[llvm] [DAGCombine] Propagate truncate to operands (PR #98666)

[Matt Arsenault via llvm-commits]

================
@@ -7,15 +7,17 @@ define amdgpu_kernel void @break_inserted_outside_of_loop(ptr addrspace(1) %out,
 ; SI:       ; %bb.0: ; %main_body
 ; SI-NEXT:    s_load_dword s0, s[2:3], 0xb
 ; SI-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, -1, 0
-; SI-NEXT:    s_waitcnt lgkmcnt(0)
-; SI-NEXT:    v_and_b32_e32 v0, s0, v0
 ; SI-NEXT:    v_and_b32_e32 v0, 1, v0
 ; SI-NEXT:    v_cmp_eq_u32_e32 vcc, 1, v0
+; SI-NEXT:    s_waitcnt lgkmcnt(0)
+; SI-NEXT:    s_bitcmp1_b32 s0, 0
----------------
arsenm wrote:

Worse 

https://github.com/llvm/llvm-project/pull/98666