ARM: match GCC's behaviour for builtins

GCC changes the CC between the user-code and the builtins based on the
value of `-target` rather than `-mfloat-abi`.  When a HF target is used,
the VFP variant of the AAPCS CC is used.  Otherwise, the AAPCS variant
is used.  In all cases, the AEABI functions use the AAPCS CC.  Adjust
the calling convention based on the target.

Resolves PR30543!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@291909 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/ARM/ARMISelLowering.cpp b/lib/Target/ARM/ARMISelLowering.cpp
index 25716807d74aedc80c31897ac8261e5e337e2786..b805cc3e42bf35226cec60877a11cb1fe426616d 100644 (file)
--- a/lib/Target/ARM/ARMISelLowering.cpp
+++ b/lib/Target/ARM/ARMISelLowering.cpp
@@ -97,171 +97,6 @@ namespace {
   };
 }
 
-void ARMTargetLowering::InitLibcallCallingConvs() {
-  // The builtins on ARM always use AAPCS, irrespective of wheter C is AAPCS or
-  // AAPCS_VFP.
-  for (const auto LC : {
-           RTLIB::SHL_I16,
-           RTLIB::SHL_I32,
-           RTLIB::SHL_I64,
-           RTLIB::SHL_I128,
-           RTLIB::SRL_I16,
-           RTLIB::SRL_I32,
-           RTLIB::SRL_I64,
-           RTLIB::SRL_I128,
-           RTLIB::SRA_I16,
-           RTLIB::SRA_I32,
-           RTLIB::SRA_I64,
-           RTLIB::SRA_I128,
-           RTLIB::MUL_I8,
-           RTLIB::MUL_I16,
-           RTLIB::MUL_I32,
-           RTLIB::MUL_I64,
-           RTLIB::MUL_I128,
-           RTLIB::MULO_I32,
-           RTLIB::MULO_I64,
-           RTLIB::MULO_I128,
-           RTLIB::SDIV_I8,
-           RTLIB::SDIV_I16,
-           RTLIB::SDIV_I32,
-           RTLIB::SDIV_I64,
-           RTLIB::SDIV_I128,
-           RTLIB::UDIV_I8,
-           RTLIB::UDIV_I16,
-           RTLIB::UDIV_I32,
-           RTLIB::UDIV_I64,
-           RTLIB::UDIV_I128,
-           RTLIB::SREM_I8,
-           RTLIB::SREM_I16,
-           RTLIB::SREM_I32,
-           RTLIB::SREM_I64,
-           RTLIB::SREM_I128,
-           RTLIB::UREM_I8,
-           RTLIB::UREM_I16,
-           RTLIB::UREM_I32,
-           RTLIB::UREM_I64,
-           RTLIB::UREM_I128,
-           RTLIB::SDIVREM_I8,
-           RTLIB::SDIVREM_I16,
-           RTLIB::SDIVREM_I32,
-           RTLIB::SDIVREM_I64,
-           RTLIB::SDIVREM_I128,
-           RTLIB::UDIVREM_I8,
-           RTLIB::UDIVREM_I16,
-           RTLIB::UDIVREM_I32,
-           RTLIB::UDIVREM_I64,
-           RTLIB::UDIVREM_I128,
-           RTLIB::NEG_I32,
-           RTLIB::NEG_I64,
-           RTLIB::ADD_F32,
-           RTLIB::ADD_F64,
-           RTLIB::ADD_F80,
-           RTLIB::ADD_F128,
-           RTLIB::SUB_F32,
-           RTLIB::SUB_F64,
-           RTLIB::SUB_F80,
-           RTLIB::SUB_F128,
-           RTLIB::MUL_F32,
-           RTLIB::MUL_F64,
-           RTLIB::MUL_F80,
-           RTLIB::MUL_F128,
-           RTLIB::DIV_F32,
-           RTLIB::DIV_F64,
-           RTLIB::DIV_F80,
-           RTLIB::DIV_F128,
-           RTLIB::POWI_F32,
-           RTLIB::POWI_F64,
-           RTLIB::POWI_F80,
-           RTLIB::POWI_F128,
-           RTLIB::FPEXT_F64_F128,
-           RTLIB::FPEXT_F32_F128,
