[CodeGen][ARM] ARM runtime helper functions are not always soft-fp

[peter.smith]

Summary

The ARM Runtime ABI document (IHI0043) defines the AEABI floating point helper functions in section 4.1 Floating-point library. These functions always use the base PCS (soft-fp). However helper functions defined outside of this document such as the complex-number multiply and divide helpers are not covered by this requirement and should use hard-float PCS if the target is hard-float. All of the floating point helper functions that are explicitly soft float are expanded in the llvm ARM backend. This change makes clang not force the BuiltinCC to AAPCS for AAPCS_VFP.

With this change clang matches the behavior of gcc and the expected calling convention of the hard-float compiled compiler-rt and libgcc.a libraries as these functions do not have the attribute((pcs("aapcs"))).

fixes https://bugs.llvm.org/show_bug.cgi?id=28164
fixes test failures in compiler-rt for a hard-float target "divtc3_test.c, divdc3_test.c, divxc3_test.c"

I'm very new to clang so the way this has been fixed maybe a bit too simplistic. At present none of the builtins that produce calls in clang must use the base PCS. If clang ever needs to make a call to one of the AEABI defined functions that need base AAPCS then we'll need to identify the function when choosing a calling convention.

References:

• GCC list of functions that must be base AAPCS https://gcc.gnu.org/viewcvs/gcc/trunk/gcc/config/arm/arm.c?view=co&content-type=text%2Fplain
• Runtime ABI for the ARM Architecture http://infocenter.arm.com/help/topic/com.arm.doc.ihi0043d/IHI0043D_rtabi.pdf

[compnerd]

Can you please add a test that shows that the AEABI functions are not given the wrong CC? Also, can you show that if someone also passes -meabi=gnu with a VFP target, that we still annotate the functions with the differing CC (since that should prefer the GNU functions over the AEABI functions).

[peter.smith]

In D35538#814008, @compnerd wrote:

Can you please add a test that shows that the AEABI functions are not given the wrong CC? Also, can you show that if someone also passes -meabi=gnu with a VFP target, that we still annotate the functions with the differing CC (since that should prefer the GNU functions over the AEABI functions).

Thanks for the comment. Can I just check my understanding before I proceed? In particular I think I need a bit more information about what -meabi=gnu is supposed to do.
I think tests to make sure the the floating point helper functions always use AAPCS is a good idea and I'll start working on that. At the moment I'm trying to work out the best way of doing this as this is all done in the ARM backend in llvm. My current favorite options are to compile and check the assembly in the clang test suite, or just check that the calls to the floating point helper functions are not emitted by clang and add a test to llvm to make sure that they have the correct calling convention. I'm currently tending towards the second option.

With respect to -meabi=gnu, this was added in clang under r252463 and llvm r252462. From tracing through the flag through ARMTargetMachine.cpp and ARMISelLowering.cpp it seems like all this flag does in llvm is to use memcpy, memmove and memset instead of aeabi_memcpy, aeabi_memmove and aeabi_memset respectively. These functions always use AAPCS calling convention, but as they do not take any floating point parameters it doesn't matter either way. As it stands the option won't have any effect on the use of other library functions for example doing a double precision divide with -meabi=gnu results in a call to aeabi_ddiv. Does this match your expectation of what -meabi=gnu is supposed to do? If it doesn't can you point me at a description? All that I could find externally was the -meabi option in the GNU assembler (there is no equivalent gcc option), the value of -meabi=gnu maps to EF_ARM_ABI_UNKNOWN which is intended for generation of pre ARM ABI objects.

I'll be on vacation for the next 3 days, I'll pick this back up next week.

[peter.smith]

I've added a clang test that will check that none of the floating point helper functions defined in the Runtime ABI for the ARM Architecture are directly called by clang. Given that the calling convention decisions for these helpers are made in llvm I think the best place for a test that the correct calling convention is in llvm. I'll start working on that and post in a separate review. Please let me know if I'm not on the right lines.

