[U-Boot] [PATCH] crc32: more optimizations

[U-Boot] [PATCH] crc32: more optimizations

[Joakim Tjernlund]

Shave off yet 4 bytes and make if faster for unaligned and/or
len & 3 != 0.

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
---

 Now I am done with crc32 optimizations.

 Someone should look at porting over crc32_be from linux
 since CRC32 BE is used by bzip and it could probably
 benefit from a faster CRC routine.

 lib_generic/crc32.c |   18 +++++++++---------
 1 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/lib_generic/crc32.c b/lib_generic/crc32.c
index 737587a..2837d93 100644
--- a/lib_generic/crc32.c
+++ b/lib_generic/crc32.c
@@ -188,30 +188,30 @@ uint32_t ZEXPORT crc32_no_comp(uint32_t crc, const Bytef *buf, uInt len)
     crc = __cpu_to_le32(crc);
     /* Align it */
     if(((long)b)&3 && len) {
+	 uint8_t *p = (uint8_t *)b - 1;
 	 do {
-	      uint8_t *p = (uint8_t *)b;
-	      DO_CRC(*p++);
-	      b = (void *)p;
+	      DO_CRC(*++p); /* use pre increment for speed */
 	 } while ((--len) && ((long)b)&3 );
+	 b = (uint32_t *)p + 1;
     }
 
     save_len = len & 3;
     len = len >> 2;
     for (--b; len; --len) {
 	 /* load data 32 bits wide, xor data 32 bits wide. */
-	 crc ^= *++b; /* use pre increment below(*++b) for speed */
+	 crc ^= *++b; /* use pre increment for speed */
 	 DO_CRC(0);
 	 DO_CRC(0);
 	 DO_CRC(0);
 	 DO_CRC(0);
     }
-    b++; /* point to next byte(s) */
     len = save_len;
     /* And the last few bytes */
-    for (; len; --len) {
-	 uint8_t *p = (uint8_t *)b;
-	 DO_CRC(*p++);
-	 b = (void *)p;
+    if (len) {
+	 uint8_t *p = (uint8_t *)(b++) - 1;
+	 do {
+	      DO_CRC(*++p); /* use pre increment for speed */
+	 } while (--len);
     }
 
     return __le32_to_cpu(crc);
-- 
1.6.4.4