On Sat, Jun 22, 2002 at 01:03:34AM +0200, Mark Santcroos wrote: > On Fri, Jun 21, 2002 at 11:04:36AM -0700, Brooks Davis wrote: > > For my purposes, it would need to be seperate so you could copy the > > module and hack in a new TCP without changing the existing one. > > I understand, but you won't need to do that for the IP layer in your case. > Other people might have a reverse situation, so some hooks to both these > layers would come in handy, that was my point.
It depends on what you're trying to do. If all you want to do is mess with in-kernel TCP implementations then just hooking into the existing IP layer is sufficent. I'm also thinking that the ability to run netgraph code in a hybrid userland/kernel environment for development would be useful in which case it would be useful to be able to implement the whole network stack in netgraph nodes. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4
msg35179/pgp00000.pgp
Description: PGP signature
