X-Git-Url: https://git.tcpdump.org/tcpdump/blobdiff_plain/fb75d3cd5ad603bd255d9cdc20aeca674c6f3720..ffa1470e5c7ff0e50028d085a481dc797b0b51ed:/extract.h

diff --git a/extract.h b/extract.h
index 6aa21e87..f0479638 100644
--- a/extract.h
+++ b/extract.h
@@ -18,40 +18,109 @@
  * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
  * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
  *
- * @(#) $Header: /tcpdump/master/tcpdump/extract.h,v 1.16 2000-10-03 02:54:55 itojun Exp $ (LBL)
+ * @(#) $Header: /tcpdump/master/tcpdump/extract.h,v 1.24 2005-01-15 02:06:50 guy Exp $ (LBL)
  */
 
-/* Network to host order macros */
-
+/*
+ * Macros to extract possibly-unaligned big-endian integral values.
+ */
 #ifdef LBL_ALIGN
+/*
+ * The processor doesn't natively handle unaligned loads.
+ */
+#ifdef HAVE___ATTRIBUTE__
+/*
+ * We have __attribute__; we assume that means we have __attribute__((packed)).
+ * Declare packed structures containing a u_int16_t and a u_int32_t,
+ * cast the pointer to point to one of those, and fetch through it;
+ * the GCC manual doesn't appear to explicitly say that
+ * __attribute__((packed)) causes the compiler to generate unaligned-safe
+ * code, but it apppears to do so.
+ *
+ * We do this in case the compiler can generate, for this instruction set,
+ * better code to do an unaligned load and pass stuff to "ntohs()" or
+ * "ntohl()" than the code to fetch the bytes one at a time and
+ * assemble them.  (That might not be the case on a little-endian platform,
+ * where "ntohs()" and "ntohl()" might not be done inline.)
+ */
+typedef struct {
+	u_int16_t	val;
+} __attribute__((packed)) unaligned_u_int16_t;
+
+typedef struct {
+	u_int32_t	val;
+} __attribute__((packed)) unaligned_u_int32_t;
+
 #define EXTRACT_16BITS(p) \
-	((u_int16_t)*((u_int8_t *)(p) + 0) << 8 | \
-	(u_int16_t)*((u_int8_t *)(p) + 1))
+	((u_int16_t)ntohs(((const unaligned_u_int16_t *)(p))->val))
 #define EXTRACT_32BITS(p) \
-	((u_int32_t)*((u_int8_t *)(p) + 0) << 24 | \
-	(u_int32_t)*((u_int8_t *)(p) + 1) << 16 | \
-	(u_int32_t)*((u_int8_t *)(p) + 2) << 8 | \
-	(u_int32_t)*((u_int8_t *)(p) + 3))
-#else
+	((u_int32_t)ntohl(((const unaligned_u_int32_t *)(p))->val))
+#define EXTRACT_64BITS(p) \
+	((u_int64_t)(((u_int64_t)ntohl(((const unaligned_u_int32_t *)(p) + 0)->val)) << 32 | \
+		     ((u_int64_t)ntohl(((const unaligned_u_int32_t *)(p) + 1)->val)) << 0))
+
+#else /* HAVE___ATTRIBUTE__ */
+/*
+ * We don't have __attribute__, so do unaligned loads of big-endian
+ * quantities the hard way - fetch the bytes one at a time and
+ * assemble them.
+ */
+#define EXTRACT_16BITS(p) \
+	((u_int16_t)((u_int16_t)*((const u_int8_t *)(p) + 0) << 8 | \
+		     (u_int16_t)*((const u_int8_t *)(p) + 1)))
+#define EXTRACT_32BITS(p) \
+	((u_int32_t)((u_int32_t)*((const u_int8_t *)(p) + 0) << 24 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 1) << 16 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 2) << 8 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 3)))
+#define EXTRACT_64BITS(p) \
+	((u_int64_t)((u_int64_t)*((const u_int8_t *)(p) + 0) << 56 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 1) << 48 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 2) << 40 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 3) << 32 | \
+	             (u_int64_t)*((const u_int8_t *)(p) + 4) << 24 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 5) << 16 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 6) << 8 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 7)))
+#endif /* HAVE___ATTRIBUTE__ */
+#else /* LBL_ALIGN */
+/*
+ * The processor natively handles unaligned loads, so we can just
+ * cast the pointer and fetch through it.
+ */
 #define EXTRACT_16BITS(p) \
-	((u_int16_t)ntohs(*(u_int16_t *)(p)))
+	((u_int16_t)ntohs(*(const u_int16_t *)(p)))
 #define EXTRACT_32BITS(p) \
-	((u_int32_t)ntohl(*(u_int32_t *)(p)))
-#endif
+	((u_int32_t)ntohl(*(const u_int32_t *)(p)))
+#define EXTRACT_64BITS(p) \
+	((u_int64_t)(((u_int64_t)ntohl(*((const u_int32_t *)(p) + 0))) << 32 | \
+		     ((u_int64_t)ntohl(*((const u_int32_t *)(p) + 1))) << 0))
+#endif /* LBL_ALIGN */
 
 #define EXTRACT_24BITS(p) \
-	((u_int32_t)*((u_int8_t *)(p) + 0) << 16 | \
-	(u_int32_t)*((u_int8_t *)(p) + 1) << 8 | \
-	(u_int32_t)*((u_int8_t *)(p) + 2))
-
-/* Little endian protocol host order macros */
+	((u_int32_t)((u_int32_t)*((const u_int8_t *)(p) + 0) << 16 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 1) << 8 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 2)))
 
+/*
+ * Macros to extract possibly-unaligned little-endian integral values.
+ * XXX - do loads on little-endian machines that support unaligned loads?
+ */
 #define EXTRACT_LE_8BITS(p) (*(p))
 #define EXTRACT_LE_16BITS(p) \
-	((u_int16_t)*((u_int8_t *)(p) + 1) << 8 | \
-	(u_int16_t)*((u_int8_t *)(p) + 0))
+	((u_int16_t)((u_int16_t)*((const u_int8_t *)(p) + 1) << 8 | \
+		     (u_int16_t)*((const u_int8_t *)(p) + 0)))
 #define EXTRACT_LE_32BITS(p) \
-	((u_int32_t)*((u_int8_t *)(p) + 3) << 24 | \
-	(u_int32_t)*((u_int8_t *)(p) + 2) << 16 | \
-	(u_int32_t)*((u_int8_t *)(p) + 1) << 8 | \
-	(u_int32_t)*((u_int8_t *)(p) + 0))
+	((u_int32_t)((u_int32_t)*((const u_int8_t *)(p) + 3) << 24 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 2) << 16 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 1) << 8 | \
+		     (u_int32_t)*((const u_int8_t *)(p) + 0)))
+#define EXTRACT_LE_64BITS(p) \
+	((u_int64_t)((u_int64_t)*((const u_int8_t *)(p) + 7) << 56 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 6) << 48 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 5) << 40 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 4) << 32 | \
+	             (u_int64_t)*((const u_int8_t *)(p) + 3) << 24 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 2) << 16 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 1) << 8 | \
+		     (u_int64_t)*((const u_int8_t *)(p) + 0)))