Author: dim
Date: Sat Nov  2 16:59:53 2019
New Revision: 354255
URL: https://svnweb.freebsd.org/changeset/base/354255

Log:
  Add __isnan()/__isnanf() aliases for compatibility with glibc and CUDA
  
  Even though clang comes with a number of internal CUDA wrapper headers,
  compiling sample CUDA programs will result in errors similar to:
  
  In file included from <built-in>:1:
  In file included from 
/usr/lib/clang/9.0.0/include/__clang_cuda_runtime_wrapper.h:204:
  
/usr/home/arr/cuda/var/cuda-repo-10-0-local-10.0.130-410.48/usr/local/cuda-10.0//include/crt/math_functions.hpp:2910:7:
 error: no matching function for call to '__isnan'
    if (__isnan(a)) {
        ^~~~~~~
  /usr/lib/clang/9.0.0/include/__clang_cuda_device_functions.h:460:16: note: 
candidate function not viable: call to __device__ function from __host__ 
function
  __DEVICE__ int __isnan(double __a) { return __nv_isnand(__a); }
                 ^
  
  CUDA expects __isnan() and __isnanf() declarations to be available,
  which are glibc specific extensions, equivalent to the regular isnan()
  and isnanf().
  
  To provide these, define __isnan() and __isnanf() as aliases of the
  already existing static inline functions __inline_isnan() and
  __inline_isnanf() from math.h.
  
  Reported by:  arrowd
  PR:           241550
  MFC after:    1 week

Modified:
  head/lib/libc/gen/isnan.c
  head/lib/msun/src/math.h
  head/lib/msun/src/s_isnan.c

Modified: head/lib/libc/gen/isnan.c
==============================================================================
--- head/lib/libc/gen/isnan.c   Sat Nov  2 15:08:19 2019        (r354254)
+++ head/lib/libc/gen/isnan.c   Sat Nov  2 16:59:53 2019        (r354255)
@@ -43,6 +43,13 @@
  */
 
 #ifdef PIC
+/*
+ * Because math.h defines __isnan and __isnanf as aliases for compatibility 
with
+ * glibc and CUDA, we have to undefine them here to avoid redefinition errors.
+ */
+#undef __isnan
+#undef __isnanf
+
 __weak_reference(__isnan, isnan);
 __weak_reference(__isnanf, isnanf);
 

Modified: head/lib/msun/src/math.h
==============================================================================
--- head/lib/msun/src/math.h    Sat Nov  2 15:08:19 2019        (r354254)
+++ head/lib/msun/src/math.h    Sat Nov  2 16:59:53 2019        (r354255)
@@ -215,6 +215,12 @@ __inline_isnanl(__const long double __x)
 }
 
 /*
+ * Define the following aliases, for compatibility with glibc and CUDA.
+ */
+#define __isnan __inline_isnan
+#define __isnanf __inline_isnanf
+
+/*
  * Version 2 of the Single UNIX Specification (UNIX98) defined isnan() and
  * isinf() as functions taking double.  C99, and the subsequent POSIX revisions
  * (SUSv3, POSIX.1-2001, define it as a macro that accepts any real floating

Modified: head/lib/msun/src/s_isnan.c
==============================================================================
--- head/lib/msun/src/s_isnan.c Sat Nov  2 15:08:19 2019        (r354254)
+++ head/lib/msun/src/s_isnan.c Sat Nov  2 16:59:53 2019        (r354255)
@@ -45,6 +45,12 @@ isnan(double d)
 }
 #endif /* !PIC */
 
+/*
+ * Because math.h defines __isnanf as an alias for compatibility with glibc and
+ * CUDA, we have to undefine it here to avoid redefinition errors.
+ */
+#undef __isnanf
+
 int
 __isnanf(float f)
 {
_______________________________________________
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Reply via email to