This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 4/8] gdb.trace: Use manually-defined start labels in unavailable-dwarf-piece.exp
- From: "Ulrich Weigand" <uweigand at de dot ibm dot com>
- To: koriakin at 0x04 dot net (Marcin KoÅcielnicki)
- Cc: gdb-patches at sourceware dot org
- Date: Wed, 9 Mar 2016 18:35:06 +0100 (CET)
- Subject: Re: [PATCH 4/8] gdb.trace: Use manually-defined start labels in unavailable-dwarf-piece.exp
- Authentication-results: sourceware.org; auth=none
Marcin KoÅ?cielnicki wrote:
> On 09/03/16 18:24, Ulrich Weigand wrote:
> > Marcin KoÅ?cielnicki wrote:
> >
> >> On powerpc64, foo/bar point to a function descriptor, not to function code.
> >> Since there are no global labels pointing at the actual function code,
> >> let's make our own.
> >
> >> with_test_prefix "tracing foo" {
> >> - gdb_test "trace foo" ".*"
> >> + gdb_test "trace *foo_start_lbl" ".*"
> >
> > Well, I'd have thought that "trace foo" should simply work as-is,
> > otherwise, this isn't particularly user-friedly ...
> >
> > Is there a call to gdbarch_convert_from_func_ptr_addr missing
> > somewhere in the trace code, maybe?
>
> Don't worry, trace foo does work, but not for this particular test,
> because of the way the fake DWARF is emitted: before, the DWARF covered
> area from foo to foo_end_lbl, and thus a tracepoint set on foo worked
> just fine. However, now the DWARF covers foo_start_lbl to foo_end_lbl -
> and function entry point (which is where the tracepoint is) is not covered.
Ah, I see.
The patch is OK, then.
Thanks,
Ulrich
--
Dr. Ulrich Weigand
GNU/Linux compilers and toolchain
Ulrich.Weigand@de.ibm.com