My computer hangs totally with gentoo-sources-2.6.21 and 2.6.21-r1 in X after some compiling. I first compiled mplayer without problems (while browsing WWW with Firefox). And then I compiled it another time and it just freezed (mouse didn't move either). I had to reset using power button. Everything works fine with 2.6.20-r6 and earlier. I'll attach my .config. And please ask more info as I'm not sure what is relevant and what isn't. System: Asus P5B deluxe Intel Core 2 Duo Sata HD (WD1600YS) 2GB Ram nVidia 7900GTO + nvidia-drivers-1.0.9755-r1 Reproducible: Always
Created attachment 119689 [details] Config file for gentoo-sources-2.6.21-r1
Additional info: I use Paludis 0.24.2. So when I said "compiling" I meant "installing packages with Paludis". I'm now testing if the system freezes without Paludis. I have not yet tested with emerge.
So far it seems that my system is otherwise stable. I'll report back when I try compiling with emerge.
Same problem on my amd64 3000+ 512 Mo Ram under ck-sources-2.6.21-r2 (with _p1 too) : when running "emerge --sync" and clicking on a shortcut on the desktop it comes to a usb panic (mouse and keyboard Logitech Cordless Desktop are disabled and leds on connecting-base are lightenning very fast) and the system freezes (processor halts).
Rufus: sounds like a different issue, please file a new bug with more info after reproducing it on a supported kernel such as gentoo-sources
I tested with emerge and after the 4th recompile of mplayer it alse freezed the machine. Very odd and sadly a bit randomish.
Could you please test with the latest vanilla kernel 2.6.22-rc2? Also, please post the output of dmesg. Does the system only freeze when/after compiling, or does it freeze after some time no matter what you do? When the system freezes, do you just mean X, ie. can you still switch to another virtual console and use the keyboard?
(In reply to comment #6) > I tested with emerge and after the 4th recompile of mplayer it alse freezed the > machine. Very odd and sadly a bit randomish. > do you have ACPI Thermal zone compiled in? i found that kernel sometimes hang when cpu making hot... the best way to make kernel hang is kompile glibc...
it seems that is a linux-2.6.21 specyfic bug please look: http://bugzilla.kernel.org/show_bug.cgi?id=8385
I'm beginning to think that this is a CPU heating issue which is triggered by 2.6.21 easier than by earlier kernels: 1. Thermal zone support had no effect. 2. I was able to freeze my machine even with 2.6.20-r6 by compiling mplayer MANY times. For some odd reason 2.6.21 freezed a lot easier than 2.6.20. This is why I thought that this is a kernel issue. Maybe 2.6.21 utilizes cores more efficiently? I have no idea. 3. I was also able to crash my computer using a totally different program (a 2 thread particle simulator by myself) and at the same time untarring kernel tarball. By reducing the CPU temperature I seem to be stable now. I will mark this bug now INVALID and reopen if I get crashes even with these low temps. Sorry for taking your time and thanks for all the help. PS. I think this is not about the issue in Comment #9.
(In reply to comment #10) > I'm beginning to think that this is a CPU heating issue which is triggered by > 2.6.21 easier than by earlier kernels: > > 1. Thermal zone support had no effect. > 2. I was able to freeze my machine even with 2.6.20-r6 by compiling mplayer > MANY times. For some odd reason 2.6.21 freezed a lot easier than 2.6.20. This > is why I thought that this is a kernel issue. Maybe 2.6.21 utilizes cores more > efficiently? I have no idea. > 3. I was also able to crash my computer using a totally different program (a 2 > thread particle simulator by myself) and at the same time untarring kernel > tarball. > > By reducing the CPU temperature I seem to be stable now. I will mark this bug > now INVALID and reopen if I get crashes even with these low temps. Sorry for > taking your time and thanks for all the help. > > PS. I think this is not about the issue in Comment #9. > wirld but kernel devs confirmed this is a bug. my 2.6.20-r8 kernel not hangs even i make many compilations at once witk super pi program placet in infinite loop. cpu temp ~75C , load average 21.0 19.2 18.1 uptime 2 days and no crash...
(In reply to comment #11) > wirld but kernel devs confirmed this is a bug. It only affects very select hardware configurations, chances are that paapaa125 is not affected by it. If you are affected by that bug, watch bug 176615
I found out the actual reason (yes, I'm embarrassed): A few weeks ago I cleaned my computer case interior and also fastened the CPU fan better. Better? Well, maybe it was more securely in place but I managed to reverse the airflow... Thermalright Ultra-120 doesn't like air being pulled through it. It wants air _pushed_ through it and I didn't notice the 'polarity' of the fan. This increased the core temperature almost 10 degrees C. For some reason 2.6.21 caused more load to CPU than 2.6.20 - or maybe it was just pure coincidence. But yes, 2.6.21 is rock solid now and this bug is as INVALID as it gets ;-)
