[Bridge] Comments on the code

Wayback MachineAbout this captureCOLLECTED BY Organization: Alexa Crawls Starting in 1996, Alexa Internet has been donating their crawl data to the Internet Archive. Flowing in every day, these data are added to the Wayback Machine after an embargo period. Collection: Alexa Crawls DE Crawl data donated by Alexa Internet. This data is currently not publicly accessible TIMESTAMPSloadingLennert Buytenhekbuytenh@gnu.org
Sat, 1 Apr 2000 18:16:26 +0200 (MEST)On Fri, 31 Mar 2000 jacob@quantec.de wrote:> Some comments on the patches and on the code (sorry, I’ve forgotten it in> the other mail):>> – I had problems to compile the bridge code with my libc 2.0.7, so there> is one patch covering this. Don’t apply it blindly ;)I hope I’m not going to need autoconf….> – To get it working properly I separated is_local and is_static> completely.Okay. I’ll look at your patch.> – There are 3 namespaces to identify ports:> port index > port id> device name> IMHO port index and port id are equivalent (bijective mapping) and> to different id’s will never belong to the same device (please tell> me, if I missed something). Therefore I find it rather confusing, to> have one here and another there (e.g. showmacs vs. addif).addif enslaves an interface to a bridge, so it cannot be passed a port id,since ports which aren’t enslaved do not have a port id yet. showmacs usesport indices instead, which is perfectly logical. How on earth would youwant to enslave a port by using its port id?! Before you enslave it, itdoesn’t even have a port id!> I tried to make life easier for the users and changed br_find_port to> accept port indices and ids as well. But my main question remains: Is> it sensible to use all three in the user interface? At least the port> index gives no more information to the user (IMHO!).The indices are there as machine-locally unique netdevice identifiers. Theport IDs are there for the spanning tree protocol. I can’t remember usingdevice names apart from the bridge creation ioctl, which obviously needs adevice name.I’ve carefully designed the API, and everything is there for a reason.Lennert

[Bridge] incoming/outgoing

Wayback MachineAbout this captureCOLLECTED BY Organization: Alexa Crawls Starting in 1996, Alexa Internet has been donating their crawl data to the Internet Archive. Flowing in every day, these data are added to the Wayback Machine after an embargo period. Collection: Alexa Crawls DE Crawl data donated by Alexa Internet. This data is currently not publicly accessible TIMESTAMPSloadingLuckylucky@hera.osk.uni-bielefeld.de
Thu, 17 Feb 2000 12:17:56 +0100Quoting Janne Liimatainen…>> Hi.>> I just realized I’d need filtering on incoming interface. If I redirect> port 80 on the br0-chain with ipchains then eg. Squid doesn’t work on the> bridge.>> What good sides are there on filtering on outgoing?Why do not make this optional? Those who want incoming filtering select incoming filtering, and those whowant outgoing select outgoing?Im not a real programmer, so i dont know if its possible…– bye— Lucky