Have the "maybe an FCS" version of the Ethernet dissector take a data argument.

It's called from some places other than packet-frame.c, at least one of
which currently can't call anything else (the ATM dissector, for
VC-multiplexed bridged frames, where you don't know whether the frames
include the FCS or not), so the frame's pseudo-data doesn't necessarily
have the appropriate "FCS length" value.  Have it explicitly check the
data argument, and explicitly pass the appropriate value to it.

Ping-Bug: 9933
Change-Id: I0c75f921d25d1e2b75e476c15ff9625205036b25
Reviewed-on: https://code.wireshark.org/review/13382
Reviewed-by: Guy Harris <guy@alum.mit.edu>

1 files changed, 7 insertions, 1 deletions

diff --git a/epan/dissectors/packet-epon.c b/epan/dissectors/packet-epon.c
index c808534212..1ba9bcb285 100644
--- a/epan/dissectors/packet-epon.c
+++ b/epan/dissectors/packet-epon.c
@@ -79,6 +79,7 @@ dissect_epon(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree,
   guint       offset = 0;
   guint       dpoe_sec_byte;
   gboolean    dpoe_encrypted = FALSE;
+  struct eth_phdr eth;
 
   /* Start_of_Packet delimiter (/S/) can either happen in byte 1 or byte 2,
    * making the captured preamble either 7 or 6 bytes in length. If the
@@ -195,7 +196,12 @@ dissect_epon(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree,
     col_append_str(pinfo->cinfo, COL_INFO, " [ENCRYPTED]");
   } else {
     next_tvb = tvb_new_subset_remaining(tvb, 6+offset);
-    call_dissector(eth_handle, next_tvb, pinfo, tree);
+    /*
+     * XXX - is it guaranteed whether the capture will, or won't, have
+     * an FCS?
+     */
+    eth.fcs_len = -1;
+    call_dissector_with_data(eth_handle, next_tvb, pinfo, tree, &eth);
   }
 
   return tvb_captured_length(tvb);