Hi, So I merged xorg-x11-6.8.99.3 and tried to start it. All that happens is that the virtual terminal changes. Then I get a black screen and the fans of my G5 are spinning like crazy. I cannot change the terminal back and ssh connects also don't work. After 10 minutes I restarted my computer hard. I'll attach xorg log file and the output of startx. Markus
Created attachment 57956 [details] Xorg.0.log
Created attachment 57957 [details] startx.log
Take a look at http://xorg.freedesktop.org/wiki/DebuggingTheXserver
I updated to 6.8.99.5. Here is what I did (following the guide): start ssh connection the my G5 and doing this: # gdb X GNU gdb 6.3 Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "powerpc64-unknown-linux-gnu"...Using host libthread_db library "/lib/libthread_db.so.1". (gdb) run Starting program: /usr/bin/X warning: Unable to find dynamic linker breakpoint function. GDB will be unable to debug shared library initializers and track explicitly loaded dynamic code. warning: Breakpoint address adjusted from 0x102ac9b0 to 0x1004ce94. _XSERVTransSocketOpenCOTSServer: Unable to open socket for inet6 _XSERVTransOpen: transport open failed for inet6/localhost:0 _XSERVTransMakeAllCOTSServerListeners: failed to open listener for inet6 This is a pre-release version of the The X.Org Foundation X11. It is not supported in any way. Bugs may be filed in the bugzilla at http://bugs.freedesktop.org/. Select the "xorg" product for bugs you find in this release. Before reporting bugs in pre-release versions please check the latest version in the The X.Org Foundation "monolithic tree" CVS repository hosted at http://www.freedesktop.org/Software/xorg/ X Window System Version 6.8.99.5 Release Date: 01 May 2005 + cvs X Protocol Version 11, Revision 0, Release 6.8.99.5 Build Operating System: Linux 2.6.12-rc3 ppc64 [ELF] Current Operating System: Linux localhost 2.6.12-rc3 #1 SMP Wed Apr 27 14:41:21 Local time zone must be set--see zic ppc64 Build Date: 04 May 2005 Before reporting problems, check http://wiki.X.Org to make sure that you have the latest version. Module Loader present Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (==) Log file: "/var/log/Xorg.0.log", Time: Thu May 5 09:46:15 2005 (==) Using config file: "/root/xorg.conf" ------------------------------------------------------------------------------- At this point the fans are spinning like crazy and the display changed to vt7. The Xorg.0.log got some more lines than the last, but nothing that could help.. Last but not least the Xserver seems to run and does not crash at all (according to gdb).. Also I can login with ssh, but I cannot run any programms. Here is what it looks like if I try it: -------- # uptime RETURN <CURSOR> -------- and then no nothing. BTW: I didn't use dlloader (as the guide suggests), because this is in use.mask. Is dlloader not supported for all archs?
Honestly, I have no clue whether dlloader's supposed to work on ppc64. Ajax, do you? You won't be able to get any useful debugging stuff out of vanilla gdb with elfloader xorg -- try dlloader, and if that doesn't work, maybe static can tell you something. Alternately, patches exist for gdb to support the elfloader.
dlloader should never be use.masked. dlloader should work on any platform with a libdl, which includes ppc64.
It apparently doesn't on hppa, according to gmsoft. But that's another issue.
dlloader seems to work on ppc64. I've compiled 6.8.2-r2 with USE="dlloader" and this works just fine. I'll remove it from use.mask in the ppc64 profiles. 6.8.99.5 compiled with dlloader gives no other output using gdb.
I found the interesting bug from the upstream bugs. The problem is similar although hardware is ppc. https://bugs.freedesktop.org/show_bug.cgi?id=3318 And, I found it starting by using fbdev. FYI, On PPC64, "#define MakeDllModules YES" is always added irrespective of the state of dlloader flag. This is required in order that a module may solve a symbol. Please see Bug #56248 for details.
Created attachment 60196 [details] my xorg.conf when starting on 6.8.99.8 The status of a USE flag is the following. [ebuild R ] x11-base/xorg-x11-6.8.99.8 (-3dfx) (-3dnow) +bitmap-fonts +cjk -debug -dlloader -dmx -doc +font-server +insecure-drivers -ipv6 -minimal (-mmx) +nls +opengl +pam -sdk (-sse) -static +truetype-fonts +type1-fonts (-uclibc) -xprint +xv 0 kB
6.8.99.8 same issue.
I just compared the xc/config/cf/ folder of 6.8.2-r2 and 6.8.99.8 and I found out, that the 9360_all_6.7.99.2-ppc64-support-updates-v3.patch is no longer needed. We patched an already patched file. All patches seem to be applied upstream. Please remove the patch from patch tarball. I'll *try* to do some further debugging.
xorg-x11-6.8.99.14 same issue. spyderous: please exclude 9360_all_6.7.99.2-ppc64-support-updates-v3.patch, as it is not necessary.
Does excluding that patch actually fix things, or is it merely aesthetic?
sorry, I should have been more clear. it is just aestetic and does not fix this.
Can you try bumping the ebuild backwards to older versions to see where it stopped working? We know it didn't work at 6.8.99.3, but we don't know when it stopped. Does 6.8.99.1 work?
I've just compiled 6.8.99.1 and it works. Maybe this patches are responsilbe of the behaviour? https://bugs.freedesktop.org/show_bug.cgi?id=2064 I'll check after which snapshot they got in.
no, the patches from previous mentioned bug are already applied to 6.8.99.1, so they are not the reason...
OK, how about .2?
ok.. I finaly got some time to build xorg-x11-6.8.99.2 and test it. And it fails.. So now we know the breakage happend between .1 and .2. Now what to do next to track this problem down?
See how big the diff is and check whether anything obvious happened in roughly the right place.
Created attachment 63772 [details] xorg-x11-ppc64.patch there was a patch commited after 6.8.99.1 which added this lines and broke ppc64. So this patch "unpatches" radeon_driver.c. Now everything works like a charm!! The Changelog sais this about the patch: * programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c: Release SDA/SDL I2C lines from an asserted-low state after a DDC probe, which caused Apple Studio Display monitors to shut off after a second or so (Ben Herrenschmidt). I have an Apple Studio Display, so I'll contact benh and ask him what is going on.
Please file a bug at bugs.freedesktop.org on this as well, and set it up as a blocker on bug #1690.
here is the upstream bug: https://bugs.freedesktop.org/show_bug.cgi?id=3821
I was wrong. The breakage didn't happened from .1 to .2. I figured out that .1 was just starting because I haven't had a xorg.conf in /etc/X11/ and the auto configuration seems to have built a configuration similar to the attached one ("my xorg.conf when starting on 6.8.99.8") Should we close this bug as UPSTREAM and handle this in the upstream bug?
Yes
Markus stated upstream that 6.8.99.15-r4 works fine.
Marking fixed.
