You are not logged in.
Wilco wrote:I was a bit too fast with this. It is *very* unstable with KMS enabled, but it runs fine with xorg-server 1.6.1 with EXA (yep that's not UXA) and the new intel drivers.
UXA with KMS doesnt seem unstable here at all, just a little buggy. The only problem Ive found with it is that some of the fading effects in compiz fade improperly and look terrible as a result (they fade to white rather than fade to nothing).
Other than that though, everything seems fine. Maybe im just lucky
those problems are a known transparency bug....i dont think there is a fix for it yet. maybe it requires lots of changes here and there in order to work as it should
Offline
This are the errors I get when using UXA. It happens at random after a few minutes or in a matter of seconds when there is some opengl stuff
[drm:i915_get_vblank_counter] *ERROR* trying to get vblank count for disabled pipe 0
mtrr: no MTRR for d0000000,10000000 found
At this point all graphics are completely frozen. Without KMS, I can Ctrl+Alt+Backspace out of this. Only EXA is stable and fast for me at this point.
Offline
Wilco wrote:I was a bit too fast with this. It is *very* unstable with KMS enabled, but it runs fine with xorg-server 1.6.1 with EXA (yep that's not UXA) and the new intel drivers.
UXA with KMS doesnt seem unstable here at all, just a little buggy. The only problem Ive found with it is that some of the fading effects in compiz fade improperly and look terrible as a result (they fade to white rather than fade to nothing).
Other than that though, everything seems fine. Maybe im just lucky
Yeah this is what was happening to me.
Deleted my xorg.conf and things look much better, despite not being super smooth.
And in the midst of such perfection,
I can't help but feel diseased.
Offline
Hi there, if you want your performance back on your 945GM and xf86-video-intel 2.6+, upgrade your kernel to at least 2.6.30-rc2 (= you have to compile it yourself). The performance's even better than with xf86-video-intel 2.4.x (at least in Quake III Arena)
Last edited by grubber (2009-04-26 06:48:25)
Offline
Hi there, if you want your performance back on your 945GM and xf86-video-intel 2.6+, upgrade your kernel to at least 2.6.30-rc2 (= you have to compile it yourself). The performance's even better than with xf86-video-intel 2.4.x (at least in Quake III Arena)
I tried that shortly after the patch went in and it didn't cut it for me. But I was kind of frustrated and did not play with the settings as much as I should have... Do you use KMS or UXA alone? What does your xorg.conf look like?
Offline
I use both KMS and UXA. I don't use any special settings in my xorg.conf, see for yourself:
Section "ServerLayout"
Identifier "Xorg Configured"
Screen 0 "Screen0" 0 0
EndSection
Section "ServerFlags"
Option "AllowMouseOpenFail" "true"
Option "DontZap" "False"
EndSection
Section "Files"
ModulePath "/usr/lib/xorg/modules"
FontPath "/usr/share/fonts/misc:unscaled"
FontPath "/usr/share/fonts/misc"
FontPath "/usr/share/fonts/75dpi:unscaled"
FontPath "/usr/share/fonts/75dpi"
FontPath "/usr/share/fonts/100dpi:unscaled"
FontPath "/usr/share/fonts/100dpi"
FontPath "/usr/share/fonts/Type1"
EndSection
Section "Module"
Load "ddc" # ddc probing of monitor
Load "dbe"
Load "dri"
Load "extmod"
Load "glx"
Load "bitmap" # bitmap-fonts
Load "record"
Load "evdev"
# Load "synaptics"
EndSection
Section "Monitor"
Identifier "Monitor0"
Option "DPMS" "true"
# HorizSync 28.0 - 78.0 # Warning: This may fry very old Monitors
HorizSync 28.0 - 96.0 # Warning: This may fry old Monitors
VertRefresh 50.0 - 75.0 # Very conservative. May flicker.
