Math builtin definition tweaks.

There were missed optimizations when the system headers didn't have attributes
in place, specifically:

- Add copysign, exp2, log2, nearbyint, rint and trunc to the list.
  These are functions that get inlined by LLVM's optimizer, but only when they
  have the right attributes.
- Mark copysign, fabs, fmax, fmin and trunc const unconditionally.
  Previously these were only const with -fno-math-errno, but they never set
  errno per POSIX.

For ceil/floor/nearbyint/round I'm not aware of any implementation that sets
errno, but POSIX says it may signal overflow so I left them alone for now.

llvm-svn: 162375

1 files changed, 176 insertions, 0 deletions

diff --git a/clang/test/CodeGen/libcall-declarations.c b/clang/test/CodeGen/libcall-declarations.c
new file mode 100644
index 00000000000..46a024a3416
--- /dev/null
+++ b/clang/test/CodeGen/libcall-declarations.c
@@ -0,0 +1,176 @@
+// RUN: %clang_cc1 -triple x86_64-apple-darwin12 -S -o - -emit-llvm %s | FileCheck %s -check-prefix=CHECK-NOERRNO
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -S -o - -emit-llvm %s | FileCheck %s -check-prefix=CHECK-ERRNO
+
+// Prototypes.
+double acos(double);
+long double acosl(long double);
+float acosf(float);
+double asin(double);
+long double asinl(long double);
+float asinf(float);
+double atan(double);
+long double atanl(long double);
+float atanf(float);
+double atan2(double, double);
+long double atan2l(long double, long double);
+float atan2f(float, float);
+double ceil(double);
+long double ceill(long double);
+float ceilf(float);
+double copysign(double, double);
+long double copysignl(long double, long double);
+float copysignf(float, float);
+double cos(double);
+long double cosl(long double);
+float cosf(float);
+double exp(double);
+long double expl(long double);
+float expf(float);
+double exp2(double);
+long double exp2l(long double);
+float exp2f(float);
+double fabs(double);
+long double fabsl(long double);
+float fabsf(float);
+double floor(double);
+long double floorl(long double);
+float floorf(float);
+double fma(double, double, double);
+long double fmal(long double, long double, long double);
+float fmaf(float, float, float);
+double fmax(double, double);
+long double fmaxl(long double, long double);
+float fmaxf(float, float);
+double fmin(double, double);
+long double fminl(long double, long double);
+float fminf(float, float);
+double log(double);
+long double logl(long double);
+float logf(float);
+double log2(double);
+long double log2l(long double);
+float log2f(float);
+double nearbyint(double);
+long double nearbyintl(long double);
+float nearbyintf(float);
+double pow(double, double);
+long double powl(long double, long double);
+float powf(float, float);
+double rint(double);
+long double rintl(long double);
+float rintf(float);
+double round(double);
+long double roundl(long double);
+float roundf(float);
+double sin(double);
+long double sinl(long double);
+float sinf(float);
+double sqrt(double);
+long double sqrtl(long double);
+float sqrtf(float);
+double tan(double);
+long double tanl(long double);
+float tanf(float);
+double trunc(double);
+long double truncl(long double);
+float truncf(float);
+
+// Force emission of the declare statements.
+void *use[] = {
+  acos, acosl, acosf, asin, asinl, asinf, atan, atanl, atanf, atan2, atan2l,
+  atan2f, ceil, ceill, ceilf, copysign, copysignl, copysignf, cos, cosl, cosf,
+  exp, expl, expf, exp2, exp2l, exp2f, fabs, fabsl, fabsf, floor, floorl,
+  floorf, fma, fmal, fmaf, fmax, fmaxl, fmaxf, fmin, fminl, fminf, log, logl,
+  logf, log2, log2l, log2f, nearbyint, nearbyintl, nearbyintf, pow, powl, powf,
+  rint, rintl, rintf, round, roundl, roundf, sin, sinl, sinf, sqrt, sqrtl,
+  sqrtf, tan, tanl, tanf, trunc, truncl, truncf
+};
+
+// CHECK-NOERRNO: declare double @acos(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @acosl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @acosf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @asin(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @asinl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @asinf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @atan(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @atanl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @atanf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @atan2(double, double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @atan2l(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @atan2f(float, float) nounwind readnone
+// CHECK-NOERRNO: declare double @ceil(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @ceill(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @ceilf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @copysign(double, double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @copysignl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @copysignf(float, float) nounwind readnone
+// CHECK-NOERRNO: declare double @cos(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @cosl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @cosf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @exp(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @expl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @expf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @exp2(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @exp2l(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @exp2f(float) nounwind readnone
+// CHECK-NOERRNO: declare double @fabs(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @fabsl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @fabsf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @floor(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @floorl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @floorf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @fma(double, double, double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @fmal(x86_fp80, x86_fp80, x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @fmaf(float, float, float) nounwind readnone
+// CHECK-NOERRNO: declare double @fmax(double, double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @fmaxl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @fmaxf(float, float) nounwind readnone
+// CHECK-NOERRNO: declare double @fmin(double, double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @fminl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @fminf(float, float) nounwind readnone
+// CHECK-NOERRNO: declare double @log(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @logl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @logf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @log2(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @log2l(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @log2f(float) nounwind readnone
+// CHECK-NOERRNO: declare double @nearbyint(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @nearbyintl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @nearbyintf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @pow(double, double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @powl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @powf(float, float) nounwind readnone
+// CHECK-NOERRNO: declare double @rint(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @rintl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @rintf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @round(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @roundl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @roundf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @sin(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @sinl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @sinf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @sqrt(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @sqrtl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @sqrtf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @tan(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @tanl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @tanf(float) nounwind readnone
+// CHECK-NOERRNO: declare double @trunc(double) nounwind readnone
+// CHECK-NOERRNO: declare x86_fp80 @truncl(x86_fp80) nounwind readnone
+// CHECK-NOERRNO: declare float @truncf(float) nounwind readnone
+
+// CHECK-ERRNO: declare double @copysign(double, double) nounwind readnone
+// CHECK-ERRNO: declare x86_fp80 @copysignl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-ERRNO: declare float @copysignf(float, float) nounwind readnone
+// CHECK-ERRNO: declare double @fabs(double) nounwind readnone
+// CHECK-ERRNO: declare x86_fp80 @fabsl(x86_fp80) nounwind readnone
+// CHECK-ERRNO: declare float @fabsf(float) nounwind readnone
+// CHECK-ERRNO: declare double @fmax(double, double) nounwind readnone
+// CHECK-ERRNO: declare x86_fp80 @fmaxl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-ERRNO: declare float @fmaxf(float, float) nounwind readnone
+// CHECK-ERRNO: declare double @fmin(double, double) nounwind readnone
+// CHECK-ERRNO: declare x86_fp80 @fminl(x86_fp80, x86_fp80) nounwind readnone
+// CHECK-ERRNO: declare float @fminf(float, float) nounwind readnone
+// CHECK-ERRNO: declare double @trunc(double) nounwind readnone
+// CHECK-ERRNO: declare x86_fp80 @truncl(x86_fp80) nounwind readnone
+// CHECK-ERRNO: declare float @truncf(float) nounwind readnone