-           RTLIB::FPEXT_F32_F64,
-           RTLIB::FPEXT_F16_F32,
-           RTLIB::FPROUND_F32_F16,
-           RTLIB::FPROUND_F64_F16,
-           RTLIB::FPROUND_F80_F16,
-           RTLIB::FPROUND_F128_F16,
-           RTLIB::FPROUND_F64_F32,
-           RTLIB::FPROUND_F80_F32,
-           RTLIB::FPROUND_F128_F32,
-           RTLIB::FPROUND_F80_F64,
-           RTLIB::FPROUND_F128_F64,
-           RTLIB::FPTOSINT_F32_I32,
-           RTLIB::FPTOSINT_F32_I64,
-           RTLIB::FPTOSINT_F32_I128,
-           RTLIB::FPTOSINT_F64_I32,
-           RTLIB::FPTOSINT_F64_I64,
-           RTLIB::FPTOSINT_F64_I128,
-           RTLIB::FPTOSINT_F80_I32,
-           RTLIB::FPTOSINT_F80_I64,
-           RTLIB::FPTOSINT_F80_I128,
-           RTLIB::FPTOSINT_F128_I32,
-           RTLIB::FPTOSINT_F128_I64,
-           RTLIB::FPTOSINT_F128_I128,
-           RTLIB::FPTOUINT_F32_I32,
-           RTLIB::FPTOUINT_F32_I64,
-           RTLIB::FPTOUINT_F32_I128,
-           RTLIB::FPTOUINT_F64_I32,
-           RTLIB::FPTOUINT_F64_I64,
-           RTLIB::FPTOUINT_F64_I128,
-           RTLIB::FPTOUINT_F80_I32,
-           RTLIB::FPTOUINT_F80_I64,
-           RTLIB::FPTOUINT_F80_I128,
-           RTLIB::FPTOUINT_F128_I32,
-           RTLIB::FPTOUINT_F128_I64,
-           RTLIB::FPTOUINT_F128_I128,
-           RTLIB::SINTTOFP_I32_F32,
-           RTLIB::SINTTOFP_I32_F64,
-           RTLIB::SINTTOFP_I32_F80,
-           RTLIB::SINTTOFP_I32_F128,
-           RTLIB::SINTTOFP_I64_F32,
-           RTLIB::SINTTOFP_I64_F64,
-           RTLIB::SINTTOFP_I64_F80,
-           RTLIB::SINTTOFP_I64_F128,
-           RTLIB::SINTTOFP_I128_F32,
-           RTLIB::SINTTOFP_I128_F64,
-           RTLIB::SINTTOFP_I128_F80,
-           RTLIB::SINTTOFP_I128_F128,
-           RTLIB::UINTTOFP_I32_F32,
-           RTLIB::UINTTOFP_I32_F64,
-           RTLIB::UINTTOFP_I32_F80,
-           RTLIB::UINTTOFP_I32_F128,
-           RTLIB::UINTTOFP_I64_F32,
-           RTLIB::UINTTOFP_I64_F64,
-           RTLIB::UINTTOFP_I64_F80,
-           RTLIB::UINTTOFP_I64_F128,
-           RTLIB::UINTTOFP_I128_F32,
-           RTLIB::UINTTOFP_I128_F64,
-           RTLIB::UINTTOFP_I128_F80,
-           RTLIB::UINTTOFP_I128_F128,
-           RTLIB::OEQ_F32,
-           RTLIB::OEQ_F64,
-           RTLIB::OEQ_F128,
-           RTLIB::UNE_F32,
-           RTLIB::UNE_F64,
-           RTLIB::UNE_F128,
-           RTLIB::OGE_F32,
-           RTLIB::OGE_F64,
-           RTLIB::OGE_F128,
-           RTLIB::OLT_F32,
-           RTLIB::OLT_F64,
-           RTLIB::OLT_F128,
-           RTLIB::OLE_F32,
-           RTLIB::OLE_F64,
-           RTLIB::OLE_F128,
-           RTLIB::OGT_F32,
-           RTLIB::OGT_F64,
-           RTLIB::OGT_F128,
-           RTLIB::UO_F32,
-           RTLIB::UO_F64,
-           RTLIB::UO_F128,
-           RTLIB::O_F32,
-           RTLIB::O_F64,
-           RTLIB::O_F128,
-       })
-  setLibcallCallingConv(LC, CallingConv::ARM_AAPCS);
-}
-
 // The APCS parameter registers.
 static const MCPhysReg GPRArgRegs[] = {
   ARM::R0, ARM::R1, ARM::R2, ARM::R3
@@ -349,7 +184,22 @@ ARMTargetLowering::ARMTargetLowering(const TargetMachine &TM,
 
