Wireshark-bugs: [Wireshark-bugs] [Bug 9482] Buildbot crash output: fuzz-2013-11-27-22455.pcap

Date: Wed, 27 Nov 2013 23:19:04 +0000

Comment # 4 on bug 9482 from
(In reply to comment #3)
> I committed something similar in r53617 - before I saw this bug and the
> patch :-(

It's OK.  I had missed dissect_fc_ifcp() anyway.  Maybe it would be better to
change dissect_fc_helper() so it returns either 0 or tvb_length(tvb), depending
on whether fc_data is NULL or not, then it's only done in 1 place?

> It looks like a generic problem. A new-style dissector can never be sure
> that its data parameter contains something useful. Do we have to add a check
> everywhere?

Yes, I think so.  Any dissector expecting data should probably verify that data
!= NULL and reject the packet if it is.


You are receiving this mail because:
  • You are watching all bug changes.