Re: Xorg and multiple graphics cards
also sprach Tiago Vignatti vigna...@freedesktop.org [2009.04.29.0422 +0200]: Yeah, I'm trying to solve part of this. Great, I will definitely try to help, but it will take me a few days. I will study the bugs and reply to them, but also keep the list in the loop, okay? -- martin | http://madduck.net/ | http://two.sentenc.es/ arrogance on the part of the meritorious is even more offensive to us than the arrogance of those without merit: for merit itself is offensive. -- friedrich nietzsche spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.21.1734 +0200]:
Looks like an XAA problem. Can you try with EXA?
Option AccelMethod EXA
in your device sections.
I tried with EXA in all four sections, and the result I get is
another segfault:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f611512d790 (LWP 16444)]
dixLookupPrivate (privates=0x2739540, key=0x0) at ../../dix/privates.c:79
79 ../../dix/privates.c: No such file or directory.
in ../../dix/privates.c
#0 dixLookupPrivate (privates=0x2739540, key=0x0) at ../../dix/privates.c:79
ptr = value optimized out
#1 0x004b31c0 in xf86RandR12CreateScreenResources (pScreen=0x27392b0)
at ../../../../hw/xfree86/modes/xf86RandR12.c:756
pScrn = (ScrnInfoPtr) 0x2701440
c = value optimized out
width = value optimized out
height = value optimized out
mmWidth = value optimized out
mmHeight = value optimized out
#2 0x004ab87f in xf86CrtcCreateScreenResources (screen=0x27392b0)
at ../../../../hw/xfree86/modes/xf86Crtc.c:701
No locals.
#3 0x00433014 in main (argc=4, argv=0x7fff1d14e6d8,
envp=value optimized out) at ../../dix/main.c:325
pScreen = (ScreenPtr) 0x27392b0
i = 0
alwaysCheckForInput = {0, 1}
Backtrace:
0: /usr/bin/Xorg(xorg_backtrace+0x26) [0x4ef016]
1: /usr/bin/Xorg(xf86SigHandler+0x39) [0x476509]
2: /lib/libc.so.6 [0x7f6112bd71a0]
3: /usr/bin/Xorg(dixLookupPrivate+0x4) [0x433bd4]
4: /usr/bin/Xorg(xf86RandR12CreateScreenResources+0x50) [0x4b31c0]
5: /usr/bin/Xorg [0x4ab87f]
6: /usr/bin/Xorg(main+0x264) [0x433014]
7: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f6112bc35a6]
8: /usr/bin/Xorg [0x4325f9]
Log and xorg.conf attached.
--
martin | http://madduck.net/ | http://two.sentenc.es/
i never travel without my diary. one should always have something
sensational to read on the train.