[compnerd]

I think that the approach is reasonable. However, the test needs tweaking. The function and the call both are ambiguous in the match. Can you match the call sequence? Or explicitly the local label and then the call.

[majnemer]

lib/CodeGen/TargetInfo.cpp

5625 ↗

(On Diff #108092)

Period at the end of a sentence.

test/CodeGen/arm-float-helpers.c

32 ↗

(On Diff #108092)

Ditto.

test/CodeGen/complex-math.c

482 ↗

(On Diff #108092)

Ditto.

[peter.smith]

Thanks both for the comments, I've added the missing punctuation and have expanded the tests so that we anchor the check to the function and check for the operation that will be lowered by the back-end.

[peter.smith]

I've created D35904 to cover adding a test for LLVMs use of the runtime helper functions.

cfe/trunk/lib/CodeGen/TargetInfo.cpp

void ARMABIInfo::setCCs() {
   // they'd just match what LLVM will infer from the triple.
   llvm::CallingConv::ID abiCC = getABIDefaultCC();
   if (abiCC != getLLVMDefaultCC())
     RuntimeCC = abiCC;
 
   // AAPCS apparently requires runtime support functions to be soft-float, but
   // that's almost certainly for historic reasons (Thumb1 not supporting VFP
   // most likely). It's more convenient for AAPCS16_VFP to be hard-float.
-  switch (getABIKind()) {
-  case APCS:
-  case AAPCS16_VFP:
-    if (abiCC != getLLVMDefaultCC())
+
+  // The Run-time ABI for the ARM Architecture section 4.1.2 requires
+  // AEABI-complying FP helper functions to use the base AAPCS.
+  // These AEABI functions are expanded in the ARM llvm backend, all the builtin
+  // support functions emitted by clang such as the _Complex helpers follow the
+  // abiCC.
       BuiltinCC = abiCC;
-    break;
-  case AAPCS:
-  case AAPCS_VFP:
-    BuiltinCC = llvm::CallingConv::ARM_AAPCS;
-    break;
-  }
 }
 
 ABIArgInfo ARMABIInfo::classifyArgumentType(QualType Ty,
                                             bool isVariadic) const {
   // 6.1.2.1 The following argument types are VFP CPRCs:
   //   A single-precision floating-point type (including promoted
   //   half-precision types); A double-precision floating-point type;
   //   A 64-bit or 128-bit containerized vector type; Homogeneous Aggregate

