Wireshark-bugs: [Wireshark-bugs] [Bug 9079] Buildbot crash output: fuzz-2013-08-27-29828.pcap

Date: Fri, 30 Aug 2013 18:16:45 +0000

Comment # 4 on bug 9079 from
(In reply to comment #3)
> (In reply to comment #2)
> > Looks like a pretty standard long-loop bug:
> > - iNbrRecords value is enormous
> > - none of the flags are set, so iOffsetPMR never advances and we never run
> > off  the end of the TVB
> > 
> > I am out of time tonight, somebody who knows the protocol better might know
> > the right fix.
> 
> I tried to simply jump out of the loop if we don't advance, I'll commit this
> change later today.

I tried a fix in r51603.

If none of the flags are set, we know up front that each run through the loop
will not dissect anything and will leave us pointing at the same data. I would
assume that can't happen as part of a valid message.

If I'm wrong, please feel free to correct my fix.


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