[Greedy RegAlloc] Add a test for single block split with statepoint uses.

The test shows redundant usage of callee saved register in case
basic block has only one non-statepoint use.

llvm/test/CodeGen/X86/statepoint-split-single-block.ll

@@ -0,0 +1,81 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -max-registers-for-gc-values=256 -use-registers-for-deopt-values=true -code-model=large -fixup-allow-gcptr-in-csr=true < %s | FileCheck %s
+
+; The test checks that Greedy register allocator should not split single basic block
+; if it has only one non-statepoint use. Otherwise we may a redundant register usage.
+
+target datalayout = "e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-win64"
+
+define ptr addrspace(1) @foo(ptr addrspace(1) %arg) gc "statepoint-example" {
+; CHECK-LABEL: foo:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pushq %rbp
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    pushq %r15
+; CHECK-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-NEXT:    pushq %r14
+; CHECK-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NEXT:    pushq %r13
+; CHECK-NEXT:    .cfi_def_cfa_offset 40
+; CHECK-NEXT:    pushq %r12
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    pushq %rbx
+; CHECK-NEXT:    .cfi_def_cfa_offset 56
+; CHECK-NEXT:    pushq %rax
+; CHECK-NEXT:    .cfi_def_cfa_offset 64
+; CHECK-NEXT:    .cfi_offset %rbx, -56
+; CHECK-NEXT:    .cfi_offset %r12, -48
+; CHECK-NEXT:    .cfi_offset %r13, -40
+; CHECK-NEXT:    .cfi_offset %r14, -32
+; CHECK-NEXT:    .cfi_offset %r15, -24
+; CHECK-NEXT:    .cfi_offset %rbp, -16
+; CHECK-NEXT:    movq %rdi, (%rsp) # 8-byte Spill
+; CHECK-NEXT:    movabsq $nocsr, %rax
+; CHECK-NEXT:    callq *%rax
+; CHECK-NEXT:    movabsq $bar, %rax
+; CHECK-NEXT:    movq (%rsp), %rbx # 8-byte Reload
+; CHECK-NEXT:    movq %rbx, %rdi
+; CHECK-NEXT:    callq *%rax
+; CHECK-NEXT:  .Ltmp0:
+; CHECK-NEXT:    movq %rbx, (%rsp) # 8-byte Spill
+; CHECK-NEXT:    movabsq $nocsr, %rax
+; CHECK-NEXT:    callq *%rax
+; CHECK-NEXT:    movq (%rsp), %rax # 8-byte Reload
+; CHECK-NEXT:    addq $8, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 56
+; CHECK-NEXT:    popq %rbx
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    popq %r12
+; CHECK-NEXT:    .cfi_def_cfa_offset 40
+; CHECK-NEXT:    popq %r13
+; CHECK-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NEXT:    popq %r14
+; CHECK-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-NEXT:    popq %r15
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    popq %rbp
+; CHECK-NEXT:    .cfi_def_cfa_offset 8
+; CHECK-NEXT:    retq
+entry:
+  ; force spill %arg on stack.
+  call void @nocsr()
+  br label %do_call
+
+do_call:
+  ; Basic block with two use intructions inside:
+  ; copy to rdi as an argument to call and use in statepoint instruction as gc-live and deopt value.
+  %statepoint_token = call token (i64, i32, ptr, i32, i32, ...) @llvm.experimental.gc.statepoint.p0(i64 2882400000, i32 0, ptr elementtype(void (ptr addrspace(1))) @bar, i32 1, i32 0, ptr addrspace(1) %arg, i32 0, i32 0) [ "deopt"(ptr addrspace(1) %arg), "gc-live"(ptr addrspace(1) %arg) ]
+  %arg.reloc = call coldcc ptr addrspace(1) @llvm.experimental.gc.relocate.p1(token %statepoint_token, i32 0, i32 0)
+  br label %next
+
+next:
+  ; force spill %arg.reloc on stack.
+  call void @nocsr()
+  ret ptr addrspace(1) %arg.reloc
+}
+
+declare void @nocsr() "no_callee_saved_registers"
+declare void @bar(ptr addrspace(1))
+declare token @llvm.experimental.gc.statepoint.p0(i64 immarg, i32 immarg, ptr, i32 immarg, i32 immarg, ...)
+declare ptr addrspace(1) @llvm.experimental.gc.relocate.p1(token, i32 immarg, i32 immarg)