**** BEGIN LOGGING AT Wed Nov 12 02:59:58 2008
Nov 12 04:56:22 <rwhitby> anyone seeing this failure in xrelayd? xrelayd.c:(.text+0x14f4): undefined reference to `x509write_keyfile'
Nov 12 11:49:32 <DagMoller> hi folks
Nov 12 11:49:39 <wrt> welcome to #openwrt-devel!
Nov 12 11:49:40 <DagMoller> who is working on ar7port?
Nov 12 12:01:50 <[Fate]> DagMoller https://dev.openwrt.org/wiki/platforms
Nov 12 12:02:47 <DagMoller> matteo, are you working on ar7?
Nov 12 12:02:52 <DagMoller> [Fate], thx
Nov 12 12:03:05 <matteo> DagMoller: yes, but the serial bug is killing my development
Nov 12 12:03:12 <matteo> i read _nothing_ from teh serial
Nov 12 12:03:37 <DagMoller> so, i'm working on ar7 tiatn...
Nov 12 12:03:57 <matteo> teh DSL?
Nov 12 12:04:00 <DagMoller> and at moment, unit boots openwrt sucessful
Nov 12 12:04:08 <DagMoller> matteo, no, WRTP54g
Nov 12 12:04:30 <DagMoller> matteo, we got LAN, WAN and WIFI working
Nov 12 12:04:41 <matteo> great
Nov 12 12:04:43 <matteo> with acx?
Nov 12 12:04:47 <DagMoller> yes
Nov 12 12:05:10 <DagMoller> but wifi need to coorect reset_bit
Nov 12 12:05:22 <DagMoller> lonestar discover that bit is 14
Nov 12 12:05:53 <DagMoller> but i need to undertand what is cpmac and what is vlynq
Nov 12 12:06:37 <matteo> vlynq is a bus developed by TI
Nov 12 12:07:00 <matteo> cpmac is an eth driver for the TI eth which uses the vlynq bus
Nov 12 12:07:12 <matteo> theyr wifi card uses vlynq too
Nov 12 12:07:23 <DagMoller> yes, wifi is in vlynq0
Nov 12 12:08:02 <DagMoller> the reset ocurs in vlynq or cpmac or both?
Nov 12 12:09:05 <DagMoller> matteo, take a look http://openwrt.pastebin.com/m7d5c9b38
Nov 12 12:09:51 <matteo> what's titan?
Nov 12 12:10:03 <DagMoller> a derivation of ar7 o think
Nov 12 12:10:18 <DagMoller> titan board
Nov 12 12:10:50 <DagMoller> matteo, the wifi only works after a cold boot
Nov 12 12:10:55 <matteo> DagMoller: i think i added that reset
Nov 12 12:11:00 <matteo> to avoid an IRQ storm
Nov 12 12:11:44 <DagMoller> i found cpmac reset, but none to vlynq
Nov 12 12:13:38 <DagMoller> matteo, in target/linux/ar7/files/include/asm-mips/ar7/ar7.h i found AR7_RESET_BIT_CPMAC_LO
Nov 12 12:14:29 <DagMoller> but in target/linux/ar7/files/arch/mips/ar7/platform.c cpmac_{low,high}_data .reset_bit is hardcoded
Nov 12 12:14:34 <DagMoller> its wright?
Nov 12 12:17:34 <matteo> mmm
Nov 12 12:18:19 <DagMoller> and other problem is that mii-tool always says that link is down on eth
Nov 12 12:18:30 <DagMoller> but eth works
Nov 12 12:18:42 <matteo> DagMoller: i know
Nov 12 12:18:52 <matteo> that's why ia had the temporary_fix
Nov 12 12:18:57 <matteo> it brings the iface up
Nov 12 12:19:11 <matteo> eg. it starts teh netif queue
Nov 12 12:19:15 <matteo> even if it's up
Nov 12 12:19:28 <matteo> and the reset bit it's hardcoded, right
Nov 12 12:22:31 <DagMoller> matteo, you know when vlynq reset ocurs?
Nov 12 12:23:10 <matteo> no
Nov 12 12:23:16 <matteo> ejka wrote all the vlynq code
Nov 12 12:38:38 <danage> DagMoller: sn9 was fiddling with it too, and you could ask loswillios as well
Nov 12 12:56:50 <DagMoller> danage, long time sn9 no comes online
Nov 12 13:15:03 <danage> i know
Nov 12 13:15:19 <danage> maybe he fried all his routers?