# VertRefresh 50.0 - 62.0 # Extreme conservative. Will flicker. TFT default.
Modeline "1024x768" 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync
Modeline "1024x768" 54.16 1024 1048 1184 1344 768 771 777 806 -hsync -vsync
Modeline "800x600" 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync
Modeline "640x480" 25.18 640 656 752 800 480 490 492 525 -hsync -vsync
EndSection
# Auto-generated by Archie mkxcfg
Section "Device"
Identifier "Card0"
Driver "intel"
VendorName "All"
BoardName "All"
EndSection
Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
DefaultColorDepth 24
SubSection "Display"
Depth 24
Modes "1024x768" "800x600" "640x480"
EndSubSection
EndSection
Section "DRI"
Mode 0666
EndSection
Offline
I use both KMS and UXA. I don't use any special settings in my xorg.conf
Neat, I guess I have to try again. Now all I need is some time for play...
Offline
Another weird error.
It seems to be drawing my desktop at the wrong resolution. But twice. I don't know how to describe it, but my system tray and whatnot aren't using the rightmost 200 pixels of my screen.
However, I can move windows over there, the wallpaper goes over there, so it functions, but it's not drawn right. If I maximize a window that's sitting over on the left, it only uses the leftmost 3/4 of the screen. If the window is over on the right, when maximized it uses up all the screen. I'll have to re-do the instructions and take some screenshots. It's VERY strange.
And in the midst of such perfection,
I can't help but feel diseased.
Offline
Another weird error.
It seems to be drawing my desktop at the wrong resolution. But twice. I don't know how to describe it, but my system tray and whatnot aren't using the rightmost 200 pixels of my screen.
However, I can move windows over there, the wallpaper goes over there, so it functions, but it's not drawn right. If I maximize a window that's sitting over on the left, it only uses the leftmost 3/4 of the screen. If the window is over on the right, when maximized it uses up all the screen. I'll have to re-do the instructions and take some screenshots. It's VERY strange.
I had some similar problem too. Just to be sure it's the same, is this solved if you kill X and restart it? I had to compare Xorg.0.log with Xorg.0.log.old to see what was the difference between the two sessions. It appeared that the first session was trying another output (TV output, IIRC). I had to disable this output in xorg.conf and everything went right.
Offline
I haven't tried. I don't have an xorg.conf whatsoever right now. I'll poke around perhaps.
And in the midst of such perfection,
I can't help but feel diseased.
Offline
For me both EXA and UXA gives me very slow graphics on 2.6.29 (I can tell EXA is working because OpenGL apps are not redirected). I can get about 100fps in GlxGears.
With 2.6.28 with UXA, everything is quick as possible and the performance is great (even tough it has the transparency bug).
I'm staying with 2.6.28 for now.
Offline
Yay! I can confirm grubber's experience on a
$ lspci | grep -i vga
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)
I compiled kernel 2.26.30-rc3 (released a few hours ago) with the intel drm built into the kernel. I have completely deleted xorg.conf. Now my system runs speedily on KMS+UXA, with KWin's desktop effects enabled. I get nearly the same frame rate as with xorg-server 1.4 and xf86-video-intel 2.3.2. So far no show stoppers.
I am using the 2.7.0 driver from [testing]. The 2.6.3 driver is not significantly slower, but it seems to have some KMS-related bugs, among them X server freezes when enabling vsync.
Last edited by mutlu_inek (2009-04-28 05:43:20)
Offline
I compiled kernel 2.26.30-rc3 (released a few hours ago) with the intel drm built into the kernel. I have completely deleted xorg.conf. Now my system runs speedily on KMS+UXA, with KWin's desktop effects enabled. I get nearly the same frame rate as with xorg-server 1.4 and xf86-video-intel 2.3.2. So far no show stoppers.
I am using the 2.7.0 driver from [testing]. The 2.6.3 driver is not significantly slower, but it seems to have some KMS-related bugs, among them X server freezes when enabling vsync.
That sounds interesting. Can you post a PKGBUILD if you have one? Cheers!
Offline
Well I don't feel like compiling anything, but guess I can wait until .30 hits testing if it works!
And in the midst of such perfection,
I can't help but feel diseased.
Offline
Well I don't feel like compiling anything, but guess I can wait until .30 hits testing if it works!
would you mind posting your xorg log. your problem is so similar to the one i had that i'm quite sure this can be easily solved (and without building any kernel).