-- oscar wilde
spamtraps: madduck.bo...@madduck.net
Xorg.1.log.bz2
Description: Binary data
# /etc/X11/xorg.conf (xorg X Window System server configuration file)
#
# This file was generated by dexconf, the Debian X Configuration tool, using
# values from the debconf database.
#
# Edit this file with caution, and see the /etc/X11/xorg.conf manual page.
# (Type man /etc/X11/xorg.conf at the shell prompt.)
#
# This file is automatically updated on xserver-xorg package upgrades *only*
# if it has not been modified since the last upgrade of the xserver-xorg
# package.
#
# If you have edited this file but would like it to be automatically updated
# again, run the following commands:
#
# cp /etc/X11/xorg.conf /etc/X11/xorg.conf.custom
# sudo sh -c 'md5sum /etc/X11/xorg.conf /var/lib/xfree86/xorg.conf.md5sum'
# sudo dpkg-reconfigure xserver-xorg
### begin triple-head config
Section ServerLayout
IdentifierTriple Head
Screen 0Screen[0]
Screen 1Screen[1] RightOf Screen[0]
Screen 2Screen[2] RightOf Screen[1]
InputDevice Generic Keyboard
InputDevice Configured Mouse
InputDevice Wacom stylus SendCoreEvents
InputDevice Wacom eraser SendCoreEvents
InputDevice Wacom cursor SendCoreEvents
Option Clone off
Option Xinerama yes
EndSection
#Section Files
# ModulePath/usr/local/lib/xorg/modules,/usr/lib/xorg/modules
#EndSection
Section Module
Load wacom
EndSection
Section ServerFlags
Option DontZap yes
Option AllowDeactivateGrabs yes
Option AllowClosedownGrabs yes
EndSection
Section Device
Identifier Radeon 9250[0]
Driver radeon
BusID PCI:0:12:0
Screen 0
# Option HW Cursor off
# Option MergedFB true
Option MergedFB off
Option AccelMethod EXA
EndSection
Section Device
Identifier Radeon 9250[1]
Driver radeon
BusID PCI:0:12:1
Screen 1
# Option HW Cursor off
# Option MergedFB true
Option MergedFB off
Option AccelMethod EXA
EndSection
Section Device
Identifier Radeon 9200[0]
Driver radeon
BusID PCI:1:0:0
Screen 0
# Option MergedFB true
# Option MonitorLayout TMDS, CRT
# Option CRT2Position RightOf
# Option MetaModes 1280x1024-1280x1024 1152x864-1152x864
1024x768-1024x768 800x600-800x600 640x480-640x480
Option MergedFB off
Option AccelMethod EXA
EndSection
Section Device
Identifier Radeon 9200[1]
Driver radeon
BusID PCI:1:0:0
Screen 1
# Option MergedFB true
# Option MonitorLayout TMDS, CRT
# Option CRT2Position RightOf
# Option MetaModes 1280x1024-1280x1024 1152x864-1152x864
1024x768-1024x768 800x600-800x600 640x480-640x480
Option MergedFB off
Option AccelMethod EXA
EndSection
Section Monitor
Identifier MRM B18XA
HorizSync 24-80
Re: Xorg and multiple graphics cards
also sprach martin f krafft madd...@madduck.net [2009.04.28.1806 +0200]: I tried with EXA in all four sections, and the result I get is another segfault: Julien pointed me to commit faf7dfa099f5b42a703313fbd1bf8afdad07a179, which seems to solve that problem. Unfortunately, like the hydra, this gave way to more problems. Remember I have a dual-card triple-head Zaphod setup, so the heads are 0-0, 0-1, and 1-0, where the latter is the first and only head on the second card. Both cards are Radeon 9200: 00:0c.1 Display controller: ATI Technologies Inc RV280 [Radeon 9200 PRO] (Secondary) (rev 01) 01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01) 1. The 1-0 head has plenty of weird pixelation errors, making fonts only barely readable 2. X (as well as console-setup on tty) seem to believe the screens are taller and place the bottom edge below the physical edge, meaning I cannot see the bottom centimetre or so. Auto-adjust does not fix this on any of the heds. 3. When moving the mouse from head to head, the old pointer is left on the edge of the head we just left, and stays there while a new pointer moves about on the entered head. When moving back, the new pointer stays at the edge, and the old pointer is resumed. 4. Switching to tty1 causes screen corruption and the tty is actually never displayed. Switching to X on :0 (I used :1 to test) restores heads 0-0 and 1-0, but head 0-1 stays black -- until I reboot. It feels like progress, but it's not where I need to be yet... I now have a good feeling of the software and can easily test commits, so if you have anything... Cheers, -- martin | http://madduck.net/ | http://two.sentenc.es/ plan to be spontaneous tomorrow. spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
