xorg-server 1.6.0 moves to extra

Dieter@be · 2009-04-11 19:30:32

jose1711 wrote:

i was a bit hoping that with xorg server upgrade the artefact i'm seeing in opengl apps would go away, but bad luck..
http://img515.imageshack.us/img515/3836/ati.th.png
jose, ati mobility 9000 - ati rv250 and opensource driver (dell latitude d600)

The cause of this is very likely bad memory on your graphics card.

skottish · 2009-04-11 20:39:11

bender02 wrote:

No it wasn't a bad advice. *In this case!* The point skottish is trying to make (I think) is that one should not automatically reply in this way:
-> OP: blablabla the world is collapsing, since i get " ... exists in filesystem. Errors occured, no packages upgraded" from pacman.
-> Yea sure I get it all the time, "pacman -Syuf" solves all your problems and fixes you a girlfriend as well.
In cases when it's advised by the devs (like in this case in the news), it's OK to use the 'f' switch with pacman. Otherwise *not*, and especially *not* if you're slightly new to (arch) linux. And a much worse sin is to reply in the forums like I described above, without giving a reference to the devs, or explaining in detail why is it OK in this particular case.
EDIT: typos... sometimes I get heated up and make mystakes.

Thank you. Yes and no. I'll always post something like this when someone suggests the force switch. In this case I looked at the most recent announcement before saying what I said. Since a developer did say go ahead and do this previously, I jumped the gun a bit. In other words, I should have said that it's alright this time. I was generically correct but incidentally wrong.

jose1711 · 2009-04-11 21:48:54

Dieter@be wrote:

jose1711 wrote:
i was a bit hoping that with xorg server upgrade the artefact i'm seeing in opengl apps would go away, but bad luck..
http://img515.imageshack.us/img515/3836/ati.th.png
jose, ati mobility 9000 - ati rv250 and opensource driver (dell latitude d600)
The cause of this is very likely bad memory on your graphics card.

i somewhat doubt this is a memory problem due to the fact that:

a) fullscreen openGL applications do not have this issue
b) it only manifests itself when i switch a virtual screen and then come back. sometimes the garbage is gone when i run another opengl application (e. g. glxgears). i also do not remember seeing any issues on my windows installation, though i never had windows installed for a long time.

jose

darehanl · 2009-04-12 02:04:44

Wasn't really sure where to post this, but shouldn't the Wiki be updated on the DontZap option?

Section "ServerFlags"
    Option "DontZap" "false"
EndSection

I was reading through for a faster intel xorg.conf, and I see the wiki advising Ctrl-Alt-Del quite frequently when configuring Xorg. Most obviously in "4.3.5 C: Test X" and "4.4 Simple baseline X test(if necessary)" in the Beginners Guide.

<strike>I have my own rants against X.org for this decision, but upstream wins.</strike>

banshee28 · 2009-04-12 02:05:50

Just wanted to say that I updated to this new xorg and did a full sys update including the new kernel26 2.6.29.1-3.

I had to first update xorg server separately then everything worked fine, reboot and all

Edit: maybe its just me but my system feels a good bit more responsive since this upgrade?

Last edited by banshee28 (2009-04-12 02:07:50)

theringmaster · 2009-04-12 04:02:41

banshee28 wrote:

Just wanted to say that I updated to this new xorg and did a full sys update including the new kernel26 2.6.29.1-3.
I had to first update xorg server separately then everything worked fine, reboot and all
Edit: maybe its just me but my system feels a good bit more responsive since this upgrade?

i'm finding that there is some lag when I scroll a page in firefox.

Anyone else have jittery scrolling in firefox since upgrading their xorg-server?

