know which chipset model we discuss.

Secondly, please describe a pattern of crash. What do you usually do when the system cracks?

Thirdly, you have to wait until tomorrow. Can we talk after 1 PM, EEST.

Until then, prepare as much information as:

/var/log/dmesg, /var/log/messages, /var/log/syslog, /var/log/Xorg.0.log and what lspci say related to VGA. Use the LQ mail-to-me, for this data, if there is too much data to post it.

brixtoncalling · 04-22-2011, 06:03 PM

Your help is much appreciated. I'm going to put the information here since I guess other people have similar problems. There's no hurry by the way. I have the 2.14 now, so at least my system is stable again! (Xorg.0.log is attached)
Thanks.

lspci:

Code:

VGA compatible controller: Intel Corporation Core Processor Integrated Graphics Controller (rev 02)

dmesg (just the important stuff):

Code:

[    1.230608] vgaarb: device added: PCI:0000:00:02.0,decodes=io+mem,owns=io+mem,locks=none
[    1.230757] vgaarb: loaded
0xffffffff
[    8.622597] intel ips 0000:00:1f.6: CPU TDP doesn't match expected value (found 25, expected 29)
[    8.622785] intel ips 0000:00:1f.6: PCI INT A -> GSI 21 (level, low) -> IRQ 21
[    8.623051] intel ips 0000:00:1f.6: failed to get i915 symbols, graphics turbo disabled
[    8.623316] intel ips 0000:00:1f.6: IPS driver initialized, MCP temp limit 90
[   11.317083] lib80211: common routines for IEEE802.11 drivers
[   11.450965] agpgart-intel 0000:00:00.0: Intel HD Graphics Chipset
[   11.451244] agpgart-intel 0000:00:00.0: detected gtt size: 524288K total, 262144K mappable
[   11.452161] agpgart-intel 0000:00:00.0: detected 131072K stolen memory, trimming to 32768K
[   11.476821] agpgart-intel 0000:00:00.0: AGP aperture is 256M @ 0xa0000000
[   11.522185] [drm] Initialized drm 1.1.0 20060810
[   11.784450] [drm] MTRR allocation failed.  Graphics performance may suffer.
[   11.851671] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem
[   11.960167] fb0: inteldrmfb frame buffer device
[   12.273002] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[   13.820821] intel ips 0000:00:1f.6: i915 driver attached, reenabling gpu turbo
[ 1388.002335] intel ips 0000:00:1f.6: MCP limit exceeded: Avg power 36235, limit 35000
[ 1732.797073] intel ips 0000:00:1f.6: MCP limit exceeded: Avg power 36248, limit 35000
[ 2207.517550] intel ips 0000:00:1f.6: MCP limit exceeded: Avg power 37179, limit 35000
[ 2342.437287] intel ips 0000:00:1f.6: MCP limit exceeded: Avg power 37366, limit 35000
[ 2632.264689] intel ips 0000:00:1f.6: MCP limit exceeded: Avg power 35145, limit 35000
[ 3821.556788] intel ips 0000:00:1f.6: MCP limit exceeded: Avg power 35009, limit 35000

Darth Vader · 04-24-2011, 09:33 AM

brixtoncalling,

First, what is wrong with your MTTR support? You have BIOS problems or an incorrect BIOS setup?

Reason:

Code:

[   11.784450] [drm] MTRR allocation failed.  Graphics performance may suffer.

Secondly, looks like your PageFlip-ing don't works properly. I recommend you to disable it, using:

Code:

Option        "EnablePageFlip" "false"

Reason:

Code:

[  9172.840] (WW) intel(0): I830DRI2FlipEventHandler: Pageflip completion has impossible msc 546932 < target_msc 546933
[  9172.857] (WW) intel(0): I830DRI2FlipEventHandler: Pageflip completion has impossible msc 546933 < target_msc 546934
[  9172.907] (WW) intel(0): I830DRI2FlipEventHandler: Pageflip completion has impossible msc 546936 < target_msc 546937
[  9172.924] (WW) intel(0): I830DRI2FlipEventHandler: Pageflip completion has impossible msc 546937 < target_msc 546938
[  9172.941] (WW) intel(0): I830DRI2FlipEventHandler: Pageflip completion has impossible msc 546938 < target_msc 546939

brixtoncalling · 04-24-2011, 11:02 AM

I don't know what causes the MTRR errors. I tried putting enable_mtrr_cleanup mtrr_spare_reg_nr=1 in my lilo.conf but it didn't solve the problem. I don't think there is anything in my BIOS that needs to be changed. Do you have any other ideas?

I tried Option "EnablePageFlip" "false" but get this in my Xorg.0.log:

Code:

(WW) intel(0): Option "EnablePageFlip" is not use

Darth Vader · 04-25-2011, 01:28 PM

Quote:

Originally Posted by brixtoncalling

I don't know what causes the MTRR errors. I tried putting enable_mtrr_cleanup mtrr_spare_reg_nr=1 in my lilo.conf but it didn't solve the problem. I don't think there is anything in my BIOS that needs to be changed. Do you have any other ideas?

I tried Option "EnablePageFlip" "false" but get this in my Xorg.0.log:

Code:

(WW) intel(0): Option "EnablePageFlip" is not use

First we will focus on the problem of MTRR.

I need full dump of dmesg and 'lspci-vvv'.

Second ... Sorry, my bad! This was the Radeon variant ...

Try this:

Code:

Option "PageFlip" "off"

brixtoncalling · 04-26-2011, 08:50 AM

Quote:

Originally Posted by Darth Vader

First we will focus on the problem of MTRR.

I need full dump of dmesg and 'lspci-vvv'.

Second ... Sorry, my bad! This was the Radeon variant ...

Try this:

Code:

Option "PageFlip" "off"

"PageFlip" is also not used

I attached my lscpi and dmesg (and my mtrr layout) ... Thanks again for taking the time to look at these.

Darth Vader · 04-26-2011, 10:25 AM

@brixtoncalling

Well, run this script and paste back it's output.

http://ftp.darkstarlinux.ro/people/darth/fixmtrr.sh