This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: RFC: next/finish/etc -vs- exceptions
- From: Tom Tromey <tromey at redhat dot com>
- To: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Wed, 01 Dec 2010 14:40:25 -0700
- Subject: Re: RFC: next/finish/etc -vs- exceptions
- References: <m38w2a236l.fsf@fleche.redhat.com> <20101125075847.GA19270@host0.dyn.jankratochvil.net> <m3fwuisqxb.fsf@fleche.redhat.com>
>>>>> "Tom" == Tom Tromey <tromey@redhat.com> writes:
Jan> BTW the testcase does not work on neither ppc32 nor on ppc64.
Tom> Thanks, I will investigate.
On PPC32 the problem was just that the test case was not robust against
differences between compilers on different arches. In particular,
different versions of gcc seem to give different (sometimes even quite
weird) line numbers to the landing point of an exception.
I fixed this by rewriting the test case to be more robust against these
kinds of differences.
On PPC64, as you surmised, the problem was related to function
descriptors. In particular, create_exception_master_breakpoint
was calling lookup_minimal_symbol_text -- but that failed on PPC64.
FWIW I expect that the longjmp code also has a problem here, but again
probably masked by glibc PC mangling.
I think I've addressed all the comments. I'll send an updated patch
tomorrow.
Tom