Clarify what the special cases in wtap_wtap_encap_to_pcap_encap() are

doing.

svn path=/trunk/; revision=51209

diff --git a/wiretap/pcap-common.c b/wiretap/pcap-common.c
index 7b3879dc6ca2c20dd14ba418da487fd22ab4d86e..c5013a9dc6574e6a9a8db2079850b8defdf91344 100644 (file)
--- a/wiretap/pcap-common.c
+++ b/wiretap/pcap-common.c
@@ -636,7 +636,6 @@ wtap_wtap_encap_to_pcap_encap(int encap)
 
        case WTAP_ENCAP_FDDI:
        case WTAP_ENCAP_FDDI_BITSWAPPED:
-       case WTAP_ENCAP_NETTL_FDDI:
                /*
                 * Special-case WTAP_ENCAP_FDDI and
                 * WTAP_ENCAP_FDDI_BITSWAPPED; both of them get mapped
@@ -644,12 +643,25 @@ wtap_wtap_encap_to_pcap_encap(int encap)
                 * order in the FDDI MAC addresses is wrong; so it goes
                 * - libpcap format doesn't record the byte order,
                 * so that's not fixable).
+                *
+                * The pcap_to_wtap_map[] table will only have an
+                * entry for one of the above, which is why we have
+                * to special-case them.
+                */
+               return 10;      /* that's DLT_FDDI */
+
+       case WTAP_ENCAP_NETTL_FDDI:
+               /*
+                * This will discard the nettl information, as that's
+                * in the pseudo-header.
+                *
+                * XXX - what about Ethernet and Token Ring?
                 */
                return 10;      /* that's DLT_FDDI */
 
        case WTAP_ENCAP_FRELAY_WITH_PHDR:
                /*
-                * Do the same with Frame Relay.
+                * This will discard the pseudo-header information.
                 */
                return 107;
 
@@ -657,21 +669,9 @@ wtap_wtap_encap_to_pcap_encap(int encap)
                /*
                 * Map this to DLT_IEEE802_11, for now, even though
                 * that means the radio information will be lost.
-                * Once tcpdump support for the BSD radiotap header
-                * is sufficiently widespread, we should probably
-                * use that, instead - although we should probably
-                * ultimately just have WTAP_ENCAP_IEEE_802_11
-                * as the only Wiretap encapsulation for 802.11,
-                * and have the pseudo-header include a radiotap-style
-                * list of attributes.  If we do that, though, we
-                * should probably bypass the regular Wiretap code
-                * when writing out packets during a capture, and just
-                * do the equivalent of a libpcap write (unfortunately,
-                * libpcap doesn't have an "open dump by file descriptor"
-                * function, so we can't just use "pcap_dump()"), so
-                * that we don't spend cycles mapping from libpcap to
-                * Wiretap and then back to libpcap.  (There are other
-                * reasons to do that, e.g. to handle AIX libpcap better.)
+                * We should try to map those values to radiotap
+                * values and write this out as a radiotap file,
+                * if possible.
                 */
                return 105;
        }