On Tue, Apr 28, 2009 at 1:47 PM, martin f krafft madd...@madduck.net wrote: also sprach martin f krafft madd...@madduck.net [2009.04.28.1806 +0200]: I tried with EXA in all four sections, and the result I get is another segfault: Julien pointed me to commit faf7dfa099f5b42a703313fbd1bf8afdad07a179, which seems to solve that problem. Unfortunately, like the hydra, this gave way to more problems. Remember I have a dual-card triple-head Zaphod setup, so the heads are 0-0, 0-1, and 1-0, where the latter is the first and only head on the second card. Both cards are Radeon 9200: 00:0c.1 Display controller: ATI Technologies Inc RV280 [Radeon 9200 PRO] (Secondary) (rev 01) 01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01) 1. The 1-0 head has plenty of weird pixelation errors, making fonts only barely readable 2. X (as well as console-setup on tty) seem to believe the screens are taller and place the bottom edge below the physical edge, meaning I cannot see the bottom centimetre or so. Auto-adjust does not fix this on any of the heds. 3. When moving the mouse from head to head, the old pointer is left on the edge of the head we just left, and stays there while a new pointer moves about on the entered head. When moving back, the new pointer stays at the edge, and the old pointer is resumed. 4. Switching to tty1 causes screen corruption and the tty is actually never displayed. Switching to X on :0 (I used :1 to test) restores heads 0-0 and 1-0, but head 0-1 stays black -- until I reboot. It feels like progress, but it's not where I need to be yet... I now have a good feeling of the software and can easily test commits, so if you have anything... It won't work until your secondary card is posted. Int10 posting of secondary cards doesn't work with libpciaccess. The non-int10 post code in the ddx is incomplete and disabled. You can try it with the attached patch, but it likely won't make a difference. Your best bet short term is to use xserver 1.4.x. Alex legacy_init.diff Description: application/mbox ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.28.1957 +0200]: It won't work until your secondary card is posted. Int10 posting of secondary cards doesn't work with libpciaccess. The non-int10 post code in the ddx is incomplete and disabled. You also explained on IRC: Normally when you boot, the system bios initializes your video card by executing a routine in the video bios. That's what we call posting. So in your case, the non-int10 post code in the ddx needs to be fixed, or we need to make int10 stuff work for seconday cards with libpciaccess. You can try it with the attached patch, but it likely won't make a difference. It didn't. :( Your best bet short term is to use xserver 1.4.x. Any idea how long that short term is? 1.4.x works fine for me at the moment, but I'd rather not build on that basis for indeterminate time... -- martin | http://madduck.net/ | http://two.sentenc.es/ our destiny exercises its influence over us even when, as yet, we have not learned its nature; it is our future that lays down the law of our today. - friedrich nietzsche spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
On Tue, Apr 28, 2009 at 3:07 PM, martin f krafft madd...@madduck.net wrote: also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.28.1957 +0200]: It won't work until your secondary card is posted. Int10 posting of secondary cards doesn't work with libpciaccess. The non-int10 post code in the ddx is incomplete and disabled. You also explained on IRC: Normally when you boot, the system bios initializes your video card by executing a routine in the video bios. That's what we call posting. So in your case, the non-int10 post code in the ddx needs to be fixed, or we need to make int10 stuff work for seconday cards with libpciaccess. You can try it with the attached patch, but it likely won't make a difference. It didn't. :( Your best bet short term is to use xserver 1.4.x. Any idea how long that short term is? 1.4.x works fine for me at the moment, but I'd rather not build on that basis for indeterminate time... Hard to say, depends on when the work gets done. The universal fix would be to get int10 and vga stuff working with libpciaccess, which on Linux at least, requires a kernel vga arbiter. Not sure about other OSs. radeon specific fixes would be to fix the non-int10 post code in the radeon driver or switch to the kms/mm enabled radeon drm and fix up any multi-card issues that may be lurking there. Alex ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.28.2122 +0200]: Hard to say, depends on when the work gets done. The universal fix would be to get int10 and vga stuff working with libpciaccess, which on Linux at least, requires a kernel vga arbiter. Not sure about other OSs. radeon specific fixes would be to fix the non-int10 post code in the radeon driver or switch to the kms/mm enabled radeon drm and fix up any multi-card issues that may be lurking there. Let me know if I can help, even though I better stay away from actual code or hardware. -- martin | http://madduck.net/ | http://two.sentenc.es/ it usually takes more than three weeks to prepare a good impromptu speech. -- mark twain spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