Tri Nguyen · 2009-04-12 04:56:46

Anyone can help. 3D applications works fine before the update. But after the update any OpenGL application crash with those message. I'm using ATI Radeon Xpress 1150 with xf86-video-ati driver. Compositing works fine.

glxgears:

glxgears: main/context.c:941: check_context_limits: Assertion `ctx->Const.MaxTextureCoordUnits <= ctx->Const.MaxTextureImageUnits' failed.
Aborted

glxinfo:

name of display: :0.0
glxinfo: main/context.c:941: check_context_limits: Assertion `ctx->Const.MaxTextureCoordUnits <= ctx->Const.MaxTextureImageUnits' failed.
Aborted

mcmillan · 2009-04-12 05:46:40

I'll always post something like this when someone suggests the force switch. In this case I looked at the most recent announcement before saying what I said. Since a developer did say go ahead and do this previously, I jumped the gun a bit. In other words, I should have said that it's alright this time. I was generically correct but incidentally wrong.

I understand, that's pretty much what I figured you were thinking, and I'll do the same thing when I see advice being given to just blindly pacman -Syuf, I was being a tad snarky in pointing out that this was a case where it was advised to force things. The particular suggestion of remove then install just got me thinking since I tend to back up than force when I'm uncertain, so I was trying to think how those might result in different behavior. My thought is in most cases either should work, though if there's problems just restoring a backup file might cause more problems if there's some kind of version incompatibility.

jiu · 2009-04-12 07:21:45

Hi,
The first post wasn't very clear: who should use xorg-video-intel and who should use xorg-video-intel-legacy? The way I understood it, xorg-video-intel applies only to those who can use kernel modesetting, ie those with an intel i915 chipset. For everyone else, the -legacy driver is the proper choice. Please clarify and correct me if necessary.

dhave · 2009-04-12 08:35:56

Update went fine on my system (nvidia Quadro FX 570M in a Thinkpad T61p). Thanks for a smooth transition. I was a little apprehensive, as I recall a pretty disastrous xorg update on Gentoo a couple of years ago.

I pretty much always find that with Arch, important updates will either work right the first time or there's enough brainpower in the community to get things fixed quickly, which is pretty good considering the number of variables.

bernarcher · 2009-04-12 09:51:04

mcmillan wrote:

... and I'll do the same thing when I see advice being given to just blindly pacman -Syuf ...

Now, if I understand things right, this would be really dangerous, as -Syuf attempts to force install everything (even unnoticed if you like me tend to rather skim long messages lists from pacman).

The advice was however to
- First refresh the databases: pacman -Sy
- Then force install xorg-server (and possible prerequisites) only: pacman -Sf xorg-server
- Finally do a normal system upgrade: pacman -Su

ngoonee · 2009-04-12 13:18:00

Just upgraded, haven't restarted yet (doing something else), but a quick question, after the upgrade I've got 3 'installed as dependencies but are no longer required' packages:-
glproto
libxkbui
xorg-utils

Are they safe to remove? I'm thinking they're not depended on anymore because their functionality no longer requires external libs (the first two)? What about xorg-utils?

VPeric · 2009-04-12 14:07:56

Alright, so I've got a Latitude 110L, which is supposed to have an "910GML" graphics card (as far as I can see/figure out), which is, I believe, supposed to support kernel mode setting, but I've noticed no changes after rebooting with the new xorg-server.

Is there an option I'm supposed to enable, or does the card just not support KMS? Thanks!

theringmaster · 2009-04-12 14:57:43

ngoonee wrote:

Just upgraded, haven't restarted yet (doing something else), but a quick question, after the upgrade I've got 3 'installed as dependencies but are no longer required' packages:-
glproto
libxkbui
xorg-utils
Are they safe to remove? I'm thinking they're not depended on anymore because their functionality no longer requires external libs (the first two)? What about xorg-utils?

keep them, they are still needed. xorg-utils was one of the packages that was upgraged as well.

Arkane · 2009-04-12 16:22:17

jiu wrote:

Hi,
The first post wasn't very clear: who should use xorg-video-intel and who should use xorg-video-intel-legacy? The way I understood it, xorg-video-intel applies only to those who can use kernel modesetting, ie those with an intel i915 chipset. For everyone else, the -legacy driver is the proper choice. Please clarify and correct me if necessary.

