home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.questions:13853 comp.unix.pc-clone.32bit:505
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!destroyer!ncar!vexcel!copper!slate!mbarkah
- From: mbarkah@slate.mines.colorado.edu (Ade Barkah)
- Newsgroups: comp.unix.questions,comp.unix.pc-clone.32bit
- Subject: Re: need to make a program dump core on a signal
- Message-ID: <1992Nov23.181302.64139@slate.mines.colorado.edu>
- Date: 23 Nov 92 18:13:02 GMT
- References: <1992Nov23.145755.20779@nervous.com>
- Sender: mbarkah@slate.mines.colorado.edu (Ade Barkah)
- Followup-To: comp.unix.pc-clone.32bit
- Organization: Colorado School of Mines
- Lines: 30
-
- pizzi@nervous.com (Riccardo Pizzi) writes:
- : I'm trying to make a program dump core upon reception of SIGUSR1
-
- : Unfortunately, the abort(3) call doesn't seem to work.
- : Yes, I verified that:
- : 1) the current working dir is writable
- : 2) uid and euid of the program are the same
- : 3) there is no shell-imposed coresize limit
- : Still the abort() call doesn't create a core dump.
- :
- : Is there another way to make a program dump core?
-
- You also need to check if the kernel parameters allow core dumps.
- In svr4 this would be the S/HCORLIM setting. A SCORLIM setting of 0
- prevents the creation of core dumps.
-
- Otherwise, you may need to manually cause a dump. Replace abort()
- with a delayer, and you can come along and issue a kill -x statement
- to the process, where x being any one of those signals which
- causes core to be dumped (SIGILL or SIGSEGV or something.)
- Your system or compiler or shell might be trapping or ignoring
- SIGABRT. (Or, you can use signal/raise to simulate the other
- core-dumping signals.)
-
- -Regards,
-
- -Ade.
- --
- Internet : mbarkah@slate.mines.colorado.edu (NeXT Mailable)
- CompuServe: 74160,3404
-