You are not logged in.
Hmmm, I had to revert to version 2.6.1 of intel driver, since this is the last version that lets me use xv overlay to watch videos -- watching videos without xv overlay results in serious tearing on my card (945GME). The current driver (and 2.6.2 driver as well) results in a green screen when trying to play videos using xv overlay, but that's unfortunately not all of it -- when exiting X after a green screen incident the monitor just turns black and I have to make a hard reset to get the system working again.
Anyone else experiencing a problem like this? I wonder if it's worth a bug report on the freedesktop bugtracker...
EDIT: Bug report for this issue already exists: https://bugs.freedesktop.org/show_bug.cgi?id=20416
Last edited by fwojciec (2009-03-08 23:23:10)
Offline
After upgrading my XVideo extension stopped working... I'm using a intel i830. Has anybody an idea?
EDIT: Sorry... I'm too slow
Last edited by JoyFM (2009-03-08 22:59:21)
Offline
Hmmm, I had to revert to version 2.6.1 of intel driver, since this is the last version that lets me use xv overlay to watch videos -- watching videos without xv overlay results in serious tearing on my card (945GME). The current driver (and 2.6.2 driver as well) results in a green screen when trying to play videos using xv overlay, but that's unfortunately not all of it -- when exiting X after a green screen incident the monitor just turns black and I have to make a hard reset to get the system working again.
Anyone else experiencing a problem like this? I wonder if it's worth a bug report on the freedesktop bugtracker...
As I experienced the same thing with another intel card (830M) it seems to be a bug!
Offline
I've been using the xf86-video-intel without the disable-reuse.patch now for a couple of hours, and I have no artifacts or random freezes (except the switch vt one). Regular use, GM965, UXA and KWin compositing.
I can't tell about memory leaks. I don't know how to measure that...
Last edited by Xarathur (2009-03-09 00:14:12)
Offline
hoschi wrote:congyiwu wrote:I'm having the same issue with GM965 too! Different combinations of the xserver 1.6 release candidates, intel 1.6 drivers, and libdrm packages from Debian experimental had the same issue intermittently with VT switching and suspending for at least a month on my laptop. This was before lenny got released. Not sure if it's fixed now in Debian, since I switched to arch after I got tired of all the lenny release drama, and stagnation in sid.
Same probleme here, the screen just freezes.
We have got back the performance, now they have taken our ability to switch vt or suspendGreat job intel, great job...
You are using testing packages. you can't expect it to work properly.
Yes an No and it is NOT an Arch problem.
If someone with a graphics-card from intel use the stable branch, he/she will lost around 50% of the graphic performance, because Xorg doesn't support TTM. If someone with a graphics-card from intel use the testing branch, he/she will lost ability to switch vt or suspend, but TTM works and 100% performance are available.
That's great, isn't it?
The only one solution with full performance and full functionality is using the old xorg-server 1.4(!), with the old drivers. The early upgrade to xorg-server 1.5 in stable branch of Arch was not a good idea, but the current problems in testing and stable are just caused by Intel and Xorg.
And as you can see, I don't blame Arch for not correct working software in the testing branch. I blame Intel and Xorg for a long lasting bad work over the last half year. Of course I now about the bad developer situation at Xorg, but we now everything worked well with the old Xorg-Server 1.4, so the current bad situation is simply not necessary.
I have spend money for the graphics-card from Intel, so I think it is my good right to blame Intel?
If I am honest, I am really angry. So, sorry if I am looking unhappy
b9anders wrote:hoschi wrote:Same probleme here, the screen just freezes.
We have got back the performance, now they have taken our ability to switch vt or suspendGreat job intel, great job...
You are using testing packages. you can't expect it to work properly.
They are testing in Arch, but consdered stable by Intel and X.org. So we CAN expect they will work fine.
BTW. There are some rumors that using 2.6.29 kernel solves the problem.
Right.
How can I see what I am using EXA or UXA (no xorg.conf present)?
How can I switch?
Option "AccelMethod" "UXA"
?
Last edited by hoschi (2009-03-09 11:06:48)
Offline
How can I see what I am using EXA or UXA (no xorg.conf present)?
How can I switch?
What you are using you can check in xorg logfile.
Default (no options) is EXA for 2.6 series drivers.
To change you have to:
Option "AccelMethod" "UXA"
or
Option "AccelMethod" "XAA"
in you xorg.conf file.
Last edited by blasse (2009-03-09 13:07:19)
Proud ex-maintainer of firefox-pgo
Offline
945gm here (on eeepc 901) and I can confirm what Inkaine says... with UXA I get a sudden freeze (altough the PC continues to work, I can here sounds from my speakers). I can cahnge VT to kill X, just returning to VT7 does not help. Btw with EXA performance is somewhat better than using xorg 1.5.
