From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0rc2) Gecko/20020510 Description of problem: When installing or compiling programs with such options as "configure", "make", "make install" or "gcc x -o x", I get segmentation faults. I would have just said it was gcc, but make is powered by perl, so I don't know which is having the problem. I have tried running simple programs and complex ones. I sometimes get the segmentation fault very randomly. I recieved it once when rebooting and the if-cfg script generated a segmentation fault, the network card did not come up. However when I logged in I started /etc/init.d/network and it faulted, and I did it once more and it worked just fine. Even while using the uptime command it is random. I wasn't really bothered until when trying to compile apache I couldn't even get through the configure script without it faulting. I was able to get through the configure script of openssl, but could not run the makefile with any success. Version-Release number of selected component (if applicable): kernel-2.4.21-1.1931.2.349.2.2.ent How reproducible: Sometimes Steps to Reproduce: 1.Works everytime I try to configure httpd-2.0.46 2. unzip and cd 3. ./configure (randomly get segmentation faults, at different points.) 4. I have not reinstalled the AS edition 5. in other situations (running uptime, it is completly random) Actual Results: sometime it works and sometimes it stops with segmentation faults all the way. Expected Results: I would expect to never get a seg fault from a binary like uptime, I usually don't have problems with configure either. Additional info: system dose not crash. no cores are dumped. sometimes outputs message (from configure script) that C compiler could not create executables, but this is false, because I have created c and c++ executables on this same box, before and after this error message.
is there anything usefull in dmesg or the logs ?
Created attachment 93561 [details] says kernel bug at line 158 fault.c output of /var/log/messages sorry I didn't send this sooner, was checking email, not bugzilla. never did this before, getting used to it :) thanks, -sam
Created attachment 93562 [details] looked at the fault.c and realized that I didn't give enough info looking at the fault.c code, it writes a printk() messages, this messgage is exactly the same for all instances when it appears to crash. keeps saying that "invalid kernel-mode pagefault 0! [addr:00000000, eip:021389ef]" I included the whole messages file for you, active since the computer has been up for now 5 days. thanks, -sam. ps happy to provide any more information that you request.
this looks like the fingerprint of a crash we already fixed; a fixed kernel is available for the RHEL beta via RHN