Hi, martin f krafft escreveu: 3. When moving the mouse from head to head, the old pointer is left on the edge of the head we just left, and stays there while a new pointer moves about on the entered head. When moving back, the new pointer stays at the edge, and the old pointer is resumed. Yeah, I saw it here as well. But this is usually easy to fix. Find which commit caused the bug with a binary search in the tree (man git-bisect). I'd guess something inside mipointer.c Cheers, Tiago ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
Alex Deucher escreveu: On Tue, Apr 28, 2009 at 3:07 PM, martin f krafft madd...@madduck.net wrote: Your best bet short term is to use xserver 1.4.x. Any idea how long that short term is? 1.4.x works fine for me at the moment, but I'd rather not build on that basis for indeterminate time... Hard to say, depends on when the work gets done. The universal fix would be to get int10 and vga stuff working with libpciaccess, which on Linux at least, requires a kernel vga arbiter. Not sure about other OSs. radeon specific fixes would be to fix the non-int10 post code in the radeon driver or switch to the kms/mm enabled radeon drm and fix up any multi-card issues that may be lurking there. Yeah, I'm trying to solve part of this. Martin, if you want to help in some way, test this patch and see if it works, please: http://lists.x.org/archives/xorg-devel/2009-April/000794.html Thank you, Tiago ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
On Tue, 28 Apr 2009, Tiago Vignatti wrote: Alex Deucher escreveu: On Tue, Apr 28, 2009 at 3:07 PM, martin f krafft madd...@madduck.net wrote: Your best bet short term is to use xserver 1.4.x. Any idea how long that short term is? 1.4.x works fine for me at the moment, but I'd rather not build on that basis for indeterminate time... Hard to say, depends on when the work gets done. The universal fix would be to get int10 and vga stuff working with libpciaccess, which on Linux at least, requires a kernel vga arbiter. Not sure about other OSs. radeon specific fixes would be to fix the non-int10 post code in the radeon driver or switch to the kms/mm enabled radeon drm and fix up any multi-card issues that may be lurking there. Yeah, I'm trying to solve part of this. Martin, if you want to help in some way, test this patch and see if it works, please: http://lists.x.org/archives/xorg-devel/2009-April/000794.html Can I point out that https://bugs.freedesktop.org/show_bug.cgi?id=20816 and its dependencies are the appropriate bugs for the issue? Not that I want to discourage mailing list comment, or anything, but merely to point any relevant people at what's already been done, and what hasn't. I'm hoping, sometime within the next month or two, to be temporarily in a position where I can at least do some testing on this; I'll have access to a dual NVidia/SiS system that I can install random stuff on. Thanks all for your time... - | Name: Tim Nelson | Because the Creator is,| | E-mail: wayl...@wayland.id.au| I am | - BEGIN GEEK CODE BLOCK Version 3.12 GCS d+++ s+: a- C++$ U+++$ P+++$ L+++ E- W+ N+ w--- V- PE(+) Y+++ PGP-+++ R(+) !tv b++ DI D G+ e++ h! y- -END GEEK CODE BLOCK- ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
Hi Timothy, Timothy S. Nelson escreveu: Can I point out that https://bugs.freedesktop.org/show_bug.cgi?id=20816 and its dependencies are the appropriate bugs for the issue? Not that I want to discourage mailing list comment, or anything, but merely to point any relevant people at what's already been done, and what hasn't. Yes, please feed that bug. And if you like to do that, I also have a list here of related bugs :) https://bugs.freedesktop.org/show_bug.cgi?id=18160 https://bugs.freedesktop.org/show_bug.cgi?id=18839 https://bugs.freedesktop.org/show_bug.cgi?id=18321 https://bugs.freedesktop.org/show_bug.cgi?id=20849 https://bugzilla.redhat.com/show_bug.cgi?id=454864 Cheers, Tiago ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
