[PATCH 1/1] compat: modernize and simplify byte swapping functions

[Rostislav Krasny <rostiprodev@gmail.com>]

Replace manual bit operations with memcpy + network functions for better
maintainability. Add missing 16-bit network byte order conversion.

Key improvements:
- Simplify the get_be*() and put_be*() functions
- Add bswap16() macro with automatic compiler optimization:
  * GCC/Clang: __builtin_bswap16() intrinsic
  * MSVC: _byteswap_ushort() intrinsic
- Add default_bswap16() static inline function with manual fallback implementation
- Rename default_swab32() to default_bswap32() for naming consistency
- Add ntohs() and htons() macros for complete 16/32/64-bit network conversion
- Add put_be16() function for API completeness alongside existing put_be*() and
  get_be*() functions
- Performance improvements (GCC 15.2.1, Clang 21.1.7):
  * on x86-64 with -O0 4.2x faster (GCC), 3.7x faster (Clang)
  * on x86-64 with -O1 4x faster (GCC), identical (Clang)
  * on x86-64 with -O2 identical (GCC), 1.8x faster (Clang)

Signed-off-by: Rostislav Krasny <rostiprodev@gmail.com>
---
 compat/bswap.h | 74 ++++++++++++++++++++++++++++++--------------------
 1 file changed, 44 insertions(+), 30 deletions(-)

diff --git a/compat/bswap.h b/compat/bswap.h
index 28635ebc69..f9954ef090 100644
--- a/compat/bswap.h
+++ b/compat/bswap.h
@@ -9,10 +9,15 @@
  */
 
 /*
- * Default version that the compiler ought to optimize properly with
+ * Default versions that the compiler ought to optimize properly with
  * constant values.
  */
-static inline uint32_t default_swab32(uint32_t val)
+static inline uint16_t default_bswap16(uint16_t val)
+{
+	return ((val & 0xff00) >> 8) | ((val & 0x00ff) << 8);
+}
+
+static inline uint32_t default_bswap32(uint32_t val)
 {
 	return (((val & 0xff000000) >> 24) |
 		((val & 0x00ff0000) >>  8) |
@@ -40,6 +45,7 @@ static inline uint64_t default_bswap64(uint64_t val)
 # define __has_builtin(x) 0
 #endif
 
+#undef bswap16
 #undef bswap32
 #undef bswap64
 
@@ -47,6 +53,7 @@ static inline uint64_t default_bswap64(uint64_t val)
 
 #include <stdlib.h>
 
+#define bswap16(x) _byteswap_ushort(x)
 #define bswap32(x) _byteswap_ulong(x)
 #define bswap64(x) _byteswap_uint64(x)
 
@@ -54,8 +61,9 @@ static inline uint64_t default_bswap64(uint64_t val)
 #define GIT_BIG_ENDIAN 4321
 #define GIT_BYTE_ORDER GIT_LITTLE_ENDIAN
 
-#elif __has_builtin(__builtin_bswap32) && __has_builtin(__builtin_bswap64)
+#elif __has_builtin(__builtin_bswap16) && __has_builtin(__builtin_bswap32) && __has_builtin(__builtin_bswap64)
 
+#define bswap16(x) __builtin_bswap16((x))
 #define bswap32(x) __builtin_bswap32((x))
 #define bswap64(x) __builtin_bswap64((x))
 
@@ -98,24 +106,36 @@ static inline uint64_t default_bswap64(uint64_t val)
 
 #endif
 
+#undef ntohs
+#undef htons
 #undef ntohl
 #undef htonl
 #undef ntohll
 #undef htonll
 
 #if GIT_BYTE_ORDER == GIT_BIG_ENDIAN
+# define ntohs(x) (x)
+# define htons(x) (x)
 # define ntohl(x) (x)
 # define htonl(x) (x)
 # define ntohll(x) (x)
 # define htonll(x) (x)
 #else
 
+# if defined(bswap16)
+#  define ntohs(x) bswap16(x)
+#  define htons(x) bswap16(x)
+# else
+#  define ntohs(x) default_bswap16(x)
+#  define htons(x) default_bswap16(x)
+# endif
+
 # if defined(bswap32)
 #  define ntohl(x) bswap32(x)
 #  define htonl(x) bswap32(x)
 # else
-#  define ntohl(x) default_swab32(x)
-#  define htonl(x) default_swab32(x)
+#  define ntohl(x) default_bswap32(x)
+#  define htonl(x) default_bswap32(x)
 # endif
 
 # if defined(bswap64)
@@ -129,47 +149,41 @@ static inline uint64_t default_bswap64(uint64_t val)
 
 static inline uint16_t get_be16(const void *ptr)
 {
-	const unsigned char *p = ptr;
-	return	(uint16_t)p[0] << 8 |
-		(uint16_t)p[1] << 0;
+	uint16_t n;
+	memcpy(&n, ptr, sizeof n);
+	return ntohs(n);
 }
 
 static inline uint32_t get_be32(const void *ptr)
 {
-	const unsigned char *p = ptr;
-	return	(uint32_t)p[0] << 24 |
-		(uint32_t)p[1] << 16 |
-		(uint32_t)p[2] <<  8 |
-		(uint32_t)p[3] <<  0;
+	uint32_t n;
+	memcpy(&n, ptr, sizeof n);
+	return ntohl(n);
 }
 
 static inline uint64_t get_be64(const void *ptr)
 {
-	const unsigned char *p = ptr;
-	return	(uint64_t)get_be32(&p[0]) << 32 |
-		(uint64_t)get_be32(&p[4]) <<  0;
+	uint64_t n;
+	memcpy(&n, ptr, sizeof n);
+	return ntohll(n);
+}
+
+static inline void put_be16(void *ptr, uint16_t value)
+{
+	uint16_t n = htons(value);
+	memcpy(ptr, &n, sizeof n);
 }
 
 static inline void put_be32(void *ptr, uint32_t value)
 {
-	unsigned char *p = ptr;
-	p[0] = (value >> 24) & 0xff;
-	p[1] = (value >> 16) & 0xff;
-	p[2] = (value >>  8) & 0xff;
-	p[3] = (value >>  0) & 0xff;
+	uint32_t n = htonl(value);
+	memcpy(ptr, &n, sizeof n);
 }
 
 static inline void put_be64(void *ptr, uint64_t value)
 {
-	unsigned char *p = ptr;
-	p[0] = (value >> 56) & 0xff;
-	p[1] = (value >> 48) & 0xff;
-	p[2] = (value >> 40) & 0xff;
-	p[3] = (value >> 32) & 0xff;
-	p[4] = (value >> 24) & 0xff;
-	p[5] = (value >> 16) & 0xff;
-	p[6] = (value >>  8) & 0xff;
-	p[7] = (value >>  0) & 0xff;
+	uint64_t n = htonll(value);
+	memcpy(ptr, &n, sizeof n);
 }
 
 #endif /* COMPAT_BSWAP_H */
-- 
2.52.0