[rfc] [05/09] Get rid of current_gdbarch (macros in mn10300 target)

[rfc] [05/09] Get rid of current_gdbarch (macros in mn10300 target)

[Markus Deuling]

[-- Attachment #1: Type: text/plain, Size: 466 bytes --]

Hi, 

this patch replaces AM33 macro from n10300 target and then replaces current_gdbarch
by appropriate methods. 

Tested building with --target=mn10300. Ok to commit?


ChangeLog:

	* mn10300-tdep.c (set_reg_offsets,mn10300_analyze_prologue): Replace
	AM33 by its expression. Use get_frame_arch to get at the current
	architecture by frame_info.
	* mn10300-tdep.h (AM33_MODE): Remove.


-- 
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com





[-- Attachment #2: diff-mn10300 --]
[-- Type: text/plain, Size: 1420 bytes --]

diff -urpN src/gdb/mn10300-tdep.c dev2/gdb/mn10300-tdep.c
--- src/gdb/mn10300-tdep.c	2007-11-05 05:32:22.000000000 +0100
+++ dev2/gdb/mn10300-tdep.c	2007-11-05 09:10:03.000000000 +0100
@@ -328,7 +328,7 @@ set_reg_offsets (struct frame_info *fi, 
 
   trad_frame_set_this_base (cache, base);
 
-  if (AM33_MODE == 2)
+  if (gdbarch_tdep (get_frame_arch (fi))->am33_mode == 2)
     {
       /* If bit N is set in fpregmask, fsN is saved on the stack.
 	 The floating point registers are saved in ascending order.
@@ -384,7 +384,7 @@ set_reg_offsets (struct frame_info *fi, 
       trad_frame_set_reg_addr (cache, E_D2_REGNUM, base + offset);
       offset += 4;
     }
-  if (AM33_MODE)
+  if (gdbarch_tdep (get_frame_arch (fi))->am33_mode)
     {
       if (movm_args & movm_exother_bit)
         {
@@ -603,7 +603,7 @@ mn10300_analyze_prologue (struct frame_i
 	goto finish_prologue;
     }
 
-  if (AM33_MODE == 2)
+  if (gdbarch_tdep (get_frame_arch (fi))->am33_mode == 2)
     {
       /* Determine if any floating point registers are to be saved.
 	 Look for one of the following three prologue formats:
diff -urpN src/gdb/mn10300-tdep.h dev2/gdb/mn10300-tdep.h
--- src/gdb/mn10300-tdep.h	2007-08-23 20:08:36.000000000 +0200
+++ dev2/gdb/mn10300-tdep.h	2007-11-05 09:10:12.000000000 +0100
@@ -77,4 +77,3 @@ struct gdbarch_tdep
   int am33_mode;
 };
 
-#define AM33_MODE (gdbarch_tdep (current_gdbarch)->am33_mode)

Re: [rfc] [05/09] Get rid of current_gdbarch (macros in mn10300 target)

[Ulrich Weigand]

Markus Deuling wrote:

> 	* mn10300-tdep.c (set_reg_offsets,mn10300_analyze_prologue): Replace
> 	AM33 by its expression. Use get_frame_arch to get at the current
> 	architecture by frame_info.
> 	* mn10300-tdep.h (AM33_MODE): Remove.

Sorry, but this is not OK.  Unfortunately, the mn10300_analyze_prologue
routine is called both from within the unwinder (where we always have
a frame), *and* from mn10300_skip_prologue, where frame information is
not available.  Thus "fi" may actually be NULL.

Bye,
Ulrich

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

Re: [rfc] [05/09] Get rid of current_gdbarch (macros in mn10300 target)

[Markus Deuling]

[-- Attachment #1: Type: text/plain, Size: 1062 bytes --]

Ulrich Weigand schrieb:
> Markus Deuling wrote:
> 
>> 	* mn10300-tdep.c (set_reg_offsets,mn10300_analyze_prologue): Replace
>> 	AM33 by its expression. Use get_frame_arch to get at the current
>> 	architecture by frame_info.
>> 	* mn10300-tdep.h (AM33_MODE): Remove.
> 
> Sorry, but this is not OK.  Unfortunately, the mn10300_analyze_prologue
> routine is called both from within the unwinder (where we always have
> a frame), *and* from mn10300_skip_prologue, where frame information is
> not available.  Thus "fi" may actually be NULL.

Ok, I see. Your're right. I attached a patch which just replaces AM33_MODe macro to
get rid of current_gdbarch in mn10300-tdep.h file. Later on I'll come up with a patch
to get rid of the newly introduced current_gdbarch's in mn10300-tdep.c.

Tested by building with --target=mn10300. Ok to commit?

ChangeLog:

	* mn10300-tdep.c (set_reg_offsets,mn10300_analyze_prologue): Replace
	AM33 by its expression.
	* mn10300-tdep.h (AM33_MODE): Remove.


-- 
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com


[-- Attachment #2: diff-mn10300 --]
[-- Type: text/plain, Size: 1401 bytes --]

diff -urpN src/gdb/mn10300-tdep.c dev/gdb/mn10300-tdep.c
--- src/gdb/mn10300-tdep.c	2007-11-05 05:32:22.000000000 +0100
+++ dev/gdb/mn10300-tdep.c	2007-11-06 10:37:39.000000000 +0100
@@ -328,7 +328,7 @@ set_reg_offsets (struct frame_info *fi, 
 
   trad_frame_set_this_base (cache, base);
 
-  if (AM33_MODE == 2)
+  if (gdbarch_tdep (current_gdbarch)->am33_mode == 2)
     {
       /* If bit N is set in fpregmask, fsN is saved on the stack.
 	 The floating point registers are saved in ascending order.
@@ -384,7 +384,7 @@ set_reg_offsets (struct frame_info *fi, 
       trad_frame_set_reg_addr (cache, E_D2_REGNUM, base + offset);
       offset += 4;
     }
-  if (AM33_MODE)
+  if (gdbarch_tdep (current_gdbarch)->am33_mode)
     {
       if (movm_args & movm_exother_bit)
         {
@@ -603,7 +603,7 @@ mn10300_analyze_prologue (struct frame_i
 	goto finish_prologue;
     }
 
-  if (AM33_MODE == 2)
+  if (gdbarch_tdep (current_gdbarch)->am33_mode == 2)
     {
       /* Determine if any floating point registers are to be saved.
 	 Look for one of the following three prologue formats:
diff -urpN src/gdb/mn10300-tdep.h dev/gdb/mn10300-tdep.h
--- src/gdb/mn10300-tdep.h	2007-08-23 20:08:36.000000000 +0200
+++ dev/gdb/mn10300-tdep.h	2007-11-06 10:38:52.000000000 +0100
@@ -77,4 +77,3 @@ struct gdbarch_tdep
   int am33_mode;
 };
 
-#define AM33_MODE (gdbarch_tdep (current_gdbarch)->am33_mode)

Re: [rfc] [05/09] Get rid of current_gdbarch (macros in mn10300 target)

[Ulrich Weigand]

Markus Deuling wrote:

> Ok, I see. Your're right. I attached a patch which just replaces AM33_MODe macro to
> get rid of current_gdbarch in mn10300-tdep.h file. Later on I'll come up with a patch
> to get rid of the newly introduced current_gdbarch's in mn10300-tdep.c.

I don't see how this patch alone really helps much; we don't we wait for the
real solution and check it in in one go?

Bye,
Ulrich

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