On Thu, Apr 16, 2009 at 1:33 PM, martin f krafft madd...@madduck.net wrote: also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.16.1908 +0200]: Can you get a full backtrace with gdb? http://wiki.x.org/wiki/Development/Documentation/ServerDebugging Wow, you guys *really* make it easy for people, especially on Debian. Thanks! Looks like an XAA problem. Can you try with EXA? Option AccelMethod EXA in your device sections. Alex ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.15.1859 +0200]: That's what zaphod mode is. What sort of problem do you get when you use that config against the newer driver? Using the following software on Debian sid: ii xserver-xorg1:7.3+18the X.Org X server ii xserver-xorg-co 2:1.4.2-11 Xorg X server - core server ii xserver-xorg-vi 1:6.9.0-1+lenny X.Org X server -- ATI display driver wrapper and the first xorg.conf included in this post http://lists.freedesktop.org/archives/xorg/2009-April/045107.html with ModulePath commented, such that it uses the Debian-installed driver, instead of the one I custom-compiled and installed into /usr/local: 0. the server starts up 1. it displays some Radeon identification string on both heads connected to the primary card, while the head on the secondary card stays black 2. the identification string changes into the sequence ♀♀#♀#♀#♀#♀#♀, which stays there forever, but the screens eventually go black 3. the log (attached) informs me of a segfault with this backtrace: Backtrace: 0: /usr/bin/X(xf86SigHandler+0x6a) [0x47898a] 1: /lib/libc.so.6 [0x7f8e894b1190] 2: /usr/bin/X(RRCrtcSetRotations+0) [0x525c90] 3: /usr/bin/X(xf86RandR12SetRotations+0x77) [0x4b90a7] 4: /usr/bin/X(xf86CrtcScreenInit+0xad) [0x4b45ad] 5: /usr/lib/xorg/modules/drivers//radeon_drv.so(RADEONScreenInit+0x1bd8) [0x7f8e8701ef38] 6: /usr/bin/X(AddScreen+0x249) [0x436209] 7: /usr/bin/X(InitOutput+0x277) [0x46ab27] 8: /usr/bin/X(main+0x2c6) [0x4369f6] 9: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f8e8949d5a6] 10: /usr/bin/X [0x435e99] Fatal server error: Caught signal 11. Server aborting This looks similar to the result I got 14 months ago: http://archive.netbsd.se/?ml=xorga=2008-02m=6443010 Thanks for your time and help, -- martin | http://madduck.net/ | http://two.sentenc.es/ a c programmer asked whether computers have buddha's nature. as the answer, the master did rm -rf on the programmer's home directory. and then the c programmer became enlightened... spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
On 4/16/09, martin f krafft madd...@madduck.net wrote: also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.15.1859 +0200]: That's what zaphod mode is. What sort of problem do you get when you use that config against the newer driver? Using the following software on Debian sid: ii xserver-xorg1:7.3+18the X.Org X server ii xserver-xorg-co 2:1.4.2-11 Xorg X server - core server ii xserver-xorg-vi 1:6.9.0-1+lenny X.Org X server -- ATI display driver wrapper and the first xorg.conf included in this post http://lists.freedesktop.org/archives/xorg/2009-April/045107.html with ModulePath commented, such that it uses the Debian-installed driver, instead of the one I custom-compiled and installed into /usr/local: 0. the server starts up 1. it displays some Radeon identification string on both heads connected to the primary card, while the head on the secondary card stays black 2. the identification string changes into the sequence ♀♀#♀#♀#♀#♀#♀, which stays there forever, but the screens eventually go black 3. the log (attached) informs me of a segfault with this backtrace: Backtrace: 0: /usr/bin/X(xf86SigHandler+0x6a) [0x47898a] 1: /lib/libc.so.6 [0x7f8e894b1190] 2: /usr/bin/X(RRCrtcSetRotations+0) [0x525c90] 3: /usr/bin/X(xf86RandR12SetRotations+0x77) [0x4b90a7] 4: /usr/bin/X(xf86CrtcScreenInit+0xad) [0x4b45ad] 5: /usr/lib/xorg/modules/drivers//radeon_drv.so(RADEONScreenInit+0x1bd8) [0x7f8e8701ef38] 6: /usr/bin/X(AddScreen+0x249) [0x436209] 7: /usr/bin/X(InitOutput+0x277) [0x46ab27] 8: /usr/bin/X(main+0x2c6) [0x4369f6] 9: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f8e8949d5a6] 10: /usr/bin/X [0x435e99] Fatal server error: Caught signal 11. Server aborting Does it work any better with xserver 1.5 or newer? IIRC, there was a bug where RRCrtcSetRotations() dereferenced pScrn-pScreen