If I understood it correctly, these drivers both support the same hardware. The xf86-video-intel-legacy driver is an old version of xf86-video-intel (patched to work with the current Xorg), for the people who find things worked better a year ago before all the intel architecture change mess started.

linfan · 2009-04-12 19:18:50

For me the whole xorg-server upgrade is for the worse, so I have downgraded and put it in IgnorePkg

IgnorePkg = xf86-input-evdev xf86-input-synaptics xf86-video-vesa xorg-server
xorg-utils

Especially the synaptics almost drove me crazy with a lot of boxes popping up from nowhere all over desktop as soon as I just put a finger on the touchpad. Copying and pasting also lived its own life and did not at all obey the cursor.

I already noticed this in testing, why I chose not to install gnome from testing as it dragged in this highly unwelcomed xorg upgrade.

I hope I am not the only one with these side effects.

joaca_rj · 2009-04-12 21:00:16

I'm having some issues with the new xorg and intel driver ( arch64, intel 965GM, kms-enabled-kernel, without xorg.conf)

first, blender3d is screwed, with corrupted screen....:(

Some problems with openbox just like the one described in this post (http://bbs.archlinux.org/viewtopic.php?id=56261)... it also happens in fluxbox.... oddly, if I quit openbox then "startx" back the error vanishes....

Other 3d apps are running quite nicer than before, even google sketchup 7 runs in wine ( it didn't before), games are running a lot faster, stellarium and google earth don't show corrupted fonts in 3d display anymore....:)

(sorry for poor english)

EDIT : Typo

Last edited by joaca_rj (2009-04-12 21:02:43)

kensai · 2009-04-12 21:12:04

Here, the new xorg-server works like a dream, I don't have to add or edit any synaptics rules anymore, now my touchpad works automagically. Nice work JGC.

jiu · 2009-04-12 21:43:14

first, blender3d is screwed, with corrupted screen....sad

I'm having the same problem here. thinking of downgrading as well.

As for google earth, using the AUR provided version, I'm getting a strange crash at startup which could be unrelated to the update, the console output follows:
Warning: Unable to create prefs directory '/home/jacques/.googleearth'. Le fichier existe.
Failed to initialize GEM. Falling back to classic.
do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
Try adjusting the vblank_mode configuration parameter.
./googleearth-bin: symbol lookup error: /usr/lib/libssl.so.0.9.8: undefined symbol: EVP_MD_CTX_set_flags

[EDIT] Ok, the problem with gg earth is solved, now it runs but it's very flickery although when it doesn't flicker it runs pretty smoothly

Last edited by jiu (2009-04-12 23:20:02)

mudrii · 2009-04-12 22:13:08

One way dealing with x-org update error

xorg-server: /usr/lib/xorg/modules/extensions/libdri.so exists in filesystem
xorg-server: /usr/lib/xorg/modules/libwfb.so exists in filesystem

Is just remove 'move' trouble some simlinks
sudo rm /usr/lib/xorg/modules/extensions/libdri.so
sudo rm /usr/lib/xorg/modules/libwfb.so

and should work after this.

Last edited by mudrii (2009-04-12 22:13:41)

sokuban · 2009-04-12 23:02:46

Hmm, the legacy intel driver isn't good for me. I'm getting about 300fps on glxgears. The normal intel driver gives me about 100, so legacy is better. (If I recall correctly I got around 500fps with the 1.5.3 xorg, so I don't know what's up.)

The weird thing is, right after I updated, I got about 800fps. Then I was playing with my xorg.conf file for a bit and now I'm stuck at 300fps. Resetting to the old configurations doesn't seem to solve anything either. I think this is pretty weird.

Also whenever I run glxgears, I get this message:

Failed to initialize GEM.  Falling back to classic.

Which reminds me of a message I used to get last year, right when all the horror started, except this time it failed to initialize GEM instead of TTM.

Is there any way to get TTM working on the new kernel?

Also are there older legacy drivers available? My card worked best with even older drivers. (Namely the i810)

EDIT: Xorg related this time. Am I the only person who can't ctrl-alt-bksp anymore to kill X?

