On 20 Nov 2002, Jason Dixon wrote: > I'm still trying to isolate exactly which PF reassembly configuration(s) > can't handle the test igmp DoS I'm sending at it in the "lab". I've > currently got frags and state limited to 65k each. I've tried both crop > and drop-ovl, both (much less full reassemble) experience kernel panics > after various amounts of effort. Generally between 2000 and 10000 igmp > packets. Packets were sent via igmpofdeath, which fragments and spoofs > packets to the destination. Note that the trace below is different than > the one I get from full reassembly tests. > > Should I forward this over to tech@ or just leave it here?
Ideally you use sendbug(1), so the bug report gets into the GNAT bug tracking system. Read http://www.openbsd.org/report.html for more info. The GNAT is disabled on www.openbsd.org, but the reports seem to be read. To be sure CC the report to [EMAIL PROTECTED] as well; that's what I do. BTW the crash seems to be unrelated to the pool exhaustion bug. So I guess it's a different bug. Cheers, Dries -- Dries Schellekens email: [EMAIL PROTECTED]
