[X86] Add test to show incorrect ordering of flt.rounds intrinsic relative to calls...

We don't order flt.rounds intrinsics relative to side effecting
operations in SelectionDAG. And we CSE multiple calls because of
this.

diff --git a/llvm/test/CodeGen/X86/flt-rounds.ll b/llvm/test/CodeGen/X86/flt-rounds.ll
index eb640f5e8c6952e9d8d0570b9119a2e6cbfe6150..22101cc7343d9860766c802f1753b05215478c4d 100644 (file)
--- a/llvm/test/CodeGen/X86/flt-rounds.ll
+++ b/llvm/test/CodeGen/X86/flt-rounds.ll
@@ -34,3 +34,116 @@ define i32 @test_flt_rounds() nounwind {
   %1 = call i32 @llvm.flt.rounds()
   ret i32 %1
 }
+
+; Make sure we preserve order with fesetround.
+define i32 @multiple_flt_rounds() nounwind {
+; X86-LABEL: multiple_flt_rounds:
+; X86:       # %bb.0: # %entry
+; X86-NEXT:    pushl %ebx
+; X86-NEXT:    pushl %esi
+; X86-NEXT:    subl $20, %esp
+; X86-NEXT:    movl $1024, (%esp) # imm = 0x400
+; X86-NEXT:    calll fesetround
+; X86-NEXT:    fnstcw {{[0-9]+}}(%esp)
+; X86-NEXT:    movzwl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    shrl $9, %ecx
+; X86-NEXT:    andb $6, %cl
+; X86-NEXT:    movl $45, %esi
+; X86-NEXT:    # kill: def $cl killed $cl killed $ecx
+; X86-NEXT:    shrl %cl, %esi
+; X86-NEXT:    andl $3, %esi
+; X86-NEXT:    xorl %ebx, %ebx
+; X86-NEXT:    cmpl $3, %esi
+; X86-NEXT:    setne %bl
+; X86-NEXT:    movl $0, (%esp)
+; X86-NEXT:    calll fesetround
+; X86-NEXT:    movl $3072, (%esp) # imm = 0xC00
+; X86-NEXT:    calll fesetround
+; X86-NEXT:    cmpl $1, %esi
+; X86-NEXT:    leal 1(%ebx), %eax
+; X86-NEXT:    je .LBB1_2
+; X86-NEXT:  # %bb.1: # %entry
+; X86-NEXT:    movl %eax, %ebx
+; X86-NEXT:  .LBB1_2: # %entry
+; X86-NEXT:    sbbl $-1, %ebx
+; X86-NEXT:    movl $2048, (%esp) # imm = 0x800
+; X86-NEXT:    calll fesetround
+; X86-NEXT:    xorl %ecx, %ecx
+; X86-NEXT:    cmpl $2, %esi
+; X86-NEXT:    setne %cl
+; X86-NEXT:    negl %ecx
+; X86-NEXT:    xorl %eax, %eax
+; X86-NEXT:    cmpl %ecx, %ebx
+; X86-NEXT:    setne %al
+; X86-NEXT:    addl $20, %esp
+; X86-NEXT:    popl %esi
+; X86-NEXT:    popl %ebx
+; X86-NEXT:    retl
+;
+; X64-LABEL: multiple_flt_rounds:
+; X64:       # %bb.0: # %entry
+; X64-NEXT:    pushq %rbp
+; X64-NEXT:    pushq %r14
+; X64-NEXT:    pushq %rbx
+; X64-NEXT:    subq $16, %rsp
+; X64-NEXT:    movl $1024, %edi # imm = 0x400
+; X64-NEXT:    callq fesetround
+; X64-NEXT:    fnstcw {{[0-9]+}}(%rsp)
+; X64-NEXT:    movzwl {{[0-9]+}}(%rsp), %ecx
+; X64-NEXT:    shrl $9, %ecx
+; X64-NEXT:    andb $6, %cl
+; X64-NEXT:    movl $45, %ebx
+; X64-NEXT:    # kill: def $cl killed $cl killed $ecx
+; X64-NEXT:    shrl %cl, %ebx
+; X64-NEXT:    andl $3, %ebx
+; X64-NEXT:    xorl %r14d, %r14d
+; X64-NEXT:    cmpl $3, %ebx
+; X64-NEXT:    setne %r14b
+; X64-NEXT:    xorl %edi, %edi
+; X64-NEXT:    callq fesetround
+; X64-NEXT:    leal 1(%r14), %ebp
+; X64-NEXT:    movl $3072, %edi # imm = 0xC00
+; X64-NEXT:    callq fesetround
+; X64-NEXT:    cmpl $1, %ebx
+; X64-NEXT:    cmovel %r14d, %ebp
+; X64-NEXT:    sbbl $-1, %ebp
+; X64-NEXT:    movl $2048, %edi # imm = 0x800
+; X64-NEXT:    callq fesetround
+; X64-NEXT:    xorl %ecx, %ecx
+; X64-NEXT:    cmpl $2, %ebx
+; X64-NEXT:    setne %cl
+; X64-NEXT:    negl %ecx
+; X64-NEXT:    xorl %eax, %eax
+; X64-NEXT:    cmpl %ecx, %ebp
+; X64-NEXT:    setne %al
+; X64-NEXT:    addq $16, %rsp
+; X64-NEXT:    popq %rbx
+; X64-NEXT:    popq %r14
+; X64-NEXT:    popq %rbp
+; X64-NEXT:    retq
+entry:
+  %call = tail call i32 @fesetround(i32 1024)
+  %0 = tail call i32 @llvm.flt.rounds()
+  %cmp = icmp ne i32 %0, 3
+  %spec.select = zext i1 %cmp to i32
+  %call1 = tail call i32 @fesetround(i32 0)
+  %1 = tail call i32 @llvm.flt.rounds()
+  %cmp2 = icmp eq i32 %1, 1
+  %inc4 = select i1 %cmp, i32 2, i32 1
+  %errs.1 = select i1 %cmp2, i32 %spec.select, i32 %inc4
+  %call6 = tail call i32 @fesetround(i32 3072)
+  %2 = tail call i32 @llvm.flt.rounds()
+  %cmp7 = icmp ne i32 %2, 0
+  %inc9 = zext i1 %cmp7 to i32
+  %spec.select22 = add nuw nsw i32 %errs.1, %inc9
+  %call11 = tail call i32 @fesetround(i32 2048)
+  %3 = tail call i32 @llvm.flt.rounds()
+  %cmp12 = icmp ne i32 %3, 2
+  %inc14.neg = sext i1 %cmp12 to i32
+  %cmp16 = icmp ne i32 %spec.select22, %inc14.neg
+  %cond = zext i1 %cmp16 to i32
+  ret i32 %cond
+}
+
+; Function Attrs: nounwind
+declare dso_local i32 @fesetround(i32) local_unnamed_addr #1