xf86-video-intel 2.7.1 now they've done it

combuster · 2009-05-13 11:19:03

2.7.1 released. Yay...

pacman -R xf86-video-intel

download
./configure
make
make install

reboot

Black screen with white cursor on the top right edge of the screen, typing username and password on blind, hd activity, gnome started, picture stays the same, white cursor, black screen. KMS is enabled...

reboot, boot with my 2.6.30-rc4 custom kernel, same problem...

reboot with live arch usb, mount /dev/sda1, disable gdm in rc.conf, restart, make uninstall, pacman -S xf86-video-intel, reboot, Yay!!!

Disaster, I just wonder if anybody else have the same problem...

combuster · 2009-05-13 11:41:18

Oops sorry, now I've noticed that default instalation path is /usr/local... gonna recompile with --prefix=/usr

combuster · 2009-05-13 11:50:17

It works, sorry but I'm so frustrated with intel drivers that first idea that cross my mind is that it's another bug and not my mistake... I hope that random freezing is fixed as they said it would with this driver version. I expect it soon in extra...

Army · 2009-05-13 12:36:24

You should have simply changed the PKGBUILD via abs

eldragon · 2009-05-13 12:58:34

Army wrote:

You should have simply changed the PKGBUILD via abs

changing the PKGBULD and removing the patching section (apparently patches are already applied to 2.7.1) builds. i just cannot find the change log runs like 2.7.0 here... lets see how stable it is..

EDIT: i found the changelog (OHH SURPRISE!!) within the source code.

EDIT2: and here it is:

Bug fixes since 2.7.0:

    * KMS: Hook up output properties for RANDR, (this allows output
      properties to be controlled in the KMS case just as in the UMS
      case). [Zhenyu Wang <zhenyu.z.wang@intel.com>]

    * Fix multiplication error when computing required batch space.
      This could fix any number of cases where the driver did
      inexplicable things (due to having computed the wrong
      size). [Keith Packard <keithp@keithp.com>]

    * Hold reference to video binding table until all rects are
      painted. This prevent general chaos in the buffer
      manager. [Keith Packard <keithp@keithp.com>]

    * Split i915 textured video commands to fit into batch
      buffers. Video and 3D setup commands share the same batch
      buffer, so without this fix, various problems could occur when
      video and 3D clients were both heavily active at the same
      time. [Keith Packard <keithp@keithp.com>]

    * Fix crash with XV with large virtual display (> 2049). [Albert
      Damen <albrt@gmx.net>]

    * Provide missing value to 3D_STATE_VERTEX_BUFFERS command. We
      don't know that this was causing any problem, but the change
      does bring the driver into conformance with what the
      specification says the hardware requires here. [Keith Packard
      <keithp@keithp.com>]

considering the ammount of patch complaining there was, i believe there are 0 changes since 2.7.0 from the repos.

Last edited by eldragon (2009-05-13 13:03:04)

combuster · 2009-05-13 13:11:32

For now no freezes at least. Same crappy performance but they announced that the performance will stay the same and that these are only bugfixes... Still slow @ repainting message boxes...

mighelone · 2009-05-13 13:13:59

Which problems do you have with intel driver?

With the last update, the X server freezes periodically after 1 or 2 minutes of work, and I need to switch from X server to console using CTRL-ALT-F1 to work again.

I hope to solve this problem with 2.7.1!!!!

zenlord · 2009-05-13 14:40:08

I think it is very hard for everyone developping in the area of graphics (drivers, subsystems etc.) because of all the new stuff that is being incorporated as we speak. Everyone is developping towards a common goal, but somewhere along the way breakage has to happen due to the massive changes (XAA, EXA, UXA - hal - KMS - memory management - randr - etc.)

I believe this is all worth it and trust that the packages in core and extra are functional enough - it's your choice to for testing, so you should be expecting things to break...

fphillips · 2009-05-13 14:48:15

It's true - 2.7.0-3 = 2.7.1 minus some documentation changes. Just look at the patch it contains in svn.

