This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: frame_register_unwind(): "frame != NULL" assertion failure
- From: Kevin Buettner <kevinb at redhat dot com>
- To: Daniel Jacobowitz <drow at mvista dot com>, Kevin Buettner <kevinb at redhat dot com>
- Cc: Andrew Cagney <ac131313 at redhat dot com>, gdb at sources dot redhat dot com
- Date: Thu, 13 Feb 2003 14:35:27 -0700
- Subject: Re: frame_register_unwind(): "frame != NULL" assertion failure
- References: <1030213212349.ZM2427@localhost.localdomain> <20030213212904.GA14115@nevyn.them.org>
On Feb 13, 4:29pm, Daniel Jacobowitz wrote:
> Subject: Re: frame_register_unwind(): "frame != NULL" assertion failure
> On Thu, Feb 13, 2003 at 02:23:50PM -0700, Kevin Buettner wrote:
> > Andrew,
> >
> > Any ideas about what to do about this assertion failure? This is
> > occurring in a mips-linux native gdb built using current sources.
> > (Okay, they're a few hours old by now...) I was debugging gdb with
> > itself and was running to a breakpoint placed on main().
> >
> > #0 internal_error (
> > file=0x7a1a60 "/home/devel/kevinb/sourceware-mips64/src.baseline/gdb/frame.c", line=187, string=0x7a1aa0 "%s%sAssertion `%s' failed.")
> > at /home/devel/kevinb/sourceware-mips64/src.baseline/gdb/utils.c:800
> > #1 0x00577c94 in frame_register_unwind (frame=0x0, regnum=187,
> > optimizedp=0x7a1b20, lvalp=0x7fff68d4, addrp=0x7fff68d8,
> > realnump=0x7fff68e0, bufferp=0x7fff68c0)
> > at /home/devel/kevinb/sourceware-mips64/src.baseline/gdb/frame.c:187
> > #2 0x004f8534 in read_next_frame_reg (fi=0x7fff68c0, regno=29)
> > at /home/devel/kevinb/sourceware-mips64/src.baseline/gdb/mips-tdep.c:1592
>
> That backtrace must be inaccurate? read_next_frame_reg just passes fi
> to frame_register_unwind...
I'll try to get a better one. gdb was compiled with -O2 and I see:
outer-gdb> down
#2 0x004f8534 in read_next_frame_reg (fi=0x7fff68c0, regno=29)
at /home/devel/kevinb/sourceware-mips64/src.baseline/gdb/mips-tdep.c:1592
1592 frame_register_unwind (fi, regno, &optimized, &lval, &addr, &realnum,
outer-gdb> info address fi
Symbol "fi" is an argument in register s1.
I'll recompile without -O2 and post a new backtrace...
Kevin