Re: patch to omit the frame pointer from leaf functions (version 2)

[John Wehle <john at feith dot com>]

> Check out FRAME_POINTER_REQUIRED for the x86 though.  I'd bet it claims that
> a frame pointer is needed for non-leaf functions.
> 
> I don't know about the m88k.

You are right.  In both cases FRAME_POINTER_REQUIRED checks leaf_function_p.

>>      What targets (other the i386) are unabled to debug code
>>      which has been optimized by this patch?
> h8, m68k, and probably just about every port that does not define
> ACCUMULATE_OUTGOING_ARGS.

Maybe the thing to do is to fix gdb so it handles frameless leaf functions
for the i386, make the optimization dependent on CAN_DEBUG_LEAF_WITHOUT_FP,
and have i386.h define this new macro.

>> *** gcc/config/pa/pa.c.ORIGINAL	Sun Oct 11 13:57:29 1998
>> --- gcc/config/pa/pa.c	Sat Oct 17 01:05:22 1998
>> *************** basereg_operand (op, mode)
>> *** 5710,5716 ****
>>     /* While it's always safe to index off the frame pointer, it's not
>>        always profitable, particularly when the frame pointer is being
>>        eliminated.  */
>> !   if (! flag_omit_frame_pointer && op == frame_pointer_rtx)
>>       return 1;
>>   
>>     /* The only other valid OPs are pseudo registers with
>> --- 5710,5717 ----
>>     /* While it's always safe to index off the frame pointer, it's not
>>        always profitable, particularly when the frame pointer is being
>>        eliminated.  */
>> !   if (! (flag_omit_frame_pointer || current_function_sp_is_unchanging)
>> !       && op == frame_pointer_rtx)
>>       return 1;
>
> This is undesirable.  The reason we don't allow frame pointers as a base reg
> is because we get worse code when the frame pointer is eliminated to the
> stack pointer.

My change to reload1.c causes the frame pointer to be eliminated whenever 
current_function_sp_is_unchanging is set.  The comment says that using
the frame pointer as a base reg in this situation is undesirable.  My
change to pa.c was to realize that the frame pointer is eliminated if
either flag_omit_frame_pointer or current_function_sp_is_unchanging is
set so using the frame pointer as a base register should only be done if
! (flag_omit_frame_pointer || current_function_sp_is_unchanging).
Did I miss understand something?

Note that the issue is probably moot since:

  1) pa defines CAN_DEBUG_WITHOUT_FP which means that
     flag_omit_frame_pointer is set anytime that optimization
     is being used and current_function_sp_is_unchanging
     is only set when life_analysis_1 runs which means
     that optimization is being used.

  2) The optimization will probably end up being conditional
     on CAN_DEBUG_LEAF_WITHOUT_FP which means that it will
     not initially affect any port besides the i386.

-- John
-------------------------------------------------------------------------
|   Feith Systems  |   Voice: 1-215-646-8000  |  Email: john@feith.com  |
|    John Wehle    |     Fax: 1-215-540-5495  |                         |
-------------------------------------------------------------------------