Strange hard lock problems with 2.6.6 and compiles

[vlack]

I've just installed Gentoo [system... VIA chipset motherboard; Athlon 2500+ no overclockage; 1GB RAM; Promise SATA controller (libata); 160GB drive; reiserfs (v3): /, /usr, /var, /tmp; xfs: /home; nVidia GeForce 2].

Compiling works, because I can compile some stuff. However, whenever I try to compile anything big (samba, kde... although xfree worked just fine...) I get a HUGE number of "Bad page state in..." and "Badness in"... errors, and the system hard locks. In fact, from the looks of the erros, syslog-ng dies in the middle of the error (take a look at the last line... it dies in the middle, or something. heh.).
- http://vlack.com/crashexample (log file, ONE instance of this error, 4.2MB)
- http://vlack.com/crashexample.gz (same thing, gzip'd, 110k)).

My CFLAGS are fairly conservative, but is the problem with them?

[NeddySeagoon]

vlack,

Its not a kernel issue - a lot of peole would have found it by now.
It may be a hardware issue. Boot the liveCD and run memtest.
It could be a temperature or PSU problem too. The system needs more power and gets hotter when its working hard.

Which version of ReiserFS are you using?
4 is only for the very brave.

Do you overclock? If so, go back to stock.

Your 1Gb of memory is two sticks of the same brand and speed rating?
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

[vlack]

- Oops, forgot to mention that. I actually did run memtest, and it found no errors in a single pass.
- My RAM is 2 256 sticks and 1 512 stick of mushkin brand RAM, of the same speed rating.
- edit: I can take out all but the 512 and try again, if you think I should, although if memtest said no errors, I don't know if that would be worth it or not. <shrug>.
- I am using reiser3
- I do not overclock.


I may be getting a new case soon, so that'll hopefully either solve the problem or eleminate the possibility of PSU/heat issues.

[robmoss]

Let memtest86 run for 10 passes. Most bad memory only shows errors after 4 or 5 passes; let it run for 10 just to be sure.

Also, try cpuburn to see if it's your CPU overheating.

Finally, don't forget to clear your ccache before each attempt, otherwise there's no point even bothering trying to fix the compile issues you're having. In fact, I'd be tempted to put "-ccache" in FEATURES (it's enabled by default, so you need to explicitly put it in there like that) until I'd fixed that problem.

Is all well with 2.6.5?
_________________
Reality is for those who can't face Science Fiction.

emerge -U will kill your Gentoo
ecatmur, Lord of Portage Bash Scripts

[vlack]

- I'll let memtest86 run all night tonight, and get back to you in the morning.
- Running cpuburn from Knoppix now; thanks for the tricky tip.
- I have temporarily disabled ccache too, although I didn't really think that this was a problem. It's not a "compile problem" per se (I don't think), because compiling some stuff still works. Whatever causes this problem, though, just seems to be triggered by compiles.
- 2.6.5-gentoo seems to work ok, but I switched back to vanilla 2.6.6 "just because". I'm planning on trying some different kernels later, when I'm not having bigger problems ...

[vlack]

OK, memtest did run all night, and it found errors Thanks for the help, I guess next time I'll let memtest run longer than just one pass. If I have more problems I'll report back.

EDIT: Also, the KDE build is humming along quite nicely , since I removed the 2 265 sticks. yay!

[til]

I have a similar problem:
BUT: I did memtest and my ram passed without any errors. Furthermore I can compile the packages (e.g. binutils) in Knoppix without problems.

Does anybody know what to do now? ATM I'm using love-sources-2.6.7-rc2-love2 - but the problem exists already since love-2.6.5

I don't believe, that it is a kernel prob - although I could compile with older kernels (2.6.3) without hard locks, etc.