When ld, called by collect2, crashes without a core dump, collect2 says collect2: ld terminated with signal 11 [Segmentation fault], core dumped But when I set "ulimit -c unlimited" and repeat the step, so that I get a core dump, collect2 says merely collect2: ld terminated with signal 11 [Segmentation fault] Release: 3.1 Environment: System: Linux linuix 2.4.18-4GB #1 Wed Mar 27 13:57:05 UTC 2002 i686 unknown Architecture: i686 host: i686-pc-linux-gnu build: i686-pc-linux-gnu target: sparc64-unknown-linux-gnu configured with: /packages2/gcc-3.1/configure --enable-shared --disable-nls --target=sparc64-linux --prefix=/cross/sparc64-linux-tools --with-local-prefix=/cross/sparc64-linux/local --with-as=/cross/sparc64-linux-tools/bin/sparc64-linux-as --with-gnu-as --with-ld=/cross/sparc64-linux-tools/bin/sparc64-linux-ld --with-gnu-ld --enable-languages=c How-To-Repeat: Give a link command line that makes ld crash. Once with "ulimit -c 0", once with "ulimit -c unlimited".
Fix: In collect2.c, function collect_wait, change status & 0200 ? "" : ", core dumped" to status & 0200 ? ", core dumped" : ""
From: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> To: bangerth@dealii.org, bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org, gcc-prs@gcc.gnu.org, ghazi@gcc.gnu.org, nobody@gcc.gnu.org Cc: Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: Mon, 24 Mar 2003 22:15:46 -0500 (EST) > From: bangerth@dealii.org > > Kaveh, you inserted the lines in question in this PR some > 3.5 years ago. Would you mind to take a brief look at this > report and its 1-line patch? > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6955 Eh? I don't recall filing or modifying this PR. How did you determine I was responsible? Anyway, the patch (or suggested change) in there looks technically correct to me. But I would instead do something like the patch below. (I'm not sure how to test it since I don't have code any handy that causes ld to crash.) --Kaveh 2003-03-24 Kaveh R. Ghazi <ghazi@caip.rutgers.edu> PR other/6955 * collect2.c (collect_wait): Use WCOREDUMP and fix output message. * system.h (WCOREDUMP, WCOREFLG): Define if necessary. diff -rup orig/egcc-3.2-CVS20030323/gcc/collect2.c egcc-3.2-CVS20030323/gcc/collect2.c --- orig/egcc-3.2-CVS20030323/gcc/collect2.c Fri Aug 16 16:02:01 2002 +++ egcc-3.2-CVS20030323/gcc/collect2.c Mon Mar 24 22:02:35 2003 @@ -1522,7 +1522,7 @@ collect_wait (prog) int sig = WTERMSIG (status); error ("%s terminated with signal %d [%s]%s", prog, sig, strsignal(sig), - status & 0200 ? "" : ", core dumped"); + WCOREDUMP(status) ? ", core dumped" : ""); collect_exit (FATAL_EXIT_CODE); } diff -rup orig/egcc-3.2-CVS20030323/gcc/system.h egcc-3.2-CVS20030323/gcc/system.h --- orig/egcc-3.2-CVS20030323/gcc/system.h Tue May 21 19:44:39 2002 +++ egcc-3.2-CVS20030323/gcc/system.h Mon Mar 24 21:59:44 2003 @@ -266,6 +266,12 @@ extern int errno; #ifndef WSTOPSIG #define WSTOPSIG WEXITSTATUS #endif +#ifndef WCOREDUMP +#define WCOREDUMP(S) ((S) & WCOREFLG) +#endif +#ifndef WCOREFLG +#define WCOREFLG 0200 +#endif /* The HAVE_DECL_* macros are three-state, undefined, 0 or 1. If they are defined to 0 then we must provide the relevant declaration
Responsible-Changed-From-To: unassigned->ghazi Responsible-Changed-Why: Kaveh, you inserted the lines in question in this PR some 3.5 years ago. Would you mind to take a brief look at this report and its 1-line patch? Thanks Wolfgang
From: Wolfgang Bangerth <bangerth@ticam.utexas.edu> To: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> Cc: bruno@clisp.org, <gcc-bugs@gcc.gnu.org>, <gcc-gnats@gcc.gnu.org>, <gcc-patches@gcc.gnu.org> Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: Tue, 25 Mar 2003 09:40:30 -0600 (CST) > > Kaveh, you inserted the lines in question in this PR some > > 3.5 years ago. Would you mind to take a brief look at this > > report and its 1-line patch? > > Eh? I don't recall filing or modifying this PR. How did you > determine I was responsible? I _made_ you responsible :-) You weren't in any legal sense, I just looked at the output of "cvs annotate" to find out that you were the one who checked in the lines of question. This is by now one of the less successless ways to get patches in PRs reviewed ;-) At any rate, if the patch goes in, drop me a note and I'll be happy to close the report. Thanks for the quick reply! Wolfgang ------------------------------------------------------------------------- Wolfgang Bangerth email: bangerth@ticam.utexas.edu www: http://www.ticam.utexas.edu/~bangerth/
