Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!bunyip.cc.uq.oz.au!harbinger.cc.monash.edu.au!yeshua.marcam.com!usc!howland.reston.ans.net!gatech!pitt!w2xo!durham From: durham@w2xo.pgh.pa.us (Jim Durham) Newsgroups: comp.os.386bsd.bugs Subject: gcc bug on FreeBSD replies posting Keywords: replies FreeBSD gcc bug Message-ID: <11@w2xo.pgh.pa.us> Date: 16 Dec 93 04:11:18 GMT Lines: 94 Here is my original posting about the gcc bug. >Gcc crashes with signal 10 and signal 11. These are "BUS ERROR" >and "SEGMENTATION VIOLATION". There seems to be no logical pattern to >these occurences. Sometimes , while compiling a particular C >source file, gcc will die with signal 10, the, upon trying again on the >same source file, it will die with signal 11, then , on another try, it >will compile. Sounds like an unitialized pointer somewhere in gcc, for >a guess. Anyone else had this problem? And here are the answers I received. Thanks to all who replied. (Lots of stuff deleted here, the answers and enough context to figure out what is going on (hopefully) remain......... ****************************************************************************** From: csshah@sunvis1.vislab.olemiss.edu (Viren R. Shah) I'm having the same problem with gcc (i'm running FreeBSD-current). I figure that my sources got messed up when i tried to upgrade -- and now some of it is prob. with shared libs (thats a wild guess) -- i checked all the binaries but they seem to be ok. So, i haven't a clue :-( From: combssf@salem.ge.com (Stephen F. Combs) Jim, I had the same kind of problems (random sig 10's & 11's) and went bannanas trying to find the problem. It turned out to be a defective cache chip on my motherboard (If you can, try disabling cache and then rerunning the compile(s). When I did that everything ran to completion, albet a tad slow!). When I sent the m/b back to the manufacturer, they at first said there was NO problem (upon questioning, they had used a DOS based memory testing program which DIDN'T fully exersize the cache). When they put the m/b on an IN-Circuit board tester, the cache failed big-time! From: csshah@sunvis1.vislab.olemiss.edu (Viren R. Shah) > >Hello Viren, > >I got two answers back so far on the gcc problem. Both say that >this occurs with a 4 meg system but goes away at 8 megs. > thanx. However, since the system i'm running FreeBSD on has 16M of RAM, i doubt that that's the reason my system isn't working :-( I asked on the freebsd-mailing list, and no one seems to be able to figure out my problem, so as a last resort i'm prob. gonna reinstall the binaries and source trees (hoping that the prob. was caused by something that i did wrong when updating the sources -- tho i doubt that. >Now, since BSD has virtual memory and demand paging, I wonder >why that would be? > It's a bug in the FreeBSD. i think are going to fix it in the next release, but you might wanna ask around. From max.IN-Berlin.DE!berry@methan.chemie.fu-berlin.de Mon Dec 13 03:25:49 1993 On a 4MB machine of mine I have also seen this problem. After upgrading this machine to 16MB all went well. My other 16MB and 8MB machines never had problems like this (i486 and i386+i387 machines). -- Stefan (berry@max.IN-Berlin.DE) From cg@FIMP01.fim.uni-linz.ac.at Wed Dec 15 01:09:21 1993 Hello Jim! I think it's a bug and not a feature of gcc and/or the OS. This bug occured the first time to me when I compiled gcc 1.40 (with Interactive's ISC 3.2). I think it's a bug in the intel part of the compiler, as all versions of gcc work on our ULTRIX machines; but we have troubles on all PCs with *BSD. If you'll receive further information about this bug, please let me know about it! Many thanks, Christian. From: jkh@whisker.lotus.ie (Jordan K. Hubbard) Are you running on a 4MB system? FreeBSD 1.0 has known problems with 4MB systems. Your solution is to wait for 1.1, or get another 4MB of memory! :-) Jordan From: nate@bsd.coe.montana.edu (Nate Williams) How much memory do you have? If you have more than 4MB, this sounds like a bad motherboard/cache problem.