Oh, and when I run opera from a terminal it says I do not have any "X shared memory" extension and the webpage rendering is slower than usual. It was like that also with xorg 1.5, while it didn't complain (and it was way faster) with xorg 1.4
Offline
how do I set the mouse speed and acceleration in 1.6.0?
because the 'xset m' command doesn't work....
i've also noticed this new output in my xorg log
(**) <default pointer>: (accel) keeping acceleration scheme 1
(**) <default pointer>: (accel) filter chain progression: 2.00
(**) <default pointer>: (accel) filter stage 0: 20.00 ms
(**) <default pointer>: (accel) set acceleration profile 0
?
Offline
On my Acer Aspire One with Intel 945GME graphic chipset xorg-server 1.6 works great with the EXA rendering method. UXA is very slow and don't work with compositing. Indeed UXA consumes much fewer memory than EXA, wich would be ideal on a netbook, so i'm waiting for UXA to get better.
I am experiencing a strange issue since this update. I'm using qingy for logging in. It just runs startxfce4 after login. Butmy xfce session can't start compiz anymore. Compiz-manager says "xset q doesn't reveal the location of the log file. Using fallback /var/log/Xorg.0.log" But Xorg.0.log doesn't exist. If I run startxfce4 from commandline everything is OK.
Can this have to do with xset?
EDIT: I investigated this better by re-installing old xorg-server. Then xset reported the log at /var/log/Xorg.1.log. So the cleanest solution to this problem would be fixing xset. Does anybody know how to do that?
Last edited by Slikkie (2009-03-09 20:50:44)
Offline
@Slikkie:
For the UXA on the 945gm to work properly it is necessary to use the 2.6.29 kernel, are you using that? That said It doesn't work for me neither (using 2.6.29rc7), so I think we can only wait...
Offline
Guys... I think I found one major problem regarding the freezes, that some of you have with this new constellation of xf86-video-intel (including mesa etc.) and xorg-server 1.6.
I never had any freezing using uxa and the new up-to-date components in testing with my x3100 graphics adapter... While I was playing around with some settings in the xorg.conf file my notebook also freezed.
I can't tell what exactly the problem is, but I can tell you, that it has something to do with one of the following modules: record, freestyle and GLcore. Deactivating all of them should - for sure - fix the freezing problems...
EDIT: I believe it is the freestyle module...
Last edited by Dr4go (2009-03-10 00:45:47)
Offline
how do I set the mouse speed and acceleration in 1.6.0?
because the 'xset m' command doesn't work....i've also noticed this new output in my xorg log
(**) <default pointer>: (accel) keeping acceleration scheme 1 (**) <default pointer>: (accel) filter chain progression: 2.00 (**) <default pointer>: (accel) filter stage 0: 20.00 ms (**) <default pointer>: (accel) set acceleration profile 0
?
The xset m command works fine for me. Try some of the new hal options. My policy:
<match key="info.product" string="Razer DeathAdder">
<!--<merge key="input.x11_options.Sensitivity" type="string">0.1</merge>-->
<merge key="input.x11_options.Emulate3Buttons" type="string">False</merge>
<!--<merge key="input.x11_options.AdaptiveDeceleration" type="string">4</merge>-->
<merge key="input.x11_options.ConstantDeceleration" type="string">1.5</merge>
</match>
You can uncomment adaptivedeceleration and see if setting that helps you. Don't forget to restart hal and X.
I also set "xset m 1 0", though I've found anything >1 in the first value works for going faster.
Cthulhu For President!
Offline
gav616 wrote:how do I set the mouse speed and acceleration in 1.6.0?
because the 'xset m' command doesn't work....i've also noticed this new output in my xorg log
(**) <default pointer>: (accel) keeping acceleration scheme 1 (**) <default pointer>: (accel) filter chain progression: 2.00 (**) <default pointer>: (accel) filter stage 0: 20.00 ms (**) <default pointer>: (accel) set acceleration profile 0
?
The xset m command works fine for me. Try some of the new hal options. My policy:
<match key="info.product" string="Razer DeathAdder"> <!--<merge key="input.x11_options.Sensitivity" type="string">0.1</merge>--> <merge key="input.x11_options.Emulate3Buttons" type="string">False</merge> <!--<merge key="input.x11_options.AdaptiveDeceleration" type="string">4</merge>--> <merge key="input.x11_options.ConstantDeceleration" type="string">1.5</merge> </match>
You can uncomment adaptivedeceleration and see if setting that helps you. Don't forget to restart hal and X.
