Re: [2/5] C-SKY port: Backend implementation

[Jeff Law <law at redhat dot com>]

On 07/27/2018 07:49 PM, Sandra Loosemore wrote:
> On 07/26/2018 12:06 AM, 瞿仙淼 wrote:
>>
>> I wrote a case to reproduce this problem on C-SKY. C code is as follows:
>> -----------------------------------------------------------------------
>> int e1, e2;
>>
>> void func (int a, int b, int c, int d, int f, int g)
>> {
>>    e1 = a > b ? f : g;
>>    e2 = a > b ? c : d;
>>
>>    return;
>> }
>> -----------------------------------------------------------------------
>>
>> compile to assembler with option “-O3 -S” :
>> -----------------------------------------------------------------------
>> func:
>>    cmplt a1, a0
>>    ld.w  t1, (sp, 0)
>>    ld.w  t0, (sp, 4)
>>    movt  t0, t1
>>    cmplt a1, a0
>>    movt  a3, a2
>>    lrw a1, e2
>>    lrw a2, e1
>>    st.w  a3, (a1, 0)
>>    st.w  t0, (a2, 0)
>>    rts
>> -----------------------------------------------------------------------
>> There is an extra “cmplt a1, a0" in the above code without cse_cc.
>> This situation mainly occurs when a relatively short branch jump is
>> converted into a conditional execution instruction. And the CSE pass
>> can not reduce the same conditional comparison instruction . Here is
>> the rtx sequence after “cse2” pass.
>>
>> -----------------------------------------------------------------------
>> (insn 28 13 29 2 (set (reg:CC 33 c)
>>          (gt:CC (reg/v:SI 77 [ a ])
>>              (reg/v:SI 78 [ b ]))) func.c:5 1099 {*cmpgtsi}
>>       (nil))
>> (insn 29 28 30 2 (set (reg/v:SI 82 [ g ])
>>          (if_then_else:SI (eq (reg:CC 33 c)
>>                  (const_int 0 [0]))
>>              (reg/v:SI 82 [ g ])
>>              (reg/v:SI 81 [ f ]))) func.c:5 983 {movf}
>>       (expr_list:REG_DEAD (reg/v:SI 81 [ f ])
>>          (expr_list:REG_DEAD (reg:CC 33 c)
>>              (nil))))
>> (insn 30 29 31 2 (set (reg:CC 33 c)
>>          (gt:CC (reg/v:SI 77 [ a ])
>>              (reg/v:SI 78 [ b ]))) func.c:5 1099 {*cmpgtsi}
>>       (expr_list:REG_DEAD (reg/v:SI 78 [ b ])
>>          (expr_list:REG_DEAD (reg/v:SI 77 [ a ])
>>              (nil))))
>> (insn 31 30 18 2 (set (reg/v:SI 80 [ d ])
>>          (if_then_else:SI (eq (reg:CC 33 c)
>>                  (const_int 0 [0]))
>>              (reg/v:SI 80 [ d ])
>>              (reg/v:SI 79 [ c ]))) func.c:5 983 {movf}
>>       (expr_list:REG_DEAD (reg/v:SI 79 [ c ])
>>          (expr_list:REG_DEAD (reg:CC 33 c)
>>              (nil))))
>> -----------------------------------------------------------------------
>>
>> It doesn't seem to check the same conditional comparison instruction
>> .So I wrote this to solve this problem, but I am not sure if this is
>> the best way : )
>>
>> PS, the same conditional comparison instruction cannot be reduced with
>> the latest version gcc with C-SKY because I just insert the “cse_cc”
>> after “cse1”, when I insert after “cse2”, this problem can be solved
>> very well.
> 
> Thanks, this is very helpful.  I've verified this and I'm moving the
> pass as you suggest, adding a more detailed comment to the source to
> explain what the pass is for, and adding your test case to the
> testsuite.  This will be included when I resubmit the patches to address
> other review comments too.
> 
> Jeff, does that adequately address your concerns about whether the pass
> is useful?
I think the pass is papering over problems elsewhere (see my most other
reply on this thread).  I do think it would be useful to take that code
and create a test based on it.  I suspect you'll want to verify multiple
GT expressions prior to CSE2 and that after CSE2 you have a single GT
expression.  Presumably it'd be in the csky specific test.

jeff