Offline
OK, having amassed a couple weeks' worth of experience running the legacy driver with the 945, 2.6.29 kernel, and X 1.6.1, I still see no movie problems, and still see no X memory leaks; and I'm still using the hal for my touchpad and have no problems there either. Again, I'm not a gamer, but everything short of that looks pretty good. It sounds like 2.6.30 should take care of most of my concerns, so I'll wait for it.
Offline
ugh, confusement.
Has anybody here tried to play around with DRI2? Everybody seems to be using DRI
I'm currently running:
kernel26 2.6.29.1-4
xf86-video-intel 2.6.3.3
2884 frames in 5.0 seconds = 576.795 FPS
2883 frames in 5.0 seconds = 576.435 FPS
Nothing special in my xorg.conf yet, still playing with options. No KMS.
Section "ServerLayout"
Identifier "X.org Configured"
Screen 0 "Screen0" 0 0
InputDevice "Mouse0" "CorePointer"
InputDevice "Keyboard0" "CoreKeyboard"
EndSectionSection "Files"
ModulePath "/usr/lib/xorg/modules"
FontPath "/usr/share/fonts/misc"
FontPath "/usr/share/fonts/100dpi:unscaled"
FontPath "/usr/share/fonts/75dpi:unscaled"
FontPath "/usr/share/fonts/TTF"
FontPath "/usr/share/fonts/Type1"
EndSectionSection "Module"
Load "extmod"
Load "dbe"
Load "dri"
Load "record"
# Load "dri2"
Load "glx"
EndSectionSection "InputDevice"
Identifier "Keyboard0"
Driver "kbd"
Option "XkbLayout" "us"
EndSectionSection "InputDevice"
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/input/mice"
Option "ZAxisMapping" "4 5 6 7"
EndSectionSection "Monitor"
Identifier "Monitor0"
VendorName "Monitor Vendor"
ModelName "Monitor Model"
EndSectionSection "Device"
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz"
### [arg]: arg optional
#Option "NoAccel" # [<bool>]
#Option "SWcursor" # [<bool>]
#Option "ColorKey" # <i>
#Option "CacheLines" # <i>
#Option "Dac6Bit" # [<bool>]
#Option "DRI" # [<bool>]
#Option "NoDDC" # [<bool>]
#Option "ShowCache" # [<bool>]
#Option "XvMCSurfaces" # <i>
#Option "PageFlip" # [<bool>]
Identifier "Card0"
Driver "intel"
VendorName "Intel Corporation"
BoardName "Mobile GME965/GLE960 Integrated Graphics Controller"
BusID "PCI:0:2:0"
EndSectionSection "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
SubSection "Display"
Viewport 0 0
Depth 24
EndSubSection
EndSectionSection "ServerFlags"
Option "AutoAddDevices" "False"
EndSection
Compiz works well and stable at about 70 FPS, animations are fine.
Rebooting right now to do some more testing
*edit*
and back. DRI + UXA / EXA or DRI2 + UXA seem to perform mostly the same in glxgears and compiz benchmarking. It's all verry stable and smooth for me.
Last edited by stefanwilkens (2009-04-29 00:14:13)
Arch i686 on Phenom X4 | GTX760
Offline
For anyone wanting to know more about the Intel situation here is a very good post http://keithp.com/blogs/Sharpening_the_ … ver_Focus/
Also Phoronix is usually taking a good look at the Intel performance, such as this: http://www.phoronix.com/scan.php?page=n … &px=NzIyMA
Offline
Just compiled a GIT kernel (2.6.30-rc3-00483-g56a50ad-dirty) using the kernel26-git AUR package.
Using UXA and KMS, compositing perfomance is peachy again. Possibly a bit better even than with the 2.6.28 series. Will report back if I hit stability issues.