   setBooleanVectorContents(ZeroOrNegativeOneBooleanContent);
 
-  InitLibcallCallingConvs();
+  if (!Subtarget->isTargetDarwin() && !Subtarget->isTargetIOS() &&
+      !Subtarget->isTargetWatchOS()) {
+    const auto &E = Subtarget->getTargetTriple().getEnvironment();
+
+    bool IsHFTarget = E == Triple::EABIHF || E == Triple::GNUEABIHF ||
+                      E == Triple::MuslEABIHF;
+    // Windows is a special case.  Technically, we will replace all of the "GNU"
+    // calls with calls to MSVCRT if appropriate and adjust the calling
+    // convention then.
+    IsHFTarget = IsHFTarget || Subtarget->isTargetWindows();
+
+    for (int LCID = 0; LCID < RTLIB::UNKNOWN_LIBCALL; ++LCID)
+      setLibcallCallingConv(static_cast<RTLIB::Libcall>(LCID),
+                            IsHFTarget ? CallingConv::ARM_AAPCS_VFP
+                                       : CallingConv::ARM_AAPCS);
+  }
 
   if (Subtarget->isTargetMachO()) {
     // Uses VFP for Thumb libfuncs if available.

diff --git a/lib/Target/ARM/ARMISelLowering.h b/lib/Target/ARM/ARMISelLowering.h
index 7a7f91f4d3c460981ee272a4d9500bb7a440c68c..84c6eb845bb898cea953b9f804a611604a5202f6 100644 (file)
--- a/lib/Target/ARM/ARMISelLowering.h
+++ b/lib/Target/ARM/ARMISelLowering.h
@@ -538,8 +538,6 @@ class InstrItineraryData;
 
     bool HasStandaloneRem = true;
 
-    void InitLibcallCallingConvs();
-
     void addTypeForNEON(MVT VT, MVT PromotedLdStVT, MVT PromotedBitwiseVT);
     void addDRTypeForNEON(MVT VT);
     void addQRTypeForNEON(MVT VT);

diff --git a/test/CodeGen/Thumb2/float-intrinsics-double.ll b/test/CodeGen/Thumb2/float-intrinsics-double.ll
index 9e94990838299baad3c9f6d5ecce2cdb866a3b81..657d1b172da9854470684a27da7c4721c9746ce9 100644 (file)
--- a/test/CodeGen/Thumb2/float-intrinsics-double.ll
+++ b/test/CodeGen/Thumb2/float-intrinsics-double.ll
@@ -18,7 +18,7 @@ declare double     @llvm.powi.f64(double %Val, i32 %power)
 define double @powi_d(double %a, i32 %b) {
 ; CHECK-LABEL: powi_d:
 ; SOFT: {{(bl|b)}} __powidf2
-; HARD: bl __powidf2
+; HARD: b __powidf2
   %1 = call double @llvm.powi.f64(double %a, i32 %b)
   ret double %1
 }

