Re: [PATCH/i386newframe] Add asssertion to CFI frame unwinder

[Elena Zannoni <ezannoni@redhat.com>]

Elena Zannoni writes:
 > Mark Kettenis writes:
 >  > This assertion can unfortunately be triggered.  Don't know how to fix
 >  > the bug yet, but this makes clear there is a bug.
 >  > 
 >  > Mark
 >  > 
 >  > Index: ChangeLog
 >  > from  Mark Kettenis  <kettenis@gnu.org>
 >  >  
 >  > 	* dwarf-frame.c (dwarf2_frame_cache): Assert that we have a FDE.
 >  > 
 >  > Index: dwarf-frame.c
 >  > ===================================================================
 >  > RCS file: /cvs/src/src/gdb/Attic/dwarf-frame.c,v
 >  > retrieving revision 1.1.2.6
 >  > diff -u -p -r1.1.2.6 dwarf-frame.c
 >  > --- dwarf-frame.c 23 May 2003 20:18:32 -0000 1.1.2.6
 >  > +++ dwarf-frame.c 29 May 2003 15:47:12 -0000
 >  > @@ -483,6 +483,7 @@ dwarf2_frame_cache (struct frame_info *n
 >  >  
 >  >    /* Find the correct FDE.  */
 >  >    fde = dwarf2_frame_find_fde (&fs->pc);
 >  > +  gdb_assert (fde != NULL);
 >  >  
 >  >    /* Extract any interesting information from the CIE.  */
 >  >    fs->data_align = fde->cie->data_alignment_factor;
 > 
 > 
 > I am running into this in several cases running the testsuite on x86.
 > For instance for me callfuncs.exp just doesn't do anything, because
 > the assertion is triggered at startup.
 > what exactly is happening? 
 > 
 > thanks
 > elena

Well, this is a problem with the infimous separate debug info. The
.debug_blah sections are in one objfile (the .debug one), while the
.eh_frame is in another (the stripped executable). However, there is
an empty .eh_frame in the .debug file as well. This one is the one
that screws things up. I'll post a patch separately.

elena