From: Wolfgang Bangerth <bangerth@ticam.utexas.edu> To: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> Cc: bruno@clisp.org, <gcc-bugs@gcc.gnu.org>, <gcc-gnats@gcc.gnu.org>, <gcc-patches@gcc.gnu.org> Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: Tue, 25 Mar 2003 11:00:33 -0600 (CST) > > I _made_ you responsible :-) You weren't in any legal sense, I just looked > > at the output of "cvs annotate" to find out that you were the one who > > checked in the lines of question. > > I hate to be a pedant, but now honor is at stake. :-) Uh, this is getting out of hand -- I just wanted to get someone to review the patch... But then we both live in a country where the end seems to justify the means, so I would say I succeeded and my approach was justified ;-) (No WMD found while searching CVS, though.) And, by the way -- note my ingenious (and mostly unconscious) choice of words: I just said that you "checked in" these lines, not "broke" or "botched up"... > I honestly don't think I introduced this error. Looking back at the > cvs logs, it came in through one of the many infamous gcc2 merges. It > came in over four years ago: "Merge in gcc2 snapshot 19980929". Well, the line we are talking about is collect2.c:1511. This line has time stamp 1999-11-25, CVS version 1.87, which was your commit. Now, I acknowledge that if I had checked more carefully, then I'd have seen that your patch only preserved existing (wrong) behavior, but I haven't found a way to get the equivalent of cvs annotate of past versions to see who really introduced this. At any rate, I take back any non-intended blame, hereby fully rehabilitate your honor and assume all guilt myself. Let the patch go in! (I think it qualifies as "obvious".) Thanks Wolfgang ------------------------------------------------------------------------- Wolfgang Bangerth email: bangerth@ticam.utexas.edu www: http://www.ticam.utexas.edu/~bangerth/
From: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> To: bangerth@ticam.utexas.edu Cc: bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: Tue, 25 Mar 2003 11:43:25 -0500 (EST) > From: Wolfgang Bangerth <bangerth@ticam.utexas.edu> > > > > Kaveh, you inserted the lines in question in this PR some > > > 3.5 years ago. Would you mind to take a brief look at this > > > report and its 1-line patch? > > > > Eh? I don't recall filing or modifying this PR. How did you > > determine I was responsible? > > I _made_ you responsible :-) You weren't in any legal sense, I just looked > at the output of "cvs annotate" to find out that you were the one who > checked in the lines of question. This is by now one of the less > successless ways to get patches in PRs reviewed ;-) I hate to be a pedant, but now honor is at stake. :-) I honestly don't think I introduced this error. Looking back at the cvs logs, it came in through one of the many infamous gcc2 merges. It came in over four years ago: "Merge in gcc2 snapshot 19980929". http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/collect2.c.diff?r1=1.62&r2=1.63 -- Kaveh R. Ghazi ghazi@caip.rutgers.edu
From: Mike Stump <mrs@apple.com> To: Wolfgang Bangerth <bangerth@ticam.utexas.edu> Cc: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>, bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: Tue, 25 Mar 2003 12:59:09 -0800 On Tuesday, March 25, 2003, at 09:00 AM, Wolfgang Bangerth wrote: > Well, the line we are talking about is collect2.c:1511. This line has > time > stamp 1999-11-25, CVS version 1.87, which was your commit. Now, I > acknowledge that if I had checked more carefully, then I'd have seen > that > your patch only preserved existing (wrong) behavior, but I haven't > found a > way to get the equivalent of cvs annotate of past versions to see who > really introduced this. cvs diff will tell you. From there, cvs log and look at the comment and the date of the change, and from there, it's back the the ChangeLog entry. Also, the old gcc RCS files can be found on the web site, and can be used to track down things past an old gcc import.