Last edited by sokuban (2009-04-12 23:03:38)

ngoonee · 2009-04-13 00:33:08

sokuban, you may need to add DontZap as per the Ubuntu wiki https://wiki.ubuntu.com/X/Config/DontZap

I'm pretty sure I saw it mentioned elsewhere here on the Arch forums. I'd test it out, but am in the midst of compiling, so, you know...

uncholowapo · 2009-04-13 00:36:08

I love the new X server! It took away the annoying zlib error and compiz finally works with my intel graphics controller! Plus the new intel driver increased my fps in glxgears by over 300 fps.

sokuban · 2009-04-13 01:56:21

ngoonee wrote:

sokuban, you may need to add DontZap as per the Ubuntu wiki https://wiki.ubuntu.com/X/Config/DontZap
I'm pretty sure I saw it mentioned elsewhere here on the Arch forums. I'd test it out, but am in the midst of compiling, so, you know...

That works... but the screen flashes about 5 times before X loads up again. It seems X flashes more and more on startup as the version number increases. There was once a time when it never flashed, and X restarts would be almost instantaneous. But those times have come and gone. (I should note that closing the window manager doesn't cause the screen to flash 5 times—only about 3, which has been the norm for 1.5.3)

All these problems (no, this doesn't actually bother me so much, I'm just grouchy about the intel situation) sometimes make me want to just go to debian or slackware sometimes. Though I tried debian one time but I wasn't happy.

If only there was an Arch-style non bleeding edge distro. (I hope that isn't an oxymoron.)

Last edited by sokuban (2009-04-13 01:56:53)

mrunion · 2009-04-13 02:13:47

Test Blender3D here and no corruption after the upgrade.

Arch Linux x86_64
nVidia 8400 GS
Quad-core 8200
6GB RAM

Arch Linux

#26 2009-04-11 19:30:32

Re: xorg-server 1.6.0 moves to extra

#27 2009-04-11 20:39:11

Re: xorg-server 1.6.0 moves to extra

#28 2009-04-11 21:48:54

Re: xorg-server 1.6.0 moves to extra

#29 2009-04-12 02:04:44

Re: xorg-server 1.6.0 moves to extra

#30 2009-04-12 02:05:50

Re: xorg-server 1.6.0 moves to extra

#31 2009-04-12 04:02:41

Re: xorg-server 1.6.0 moves to extra

#32 2009-04-12 04:56:46

Re: xorg-server 1.6.0 moves to extra

#33 2009-04-12 05:46:40

Re: xorg-server 1.6.0 moves to extra

#34 2009-04-12 07:21:45

Re: xorg-server 1.6.0 moves to extra

#35 2009-04-12 08:35:56

Re: xorg-server 1.6.0 moves to extra

#36 2009-04-12 09:51:04

Re: xorg-server 1.6.0 moves to extra

#37 2009-04-12 13:18:00

Re: xorg-server 1.6.0 moves to extra

#38 2009-04-12 14:07:56

Re: xorg-server 1.6.0 moves to extra

#39 2009-04-12 14:57:43

Re: xorg-server 1.6.0 moves to extra

#40 2009-04-12 16:22:17

Re: xorg-server 1.6.0 moves to extra

#41 2009-04-12 19:18:50

Re: xorg-server 1.6.0 moves to extra

#42 2009-04-12 21:00:16

Re: xorg-server 1.6.0 moves to extra

#43 2009-04-12 21:12:04

Re: xorg-server 1.6.0 moves to extra

#44 2009-04-12 21:43:14

Re: xorg-server 1.6.0 moves to extra

#45 2009-04-12 22:13:08

Re: xorg-server 1.6.0 moves to extra

#46 2009-04-12 23:02:46

Re: xorg-server 1.6.0 moves to extra

#47 2009-04-13 00:33:08

Re: xorg-server 1.6.0 moves to extra

#48 2009-04-13 00:36:08

Re: xorg-server 1.6.0 moves to extra

#49 2009-04-13 01:56:21

Re: xorg-server 1.6.0 moves to extra

#50 2009-04-13 02:13:47

Re: xorg-server 1.6.0 moves to extra

Board footer