[llvm] [DAG] Fold mul 0 -> 0 when expanding mul into parts. (PR #168780)

[David Green via llvm-commits]

================
@@ -279,11 +279,11 @@ define i32 @si32_100(i32 %a, i32 %b) {
 ; CHECK-SD-LABEL: si32_100:
 ; CHECK-SD:       // %bb.0: // %entry
 ; CHECK-SD-NEXT:    mov w8, #34079 // =0x851f
-; CHECK-SD-NEXT:    mov w9, #100 // =0x64
 ; CHECK-SD-NEXT:    movk w8, #20971, lsl #16
 ; CHECK-SD-NEXT:    smull x8, w0, w8
-; CHECK-SD-NEXT:    asr x8, x8, #37
-; CHECK-SD-NEXT:    add w8, w8, w8, lsr #31
+; CHECK-SD-NEXT:    asr x9, x8, #37
+; CHECK-SD-NEXT:    add x8, x9, x8, lsr #63
----------------
davemgreen wrote:

It should not be any slower (and could theoretically start the shift before finishing the asr, although I'm not sure that is something that happens in practice).

https://github.com/llvm/llvm-project/pull/168780