[clang] [HIP] fix host min/max in header (PR #82956)

[Yaxun Liu via cfe-commits]

================
@@ -1306,15 +1306,68 @@ float min(float __x, float __y) { return __builtin_fminf(__x, __y); }
 __DEVICE__
 double min(double __x, double __y) { return __builtin_fmin(__x, __y); }
 
-#if !defined(__HIPCC_RTC__) && !defined(__OPENMP_AMDGCN__)
-__host__ inline static int min(int __arg1, int __arg2) {
-  return __arg1 < __arg2 ? __arg1 : __arg2;
+// Define host min/max functions.
+#if !defined(__HIPCC_RTC__) && !defined(__OPENMP_AMDGCN__) &&                  \
+    !defined(__HIP_NO_HOST_MIN_MAX_IN_GLOBAL_NAMESPACE__)
+
+#pragma push_macro("DEFINE_MIN_MAX_FUNCTIONS")
+#pragma push_macro("DEFINE_MIN_MAX_FUNCTIONS")
+#define DEFINE_MIN_MAX_FUNCTIONS(ret_type, type1, type2)                       \
+  static inline ret_type min(const type1 __a, const type2 __b) {               \
+    return (__a < __b) ? __a : __b;                                            \
+  }                                                                            \
+  static inline ret_type max(const type1 __a, const type2 __b) {               \
+    return (__a > __b) ? __a : __b;                                            \
+  }
+
+// Define min and max functions for same type comparisons
+DEFINE_MIN_MAX_FUNCTIONS(int, int, int)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned int, unsigned int, unsigned int)
+DEFINE_MIN_MAX_FUNCTIONS(long, long, long)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned long, unsigned long, unsigned long)
+DEFINE_MIN_MAX_FUNCTIONS(long long, long long, long long)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned long long, unsigned long long,
+                         unsigned long long)
+
+// Define min and max functions for all mixed type comparisons
+DEFINE_MIN_MAX_FUNCTIONS(unsigned int, int, unsigned int)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned int, unsigned int, int)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned long, long, unsigned long)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned long, unsigned long, long)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned long long, long long, unsigned long long)
+DEFINE_MIN_MAX_FUNCTIONS(unsigned long long, unsigned long long, long long)
----------------
yxsamliu wrote:

> I assume these are needed in order to avoid errors about ambiguous overload resolution when we pass signed/unsigned arguments.
> 
> Normally, if we were to use `std::min()` function, the user would have to explicitly cast arguments or use `std::min<unsigned>()` to resolve the issue.
> 
> Implicitly converting int->unsigned under the hood is probably not a good idea here as we do not know what the user needs/wants and whether it's a WAI or an error. For min/max converting a negative argument into an unsigned would probably be an error. I think we do need to force users to use one of the all-signed or all-unsigned variants here, too, same as with std::min/max.

You are right about that std::min/max does not allow mixed signed/unsigned arguments.

Unfortunately, that is how CUDA defines its host min/max functions, e.g. https://godbolt.org/z/Wxxq9dv91

Not following this could cause incompatibility and regressions.

https://github.com/llvm/llvm-project/pull/82956