From caf8a11b656afdaf12d0a3566c7d42e302bd7539 Mon Sep 17 00:00:00 2001
From: David Green <david.green@arm.com>
Date: Sun, 26 May 2019 10:30:22 +0000
Subject: [PATCH] [ARM] Promote fp16 frem

Promote fp16 frem operations on ARM to floats so they call fmodf.

Differential Revision: https://reviews.llvm.org/D62321

llvm-svn: 361713
---
 llvm/lib/Target/ARM/ARMISelLowering.cpp |  5 +++++
 llvm/test/CodeGen/ARM/fp16-fullfp16.ll  | 27 +++++++++++++++++++--------
 2 files changed, 24 insertions(+), 8 deletions(-)

diff --git a/llvm/lib/Target/ARM/ARMISelLowering.cpp b/llvm/lib/Target/ARM/ARMISelLowering.cpp
index 643d280..8e5c076 100644
--- a/llvm/lib/Target/ARM/ARMISelLowering.cpp
+++ b/llvm/lib/Target/ARM/ARMISelLowering.cpp
@@ -1142,6 +1142,11 @@ ARMTargetLowering::ARMTargetLowering(const TargetMachine &TM,
     }
   }
 
+  // FP16 often need to be promoted to call lib functions
+  if (Subtarget->hasFullFP16()) {
+    setOperationAction(ISD::FREM, MVT::f16, Promote);
+  }
+
   if (Subtarget->hasNEON()) {
     // vmin and vmax aren't available in a scalar form, so we use
     // a NEON instruction with an undef lane instead.
diff --git a/llvm/test/CodeGen/ARM/fp16-fullfp16.ll b/llvm/test/CodeGen/ARM/fp16-fullfp16.ll
index 5f14463..d5ded1f 100644
--- a/llvm/test/CodeGen/ARM/fp16-fullfp16.ll
+++ b/llvm/test/CodeGen/ARM/fp16-fullfp16.ll
@@ -57,14 +57,25 @@ define void @test_fdiv(half* %p, half* %q) {
   ret void
 }
 
-; FIXME
-;define void @test_frem(half* %p, half* %q) {
-;  %a = load half, half* %p, align 2
-;  %b = load half, half* %q, align 2
-;  %r = frem half %a, %b
-;  store half %r, half* %p
-;  ret void
-;}
+define arm_aapcs_vfpcc void @test_frem(half* %p, half* %q) {
+; CHECK-LABEL: test_frem:
+; CHECK:         .save {r4, lr}
+; CHECK-NEXT:    push {r4, lr}
+; CHECK-NEXT:    vldr.16 s2, [r1]
+; CHECK-NEXT:    vldr.16 s0, [r0]
+; CHECK-NEXT:    mov r4, r0
+; CHECK-NEXT:    vcvtb.f32.f16 s0, s0
+; CHECK-NEXT:    vcvtb.f32.f16 s1, s2
+; CHECK-NEXT:    bl fmodf
+; CHECK-NEXT:    vcvtb.f16.f32 s0, s0
+; CHECK-NEXT:    vstr.16 s0, [r4]
+; CHECK-NEXT:    pop {r4, pc}
+  %a = load half, half* %p, align 2
+  %b = load half, half* %q, align 2
+  %r = frem half %a, %b
+  store half %r, half* %p
+  ret void
+}
 
 define void @test_load_store(half* %p, half* %q) {
 ; CHECK-LABEL: test_load_store:
-- 
2.7.4