Messages in this thread | | | Date | Tue, 05 Jan 1999 07:26:18 -0800 | From | Dan Kegel <> | Subject | Re: Undocumented feature in 2.2.0p1 |
| |
Mike Perry schrieb: > I've found an undocumented feature in 2.2.0p1 IPmasqing. In the 2.0 series > of kernels, when you had a masq network and wanted to use Blizzard's > BattleNet servers for Starcraft internet play, you had to use ipautofw to > forward port 6112 tcp/udp to battle.net. Under 2.2.0p1, not only do you > not have to do this anymore...
Sounds like you have applied the "ip_masq_dloose" patch (included in 2.2.0pre1-ac2, but not yet in Linus' tree, I think). The description of this patch on http://juanjox.linuxhq.com/ is: patch-2.2.0pre1-ac1.ip_masq_dloose.gz Changes UDP stream semantics by keeping only one tunnel per source <addr,port> regardless of changes in destination addr,port (note that this matches "normal" sockets semantics).
I gave Juanjo the idea for that patch; the goal was to improve support for multiplayer games. You can read more at http://alumni.caltech.edu/~dank/peer-nat.html Using ipautofw limits you to one player on the masq'd LAN, I think, but the new code lets you have unlimited players, without any manual configuration, if the game is written properly.
> but if you DO specify autofw (using the new > ipmasqadm and autofw module) with those ports, communication with > battle.net is broken.. When you do nothing, game play is fine.
I'm suprised that using autofw causes trouble. I don't know enough about battle.net to guess what the problem is; all my testing was done with Activision games. If you don't apply 2.2.0pre1-ac1, does starcraft work when using the new ipmasqadm/autofw stuff? - Dan -- Speaking only for myself, not for my employer
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |