[PATCH] Remove quite obvious dead assignments.

[Martin Liška]

On 6/26/19 1:39 PM, Jakub Jelinek wrote:
> On Wed, Jun 26, 2019 at 12:57:15PM +0200, Martin Liška wrote:
>> --- a/gcc/asan.c
>> +++ b/gcc/asan.c
>> @@ -1713,8 +1713,8 @@ asan_emit_allocas_unpoison (rtx top, rtx bot, rtx_insn *before)
>>    rtx ret = init_one_libfunc ("__asan_allocas_unpoison");
>>    top = convert_memory_address (ptr_mode, top);
>>    bot = convert_memory_address (ptr_mode, bot);
>> -  ret = emit_library_call_value (ret, NULL_RTX, LCT_NORMAL, ptr_mode,
>> -				 top, ptr_mode, bot, ptr_mode);
>> +  emit_library_call_value (ret, NULL_RTX, LCT_NORMAL, ptr_mode,
>> +			   top, ptr_mode, bot, ptr_mode);
> 
> If you don't need the return value, you should use emit_library_call,
> not emit_library_call_value.

Done.

> 
>> --- a/gcc/cfgexpand.c
>> +++ b/gcc/cfgexpand.c
>> @@ -3044,7 +3044,7 @@ expand_asm_stmt (gasm *stmt)
>>  	      }
>>  	}
>>      }
>> -  unsigned nclobbers = clobber_rvec.length();
>> +  unsigned nclobbers;
> 
> Can you instead move the nclobbers declaration to the spot where it
> is initialized (right now the second time), and add space before (
> there too?

Sure.

> 
>> --- a/gcc/config/i386/i386-expand.c
>> +++ b/gcc/config/i386/i386-expand.c
>> @@ -16051,8 +16051,6 @@ ix86_expand_rounddf_32 (rtx operand0, rtx operand1)
>>    mhalf = expand_simple_binop (mode, MINUS, half, one, NULL_RTX,
>>  			       0, OPTAB_DIRECT);
>>  
>> -  /* Compensate.  */
>> -  tmp = gen_reg_rtx (mode);
>>    /* xa2 = xa2 - (dxa > 0.5 ? 1 : 0) */
>>    tmp = ix86_expand_sse_compare_mask (UNGT, dxa, half, false);
>>    emit_insn (gen_rtx_SET (tmp, gen_rtx_AND (mode, one, tmp)));
> 
> Is this really desirable?
> Just quick look suggests perhaps it wants to use one temporary
> for the gen_reg_rtx (mode) and use that on the lhs of SET, and another one
> as the last operand to AND holding ix86_expand_sse_compare_mask?
> Though, such problem is in most of the ix86_expand_sse_compare_mask callers.
> While ix86_expand_sse_compare_mask returns a result of gen_reg_rtx too,
> I think it is better not to reuse pseudos for different values.
> I'd say take out this change and deal with it with i386 maintainers
> separately.

Good point, I've just created:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91016

> 
>> --- a/gcc/tree-vect-stmts.c
>> +++ b/gcc/tree-vect-stmts.c
>> @@ -3483,8 +3483,7 @@ vectorizable_call (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi,
>>  			= gimple_build_call_internal_vec (ifn, vargs);
>>  		      gimple_call_set_lhs (call, half_res);
>>  		      gimple_call_set_nothrow (call, true);
>> -		      new_stmt_info
>> -			= vect_finish_stmt_generation (stmt_info, call, gsi);
>> +		      vect_finish_stmt_generation (stmt_info, call, gsi);
>>  		      if ((i & 1) == 0)
>>  			{
>>  			  prev_res = half_res;
>> @@ -3583,8 +3582,7 @@ vectorizable_call (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi,
>>  	      gcall *call = gimple_build_call_internal_vec (ifn, vargs);
>>  	      gimple_call_set_lhs (call, half_res);
>>  	      gimple_call_set_nothrow (call, true);
>> -	      new_stmt_info
>> -		= vect_finish_stmt_generation (stmt_info, call, gsi);
>> +	      vect_finish_stmt_generation (stmt_info, call, gsi);
>>  	      if ((j & 1) == 0)
>>  		{
>>  		  prev_res = half_res;
> 
> This looks wrong.
> There should have been
> SLP_TREE_VEC_STMTS (slp_node).quick_push (new_stmt_info);
> in between for slp_node, or the usual code like:
>           if (cond_for_first)
>             STMT_VINFO_VEC_STMT (stmt_info) = *vec_stmt = new_stmt_info;
>           else
>             STMT_VINFO_RELATED_STMT (prev_stmt_info) = new_stmt_info;
> 	  prev_stmt_info = new_stmt_info;
> otherwise.  In any case, I think this should be dealt with separately.

Likewise here:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91017

I'm going to install updated patch once regression tests finish for it.
I excluded changes for the PR mentioned.

Martin

> 
> 	Jakub
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Remove-quite-obvious-dead-assignments.patch
Type: text/x-patch
Size: 27533 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20190627/76378fb4/attachment.bin>