ArchArael · 2009-05-13 14:54:55

mighelone wrote:

Which problems do you have with intel driver?
With the last update, the X server freezes periodically after 1 or 2 minutes of work, and I need to switch from X server to console using CTRL-ALT-F1 to work again.
I hope to solve this problem with 2.7.1!!!!

I had a similar problem. I couldn't work at all with such instable machine.

I was so frustrated that I gave up and installed Slackware 12.2.

But Slackware is not my distro, although it is a great one, so I after a couple of days I have installed arch64.

No freezes so far. I hope it will continue this way. No idea why arch64 is more stable, though.

Last edited by ArchArael (2009-05-13 14:56:22)

Wilco · 2009-05-13 15:38:59

In my case it really helped to enable KMS. With 2.6.x KMS was buggy as hell but now I have some glitches without KMS, so maybe enabling KMS can help you?

tomd123 · 2009-05-13 15:45:44

This maybe off topic, but you know how you did pacman -R on the last intel driver? That make install will come back to haunt you

combuster · 2009-05-13 18:12:43

Why? I didn't understood you correctly...

madalu · 2009-05-13 18:33:25

Yes, this is very disappointing. The mailing list announcement states the following:

PS. The 2.7.1 release is equivalent (other than the version number) to
the 2.7.0.901 release candidate announced earlier.

It also states that:

We have verified that several of the reported bugs of GPU crashes,
(mouse continues to move, but otherwise X is totally unresponsive), are
fixed with the commit by Keith Packard in 2.7.1 to correct the
computation of the batch space required.

But I have been using 2.7.0.901 and have still been getting all sorts of crashes. Once I get to kernel 2.6.30 I can use the intel-gpu-tools described in the annoucement to report the bug:

If the crash persists, please attach the output of intel_gpu_dump
available here (and hopefully packaged in your distribution of choice
soon):
git://anongit.freedesktop.org/git/xorg/app/intel-gpu-tools
(Requires Linux kernel >= 2.6.30)

http://article.gmane.org/gmane.comp.fre … nounce/624

olovram · 2009-05-14 00:30:04

Will we be able to change screen resolution now?

combuster · 2009-05-14 05:23:52

I dont think so if you use kms, those apps that run in resolution less than your native one still run windowed in full screen...

Damnshock · 2009-05-14 10:27:16

fphillips wrote:

It's true - 2.7.0-3 = 2.7.1 minus some documentation changes. Just look at the patch it contains in svn.

Not exactly true: there's been a *huge* unnecessary code deletion.

eldragon · 2009-05-14 11:42:38

Damnshock wrote:

fphillips wrote:
It's true - 2.7.0-3 = 2.7.1 minus some documentation changes. Just look at the patch it contains in svn.
Not exactly true: there's been a *huge* unnecessary code deletion.

thats a good thing...yet performance shouldnt change (and this is the case here).

peoro · 2009-05-14 12:53:06

Freezes are still occurring to me with 2.7.1 drivers...
This issue is too annoying, I'm going to try and downgrade to xf86-video-intel-legacy...

combuster · 2009-05-14 13:06:01

I havent had freezes with 2.7.1. Furthermore with 2.7.0 sometimes my laptop screen gets blank all of a sudden and then back to normal... I've tripped out that maybie the cable is damaged but it isn't, so my guess that driver caused it...

ataraxia · 2009-05-14 13:31:54

combuster wrote:

Furthermore with 2.7.0 sometimes my laptop screen gets blank all of a sudden and then back to normal... I've tripped out that maybie the cable is damaged but it isn't, so my guess that driver caused it...

