[builtins] Deduplicate __eqsf2 and __gtsf2

[phosek]

Summary

The only difference between eqsf2 and gtsf2 is whether they return
1 or -1 on NaN. Rather than duplicating all the code, use a macro to
define the function twice and use an argument to decide whether to
negate the return value.

[mcgrathr]

Lgtm

[yln]

Hi Petr (@phosek),

this change broke 2 of our internal bots that test on older OS/Compilers: MacOS 10.13/Xcode 9.1 and MacOS 10.12/Xcode 8.3.3.

Can you make sense of the following build error?

FAILED: projects/compiler-rt/lib/builtins/CMakeFiles/clang_rt.soft_static_armv7m_macho_embedded.dir/arm/comparesf2.S.o 
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/cc  -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -Iprojects/compiler-rt/lib/builtins -I/Users/buildslave/jenkins/workspace/stash-master-sanitizer-10.13-incremental/clang/src/projects/compiler-rt/lib/builtins -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.13.sdk/usr/include/libxml2 -Iinclude -I/Users/buildslave/jenkins/workspace/stash-master-sanitizer-10.13-incremental/clang/src/include -O2 -g -DNDEBUG    -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.13.Internal.sdk -Oz -Wall -fomit-frame-pointer -ffreestanding -arch armv7m -static -mfloat-abi=soft -MD -MT projects/compiler-rt/lib/builtins/CMakeFiles/clang_rt.soft_static_armv7m_macho_embedded.dir/arm/comparesf2.S.o -MF projects/compiler-rt/lib/builtins/CMakeFiles/clang_rt.soft_static_armv7m_macho_embedded.dir/arm/comparesf2.S.o.d -o projects/compiler-rt/lib/builtins/CMakeFiles/clang_rt.soft_static_armv7m_macho_embedded.dir/arm/comparesf2.S.o -c /Users/buildslave/jenkins/workspace/stash-master-sanitizer-10.13-incremental/clang/src/projects/compiler-rt/lib/builtins/arm/comparesf2.S
<instantiation>:4:46: error: unexpected token in directive
.thumb ; .subsections_via_symbols ; .globl _ __eqsf2 ; ; .thumb_func ; _ __eqsf2:
                                             ^
/Users/buildslave/jenkins/workspace/stash-master-sanitizer-10.13-incremental/clang/src/projects/compiler-rt/lib/builtins/arm/comparesf2.S:88:1: note: while in macro instantiation
COMPARESF2_FUNCTION __eqsf2, __eqsf2_handle_nan
^
<instantiation>:4:81: error: unexpected token in argument list
.thumb ; .subsections_via_symbols ; .globl _ __eqsf2 ; ; .thumb_func ; _ __eqsf2:
                                                                                ^
/Users/buildslave/jenkins/workspace/stash-master-sanitizer-10.13-incremental/clang/src/projects/compiler-rt/lib/builtins/arm/comparesf2.S:88:1: note: while in macro instantiation
COMPARESF2_FUNCTION __eqsf2, __eqsf2_handle_nan
^
<instantiation>:4:46: error: unexpected token in directive
.thumb ; .subsections_via_symbols ; .globl _ __gtsf2 ; ; .thumb_func ; _ __gtsf2:
                                             ^
/var/folders/dk/vj0dyfmd1kb6tzqbcbmzz58m0000gn/T/comparesf2-16b20c.s:75:1: note: while in macro instantiation
COMPARESF2_FUNCTION __gtsf2, __gtsf2_handle_nan
^
<instantiation>:4:81: error: unexpected token in argument list
.thumb ; .subsections_via_symbols ; .globl _ __gtsf2 ; ; .thumb_func ; _ __gtsf2:
                                                                                ^
/var/folders/dk/vj0dyfmd1kb6tzqbcbmzz58m0000gn/T/comparesf2-16b20c.s:75:1: note: while in macro instantiation
COMPARESF2_FUNCTION __gtsf2, __gtsf2_handle_nan
^

compiler-rt/lib/builtins/arm/comparesf2.S

 //                           1 if either a or b is NaN
 //
 // Note that __lesf2( ) and __gesf2( ) are identical except in their handling of
 // NaN values.
 //
 //===----------------------------------------------------------------------===//
 
 #include "../assembly.h"
-    .syntax unified
-    .text
-    DEFINE_CODE_STATE
 
-@ int __eqsf2(float a, float b)
+    .macro COMPARESF2_FUNCTION name:req negate_nan
+@ int \name(float a, float b)
 
     .p2align 2
-DEFINE_COMPILERRT_FUNCTION(__eqsf2)
+DEFINE_COMPILERRT_FUNCTION(\name)
+
 #if defined(COMPILER_RT_ARMHF_TARGET)
     vmov r0, s0
     vmov r1, s1
 #endif
     // Make copies of a and b with the sign bit shifted off the top.  These will
     // be used to detect zeros and NaNs.
 #if defined(USE_THUMB_1)
     push    {r6, lr}
#endif
     // negative, this places 0 in r0; if a is negative and b positive, -1 is
     // placed in r0.
 #if defined(USE_THUMB_1)
     bhs     1f
     // Here if a and b have the same sign and absA < absB, the result is thus
     // b < 0 ? 1 : -1. Same if a and b have the opposite sign (ignoring Nan).
     movs    r0,         #1
     lsrs    r1,         #31
-    bne     LOCAL_LABEL(CHECK_NAN)
+    bne     LOCAL_LABEL(CHECK_NAN\@)
     negs    r0,         r0
-    b       LOCAL_LABEL(CHECK_NAN)
+    b       LOCAL_LABEL(CHECK_NAN\@)
 1:
 #else
     it lo
     mvnlo   r0,         r1, asr #31
 #endif
 