cfe/trunk/test/CodeGen/arm-float-helpers.c

+// REQUIRES: arm-registered-target
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-linux-gnueabi %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-linux-gnueabihf %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-linux-gnueabi -target-feature "+soft-float" -target-feature "+soft-float-abi" %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-linux-gnueabi -target-feature "+soft-float" %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabi %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabi -meabi gnu %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabi %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabi -target-feature "+soft-float" -target-feature "+soft-float-abi" -meabi gnu %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabi -target-feature "+soft-float" -meabi gnu %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabihf %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -o - -triple arm-none-eabihf -meabi gnu %s | FileCheck %s
+
+// The Runtime ABI for the ARM Architecture IHI0043 section 4.1.2 The
+// floating-point helper functions to always use the base AAPCS (soft-float)
+// calling convention.
+//
+// These helper functions such as __aeabi_fadd are not explicitly called by
+// clang, instead they are generated by the ARMISelLowering when they are
+// needed; clang relies on llvm to use the base AAPCS.
+//
+// In this test we check that clang is not directly calling the __aeabi_
+// functions. We rely on llvm to test that the base AAPCS is used for any
+// __aeabi_ function from 4.1.2 that is used.
+//
+// When compiled to an object file with -mfloat-abi=soft each function F
+// below should result in a call to __aeabi_F. If clang is changed to call any
+// of these functions directly the test will need to be altered to check that
+// arm_aapcscc is used.
+//
+// Note that it is only the functions in 4.1.2 that must use the base AAPCS,
+// other runtime functions such as the _Complex helper routines are not covered.
+
+float fadd(float a, float b) { return a + b; }
+// CHECK-LABEL: define float @fadd(float %a, float %b)
+// CHECK-NOT: __aeabi_fadd
+// CHECK: %add = fadd float %0, %1
+
+float fdiv(float a, float b) { return a / b; }
+// CHECK-LABEL: define float @fdiv(float %a, float %b)
+// CHECK-NOT: __aeabi_fdiv
+// CHECK: %div = fdiv float %0, %1
+
+float fmul(float a, float b) { return a * b; }
+// CHECK-LABEL: define float @fmul(float %a, float %b)
+// CHECK-NOT: __aeabi_fmul
+// CHECK: %mul = fmul float %0, %1
+
+float fsub(float a, float b) { return a - b; }
+// CHECK-LABEL: define float @fsub(float %a, float %b)
+// CHECK-NOT: __aeabi_fsub
+// CHECK: %sub = fsub float %0, %1
+
+int fcmpeq(float a, float b) { return a == b; }
+// CHECK-LABEL: define i32 @fcmpeq(float %a, float %b)
+// CHECK-NOT: __aeabi_fcmpeq
+// CHECK: %cmp = fcmp oeq float %0, %1
+
+int fcmplt(float a, float b) { return a < b; }
+// CHECK-LABEL: define i32 @fcmplt(float %a, float %b)
+// CHECK-NOT: __aeabi_fcmplt
+// CHECK: %cmp = fcmp olt float %0, %1
+
+int fcmple(float a, float b) { return a <= b; }
+// CHECK-LABEL: define i32 @fcmple(float %a, float %b)
+// CHECK-NOT: __aeabi_fcmple
+// CHECK: %cmp = fcmp ole float %0, %1
+
+int fcmpge(float a, float b) { return a >= b; }
+// CHECK-LABEL: define i32 @fcmpge(float %a, float %b)
+// CHECK-NOT: __aeabi_fcmpge
+// CHECK: %cmp = fcmp oge float %0, %1
+
+int fcmpgt(float a, float b) { return a > b; }
+// CHECK-LABEL: define i32 @fcmpgt(float %a, float %b)
+// CHECK-NOT: __aeabi_fcmpgt
+// CHECK: %cmp = fcmp ogt float %0, %1
+
+int fcmpun(float a, float b) { return __builtin_isunordered(a, b); }
+// CHECK-LABEL: define i32 @fcmpun(float %a, float %b)
+// CHECK-NOT: __aeabi_fcmpun
+// CHECK: %cmp = fcmp uno double %conv, %conv1
+
+double dadd(double a, double b) { return a + b; }
+// CHECK-LABEL: define double @dadd(double %a, double %b)
+// CHECK-NOT: __aeabi_dadd
+// CHECK: %add = fadd double %0, %1
+
+double ddiv(double a, double b) { return a / b; }
+// CHECK-LABEL: define double @ddiv(double %a, double %b)
+// CHECK-NOT: __aeabi_ddiv
+// CHECK: %div = fdiv double %0, %1
+
+double dmul(double a, double b) { return a * b; }
+// CHECK-LABEL: define double @dmul(double %a, double %b)
+// CHECK-NOT: __aeabi_dmul
+// CHECK: %mul = fmul double %0, %1
+
+double dsub(double a, double b) { return a - b; }
+// CHECK-LABEL: define double @dsub(double %a, double %b)