From: Joern Rennecke <joern.rennecke@superh.com> To: Wolfgang Bangerth <bangerth@ticam.utexas.edu> Cc: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>, bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org Subject: cvs annotate (Was: Re: other/6955: collect2 says "core dumped" when there is no core) Date: Tue, 25 Mar 2003 20:32:55 +0000 > Well, the line we are talking about is collect2.c:1511. This line has time > stamp 1999-11-25, CVS version 1.87, which was your commit. Now, I > acknowledge that if I had checked more carefully, then I'd have seen that > your patch only preserved existing (wrong) behavior, but I haven't found a > way to get the equivalent of cvs annotate of past versions to see who > really introduced this. cvs annotate -r 1.86 repeat with previous version till you find the relevant change (or hit 1.1 , at which point you'll have to resort to the old rcs archives) -- -------------------------- SuperH (UK) Ltd. 2410 Aztec West / Almondsbury / BRISTOL / BS32 4QX T:+44 1454 465658
From: Mark Mitchell <mark@codesourcery.com> To: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> Cc: bangerth@ices.utexas.edu, bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org, gdr@integrable-solutions.net Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: 30 Mar 2003 11:19:06 -0800 > OK for 3.3 and 3.2? Yes for 3.3. -- Mark Mitchell CodeSourcery, LLC mark@codesourcery.com
From: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> To: bangerth@ices.utexas.edu Cc: bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org, gdr@integrable-solutions.net, mark@codesourcery.com Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: Sun, 30 Mar 2003 14:07:12 -0500 (EST) > From: Wolfgang Bangerth <bangerth@ticam.utexas.edu> > > Let the patch go in! (I think it qualifies as "obvious".) > Thanks > Wolfgang Well, "obvious" patches still should be tested somehow. I figured out a way. After compiling gcc, I replaced "ld" with a shell script like so, alternately replacing SEGV with TERM to control whether to create a core from "ld". > #!/bin/bash > kill -SEGV $$ I verified that the current compiler's error message is wrong, and the patch fixes it. So I went ahead and installed it on the mainline as "obvious". Since this PR is a regression (albeit a minor one regarding an error message,) I'd like to get it into the branches if possible. OK for 3.3 and 3.2? http://gcc.gnu.org/ml/gcc-patches/2003-03/msg02151.html Thanks, --Kaveh -- Kaveh R. Ghazi ghazi@caip.rutgers.edu
From: ghazi@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: other/6955 Date: 30 Mar 2003 19:01:58 -0000 CVSROOT: /cvs/gcc Module name: gcc Changes by: ghazi@gcc.gnu.org 2003-03-30 19:01:58 Modified files: gcc : ChangeLog system.h collect2.c Log message: PR other/6955 * collect2.c (collect_wait): Use WCOREDUMP and fix output message. * system.h (WCOREDUMP, WCOREFLG): Define if necessary. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=1.17253&r2=1.17254 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/system.h.diff?cvsroot=gcc&r1=1.142&r2=1.143 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/collect2.c.diff?cvsroot=gcc&r1=1.146&r2=1.147
From: Gabriel Dos Reis <gdr@integrable-solutions.net> To: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> Cc: bangerth@ices.utexas.edu, bruno@clisp.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org, mark@codesourcery.com Subject: Re: other/6955: collect2 says "core dumped" when there is no core Date: 30 Mar 2003 21:44:37 +0200 "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes: [...] | Since this PR is a regression (albeit a minor one regarding an error | message,) I'd like to get it into the branches if possible. | | OK for 3.3 and 3.2? Yes, it is OK. Thank you very much for your effort. -- Gaby
From: ghazi@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: other/6955 Date: 31 Mar 2003 01:12:21 -0000 CVSROOT: /cvs/gcc Module name: gcc Branch: gcc-3_3-branch Changes by: ghazi@gcc.gnu.org 2003-03-31 01:12:20 Modified files: gcc : ChangeLog system.h collect2.c Log message: PR other/6955 * collect2.c (collect_wait): Use WCOREDUMP and fix output message. * system.h (WCOREDUMP, WCOREFLG): Define if necessary. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.16114.2.383&r2=1.16114.2.384 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/system.h.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.133&r2=1.133.4.1 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/collect2.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.142&r2=1.142.4.1
From: ghazi@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: other/6955 Date: 31 Mar 2003 01:14:02 -0000 CVSROOT: /cvs/gcc Module name: gcc Branch: gcc-3_2-branch Changes by: ghazi@gcc.gnu.org 2003-03-31 01:14:02 Modified files: gcc : ChangeLog system.h collect2.c Log message: PR other/6955 * collect2.c (collect_wait): Use WCOREDUMP and fix output message. * system.h (WCOREDUMP, WCOREFLG): Define if necessary. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_2-branch&r1=1.13152.2.657.2.295&r2=1.13152.2.657.2.296 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/system.h.diff?cvsroot=gcc&only_with_tag=gcc-3_2-branch&r1=1.119.2.1.2.1&r2=1.119.2.1.2.2 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/collect2.c.diff?cvsroot=gcc&only_with_tag=gcc-3_2-branch&r1=1.133.6.1&r2=1.133.6.2
State-Changed-From-To: open->closed State-Changed-Why: Fixed here: http://gcc.gnu.org/ml/gcc-patches/2003-03/msg02151.html