before it was set. I don't remember off hand when it was fixed however. Alex ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.16.1559 +0200]: 1. it displays some Radeon identification string on both heads connected to the primary card, while the head on the secondary card stays black 2. the identification string changes into the sequence ♀♀#♀#♀#♀#♀#♀, which stays there forever, but the screens eventually go black 3. the log (attached) informs me of a segfault with this backtrace: Backtrace: 0: /usr/bin/X(xf86SigHandler+0x6a) [0x47898a] 1: /lib/libc.so.6 [0x7f8e894b1190] 2: /usr/bin/X(RRCrtcSetRotations+0) [0x525c90] 3: /usr/bin/X(xf86RandR12SetRotations+0x77) [0x4b90a7] 4: /usr/bin/X(xf86CrtcScreenInit+0xad) [0x4b45ad] 5: /usr/lib/xorg/modules/drivers//radeon_drv.so(RADEONScreenInit+0x1bd8) [0x7f8e8701ef38] 6: /usr/bin/X(AddScreen+0x249) [0x436209] 7: /usr/bin/X(InitOutput+0x277) [0x46ab27] 8: /usr/bin/X(main+0x2c6) [0x4369f6] 9: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f8e8949d5a6] 10: /usr/bin/X [0x435e99] Fatal server error: Caught signal 11. Server aborting Does it work any better with xserver 1.5 or newer? IIRC, there was a bug where RRCrtcSetRotations() dereferenced pScrn-pScreen before it was set. I don't remember off hand when it was fixed however. I now tried: (II) Module radeon: vendor=X.Org Foundation compiled for 1.6.0, module version = 6.12.2 Module class: X.Org Video Driver ABI class: X.Org Video Driver, version 5.0 and I get the exact same behaviour described under (1) and (2), although (1) passes a lot quicker, and the backtrace is slightly different: Backtrace: 0: /usr/bin/X(xorg_backtrace+0x26) [0x4ef046] 1: /usr/bin/X(xf86SigHandler+0x39) [0x484219] 2: /lib/libc.so.6 [0x7fed7f3da190] 3: /usr/bin/X(dixLookupPrivate+0xa) [0x433b7a] 4: /usr/lib/xorg/modules//libxaa.so [0x7fed7c9615e2] 5: /usr/lib/xorg/modules/drivers//radeon_drv.so [0x7fed7d22a2c1] 6: /usr/bin/X(xf86CrtcSetModeTransform+0x129) [0x4ac0a9] 7: /usr/bin/X(xf86SetDesiredModes+0x2a9) [0x4ac939] 8: /usr/lib/xorg/modules/drivers//radeon_drv.so [0x7fed7d20ab1f] 9: /usr/bin/X(AddScreen+0x1c6) [0x432846] 10: /usr/bin/X(InitOutput+0x241) [0x46d111] 11: /usr/bin/X(main+0x20e) [0x432f5e] 12: /lib/libc.so.6(__libc_start_main+0xe6) [0x7fed7f3c65a6] 13: /usr/bin/X [0x432599] Thanks again for your time! -- martin | http://madduck.net/ | http://two.sentenc.es/ you don't sew with a fork, so I see no reason to eat with knitting needles. -- miss piggy, on eating chinese food spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.16.1908 +0200]: Can you get a full backtrace with gdb? http://wiki.x.org/wiki/Development/Documentation/ServerDebugging Wow, you guys *really* make it easy for people, especially on Debian. Thanks! Here it is: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fcac7073790 (LWP 11907)] 0x00433b7a in dixLookupPrivate (privates=0x290, key=0x7fcac22ce47c) at ../../dix/privates.c:79 79 ../../dix/privates.c: No such file or directory. in ../../dix/privates.c (gdb) bt #0 0x00433b7a in dixLookupPrivate (privates=0x290, key=0x7fcac22ce47c) at ../../dix/privates.c:79 #1 0x7fcac20a45e2 in XAAStateWrapSync (pScrn=0x1d4fdc0) at ../../../../hw/xfree86/xaa/xaaStateChange.c:323 #2 0x7fcac296d2c1 in radeon_crtc_lock (crtc=value optimized out) at ../../src/radeon.h:1429 #3 0x004ac0a9 in xf86CrtcSetModeTransform (crtc=0x1e738c0, mode=0x1e739a8, rotation=1, transform=0x0, x=value optimized out, y=0) at ../../../../hw/xfree86/modes/xf86Crtc.c:271 #4 0x004ac939 in xf86SetDesiredModes (scrn=0x1d4fdc0) at ../../../../hw/xfree86/modes/xf86Crtc.c:2497 #5 0x7fcac294db1f in RADEONScreenInit (scrnIndex=1, pScreen=0x1de1120, argc=value optimized out, argv=value optimized out) at ../../src/radeon_driver.c:3768 #6 0x00432846 in AddScreen (pfnInit=0x7fcac294bde0 RADEONScreenInit, argc=7, argv=0x7fffcf093df8) at ../../dix/main.c:702 #7 0x0046d111 in InitOutput (pScreenInfo=0x7d8760, argc=7, argv=0x7fffcf093df8) at ../../../../hw/xfree86/common/xf86Init.c:1245 #8 0x00432f5e in main (argc=7, argv=0x7fffcf093df8, envp=0x7fffcf093e38) at ../../dix/main.c:315 -- martin | http://madduck.net/ | http://two.sentenc.es/ we americans, we're a simple people... but piss us off, and we'll bomb your cities. -- robin williams, good morning vietnam spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Xorg and multiple graphics cards