+// CHECK-NOT: __aeabi_dsub
+// CHECK: %sub = fsub double %0, %1
+
+int dcmpeq(double a, double b) { return a == b; }
+// CHECK-LABEL: define i32 @dcmpeq(double %a, double %b)
+// CHECK-NOT: __aeabi_dcmpeq
+// CHECK: %cmp = fcmp oeq double %0, %1
+
+int dcmplt(double a, double b) { return a < b; }
+// CHECK-LABEL: define i32 @dcmplt(double %a, double %b)
+// CHECK-NOT: __aeabi_dcmplt
+// CHECK: %cmp = fcmp olt double %0, %1
+
+int dcmple(double a, double b) { return a <= b; }
+// CHECK-LABEL: define i32 @dcmple(double %a, double %b)
+// CHECK-NOT: __aeabi_dcmple
+// CHECK: %cmp = fcmp ole double %0, %1
+
+int dcmpge(double a, double b) { return a >= b; }
+// CHECK-LABEL: define i32 @dcmpge(double %a, double %b)
+// CHECK-NOT: __aeabi_dcmpge
+// CHECK: %cmp = fcmp oge double %0, %1
+
+int dcmpgt(double a, double b) { return a > b; }
+// CHECK-LABEL: define i32 @dcmpgt(double %a, double %b)
+// CHECK-NOT: __aeabi_dcmpgt
+// CHECK: %cmp = fcmp ogt double %0, %1
+
+int dcmpun(double a, double b) { return __builtin_isunordered(a, b); }
+// CHECK-LABEL: define i32 @dcmpun(double %a, double %b)
+// CHECK-NOT: __aeabi_dcmpun
+// CHECK: %cmp = fcmp uno double %0, %1
+
+int d2iz(double a) { return (int)a; }
+// CHECK-LABEL: define i32 @d2iz(double %a)
+// CHECK-NOT: __aeabi_d2iz
+// CHECK: %conv = fptosi double %0 to i32
+
+unsigned int d2uiz(double a) { return (unsigned int)a; }
+// CHECK-LABEL: define i32 @d2uiz(double %a)
+// CHECK-NOT: __aeabi_d2uiz
+// CHECK: %conv = fptoui double %0 to i32
+
+long long d2lz(double a) { return (long long)a; }
+// CHECK-LABEL: define i64 @d2lz(double %a)
+// CHECK-NOT: __aeabi_d2lz
+// CHECK: %conv = fptosi double %0 to i64
+
+unsigned long long d2ulz(double a) { return (unsigned long long)a; }
+// CHECK-LABEL: define i64 @d2ulz(double %a)
+// CHECK-NOT: __aeabi_d2ulz
+// CHECK: %conv = fptoui double %0 to i64
+
+int f2iz(float a) { return (int)a; }
+// CHECK-LABEL: define i32 @f2iz(float %a)
+// CHECK-NOT: __aeabi_f2iz
+// CHECK: %conv = fptosi float %0 to i32
+
+unsigned int f2uiz(float a) { return (unsigned int)a; }
+// CHECK-LABEL: define i32 @f2uiz(float %a)
+// CHECK-NOT: __aeabi_f2uiz
+// CHECK: %conv = fptoui float %0 to i32
+
+long long f2lz(float a) { return (long long)a; }
+// CHECK-LABEL: define i64 @f2lz(float %a)
+// CHECK-NOT: __aeabi_f2lz
+// CHECK: %conv = fptosi float %0 to i64
+
+unsigned long long f2ulz(float a) { return (unsigned long long)a; }
+// CHECK-LABEL: define i64 @f2ulz(float %a)
+// CHECK-NOT: __aeabi_f2ulz
+// CHECK: %conv = fptoui float %0 to i64
+
+float d2f(double a) { return (float)a; }
+// CHECK-LABEL: define float @d2f(double %a)
+// CHECK-NOT: __aeabi_d2f
+// CHECK: %conv = fptrunc double %0 to float
+
+double f2d(float a) { return (double)a; }
+// CHECK-LABEL: define double @f2d(float %a)
+// CHECK-NOT: __aeabi_f2d
+// CHECK: %conv = fpext float %0 to double
+
+double i2d(int a) { return (double)a; }
+// CHECK-LABEL: define double @i2d(i32 %a)
+// CHECK-NOT: __aeabi_i2d
+// CHECK: %conv = sitofp i32 %0 to double
+
+double ui2d(unsigned int a) { return (double)a; }
+// CHECK-LABEL: define double @ui2d(i32 %a)
+// CHECK-NOT: __aeabi_ui2d
+// CHECK: %conv = uitofp i32 %0 to double
+
+double l2d(long long a) { return (double)a; }
+// CHECK-LABEL: define double @l2d(i64 %a)
+// CHECK-NOT: __aeabi_l2d
+// CHECK: %conv = sitofp i64 %0 to double
+
+double ul2d(unsigned long long a) { return (unsigned long long)a; }
+// CHECK-LABEL: define double @ul2d(i64 %a)
+// CHECK-NOT: __aeabi_ul2d
+// CHECK: %conv = uitofp i64 %0 to double
+
+float i2f(int a) { return (int)a; }
+// CHECK-LABEL: define float @i2f(i32 %a)
+// CHECK-NOT: __aeabi_i2f
+// CHECK: %conv = sitofp i32 %0 to float
+
+float ui2f(unsigned int a) { return (unsigned int)a; }
+// CHECK-LABEL: define float @ui2f(i32 %a)
+// CHECK-NOT: __aeabi_ui2f
+// CHECK: %conv = uitofp i32 %0 to float
+
+float l2f(long long a) { return (long long)a; }
+// CHECK-LABEL: define float @l2f(i64 %a)
+// CHECK-NOT: __aeabi_l2f
+// CHECK: %conv = sitofp i64 %0 to float
+
+float ul2f(unsigned long long a) { return (unsigned long long)a; }
+// CHECK-LABEL: define float @ul2f(i64 %a)
+// CHECK-NOT: __aeabi_ul2f
+// CHECK: %conv = uitofp i64 %0 to float
+
+// Functions in section 4.1.2 not used by llvm and don't easily map directly to
+// C source code.
+// cfcmpeq
+// cfcmple
+// cfrcmple
+// cdcmpeq
+// cdcmple
+// cdrcmple
+// frsub
+// drsub