     // If a is greater in magnitude than b and both have the same sign, place
     // the sign of b in r0.  Thus, if both are negative and a < b, -1 is placed
     // in r0, which is the desired result.  Conversely, if both are positive
     // and a > b, zero is placed in r0.
 #if defined(USE_THUMB_1)
     bls     1f
     // Here both have the same sign and absA > absB.
     movs    r0,         #1
     lsrs    r1,         #31
-    beq     LOCAL_LABEL(CHECK_NAN)
+    beq     LOCAL_LABEL(CHECK_NAN\@)
     negs    r0, r0
 1:
 #else
     it hi
     movhi   r0,         r1, asr #31
 #endif
 
     // If you've been keeping track, at this point r0 contains -1 if a < b and
     // 0 if a >= b.  All that remains to be done is to set it to 1 if a > b.
     // If a == b, then the Z flag is set, so we can get the correct final value
     // into r0 by simply or'ing with 1 if Z is clear.
     // For Thumb-1, r0 contains -1 if a < b, 0 if a > b and 0 if a == b.
 #if !defined(USE_THUMB_1)
     it ne
     orrne   r0,     r0, #1
 #endif
 
     // Finally, we need to deal with NaNs.  If either argument is NaN, replace
     // the value in r0 with 1.
 #if defined(USE_THUMB_1)
-LOCAL_LABEL(CHECK_NAN):
+LOCAL_LABEL(CHECK_NAN\@):
     movs    r6,         #0xff
     lsls    r6,         #24
     cmp     r2,         r6
     bhi     1f
     cmp     r3,         r6
 1:
     bls     2f
     movs    r0,         #1
+.ifnb \negate_nan
+    negs    r0,         r0
+.endif
 2:
     pop     {r6, pc}
 #else
     cmp     r2,         #0xff000000
     ite ls
     cmpls   r3,         #0xff000000
+.ifb \negate_nan
     movhi   r0,         #1
+.else
+    movhi   r0,         #-1
+.endif
     JMP(lr)
 #endif
-END_COMPILERRT_FUNCTION(__eqsf2)
+END_COMPILERRT_FUNCTION(\name)
+    .endm
+
+    .syntax unified
+    .text
+    DEFINE_CODE_STATE
+
+COMPARESF2_FUNCTION __eqsf2
 
 DEFINE_COMPILERRT_FUNCTION_ALIAS(__lesf2, __eqsf2)
 DEFINE_COMPILERRT_FUNCTION_ALIAS(__ltsf2, __eqsf2)
 DEFINE_COMPILERRT_FUNCTION_ALIAS(__nesf2, __eqsf2)
 
 #if defined(__ELF__)
 // Alias for libgcc compatibility
 DEFINE_COMPILERRT_FUNCTION_ALIAS(__cmpsf2, __lesf2)
 #endif
 
-@ int __gtsf2(float a, float b)
-
-    .p2align 2
-DEFINE_COMPILERRT_FUNCTION(__gtsf2)
-    // Identical to the preceding except in that we return -1 for NaN values.
-    // Given that the two paths share so much code, one might be tempted to
-    // unify them; however, the extra code needed to do so makes the code size
-    // to performance tradeoff very hard to justify for such small functions.
-#if defined(COMPILER_RT_ARMHF_TARGET)
-    vmov r0, s0
-    vmov r1, s1
-#endif
-#if defined(USE_THUMB_1)
-    push    {r6, lr}
-    lsls    r2,        r0, #1
-    lsls    r3,        r1, #1
-    lsrs    r6,        r3, #1
-    orrs    r6,        r2
-    beq     1f
-    movs    r6,        r0
-    eors    r6,        r1
-1:
-    bmi     2f
-    subs    r0,        r2, r3
-2:
-    bhs     3f
-    movs    r0,        #1
-    lsrs    r1,        #31
-    bne     LOCAL_LABEL(CHECK_NAN_2)
-    negs    r0, r0
-    b       LOCAL_LABEL(CHECK_NAN_2)
-3:
-    bls     4f
-    movs    r0,         #1
-    lsrs    r1,         #31
-    beq     LOCAL_LABEL(CHECK_NAN_2)
-    negs    r0, r0
-4:
-LOCAL_LABEL(CHECK_NAN_2):
-    movs    r6,         #0xff
-    lsls    r6,         #24
-    cmp     r2,         r6
-    bhi     5f
-    cmp     r3,         r6
-5:
-    bls     6f
-    movs    r0,         #1
-    negs    r0,         r0
-6:
-    pop     {r6, pc}
-#else
-    mov     r2,         r0, lsl #1
-    mov     r3,         r1, lsl #1
-    orrs    r12,    r2, r3, lsr #1
-    it ne
-    eorsne  r12,    r0, r1
-    it pl
-    subspl  r0,     r2, r3
-    it lo
-    mvnlo   r0,         r1, asr #31
-    it hi
-    movhi   r0,         r1, asr #31
-    it ne
-    orrne   r0,     r0, #1
-    cmp     r2,         #0xff000000
-    ite ls
-    cmpls   r3,         #0xff000000
-    movhi   r0,         #-1
-    JMP(lr)
-#endif
-END_COMPILERRT_FUNCTION(__gtsf2)
+COMPARESF2_FUNCTION __gtsf2 negate_nan
 
 DEFINE_COMPILERRT_FUNCTION_ALIAS(__gesf2, __gtsf2)
 
 @ int __unordsf2(float a, float b)
 
     .p2align 2
 DEFINE_COMPILERRT_FUNCTION(__unordsf2)