[Crash-utility] [PATCH] s390: Fix stack trace code for program checks

Dave Anderson anderson at redhat.com
Wed May 4 14:13:07 UTC 2011



----- Original Message -----
> Hello Dave,
> 
> On Wed, 2011-05-04 at 09:56 -0400, Dave Anderson wrote:
> > Hi Michael,
> >
> > A couple minor points about this patch...
> >
> > I moved the two new offset_table entries to the end of
> > the structure. Since the offset_table is exported to extension
> > modules, changing members in the middle of the structure could
> > easily break a module that was compiled against an earlier
> > version. And I added displays of the new offset_table and
> > size_table entries in dump_offset_table() used by "help -o".
> 
> Fine, thanks!
> 
> > I'm not particularly enamored with the use of the zero-filled
> > "double-number" frame fields -- mainly because it's different
> > than all of the other architectures (even the s390).
> > The other arches just move the asterisk one field to the left if the
> > frame
> > number goes into double figures, keeping the remaining fields
> > lined up:
> >
> >         fprintf(fp, "%s#%d [ ...
> >              level < 10 ? " " : "",
> >
> > While I try not to interfere with architecture maintainers,
> > is there a particular reason you want to do it that way?
> 
> No problem. This is also fine with me. Should I change this or do you
> want to do that?
> 
> Michael

I just changed print_frame() to do it.  And, sorry -- I didn't mean
to twist your arm...   ;-)

Queued for 5.1.15.

Thanks,
  Dave



More information about the Crash-utility mailing list