[ARM] Reapply r296865 "[ARM] fpscr read/write intrinsics not aware of each other""

The original patch r296865 was reverted as it broke the chromium builds for
Android https://bugs.llvm.org/show_bug.cgi?id=32134, this patch reapplies
r296865 with a fix to make sure it doesn't cause the build regression.

The problem was that intrinsic selection on int_arm_get_fpscr was failing in
ISel this was because the code to manually select this intrinsic still thought
it was the version with no side-effects (INTRINSIC_WO_CHAIN) which is wrong as
it doesn't semantically match the definition in the tablegen code which says it
does have side-effects, I've fixed this by updating the intrinsic type to
INTRINSIC_W_CHAIN (has side-effects). I've also added a test for this based on
Hans original reproducer.

Differential Revision: https://reviews.llvm.org/D30645

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297137 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/include/llvm/IR/IntrinsicsARM.td b/include/llvm/IR/IntrinsicsARM.td
index 24239689a62ee71d9809850f647fd4d0063da598..18ed24be56d4e59f126deec4734432e726dc46e4 100644 (file)
--- a/include/llvm/IR/IntrinsicsARM.td
+++ b/include/llvm/IR/IntrinsicsARM.td
@@ -67,7 +67,7 @@ def int_arm_isb : GCCBuiltin<"__builtin_arm_isb">, MSBuiltin<"__isb">,
 // VFP
 
 def int_arm_get_fpscr : GCCBuiltin<"__builtin_arm_get_fpscr">,
-                       Intrinsic<[llvm_i32_ty], [], [IntrNoMem]>;
+                       Intrinsic<[llvm_i32_ty], [], []>;
 def int_arm_set_fpscr : GCCBuiltin<"__builtin_arm_set_fpscr">,
                        Intrinsic<[], [llvm_i32_ty], []>;
 def int_arm_vcvtr     : Intrinsic<[llvm_float_ty], [llvm_anyfloat_ty],

diff --git a/lib/Target/ARM/ARMISelLowering.cpp b/lib/Target/ARM/ARMISelLowering.cpp
index 4b6d0e1a34bdc309a306032d9be354c65bb51a46..6d735a3bed0aaa496c6f4007f71190e0bc5ec431 100644 (file)
--- a/lib/Target/ARM/ARMISelLowering.cpp
+++ b/lib/Target/ARM/ARMISelLowering.cpp
@@ -4913,9 +4913,10 @@ SDValue ARMTargetLowering::LowerFLT_ROUNDS_(SDValue Op,
   // The formula we use to implement this is (((FPSCR + 1 << 22) >> 22) & 3)
   // so that the shift + and get folded into a bitfield extract.
   SDLoc dl(Op);
-  SDValue FPSCR = DAG.getNode(ISD::INTRINSIC_WO_CHAIN, dl, MVT::i32,
-                              DAG.getConstant(Intrinsic::arm_get_fpscr, dl,
-                                              MVT::i32));
+  SDValue Ops[] = { DAG.getEntryNode(),
+                    DAG.getConstant(Intrinsic::arm_get_fpscr, dl, MVT::i32) };
+
+  SDValue FPSCR = DAG.getNode(ISD::INTRINSIC_W_CHAIN, dl, MVT::i32, Ops);
   SDValue FltRounds = DAG.getNode(ISD::ADD, dl, MVT::i32, FPSCR,
                                   DAG.getConstant(1U << 22, dl, MVT::i32));
   SDValue RMODE = DAG.getNode(ISD::SRL, dl, MVT::i32, FltRounds,

diff --git a/test/CodeGen/ARM/fpscr-intrinsics.ll b/test/CodeGen/ARM/fpscr-intrinsics.ll
new file mode 100644 (file)
index 0000000..64b9752
--- /dev/null
+++ b/test/CodeGen/ARM/fpscr-intrinsics.ll
@@ -0,0 +1,44 @@
+; RUN: llc < %s -O0 -mtriple=armv7-eabi -mcpu=cortex-a8 -mattr=+neon,+fp-armv8 | FileCheck %s
+; RUN: llc < %s -O3 -mtriple=armv7-eabi -mcpu=cortex-a8 -mattr=+neon,+fp-armv8 | FileCheck %s
+
+@a = common global double 0.000000e+00, align 8
+
+; Function Attrs: noinline nounwind uwtable
+define void @strtod() {
+entry:
+  ; CHECK: vmrs r{{[0-9]+}}, fpscr
+  %0 = call i32 @llvm.flt.rounds()
+  %tobool = icmp ne i32 %0, 0
+  br i1 %tobool, label %if.then, label %if.end
+
+if.then:                                          ; preds = %entry
+  store double 5.000000e-01, double* @a, align 8
+  br label %if.end
+
+if.end:                                           ; preds = %if.then, %entry
+  ret void
+}
+
+; Function Attrs: nounwind
+define void @fn1(i32* nocapture %p) local_unnamed_addr {
+entry:
+  ; CHECK: vmrs r{{[0-9]+}}, fpscr
+  %0 = tail call i32 @llvm.arm.get.fpscr()
+  store i32 %0, i32* %p, align 4
+  ; CHECK: vmsr fpscr, r{{[0-9]+}}
+  tail call void @llvm.arm.set.fpscr(i32 1)
+  ; CHECK: vmrs r{{[0-9]+}}, fpscr
+  %1 = tail call i32 @llvm.arm.get.fpscr()
+  %arrayidx1 = getelementptr inbounds i32, i32* %p, i32 1
+  store i32 %1, i32* %arrayidx1, align 4
+  ret void
+}
+
+; Function Attrs: nounwind readonly
+declare i32 @llvm.arm.get.fpscr()
+
+; Function Attrs: nounwind writeonly
+declare void @llvm.arm.set.fpscr(i32)
+
+; Function Attrs: nounwind
+declare i32 @llvm.flt.rounds()