Please do not reply to this email: if you want to comment on the bug, go to
the URL shown below and enter yourcomments there.
https://bugs.freedesktop.org/show_bug.cgi?id=6534
Summary: r200 brilinear filtering patch causes system crashes?
Product: Mesa
Version: unspecified
Platform: PC
URL: https://bugs.freedesktop.org/show_bug.cgi?id=2702
OS/Version: Linux
Status: NEW
Severity: critical
Priority: P2
Component: Drivers/DRI/r200
AssignedTo: dri-devel@lists.sourceforge.net
ReportedBy: [EMAIL PROTECTED]
CC: [EMAIL PROTECTED]
Okay... where to begin with this... it all started about a year ago. (-:
I was using DRM CVS in the hope that my card (Radeon IGP345M) would one day go
slightly faster than a snail (I later discovered the HyperZ option). After
doing a system-wide update, I began to experience seemingly random crashes. At
first, I thought they were just one-offs but they didn't go away. They would
occur once every other day on average. Then I figured it was probably a kernel
bug and I rushed to get every new release, even the release candidates, but
that didn't help. For various reasons, I began to suspect the USB, the RAM, the
hard drive but I couldn't work out what it actually was. I tried various
debugging techniques including Magic SysRq, netconsole and the NMI watchdog.
None of them shed any light on the situation.
Okay, enough of the sob story. (-; Fast forward to a couple of months ago. This
guy called Süleyman saw a post I'd made on LKML and e-mailed me. He said he
had
an almost identical machine. I have the Sony Vaio PCG-K195HP. He has the Sony
Vaio PCG-K215Z. The only difference, as far as we can see, is that his has a
faster processor and a better Radeon card (Mobility 9200). He confirmed that he
was having the same problem and so I was able to rule out the possibility of a
hardware fault.
I read about this cool thing called "git bisect" and decided I would use it to
get to the bottom of this. At first I tried bisecting the kernel but I got all
the way back to 2.6.9 and it was still crashing. It had to be something else. I
made a guess and removed the DRM modules. Several days went by. No crashes.
Aha! I converted the entire DRM CVS repository to git and began bisecting that.
To make matters complicated, the older modules wouldn't compile against 2.6.15
so I had to drop back to 2.6.11 and I also had to use very close versions of
Mesa/DRI. Getting that to compile against Xorg 7 was fun. :-S So I kept
bisecting and bisecting until eventually, I homed in on the one patch that
seemed to make all the difference. The corresponding bug report for it is bug #
2702.
I have no idea why this patch would cause the system to crash but seemingly it
does. It was time for the acid test. I tried removing this patch from a very
recent DRM version and went back to 2.6.15. It crashed. :-/ However, the nature
of the crash was different. In almost every case, the crashes I was
experiencing before caused the fans to spin up, suggesting some kind of spin
lock. Süleyman has confirmed this behaviour. But the crash I experienced after
removing the patch did not cause the fans to spin up and I subsequently tried
it several times after that. It would seem that naively removing these lines
causes some other problem.
I am now running on 2.6.11 again with the last-known-good version of DRM to
make absolutely sure that I've found the right patch. I've had 3 days uptime so
far. Tomorrow, I will try the first-known-bad version to make doubly sure that
it crashes.
Süleyman apparently managed to stop the crashes in a totally unrelated way by
disabling the ACPI "Processor" and "Thermal" modules. Maybe this can give you a
clue as to what's causing all of this? I haven't tried it myself but I may do
soon. It doesn't seem very ideal, especially since I'm running Gentoo. Don't
want this thing to melt after a few compiles. (-;
I will attach my lspci.
--
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
--
_______________________________________________
Dri-devel mailing list
Dri-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dri-devel
