Re: optimization/9015: bc segfaults when compiled with optimization

[Patrick Smith <patsmith@pobox.com>]

The following reply was made to PR optimization/9015; it has been noted by GNATS.

From: Patrick Smith <patsmith@pobox.com>
To: bangerth@dealii.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org,
	nobody@gcc.gnu.org, patsmith@pobox.com, gcc-gnats@gcc.gnu.org
Cc:  
Subject: Re: optimization/9015: bc segfaults when compiled with optimization
Date: Sat, 21 Dec 2002 00:51:59 -0500

 bangerth@dealii.org wrote:
 > Synopsis: bc segfaults when compiled with optimization
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: bangerth
 > State-Changed-When: Fri Dec 20 19:20:49 2002
 > State-Changed-Why:
 >     It is very hard for us to reproduce a problem with such little
 >     information, since we don't really know what we're looking
 >     for. Sometimes such problems go away by moving to a different
 >     platform, and then we would not be able to reproduce it.
 >     Basically, a whole program is too large for us to find a
 >     bug in.
 >     
 >     What you could do is the following: try to run bc under
 >     a debugger and try to find out which function it crashes in
 >     by looking at a stacktrace after the crash happens. Then
 >     try to isolate the function in which this happens from
 >     the rest of the program and see whether you can come up
 >     with a small testcase from which we can work on. This
 >     would ideally just include the one function in which things
 >     happen, called from a dummy main function.
 >     
 >     Thanks
 >       Wolfgang
 > 
 > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=9015
 
 As mentioned in the original report, I think the problem is with the 
 function load_code in the source file included in the bug report.  If 
 that function is compiled without optimization and the rest of bc is 
 optimized with -O2, the crash does not occur.
 
 I've already been the debugger and trying to create a small test program 
 route, without much luck.  The crash occurs in another function (addbyte 
 in the same source file).
 
 I'm currently reading up on PowerPC assembly, so I can look at the 
 generated code for load_code and figure out if it's right or not.  If 
 you want, I can send the .s file as well.
 -- 
 patsmith@pobox.com