[PowerPC] li of 0 into arg registers of unused arguments

[diggerlin]

cat > test.c

typedef signed char sb;
void foo(sb);
void __attribute__((noinline)) bar(sb sb1, sb var2, sb var3, sb var4,
                                   sb sb5)
{
  foo(sb5);
}
void __attribute__((noinline)) test() {
  bar(1, 2, 3, 4, 125);
}

bash-5.2$ clang -O2 seg.c -m32 -S -o seg.s

there is code as

.bar:
# %bb.0:                                # %entry
        mflr 0
        stwu 1, -64(1)
        mr      3, 7
        stw 0, 72(1)
        bl .foo[PR]
        nop
        addi 1, 1, 64
        lwz 0, 8(1)
        mtlr 0
        blr


.test:
# %bb.0:                                # %entry
        mflr 0
        stwu 1, -64(1)
        li 3, 0
        li 4, 0
        li 5, 0
        li 6, 0
        stw 0, 72(1)
        li 7, 125
        bl .bar
        nop

the asm code should be optimized out

        li 3, 0
        li 4, 0
        li 5, 0
        li 6, 0

[llvmbot]

@llvm/issue-subscribers-backend-powerpc

Author: zhijian lin (diggerlin)

cat > test.c

typedef signed char sb;
void foo(sb);
void __attribute__((noinline)) bar(sb sb1, sb var2, sb var3, sb var4,
                                   sb sb5)
{
  foo(sb5);
}
void __attribute__((noinline)) test() {
  bar(1, 2, 3, 4, 125);
}

bash-5.2$ /home/zhijian/llvm/dev/build/bin/ibm-clang -O2 seg.c -m32 -S -o seg.s

there is code as

.bar:
# %bb.0:                                # %entry
        mflr 0
        stwu 1, -64(1)
        mr      3, 7
        stw 0, 72(1)
        bl .foo[PR]
        nop
        addi 1, 1, 64
        lwz 0, 8(1)
        mtlr 0
        blr


.test:
# %bb.0:                                # %entry
        mflr 0
        stwu 1, -64(1)
        li 3, 0
        li 4, 0
        li 5, 0
        li 6, 0
        stw 0, 72(1)
        li 7, 125
        bl .bar
        nop

the asm code should be optimized out

        li 3, 0
        li 4, 0
        li 5, 0
        li 6, 0

[diggerlin]

=== test

Initial selection DAG: %bb.0 'test:entry'
SelectionDAG has 26 nodes:
  t1: i32 = GlobalAddress<ptr @bar> 0
  t2: i8 = undef
  t3: i8 = Constant<125>
  t9: i32 = Register $r1
      t0: ch,glue = EntryToken
    t8: ch,glue = callseq_start t0, TargetConstant:i32<56>, TargetConstant:i32<0>
  t11: ch,glue = CopyToReg t8, Register:i32 $r3, Constant:i32<0>
  t13: ch,glue = CopyToReg t11, Register:i32 $r4, Constant:i32<0>, t11:1
  t15: ch,glue = CopyToReg t13, Register:i32 $r5, Constant:i32<0>, t13:1
  t17: ch,glue = CopyToReg t15, Register:i32 $r6, Constant:i32<0>, t15:1
  t19: ch,glue = CopyToReg t17, Register:i32 $r7, Constant:i32<0><125>, t17:1
  t23: ch,glue = PPCISD::CALL_NOP t19, MCSymbol:i32, Register:i32 $r3, Register:i32 $r4, Register:i32 $r5, Register:i32 $r6, Register:i32 $r7, Register:i32 $r2, RegisterMask:Untyped, t19:1
    t24: ch,glue = callseq_end t23, TargetConstant:i32<56>, TargetConstant:i32<0>, t23:1
  t25: ch = PPCISD::RET_GLUE t24

the selectionDAG has Constant:i32<0> instead of undef:i32, it block the pass PROCESS IMPLICIT DEFS to take effect.

[diggerlin]

the PR #122741 will fix the issue.

[amy-kwan]

@diggerlin Is this issue a duplicate of #118474?

[diggerlin]

Yes, I will close it