cfe/trunk/test/CodeGen/complex-math.c

 // RUN: %clang_cc1 %s -O1 -emit-llvm -triple x86_64-unknown-unknown -o - | FileCheck %s --check-prefix=X86
 // RUN: %clang_cc1 %s -O1 -emit-llvm -triple x86_64-pc-win64 -o - | FileCheck %s --check-prefix=X86
 // RUN: %clang_cc1 %s -O1 -emit-llvm -triple i686-unknown-unknown -o - | FileCheck %s --check-prefix=X86
 // RUN: %clang_cc1 %s -O1 -emit-llvm -triple powerpc-unknown-unknown -o - | FileCheck %s --check-prefix=PPC
-// RUN: %clang_cc1 %s -O1 -emit-llvm -triple armv7-none-linux-gnueabihf -o - | FileCheck %s --check-prefix=ARM
+// RUN %clang_cc1 %s -O1 -emit-llvm -triple armv7-none-linux-gnueabi -o - | FileCheck %s --check-prefix=ARM
+// RUN: %clang_cc1 %s -O1 -emit-llvm -triple armv7-none-linux-gnueabihf -o - | FileCheck %s --check-prefix=ARMHF
 // RUN: %clang_cc1 %s -O1 -emit-llvm -triple thumbv7k-apple-watchos2.0 -o - -target-abi aapcs16 | FileCheck %s --check-prefix=ARM7K
 
 float _Complex add_float_rr(float a, float b) {
   // X86-LABEL: @add_float_rr(
   // X86: fadd
   // X86-NOT: fadd
   // X86: ret
   return a + b;
_Bool ne_float_cc(float _Complex a, float _Complex b) {
   // X86: fcmp une
   // X86: or i1
   // X86: ret
   return a != b;
 }
 
 // Check that the libcall will obtain proper calling convention on ARM
 _Complex double foo(_Complex double a, _Complex double b) {
+  // These functions are not defined as floating point helper functions in
+  // Run-time ABI for the ARM architecture document so they must not always
+  // use the base AAPCS.
+
   // ARM-LABEL: @foo(
-  // ARM: call arm_aapcscc { double, double } @__muldc3
+  // ARM: call void { double, double } @__muldc3
+
+  // ARMHF-LABEL: @foo(
+  // ARMHF: call { double, double } @__muldc3
 
   // ARM7K-LABEL: @foo(
   // ARM7K: call { double, double } @__muldc3
   return a*b;
 }