why no $pc offset in debug CFA?

7 msg	Bootstrap failure for current gcc trunk on cygw...
1 msg	volunteer to help with Fortran
2 msg	Re: A recent patch increased GCC's memory consu...
1 msg	Program za knjigovodstvo sa PDV-om
7 msg	Where is gstdint.h
1 msg	visailu
1 msg	oopsie, bootstrap comparison failure
3 msg	maybe_infinite_loop?
3 msg	How do you get the benefit of -fstrict-aliasing?
1 msg	gcc-4.3-20070420 is now available
6 msg	GCC mini-summit - Patch tracker
5 msg	tuples: initial infrastructure
11 msg	Does vectorizer support extension?
7 msg	Re: assign numbers to warnings; treat selected ...
2 msg	[M32C] Incorrect Frame information generated
3 msg	arm,gcc and dsp instructions
8 msg	New option: -fstatic-libgfortran
63 msg	GCC mini-summit
3 msg	Problem building gcc on Cygwin

why no $pc offset in debug CFA?
\ DJ Delorie (19 Apr 2007)
. \ Eric Botcazou (19 Apr 2007)
. \ Richard Henderson (19 Apr 2007)

Subject:	why no $pc offset in debug CFA?
Group:	Gcc
From:	DJ Delorie
Date:	19 Apr 2007

In dwarf2out.c : dwarf2out_frame_init we have this code:

#ifdef DWARF2_UNWIND_INFO
if (DWARF2_UNWIND_INFO)
initial_return_save (INCOMING_RETURN_ADDR_RTX);
#endif

However, gdb really needs that slot to unwind stack frames when
debugging, on targets that don't use dwarf2 unwinding (i.e. cygwin
with -gdwarf-2, and some embedded targets that still use sjlj). Why
don't we unconditionally emit this?