Re: [RFC] [2/2] Get rid of current_gdbarch in tdeps

["Ulrich Weigand" <uweigand at de dot ibm dot com>]

Markus Deuling wrote:


> Tested on x86 and via gdb_mbuild.

It would be good if you could also test a build with --enable-targets=all
(and --enable-64-bit-bfd); this is a more thorough (and even faster!)
check that all targets still compile, compared to gdb_mbuild.

(gdb_mbuild only tests one subtarget per processor family.  
--enable-targets=all build all code required for all subtargets.)

> 	* alpha-tdep.c (alpha_heuristic_proc_start)
> 	(alpha_sigtramp_register_address): Add gdbarch as parameter. Replace
> 	current_gdbarch by gdbarch.
> 
> 	(alpha_heuristic_frame_unwind_cache): Use get_frame_arch to get at the
> 	current architecture by frame_info. Update alpha_heuristic_proc_start 
> 	call.
> 
> 	(alpha_sigtramp_frame_this_id, alpha_sigtramp_frame_prev_register): Use
> 	get_frame_arch to get at the current architecture by frame_info. Update
> 	alpha_sigtramp_register_address call.
> 
> 	* arm-tdep.c (thumb_scan_prologue): Add gdbarch as parameter and replace
> 	current_gdbarch by gdbarch. Update caller.
> 	(convert_to_extended, convert_from_extended): Add endianess parameter
> 	for comparison. Update caller.
> 	(arm_extract_return_value, arm_store_return_value): Use
> 	get_regcache_arch to get at the current	architecture.
> 
> 	* cris-tdep.c (cris_register_size): Add gdbarch as parameter. Replace
> 	current_gdbarch by gdbarch. Update caller.
> 	(cris_gdb_func, move_to_preg_op, none_reg_mode_move_from_preg_op): Add
> 	gdbarch as parameter. Update caller. Replace current_gdbarch by gdbarch.
> 
> 	* h8300-tdep.c (E_PSEUDO_CCR_REGNUM, E_PSEUDO_EXR_REGNUM, BINWORD): Add
> 	gdbarch	as parameter. Update caller.	
> 	(h8300_init_frame_cache): Add gdbarch as parameter. Replace
> 	current_gdbarch by gdbarch. Update caller.
> 
> 	* hppa-tdep.c (skip_prologue_hard_way): Add gdbarch as parameter and 
> 	update caller. Replace current_gdbarch by gdbarch.
> 
> 	* m32c-tdep.c (m32c_skip_trampoline_code): Use get_frame_arch to get at
> 	the current architecture. Replace current_gdbarch by gdbarch.
> 	* m68hc11-tdep.c (m68hc11_frame_unwind_cache): Likewise.
> 	(STACK_CORRECTION, USE_PAGE_REGISTER): Replace M6811_TDEP by its
> 	expression. Add gdbarch as parameter and replace current_gdbarch with
> 	it. Update caller.
> 	(M6811_TDEP): Remove.
> 	(m68hc11_frame_prev_register): Use get_frame_arch to get at the current
> 	architecture.
> 	(m68hc11_scan_prologue): Add gdbarch as parameter. Replace 
> 	current_gdbarch by gdbarch. Update caller.
> 
> 	* m68k-tdep.c (m68k_analyze_prologue): Add gdbarch as parameter and
> 	update caller.
> 	(m68k_analyze_register_saves): Likewise. Also replace current_gdbarch
> 	by gdbarch.
> 
> 	* rs6000-tdep.c (skip_prologue): Add gdbarch as parameter and update
> 	caller. Relace current_gdbarch by gdbarch.
> 	(altivec_register_p, spe_register_p): Likewise.
> 	* ppc-tdep.h (altivec_register_p, spe_register_p): Add gdbarch as
> 	parameter.
> 	* ppc-linux-nat.c (fetch_register, store_register): Update caller of
> 	altivec_register_p and spe_register_p.
> 
> 	* score-tdep.c (score_fetch_inst): Add gdbarch as parameter. Update 
> 	caller. Replace current_gdbarch by gdbarch.
> 	(score_analyze_prologue): use get_frame_arch to get at the current
> 	architecture.	
> 
> 	* sparc-tdep.h (sparc_analyze_prologue): Add gdbarch as parameter.
> 	* sparc-tdep.c (sparc_analyze_prologue): Likewise. Replace
> 	current_gdbarch by gdbarch. Update caller.
> 	(sparc_frame_cache): Use get_frame_arch to get at the current
> 	architecture.
> 	* sparce64-tdep.c (sparc64_skip_prologue): Update call of
> 	sparc_analyze_prologue.

I've noticed a couple of occasions where you introduce lines > 80
characters.  If you fix those, the patch is OK.

Thanks,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com