Offline
i have GM965
i have been playing aound with the new intel driver thats in the repo as well as the legacy driver for about 2 days
ive tried UXA EXA XAA, you name it
i found that i get the best performance at this time with the legacy driver and this xorg.conf
Section "ServerLayout"
Identifier "Xorg Configured"
Screen "Default Screen"
InputDevice "Keyboard0" "CoreKeyboard"
# PS/2 Mouse not detected
# Serial Mouse not detected
InputDevice "USB Mouse" "CorePointer"
InputDevice "SynapticsTouchpad" "SendCoreEvents"
EndSection
Section "ServerFlags"
Option "AllowMouseOpenFail" "true"
# Option "AutoAddDevices" "False"
EndSection
Section "Files"
ModulePath "/usr/lib/xorg/modules"
FontPath "/usr/share/fonts/misc:unscaled"
FontPath "/usr/share/fonts/misc"
FontPath "/usr/share/fonts/75dpi:unscaled"
FontPath "/usr/share/fonts/75dpi"
FontPath "/usr/share/fonts/100dpi:unscaled"
FontPath "/usr/share/fonts/100dpi"
FontPath "/usr/share/fonts/PEX"
# Additional fonts: Locale, Gimp, TTF...
FontPath "/usr/share/fonts/cyrillic"
# FontPath "/usr/share/lib/X11/fonts/latin2/75dpi"
# FontPath "/usr/share/lib/X11/fonts/latin2/100dpi"
# True type and type1 fonts are also handled via xftlib, see /etc/X11/XftConfig!
FontPath "/usr/share/fonts/Type1"
FontPath "/usr/share/fonts/ttf/western"
FontPath "/usr/share/fonts/ttf/decoratives"
FontPath "/usr/share/fonts/truetype"
FontPath "/usr/share/fonts/truetype/openoffice"
FontPath "/usr/share/fonts/truetype/ttf-bitstream-vera"
FontPath "/usr/share/fonts/latex-ttf-fonts"
FontPath "/usr/share/fonts/defoma/CID"
FontPath "/usr/share/fonts/defoma/TrueType"
EndSection
Section "Module"
Load "ddc" # ddc probing of monitor
Load "dbe"
Load "dri"
Load "extmod"
Load "glx"
Load "bitmap" # bitmap-fonts
Load "type1"
Load "freetype"
Load "record"
Load "synaptics"
EndSection
Section "InputDevice"
Identifier "Keyboard0"
Driver "kbd"
Option "CoreKeyboard"
Option "XkbRules" "xorg"
Option "XkbModel" "pc105"
Option "XkbLayout" "us"
EndSection
Section "InputDevice"
Identifier "Serial Mouse"
Driver "mouse"
Option "Protocol" "Microsoft"
Option "Device" "/dev/ttyS0"
Option "Emulate3Buttons" "true"
Option "Emulate3Timeout" "70"
Option "SendCoreEvents" "true"
EndSection
Section "InputDevice"
Identifier "PS/2 Mouse"
Driver "mouse"
Option "Protocol" "auto"
Option "ZAxisMapping" "4 5"
Option "Device" "/dev/psaux"
Option "Emulate3Buttons" "true"
Option "Emulate3Timeout" "70"
Option "SendCoreEvents" "true"
EndSection
Section "InputDevice"
Identifier "USB Mouse"
Driver "mouse"
Option "Device" "/dev/input/mice"
Option "SendCoreEvents" "true"
Option "Protocol" "IMPS/2"
Option "ZAxisMapping" "4 5"
Option "Buttons" "5"
EndSection
Section "InputDevice"
Identifier "SynapticsTouchpad"
Driver "synaptics"
Option "AlwaysCore" "true" # send events to CorePointer
# Option "Device" "/dev/input/mice"
Option "Device" "/dev/psaux"
Option "Protocol" "auto-dev"
Option "SHMConfig" "false" # configurable at runtime? security risk
Option "LeftEdge" "1700" # x coord left
Option "RightEdge" "5300" # x coord right
Option "TopEdge" "1700" # y coord top
Option "BottomEdge" "4200" # y coord bottom
Option "FingerLow" "25" # pressure below this level triggers release
Option "FingerHigh" "30" # pressure above this level triggers touch
Option "MaxTapTime" "180" # max time in ms for detecting tap
Option "VertEdgeScroll" "true" # enable vertical scroll zone
Option "HorizEdgeScroll" "true" # enable horizontal scroll zone
Option "CornerCoasting" "true" # enable continuous scroll with finger in corner
Option "CoastingSpeed" "0.30" # corner coasting speed
Option "VertScrollDelta" "20" # edge-to-edge scroll distance of the vertical scroll
Option "HorizScrollDelta" "20" # edge-to-edge scroll distance of the horizontal scroll
Option "MinSpeed" "0.10" # speed factor for low pointer movement
Option "MaxSpeed" "0.60" # maximum speed factor for fast pointer movement