I get this behavior under totally different drivers (radeon, fglrx). I wonder if it's actually a bug in the server itself.

eldragon · 2009-05-14 13:38:52

ataraxia wrote:

combuster wrote:
Furthermore with 2.7.0 sometimes my laptop screen gets blank all of a sudden and then back to normal... I've tripped out that maybie the cable is damaged but it isn't, so my guess that driver caused it...
I get this behavior under totally different drivers (radeon, fglrx). I wonder if it's actually a bug in the server itself.

does it have something to do with this?

ataraxia · 2009-05-14 13:53:19

eldragon wrote:

ataraxia wrote:
combuster wrote:
Furthermore with 2.7.0 sometimes my laptop screen gets blank all of a sudden and then back to normal... I've tripped out that maybie the cable is damaged but it isn't, so my guess that driver caused it...
I get this behavior under totally different drivers (radeon, fglrx). I wonder if it's actually a bug in the server itself.
does it have something to do with this?

I don't think so. I don't run Gnome, and my DPMS is set to "Standby: 1200 Suspend: 1800 Off: 2400". When it happens, the affected display goes directly to the "off" state rather than to "standby". Even more weirdly, it seems to happen totally at random (4 times in one day, then not at all for several days. And it's only for one of my displays - the other stays on when it should.

eldragon · 2009-05-14 13:59:08

i dont think the display driver has the hability to completely turn off the display.... id check the power cord....

combuster · 2009-05-14 15:07:01

I'm not sure, I run gnome and gnome-power-manager so maybe I'm affected by this issue too..

Arch Linux

#1 2009-05-13 11:19:03

xf86-video-intel 2.7.1 now they've done it

#2 2009-05-13 11:41:18

Re: xf86-video-intel 2.7.1 now they've done it

#3 2009-05-13 11:50:17

Re: xf86-video-intel 2.7.1 now they've done it

#4 2009-05-13 12:36:24

Re: xf86-video-intel 2.7.1 now they've done it

#5 2009-05-13 12:58:34

Re: xf86-video-intel 2.7.1 now they've done it

#6 2009-05-13 13:11:32

Re: xf86-video-intel 2.7.1 now they've done it

#7 2009-05-13 13:13:59

Re: xf86-video-intel 2.7.1 now they've done it

#8 2009-05-13 14:40:08

Re: xf86-video-intel 2.7.1 now they've done it

#9 2009-05-13 14:48:15

Re: xf86-video-intel 2.7.1 now they've done it

#10 2009-05-13 14:54:55

Re: xf86-video-intel 2.7.1 now they've done it

#11 2009-05-13 15:38:59

Re: xf86-video-intel 2.7.1 now they've done it

#12 2009-05-13 15:45:44

Re: xf86-video-intel 2.7.1 now they've done it

#13 2009-05-13 18:12:43

Re: xf86-video-intel 2.7.1 now they've done it

#14 2009-05-13 18:33:25

Re: xf86-video-intel 2.7.1 now they've done it

#15 2009-05-14 00:30:04

Re: xf86-video-intel 2.7.1 now they've done it

#16 2009-05-14 05:23:52

Re: xf86-video-intel 2.7.1 now they've done it

#17 2009-05-14 10:27:16

Re: xf86-video-intel 2.7.1 now they've done it

#18 2009-05-14 11:42:38

Re: xf86-video-intel 2.7.1 now they've done it

#19 2009-05-14 12:53:06

Re: xf86-video-intel 2.7.1 now they've done it

#20 2009-05-14 13:06:01

Re: xf86-video-intel 2.7.1 now they've done it

#21 2009-05-14 13:31:54

Re: xf86-video-intel 2.7.1 now they've done it

#22 2009-05-14 13:38:52

Re: xf86-video-intel 2.7.1 now they've done it

#23 2009-05-14 13:53:19

Re: xf86-video-intel 2.7.1 now they've done it

#24 2009-05-14 13:59:08

Re: xf86-video-intel 2.7.1 now they've done it

#25 2009-05-14 15:07:01

Re: xf86-video-intel 2.7.1 now they've done it

Board footer