Return to BSD News archive
Path: sserve!manuel!munnari.oz.au!mips!mips!swrinde!cs.utexas.edu!uunet!walter!arran!rwu From: rwu@arran.bellcore.com (Ronald W Underwood) Newsgroups: comp.unix.bsd Subject: Re: Mysterious Compiler Lock-Ups Message-ID: <rwu.713544059@arran> Date: 11 Aug 92 14:40:59 GMT References: <1992Aug11.142410.13342@ra.msstate.edu> Sender: news@walter.bellcore.com Organization: Bellcore Lines: 74 Nntp-Posting-Host: arran.bellcore.com cctony@sun1.mcsr.olemiss.edu (Tony Reynolds) writes: >Here's the rough breakdown of the story: >I was going to work on the Trident chipset driver for X386-1.2E, so >I found some more disk space and grabbed all the sources. After >unpacking and applying the patches, I ran make and everything started >to compile fine. Too bad it didn't finish, though. It died just >as it was compiling XStrKeysym.c in Xlib. It dies *every time* it >tries to compile this particular file. I can move the file around >in the list of sources in the Imakefile, all to no avail. When I >say, ``it died,'' I mean the entire computer locked up. Eventually, >it rebooted on its own. >I have tried every stupid combination of tricks to get this to work. >I have tried manipulating the set-up parameters of the Opti chipset >that's in the machine, and I've tried the two kernel bugfixes. >On the other hand, if I change to mit/lib/X and type make , >then it will compile and not lock up the machine. >I have no clue as to the reason this dies. I can rebuild most other >things at will: I've remade several kernels, EMACS 18.58, some printer >utilities, so I can't see it being a case of operator headspace. >My configuration: >486/33, Opti Chipset (Local Bus, but nothing in the slot) >8MB RAM >300 MB Micropolis HD >160 MB CDC/Seagate HD >Trident VGA card >Nothing un-generic about this thing. >So, thanks in advance for any thoughts you might have about my >problem. I'm gonna save this article and use it as the basis for >a BUGNFIX... even though there's no fix as of yet. >-- >Tony Reynolds Have you hugged cctony@sun1.mcsr.olemiss.edu >Cray Network Intern your Washburn today? (601)232-7206 I, too, had this problem. What was happening to me was that I was running out of swap space (I believe the stock system only has around 5meg of swap). I built a new kernel that expected swap space on a second disk which I partitioned and disklabeled with a 32 meg swap partition and things worked a lot better. Apparently the system doesn't detect that it's out of swap but just keeps thrashing around and eventually locks up tight. Another thing I discovered in building X386 is that you can exhaust the default datasize and stacksize allocations. You get "Out of virtual memory" messages from the compiler in some spots. As root, just type the following at your csh: "limit datasize unlimited" "limit stacksize unlimited" "limit maxproc unlimited" "limit openfiles unlimited" Then start the make again. Hope this helps. Ron Underwood Distributed Systems Research Group rwu@bellcore.com (201) 829-3317 Bell Communications Research 435 South St. Room 2D-392 Morristown, NJ 07960-1961 -- Ron Underwood Distributed Systems Research Group rwu@bellcore.com (201) 829-3317