cc1 segfaults NEW

cc1 segfaults NEW

Post by Studnitzky Boa » Wed, 12 Jun 1996 04:00:00



This is another thread in the memory woes cc1 sig segvs.

I have been using Linux for a long long time now, my current Pentium
has been running it for over six months non-stop. Never had a problem.
I always use the latest kernel (2.0.0 now). Until recently I never
had a problem compiling or anything. However, about two weeks ago, I tried
compiling a new kernel and got my first segv. I managed to compile the kernel
after some re-runs.
The problem is persistent, in all compilations, i get at least 5 segvs
during the process. I must point out that I have been using the same
 board/memory/cache for over six months without a trace of a problem.

At some point I started getting signal 6 when compiling sbpcd.o, this
made me think that it may actually be a kernel problem, a bug in the pipe or
something like that. After one of the segvs, the kernel got crazy,
started dumping numbers ([<00xxxxxx>] ...) and ran havok through my disk drive.
I removed the -pipe option from the gcc command line, and guess what-
no more segvs. I don't know what it fixed but take note.

I would normally say that  this is a memory problem, a faulty simm (which
i am planning on replacing tomorrow), but, my system usually runs under
very heavy load, X + netscape + emacs + xv + network +.... and i never
ever get any problems with it. only gcc fails.

I'll have a look at kernel sources myself, but i'm no kernel hacker, so...

If you have anything to say, (and don't tell me it's a faulty sim or cache
or timing, cause i played with those),

Boaz Studnitzky

 
 
 

1. dvips segfaults with new kernel

I had a Linux Slackware installation, the one obtainable
by ftp from sunsite.unc.edu in Oct./Nov. '95 (Slackware 3.1?).
I.e., kernel 1.2.13, libc.so.5.0.9, libm.so.5.0.0, ld.so.1.7.3,
Elf system.
Recently I upgraded to kernel 2.0.30, libc.so.5.4.33, libm.so.5.0.9,
ld(-linux).so.1.8.10. The links and file placements are correct
according to the ELF-HOWTO. However, I have kept the LaTeX version
from the previous installation, including dvips.

And when I now runs dvips, it says:
     This is dvipsk 5.58a (and so on)
     Segmentation fault
Rebooting with the old kernel, however, I can still use the old
dvips (although of course other things like the new modules do not
work with the old kernel).

How comes?
And can I do anything to avoid the segfault, other than installing
LaTeX/dvips anew?

:)  Ingjald Finjord

2. ftp -s

3. New: smbfs segfault when mounting mountpoint that is in use(?)

4. error path deallocation in ibmtr.c (246ac5)

5. New: loading emu10k1-gp segfaults

6. Is slackware changing?

7. New: Zip drive parallel-port driver causes segfault in 2.5.x

8. Help: How is PATH set?

9. Segfault with "new"

10. X server segfaults when kernel compiled with new GCC 2.8.0, previously worked OK

11. gcc reports inernal compiler error: cc1 got fatal signal 11 always!?!

12. Kernel compile fails: cc1 internal compiler error ?

13. help: gcc first phase (cc1) output format