[libc] Support the string conversion methods on the GPU

[jhuber6]

Summary

This patch enables us to use the existing libc support for string
conversion functions on the GPU. This required setting the fenv_t and
long double configuration. As far as I am aware, long doubles are
converted to doubles on the GPU and the floating point environment is
just an uint32_t.

This code is still untested as we are still working out how to run the
unit tests on the GPU.

[michaelrj]

LGTM with a nit.

libc/src/__support/FPUtil/PlatformDefs.h

20

Nit: given the precedent of putting a comment with a link to the relevant documentation above, it would be nice to have a link to the AMD or Nvidia documentation saying they use long double as double.

[jhuber6]

I found documentation for AMD, for CUDA it will simply emit a warning if you use long double with cudafe.

[jhuber6]

Change to struct.

libc/config/gpu/api.td

 include "config/public_api.td"
 
 include "spec/stdc.td"
 
 def StringAPI : PublicAPI<"string.h"> {
   let Types = ["size_t"];
 }
 
 def StdlibAPI : PublicAPI<"stdlib.h"> {
   let Types = [
     "size_t",
     "__atexithandler_t",
   ];
 }
+
+def FenvAPI: PublicAPI<"fenv.h"> {
+  let Types = ["fenv_t"];
+}

libc/config/gpu/entrypoints.txt

set(TARGET_LIBC_ENTRYPOINTS
     libc.src.string.strrchr
     libc.src.string.strspn
     libc.src.string.strstr
     libc.src.string.strtok
     libc.src.string.strtok_r
 
     # stdlib.h entrypoints
     libc.src.stdlib.atoi
+    libc.src.stdlib.atof
+    libc.src.stdlib.atol
+    libc.src.stdlib.atoll
     libc.src.stdlib.atexit
+    libc.src.stdlib.strtod
+    libc.src.stdlib.strtof
+    libc.src.stdlib.strtol
+    libc.src.stdlib.strtold
+    libc.src.stdlib.strtoll
+    libc.src.stdlib.strtoul
+    libc.src.stdlib.strtoull
 
     # Only implemented in the test suite
     libc.src.stdlib.malloc
     libc.src.stdlib.aligned_alloc
     libc.src.stdlib.realloc
     libc.src.stdlib.free
 
     # errno.h entrypoints
     libc.src.errno.errno
 )
 
 set(TARGET_LLVMLIBC_ENTRYPOINTS
   ${TARGET_LIBC_ENTRYPOINTS}
 )
 

libc/config/gpu/headers.txt

 set(TARGET_PUBLIC_HEADERS
     libc.include.ctype
     libc.include.string
+    libc.include.fenv
     libc.include.errno
     libc.include.stdlib
 )

libc/docs/gpu/support.rst

 
 stdlib.h
 --------
 
 =============  =========  ============
 Function Name  Available  RPC Required
 =============  =========  ============
 atoi           |check|
+atof           |check|
+atol           |check|
+atoll          |check|
+atexit         |check|
+strtod         |check|
+strtof         |check|
+strtol         |check|
+strtold        |check|
+strtoll        |check|
+strtoul        |check|
+strtoull       |check|
 =============  =========  ============

libc/include/llvm-libc-types/fenv_t.h

typedef struct {
   unsigned char __mxcsr[4];
 } fenv_t;
 #elif defined(__arm__) || defined(_M_ARM)
 typedef struct {
   unsigned int __fpscr;
 } fenv_t;
 #elif defined(__riscv)
 typedef unsigned int fenv_t;
+#elif defined(__AMDGPU__) || defined(__NVPTX__)
+typedef unsigned int fenv_t;
 #else
 #error "fenv_t not defined for your platform"
 #endif
 
 #endif // __LLVM_LIBC_TYPES_FENV_T_H__

libc/src/__support/FPUtil/PlatformDefs.h

 #include "src/__support/macros/properties/architectures.h"
 
 #if defined(LIBC_TARGET_ARCH_IS_X86)
 #define X87_FPU
 #endif
 
 // https://developer.arm.com/documentation/dui0491/i/C-and-C---Implementation-Details/Basic-data-types
 // https://developer.apple.com/documentation/xcode/writing-arm64-code-for-apple-platforms
-#if defined(_WIN32) || defined(__arm__) ||                                     \
-    (defined(__APPLE__) && defined(__aarch64__))
+#if defined(_WIN32) || defined(__arm__) || defined(__NVPTX__) ||               \
+    defined(__AMDGPU__) || (defined(__APPLE__) && defined(__aarch64__))
 #define LONG_DOUBLE_IS_DOUBLE
 #endif
 
 #if !defined(LONG_DOUBLE_IS_DOUBLE) && defined(X87_FPU)
 #define SPECIAL_X86_LONG_DOUBLE
 #endif
 
 #endif // LLVM_LIBC_SRC_SUPPORT_FPUTIL_PLATFORM_DEFS_H