On Thursday 2010-10-07 09:40, Sven-Ola Tuecke wrote: >Am Donnerstag, 7. Oktober 2010, um 09:16:59 schrieben Sie: > >> 1.75 MB sounds like an older model, however, that "linux 2.4 use forced" >> statement is bogus for WRT54GL (4MB Flash) since at least openwrt >> 8.09/Linux 2.6.25. > >Yep, e.g. WAP54Gv2/v3. Only 2.4 Watt. And no: OpenWrt Backfire still supports a >brcm-2.4 target. You can use 2.6, but the opensourced WLAN driver does not >work well. You loose e.g. IBSS, AP mode, distance setting and (more criticial) >stability. Can't say. It does work well enough here. Sometimes I indeed lose the AP association, but that's probably because I have a long distance of ~60-80 freeland meters to the AP. It is easily remedied by calling iwconfig in a loop. Works now for over 1 3/4 years already. >> >As Jan pointed out: copy of the addresses may be stored elsewhere in the >> >form of hash keys. At least, I will place a warning in the docs and check >> >for oopses if conntrack is active. You think about a single rule either >> >in POSTROUTING or PREROUTING. May overcome above conntrack hash prob... >> >Good point - I place it on the TODO. >> >> RAWS/DNAT in Xtables-addons has this solved by using the raw and rawpost >> tables. > >No RAW table in 2.4 AFAICT. Yeah it's 2.6.6. You should update, or let a slightly better box do the 6-6 mapping ;-) >But a good hint, I'll check that. As a first >measure I placed a big fat warning at the start of the readme. Ref: > >http://map66.svn.sourceforge.net/viewvc/map66/README.html No oopses should happen. It's just that conntrack will record two different connections for what is one, and the states thus don't work reliably. -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