diff --git a/test/CodeGen/Thumb2/float-intrinsics-float.ll b/test/CodeGen/Thumb2/float-intrinsics-float.ll
index fda840d90f36fc079cecc4a037bb728599114607..847aeacd2f914fc3d94cfdaab5c9f5b59e092fd6 100644 (file)
--- a/test/CodeGen/Thumb2/float-intrinsics-float.ll
+++ b/test/CodeGen/Thumb2/float-intrinsics-float.ll
@@ -18,7 +18,7 @@ declare float     @llvm.powi.f32(float %Val, i32 %power)
 define float @powi_f(float %a, i32 %b) {
 ; CHECK-LABEL: powi_f:
 ; SOFT: bl __powisf2
-; HARD: bl __powisf2
+; HARD: b __powisf2
   %1 = call float @llvm.powi.f32(float %a, i32 %b)
   ret float %1
 }

diff --git a/test/CodeGen/Thumb2/intrinsics-cc.ll b/test/CodeGen/Thumb2/intrinsics-cc.ll
new file mode 100644 (file)
index 0000000..ab5081e
--- /dev/null
+++ b/test/CodeGen/Thumb2/intrinsics-cc.ll
@@ -0,0 +1,41 @@
+; RUN: llc -mtriple thumbv7-unknown-none-eabi -float-abi soft -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MATCH
+; RUN: llc -mtriple thumbv7-unknown-none-eabi -float-abi hard -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MISMATCH -check-prefix CHECK-TO-SOFT
+; RUN: llc -mtriple thumbv7-unknown-none-eabihf -float-abi soft -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MISMATCH -check-prefix CHECK-TO-HARD
+; RUN: llc -mtriple thumbv7-unknown-none-eabihf -float-abi hard -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MATCH
+
+; RUN: llc -mtriple thumbv7-unknown-none-gnueabi -float-abi soft -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MATCH
+; RUN: llc -mtriple thumbv7-unknown-none-gnueabi -float-abi hard -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MISMATCH -check-prefix CHECK-TO-SOFT
+; RUN: llc -mtriple thumbv7-unknown-none-gnueabihf -float-abi soft -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MISMATCH -check-prefix CHECK-TO-HARD
+; RUN: llc -mtriple thumbv7-unknown-none-gnueabihf -float-abi hard -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MATCH
+
+; RUN: llc -mtriple thumbv7-unknown-none-musleabi -float-abi soft -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MATCH
+; RUN: llc -mtriple thumbv7-unknown-none-musleabi -float-abi hard -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MISMATCH -check-prefix CHECK-TO-SOFT
+; RUN: llc -mtriple thumbv7-unknown-none-musleabihf -float-abi soft -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MISMATCH -check-prefix CHECK-TO-HARD
+; RUN: llc -mtriple thumbv7-unknown-none-musleabihf -float-abi hard -filetype asm -o - %s | FileCheck %s -check-prefix CHECK-MATCH
+
+declare float @llvm.powi.f32(float, i32)
+
+define float @f(float %f, i32 %i) {
+entry:
+  %0 = call float @llvm.powi.f32(float %f, i32 %i)
+  ret float %0
+}
+
+; CHECK-MATCH: b __powisf2
+; CHECK-MISMATCH: bl __powisf2
+; CHECK-TO-SOFT: vmov s0, r0
+; CHECK-TO-HARD: vmov r0, s0
+
+declare double @llvm.powi.f64(double, i32)
+
+define double @g(double %d, i32 %i) {
+entry:
+  %0 = call double @llvm.powi.f64(double %d, i32 %i)
+  ret double %0
+}
+
+; CHECK-MATCH: b __powidf2
+; CHECK-MISMATCH: bl __powidf2
+; CHECK-TO-SOFT: vmov d0, r0, r1
+; CHECK-TO-HARD: vmov r0, r1, d0
+