Improved linkind process, added padding generator

1 files changed, 5 insertions, 22 deletions

diff --git a/common/bytes.c b/common/bytes.c
index 1196989..af65568 100644
--- a/common/bytes.c
+++ b/common/bytes.c
@@ -1,31 +1,14 @@
 #include <apos/types.h>
 #include <apos/attrs.h>
+#include <apos/bytes.h>
+#include <apos/builtin.h>
 
-/* Functions for swapping endianness.
- *
- * Typically you wouldn't call these directly, instead using macros such as
- * le16_to_cpu, cpu_to_be32, etc. that take the system endianness into account.
- *
- * The macros cpu_to_be64 etc. use GCC's and Clang's __builtin_bswap16/32/64, but in
- * the case that the arch compiler doesn't provide the __builtin functions they're replaced
- * with __bswaphi2/si2/di2 respectively, and it is up to the developer to
- * provide these. The naming scheme uses GCC's internal naming scheme, where
- *
- * HI ~= 16 bits
- * SI ~= 32 bits
- * DI ~= 64 bits
- *
- * Here we assume this to be true in all cases, and it probably is, but in the
- * case that some architecture is different (SI ~= 64 bits or something) we define
- * these functions as __weak, allowing any arch to redefine them.
- */
-
-__weak uint16_t __bswaphi2(uint16_t u)
+__weak uint16_t __bswap16(uint16_t u)
 {
 	return (u & 0xff00) >> 8 | (u & 0x00ff) << 8;
 }
 
-__weak uint32_t __bswapsi2(uint32_t u)
+__weak uint32_t __bswap32(uint32_t u)
 {
 	return (u & 0xff000000) >> 24 |
 		(u & 0x00ff0000) >> 8 |
@@ -33,7 +16,7 @@ __weak uint32_t __bswapsi2(uint32_t u)
 		(u & 0x000000ff) << 24;
 }
 
-__weak uint64_t __bswapdi2(uint64_t u)
+__weak uint64_t __bswap64(uint64_t u)
 {
 	return (u & 0xff00000000000000ULL) >> 56	|
 		(u & 0x00ff000000000000ULL) >> 40	|