Option "AccelFactor" "0.0030" # acceleration factor for normal pointer movements
Option "VertTwoFingerScroll" "true" # vertical scroll anywhere with two fingers
Option "HorizTwoFingerScroll" "true" # horizontal scroll anywhere with two fingers
EndSection
Section "Monitor"
Identifier "Generic Monitor"
Option "DPMS"
EndSection
Section "Monitor"
Identifier "TVOutput"
Option "Ignore" "true"
EndSection
Section "Device"
Identifier "Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller"
Driver "intel"
BusID "PCI:0:2:0"
Option "monitor-TV" "TVOutput"
Option "DRI" "True"
Option "XAANoOffscreenPixmaps" "True"
Option "EnablePageFlip" "True"
Option "RenderAccel" "True"
Option "AccelMethod" "XAA"
EndSection
Section "Screen"
Identifier "Default Screen"
Device "Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller"
Monitor "Generic Monitor"
DefaultDepth 24
SubSection "Display"
Modes "1280x800"
EndSubSection
EndSection
Section "DRI"
Mode 0666
EndSection
Section "Extensions"
Option "Composite" "Enable"
EndSection
with this config you cant get XV video playback so you have to set your media player to use X11 video output
with totem just open gstreamer-properties and choose "no XV"
with vlc and mplayer its in the preferences
other then that everything works great
i can play games perfectly
i regularly play nexuiz, assault cube, cube, sauerbraten, zsnes emulator, mupen64, hexen 2, world of padman, and tremulous (all work perfectly in fluxbox, i havn't tried with compiz)
in fluxbox i use xcompmgr with the -c option only-and it works perfectly, but for games i make a short script to killall conky and xcompmgr when i start the game and restart them when i turn the game off
Last edited by tjwoosta (2009-04-29 14:46:06)
Offline
FTR, I gained about 90 FPS (glxgears) using 2.7.99.1 released earlier today.
Offline
Another weird error.
It seems to be drawing my desktop at the wrong resolution. But twice. I don't know how to describe it, but my system tray and whatnot aren't using the rightmost 200 pixels of my screen.
However, I can move windows over there, the wallpaper goes over there, so it functions, but it's not drawn right. If I maximize a window that's sitting over on the left, it only uses the leftmost 3/4 of the screen. If the window is over on the right, when maximized it uses up all the screen. I'll have to re-do the instructions and take some screenshots. It's VERY strange.
i had the same issue at first, this is caused by the tv-output
try adding this part to your xorg.conf underneath your usual monitor section (so you have two monitor sections)
Section "Monitor"
Identifier "TVOutput"
Option "Ignore" "true"
EndSection
and add this line under the "device" section
Option "monitor-TV" "TVOutput"
just take a look at the xorg.conf i posted in my previous post for a reference to what i mean
Last edited by tjwoosta (2009-04-29 16:09:14)
Offline
http://aur.archlinux.org/packages.php?ID=26053
I've created a PKGBUILD for all users experiencing the memory leak issue with 945GM, and can confirm that KMS/UXA is working on my own 945GM machine using 2.6.30-rc3 kernel.
It's all right, it's just a horse in the bathroom.
Offline
Well, I started to try that out, but with the ten million configuration options I had to bail partway through. I don't even know what "LSI MPT Fusion Max number of SG Entries" means.
And in the midst of such perfection,
I can't help but feel diseased.
Offline
I just updated my PKGBUILD and removed the arch patches, as they conflict with the rc3 patchset. I answered no to most of the preset options and built my kernel successfully (although if you don't change the memory allocator from SLUB to SLAB you might have problems installing things like OSSv4 and VirtualBox modules).
Last edited by icenine457 (2009-04-30 16:49:37)
It's all right, it's just a horse in the bathroom.
Offline