Segfault in longjmp() ?
The machine is a SMP 3.0-RELEASE box. A heavily threaded program is segfaulting in the longjmp() function. Any ideas what would cause this? Regards, Dan To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-hackers" in the body of the message
Re: Segfault in longjmp() ?
> > The machine is a SMP 3.0-RELEASE box. > > A heavily threaded program is segfaulting in the longjmp() function. > Any ideas what would cause this? > > Regards, > > Dan > > You could have trashed your jmp_buf... (i.e. you're passing bad data to longjmp().) Just a thought... - Dave Rivers - To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-hackers" in the body of the message
Re: Segfault in longjmp() ?
Dan Moschuk wrote: > > The machine is a SMP 3.0-RELEASE box. > > A heavily threaded program is segfaulting in the longjmp() function. > Any ideas what would cause this? A bug in the thread exit code. This was fixed at the "13th hour" during the release of 3.2. -- John Birrell - j...@cimlogic.com.au; j...@freebsd.org http://www.cimlogic.com.au/ CIMlogic Pty Ltd, GPO Box 117A, Melbourne Vic 3001, Australia +61 418 353 137 To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-hackers" in the body of the message
Re: Segfault in longjmp() ?
Dan Moschuk wrote: > The machine is a SMP 3.0-RELEASE box. > > A heavily threaded program is segfaulting in the longjmp() function. > Any ideas what would cause this? Libc_r or Linuxthreads? There are some known problems WRT signal handling in libc_r. I think they've been mostly fixed (with possible exception of signals interrupting signal handlers) with the changes at: ftp://ftp.pcnet.com/users/eischen/FreeBSD/uthread.tgz If you're having problems with libc_r and these changes don't solve your problem, let me know. Make sure you do a 'make clean' in src/lib/libc_r before building. Dan Eischen eisc...@vigrid.com To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-hackers" in the body of the message
