[llvm] r298273 - GlobalISel: allow quad-precision values to be dumped.

[Tim Northover via llvm-commits]

Author: tnorthover
Date: Mon Mar 20 11:52:08 2017
New Revision: 298273

URL: http://llvm.org/viewvc/llvm-project?rev=298273&view=rev
Log:
GlobalISel: allow quad-precision values to be dumped.

Otherwise the fallback path fails with an assertion on AAPCS AArch64 targets,
when "long double" is encountered.

Modified:
    llvm/trunk/lib/CodeGen/MachineInstr.cpp
    llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-fallback.ll

Modified: llvm/trunk/lib/CodeGen/MachineInstr.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/MachineInstr.cpp?rev=298273&r1=298272&r2=298273&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/MachineInstr.cpp (original)
+++ llvm/trunk/lib/CodeGen/MachineInstr.cpp Mon Mar 20 11:52:08 2017
@@ -420,6 +420,11 @@ void MachineOperand::print(raw_ostream &
       bool Unused;
       APF.convert(APFloat::IEEEsingle(), APFloat::rmNearestTiesToEven, &Unused);
       OS << "half " << APF.convertToFloat();
+    } else if (getFPImm()->getType()->isFP128Ty()) {
+      APFloat APF = getFPImm()->getValueAPF();
+      SmallString<16> Str;
+      getFPImm()->getValueAPF().toString(Str);
+      OS << "quad " << Str;
     } else {
       OS << getFPImm()->getValueAPF().convertToDouble();
     }

Modified: llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-fallback.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-fallback.ll?rev=298273&r1=298272&r2=298273&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-fallback.ll (original)
+++ llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-fallback.ll Mon Mar 20 11:52:08 2017
@@ -138,3 +138,11 @@ broken:
 continue:
   ret void
 }
+
+; Check that we fallback on invoke translation failures.
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: %vreg0<def>(s128) = G_FCONSTANT quad 2
+; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for test_quad_dump
+; FALLBACK-WITH-REPORT-OUT-LABEL: test_quad_dump:
+define fp128 @test_quad_dump() {
+  ret fp128 0xL00000000000000004000000000000000
+}