Hi folks, After upgrading to 7.4, I got an ABI mismatch when using the old 6.6 ATI driver to control the two Radeon cards I have in my machines, which are combined with Xinerama into a triple-monitor setup. I could also not compile the 6.6 tip against the new 7.4 code [0]. I was informed that X still does not support XRandR across multiple cards yet. Thus, I have two choices: 1. roll back and stay with X 7.3 + ATI 6.6 2. forget triple-head and return to dual-head (2) is clearly not an option, so I am doing (1), but I am also reminded of this thread on the issue: http://www.nabble.com/setting-up-xrandr-with-two-graphics-cards-td13835871.html and I was wondering if there's a roadmap ahead and multi-card XRandR is not too far down the line. Can anyone estimate, when it will be supported? Cheers, Martin 0. This is the compiler error I get when compiling the xf86-video-ati 6.6 branch [a5570f3] against X 1.6.0 headers: gcc -DHAVE_CONFIG_H -I. -I.. -I/usr/include/xorg -I/usr/include/pixman-1 -I/usr/include/drm -I/usr/include/X11/dri -g -O2 -MT atibus.lo -MD -MP -MF .deps/atibus.Tpo -c atibus.c -fPIC -DPIC -o .libs/atibus.o In file included from atibus.c:35: atistruct.h:287: error: expected specifier-qualifier-list before ‘pciVideoPtr’ atibus.c: In function ‘ATIClaimResources’: atibus.c:78: error: ‘struct _ATIRec’ has no member named ‘SharedVGA’ atibus.c:88: error: ‘struct _ATIRec’ has no member named ‘SharedVGA’ atibus.c:89: error: ‘struct _ATIRec’ has no member named ‘SharedVGA’ atibus.c:90: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:94: error: ‘struct _ATIRec’ has no member named ‘SharedVGA’ atibus.c:104: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:106: error: ‘struct _ATIRec’ has no member named ‘VGAWonderResources’ atibus.c:111: error: ‘struct _ATIRec’ has no member named ‘SharedAccelerator’ atibus.c:117: error: ‘struct _ATIRec’ has no member named ‘SharedAccelerator’ atibus.c:118: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:124: error: ‘struct _ATIRec’ has no member named ‘SharedAccelerator’ atibus.c:131: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:147: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:151: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c: In function ‘ATIClaimBusSlot’: atibus.c:176: error: ‘pciVideoPtr’ undeclared (first use in this function) atibus.c:176: error: (Each undeclared identifier is reported only once atibus.c:176: error: for each function it appears in.) atibus.c:176: error: expected ‘;’ before ‘pVideo’ atibus.c:178: error: ‘pVideo’ undeclared (first use in this function) atibus.c:179: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:181: warning: passing argument 4 of ‘xf86ClaimPciSlot’ from incompatible pointer type atibus.c:181: error: too many arguments to function ‘xf86ClaimPciSlot’ atibus.c:183: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:185: error: ‘struct _ATIRec’ has no member named ‘iEntity’ atibus.c:188: error: ‘struct _ATIRec’ has no member named ‘iEntity’ make[2]: *** [atibus.lo] Error 1 -- martin; (greetings from the heart of the sun.) \ echo mailto: !#^.*|tr * mailto:; n...@madduck writing a book about debian is like hitting a moving target with a champagne bottle cork. -- arky spamtraps: madduck.bo...@madduck.net digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/) ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
On 4/15/09, martin f krafft madd...@madduck.net wrote: Hi folks, After upgrading to 7.4, I got an ABI mismatch when using the old 6.6 ATI driver to control the two Radeon cards I have in my machines, which are combined with Xinerama into a triple-monitor setup. I could also not compile the 6.6 tip against the new 7.4 code [0]. I was informed that X still does not support XRandR across multiple cards yet. Thus, I have two choices: 1. roll back and stay with X 7.3 + ATI 6.6 2. forget triple-head and return to dual-head (2) is clearly not an option, so I am doing (1), but I am also reminded of this thread on the issue: zaphod style multi-head should still work with the radeon driver. Alex ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: Xorg and multiple graphics cards
