[llvm-branch-commits] [llvm] 3de89ab - [PowerPC] Do not attempt to reuse load for 64-bit FP_TO_UINT without FPCVT

[Tom Stellard via llvm-branch-commits]

Author: Nemanja Ivanovic
Date: 2020-05-19T08:31:01-07:00
New Revision: 3de89abf189e63da4edbfda509ed3d421bdba1a9

URL: https://github.com/llvm/llvm-project/commit/3de89abf189e63da4edbfda509ed3d421bdba1a9
DIFF: https://github.com/llvm/llvm-project/commit/3de89abf189e63da4edbfda509ed3d421bdba1a9.diff

LOG: [PowerPC] Do not attempt to reuse load for 64-bit FP_TO_UINT without FPCVT

We call the function that attempts to reuse the conversion without checking
whether the target matches the constraints that the callee expects. This patch
adds the check prior to the call.

Fixes: https://bugs.llvm.org/show_bug.cgi?id=43976

Differential revision: https://reviews.llvm.org/D77564

(cherry picked from commit 64b31d96dfd6c05e6d52d8798726dec60502cfde)

Added: 
    llvm/test/CodeGen/PowerPC/pr43976.ll

Modified: 
    llvm/lib/Target/PowerPC/PPCISelLowering.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index 352a05529bc9..00f59bba52e8 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -7851,9 +7851,10 @@ bool PPCTargetLowering::canReuseLoadAddress(SDValue Op, EVT MemVT,
                                             SelectionDAG &DAG,
                                             ISD::LoadExtType ET) const {
   SDLoc dl(Op);
+  bool ValidFPToUint = Op.getOpcode() == ISD::FP_TO_UINT &&
+                       (Subtarget.hasFPCVT() || Op.getValueType() == MVT::i32);
   if (ET == ISD::NON_EXTLOAD &&
-      (Op.getOpcode() == ISD::FP_TO_UINT ||
-       Op.getOpcode() == ISD::FP_TO_SINT) &&
+      (ValidFPToUint || Op.getOpcode() == ISD::FP_TO_SINT) &&
       isOperationLegalOrCustom(Op.getOpcode(),
                                Op.getOperand(0).getValueType())) {
 

diff  --git a/llvm/test/CodeGen/PowerPC/pr43976.ll b/llvm/test/CodeGen/PowerPC/pr43976.ll
new file mode 100644
index 000000000000..91722283f4ae
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/pr43976.ll
@@ -0,0 +1,59 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=powerpc64-unknown-unknown -verify-machineinstrs \
+; RUN:   -ppc-asm-full-reg-names < %s | FileCheck %s
+ at a = dso_local local_unnamed_addr global double 0.000000e+00, align 8
+
+define dso_local signext i32 @b() local_unnamed_addr #0 {
+; CHECK-LABEL: b:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    mflr r0
+; CHECK-NEXT:    std r0, 16(r1)
+; CHECK-NEXT:    stdu r1, -144(r1)
+; CHECK-NEXT:    addis r3, r2, a at toc@ha
+; CHECK-NEXT:    addis r4, r2, .LCPI0_0 at toc@ha
+; CHECK-NEXT:    lfd f0, a at toc@l(r3)
+; CHECK-NEXT:    lfs f1, .LCPI0_0 at toc@l(r4)
+; CHECK-NEXT:    li r4, 1
+; CHECK-NEXT:    sldi r4, r4, 63
+; CHECK-NEXT:    fsub f2, f0, f1
+; CHECK-NEXT:    fctidz f2, f2
+; CHECK-NEXT:    stfd f2, 128(r1)
+; CHECK-NEXT:    fctidz f2, f0
+; CHECK-NEXT:    stfd f2, 120(r1)
+; CHECK-NEXT:    ld r3, 128(r1)
+; CHECK-NEXT:    ld r5, 120(r1)
+; CHECK-NEXT:    fcmpu cr0, f0, f1
+; CHECK-NEXT:    xor r3, r3, r4
+; CHECK-NEXT:    bc 12, lt, .LBB0_1
+; CHECK-NEXT:    b .LBB0_2
+; CHECK-NEXT:  .LBB0_1: # %entry
+; CHECK-NEXT:    addi r3, r5, 0
+; CHECK-NEXT:  .LBB0_2: # %entry
+; CHECK-NEXT:    std r3, 112(r1)
+; CHECK-NEXT:    addis r3, r2, .LCPI0_1 at toc@ha
+; CHECK-NEXT:    lfd f0, 112(r1)
+; CHECK-NEXT:    lfs f1, .LCPI0_1 at toc@l(r3)
+; CHECK-NEXT:    fcfid f0, f0
+; CHECK-NEXT:    fmul f0, f0, f1
+; CHECK-NEXT:    fctiwz f0, f0
+; CHECK-NEXT:    stfd f0, 136(r1)
+; CHECK-NEXT:    lwa r3, 140(r1)
+; CHECK-NEXT:    bl g
+; CHECK-NEXT:    nop
+; CHECK-NEXT:    addi r1, r1, 144
+; CHECK-NEXT:    ld r0, 16(r1)
+; CHECK-NEXT:    mtlr r0
+; CHECK-NEXT:    blr
+entry:
+  %0 = load double, double* @a, align 8
+  %conv = fptoui double %0 to i64
+  %conv1 = sitofp i64 %conv to double
+  %mul = fmul double %conv1, 1.000000e+06
+  %conv2 = fptosi double %mul to i32
+  %call = tail call signext i32 @g(i32 signext %conv2) #0
+  ret i32 %call
+}
+
+declare signext i32 @g(i32 signext) local_unnamed_addr
+
+attributes #0 = { nounwind }