I also set "xset m 1 0", though I've found anything >1 in the first value works for going faster.
editing HAL config :S (I don't use dbus or hal daemons)
ill post a bug report for my xset not working, because it was much easier just to place a 'xset m 1/2 1' in my '.xinitrc'
Offline
@Dr4go:
I tried to comment the entire "modules" section of my xorg.conf, but the freezes are still there, guess it's a different problem with the 945gm. It COULD be also a kernel problem, because a lot of code has changed in the kernel lately for this cards...
Offline
Maybe we should wait for next release:
Intel releases 2.7 rc1 driver
Offline
editing HAL config :S (I don't use dbus or hal daemons)
No problem. They all work in xorg.conf.
Option "ConstantDeceleration" "1.5"
etc
The whole list: http://www.x.org/wiki/Development/Docum … celeration
Cthulhu For President!
Offline
gav616 wrote:editing HAL config :S (I don't use dbus or hal daemons)
No problem. They all work in xorg.conf.
Option "ConstantDeceleration" "1.5"
etc
The whole list: http://www.x.org/wiki/Development/Docum … celeration
thanks for the link,
is a shame its really confusing to me now, all I what is to disable acceleration altogether.
Offline
@Slikkie:
For the UXA on the 945gm to work properly it is necessary to use the 2.6.29 kernel, are you using that? That said It doesn't work for me neither (using 2.6.29rc7), so I think we can only wait...
I´m using 2.6.28.7. I already switched from kernel 2.6.27.7, otherwise Xorg won't start at all. 2.6.29 isn't aviable in Arch as far as I can see? For me it isn't worth the time to compile a newer kernel because EXA works fine.
Offline
Yes, I think I can be happy with EXA for now, even though knowing that my card could work better is a bit sad... btw, and a little OT, how do you manage having just a few packages from the testing repo? I didn't want to enable the repository (I have no problems with other packages) so I downloaded the xorg packages manually, but it's a pain to update them. Is it not possible to tell pacman to use testing just for some packages?
Offline
Guys... I think I found one major problem regarding the freezes, that some of you have with this new constellation of xf86-video-intel (including mesa etc.) and xorg-server 1.6.
I never had any freezing using uxa and the new up-to-date components in testing with my x3100 graphics adapter... While I was playing around with some settings in the xorg.conf file my notebook also freezed.
I can't tell what exactly the problem is, but I can tell you, that it has something to do with one of the following modules: record, freestyle and GLcore. Deactivating all of them should - for sure - fix the freezing problems...
EDIT: I believe it is the freestyle module...
Like Lazor I have the 945GM and the only one of those three modules I have is record and was getting the freezes.
Offline
Yes, I think I can be happy with EXA for now, even though knowing that my card could work better is a bit sad... btw, and a little OT, how do you manage having just a few packages from the testing repo? I didn't want to enable the repository (I have no problems with other packages) so I downloaded the xorg packages manually, but it's a pain to update them. Is it not possible to tell pacman to use testing just for some packages?
Use yaourt -Syu and then choose which packages you want to update by pressing M key and commenting by #.
Offline
i got a x4500 here and intel kernel module running. i get around 350 fps in glxgears, but warsow for example is at like 2 fps in the menu. i guess there is something wrong. here my xorg.conf
Section "Device"
Identifier "Card0"
Driver "intel"
VendorName "All"
BoardName "All"
Option "AccelMethod" "exa"
Option "MigrationHeuristic" "greedy"
EndSection
Offline
I tried the packages in testing with my X3100/965. With UXA + DRI2 the Xserver would freeze within 5 minutes of booting. With EXA it would freeze about 20 minutes after booting. These freezes also totally froze my keyboard, so ALT+CTRL+BACKSPACE did not work, so a hard reboot was required. I also made a new xorg.conf with Xorg -configure, adding my touchpad info and a few modules, including freetype and dri2. I also included AIGLX, the composite extension, and the DRI section. This didn't help either. I did not see anything of significance in the Xorg.1.log, just things loading at Xserver startup.
I needed to use my computer, so I used yaourt -Syu --downgrade, and am currently using non-testing repos only. I might try again next weekend.
Offline
It works perfectly fine here; 945GSE/GMA 950, UXA and xf86-video-intel 2.6.3-1 (-2 breaks things).
Offline
I got freezes, too, so I tried the xf86-video-intel from git. No more freezes, but the computer comes to a grinding halt (constant use of the hard drive, but not a memory leak). Also, compositing is even slower than with the 2.6 series. I'm back to xorg 1.4.
Offline