also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.15.1748 +0200]: (2) is clearly not an option, so I am doing (1), but I am also reminded of this thread on the issue: zaphod style multi-head should still work with the radeon driver. I am not sure I know what zaphod-style is. I've attached the two Xorg configuration files I use/would like to use to this mail. xorg.conf.saved, the first, is the old configuration that works only with 6.6. xorg.conf.xrandr, the second, is the configuration I pieced together using http://wiki.debian.org/XStrikeForce/HowToRandR12, but with 7.3, I cannot make it drive the second card, and with 7.4, it just results in a screen so black that the only way to fix it I've found is a complete reboot of the machine. Any help warmly appreciated, Thanks, -- martin | http://madduck.net/ | http://two.sentenc.es/ if you are walking on thin ice, you might as well dance! spamtraps: madduck.bo...@madduck.net # /etc/X11/xorg.conf (xorg X Window System server configuration file) # # This file was generated by dexconf, the Debian X Configuration tool, using # values from the debconf database. # # Edit this file with caution, and see the /etc/X11/xorg.conf manual page. # (Type man /etc/X11/xorg.conf at the shell prompt.) # # This file is automatically updated on xserver-xorg package upgrades *only* # if it has not been modified since the last upgrade of the xserver-xorg # package. # # If you have edited this file but would like it to be automatically updated # again, run the following commands: # # cp /etc/X11/xorg.conf /etc/X11/xorg.conf.custom # sudo sh -c 'md5sum /etc/X11/xorg.conf /var/lib/xfree86/xorg.conf.md5sum' # sudo dpkg-reconfigure xserver-xorg ### begin triple-head config Section ServerLayout IdentifierTriple Head Screen 0Screen[0] Screen 1Screen[1] RightOf Screen[0] Screen 2Screen[2] RightOf Screen[1] InputDevice Generic Keyboard InputDevice Configured Mouse InputDevice Wacom stylus SendCoreEvents InputDevice Wacom eraser SendCoreEvents InputDevice Wacom cursor SendCoreEvents Option Clone off Option Xinerama yes EndSection Section Files ModulePath/usr/local/lib/xorg/modules,/usr/lib/xorg/modules EndSection Section Module Load wacom EndSection Section ServerFlags Option DontZap yes Option AllowDeactivateGrabs yes Option AllowClosedownGrabs yes EndSection Section Device Identifier Radeon 9250[0] Driver radeon BusID PCI:0:12:0 Screen 0 # Option HW Cursor off # Option MergedFB true Option MergedFB off # Option AccelMethod EXA EndSection Section Device Identifier Radeon 9250[1] Driver radeon BusID PCI:0:12:1 Screen 1 # Option HW Cursor off # Option MergedFB true Option MergedFB off # Option AccelMethod EXA EndSection Section Device Identifier Radeon 9200[0] Driver radeon BusID PCI:1:0:0 Screen 0 # Option MergedFB true # Option MonitorLayout TMDS, CRT # Option CRT2Position RightOf # Option MetaModes 1280x1024-1280x1024 1152x864-1152x864 1024x768-1024x768 800x600-800x600 640x480-640x480 Option MergedFB off # Option AccelMethod EXA EndSection Section Device Identifier Radeon 9200[1] Driver radeon BusID PCI:1:0:0 Screen 1 # Option MergedFB true # Option MonitorLayout TMDS, CRT # Option CRT2Position RightOf # Option MetaModes 1280x1024-1280x1024 1152x864-1152x864 1024x768-1024x768 800x600-800x600 640x480-640x480 Option MergedFB off # Option AccelMethod EXA EndSection Section Monitor Identifier MRM B18XA HorizSync 24-80 VertRefresh 30-60 # #ModeLine1280x1024 108.00 1280 1328 1440 1688 1024 1025 1028 1066 +hsync +vsync # #ModeLine1152x768 65.00 1152 1178 1314 1472 768 771 777 806 +hsync +vsync # #ModeLine1024x768 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync # #ModeLine800x600 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync # #ModeLine640x480 25.20 640 656 752 800 480 490 492 525 -hsync -vsync Option DPMS EndSection Section Monitor Identifier Acer AL922[0] # HorizSync 24-80 # VertRefresh 56-75 Option DPMS EndSection Section Monitor Identifier Acer AL922[1] # HorizSync 24-80 # VertRefresh 56-75 Option DPMS EndSection Section Screen Identifier Screen[0] DeviceRadeon 9250[0] Monitor Acer AL922[0] DefaultDepth 24 SubSection Display Depth 24 Modes
Re: Xorg and multiple graphics cards
On 4/15/09, martin f krafft madd...@madduck.net wrote: also sprach Alex Deucher alexdeuc...@gmail.com [2009.04.15.1748 +0200]: (2) is clearly not an option, so I am doing (1), but I am also reminded of this thread on the issue: zaphod style multi-head should still work with the radeon driver. I am not sure I know what zaphod-style is. I've attached the two Xorg configuration files I use/would like to use to this mail. xorg.conf.saved, the first, is the old configuration that works only with 6.6. That's what zaphod mode is. What sort of problem do you get when you use that config against the newer driver? Alex ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
