On Sun, 18 Jul 2004, Bruce M Simpson wrote: > On Sun, Jul 18, 2004 at 03:59:02PM +0200, Willem Jan Withagen wrote: > > Coredump available.... 61Mb compressed. > > > > system contains a de card, which is correctly identified as GIANT-LOCKED. The > > card is not connected. Next time I open up the box, I'll remove it. > > Well this is no big surprise.... de(4) isn't locked properly and > probably never will be. However, locking problems with if_de can't (or shouldn't) explain the panic in question :-). We're probably looking either at a SACK problem, or a TCP locking problem. I need to review the TCP callouts, delivery, and reassembly code and make sure all is well. I also need to make sure that I really did merge all the relevant bits of the netperf patch relating to TCP -- there are some changes left, but I think they are purely layout changes. If you want to give it a spin, given some background in TCP, feel free :-). Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert_at_fledge.watson.org Principal Research Scientist, McAfee ResearchReceived on Sun Jul 18 2004 - 13:12:09 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:02 UTC