Nov 12 13:17:42 <DagMoller> dunno
Nov 12 13:21:04 <danage> DagMoller: try #ar7
Nov 12 13:21:10 <DagMoller> :D
Nov 12 13:38:33 <DagMoller> /sbin/wifi: eval: line 1: cannot create /sys/class/leds/wifi/brightness: nonexistent directory
Nov 12 13:38:34 <DagMoller> wlan0(acx): enable failed
Nov 12 13:39:12 <danage> acx is broken
Nov 12 13:39:16 <danage> try #acx100
Nov 12 13:39:53 <DagMoller> danage, i think that wifi led is missing on my patch
Nov 12 13:46:19 <DagMoller> http://openwrt.pastebin.com/m23e77b53
Nov 12 13:51:13 <DagMoller> wifi is working... but only after a cold boot
Nov 12 13:52:20 <DagMoller> matteo, danage ejka look http://openwrt.pastebin.com/m23e77b53
Nov 12 14:07:39 <danage> that would be surprising news because to my knowledge, acx driver is non-functional atm
Nov 12 14:08:55 <danage> DagMoller: AndyIL knows a lot about ar7 hardware too afaik
Nov 12 14:09:16 <DagMoller> danage, i configure then wlan0, and in my coputer, execute a scan, and found the router
Nov 12 14:10:06 <danage> try transmitting data
Nov 12 14:10:19 <DagMoller> danage, try it later
Nov 12 17:22:33 <russell_> nbd: is there anything more i can be doing to help figure out the soekris problem?
Nov 12 17:49:31 <CIA-8> juhosg * r13192 /trunk/ (42 files in 27 dirs):
Nov 12 17:49:31 <CIA-8> [kernel] update to 2.6.25.20, 2.6.26.8, 2.6.27.5 and refresh patches
Nov 12 17:49:31 <CIA-8>  * based on a patch by Hauke Mehrtens
Nov 12 17:49:31 <CIA-8>  * closes #4193
Nov 12 18:08:35 <CIA-8> juhosg * r13193 /trunk/target/linux/ar71xx/files/drivers/net/ag71xx/ (ag71xx_main.c ag71xx_mdio.c): [ar71xx] cosmetic changes in the ag71xx driver
Nov 12 18:37:39 <nbd> russell_: you could try to make a patch that reverts the changes in the kernel version that broke it
Nov 12 18:37:58 <nbd> russell_: (make that patch against current trunk)
Nov 12 18:38:03 <nbd> see if that works
Nov 12 18:38:07 <nbd> and if it does, reduce the patch
Nov 12 18:38:09 <nbd> until it no longer works
Nov 12 18:38:28 <russell_> is there an easy way to create revert patches from git?
Nov 12 18:38:48 <nbd> git-revert
Nov 12 18:38:54 <russell_> heh.
Nov 12 18:38:54 <nbd> iirc
Nov 12 18:40:03 <CyrusFF> for patches i'd rather say: git diff commit commit^
Nov 12 18:40:25 <russell_> there are a bunch of patches that went into the 2.6.25.4 -> 2.6.25.6 update, which i've looked at and can pretty safely exclude some of.  i'll see if i can create the reverts and try them one at a time.
Nov 12 18:40:36 <russell_> CyrusFF: i'll try that too, thanks
Nov 12 18:40:45 <nbd> git-revert has the advantage of rebasing the revert commit to your current head
Nov 12 18:40:55 <nbd> so if you get conflicts you can resolve them the git way
Nov 12 19:13:48 <russell_> where should i put the revert patch so that it gets noticed by the build system and the kernel gets rebuilt?
Nov 13 00:55:05 <x-alina> Is there a in the Buildroot a variable for downloading files like $(WGET) or so?
Nov 13 01:35:59 <nbd> you should not do wget manually in a package makefile
Nov 13 01:36:05 <nbd> that's what the Download template is for
Nov 13 01:36:14 <nbd> so that the build system takes care of mirroring and downloading for you
Nov 13 01:36:56 <nbd> x-alina: package/libertas/Makefile is an example
Nov 13 01:41:59 <x-alina> nbd: Okay, thanks i will fix this tomorrow. Is there a wrapper for unzip like $(UNZIP), or can it handle the build system too?
Nov 13 01:48:11 <nbd> i think there is no wrapper for that yet
Nov 13 02:19:37 <LetoTo1> fyi: if you enable via the GUI that it should run /etc/firewall.user, it does not do it.
Nov 13 02:20:20 <LetoTo1> i am also not a fan of the huge number of chains, and ended up hacking it to run iptables -I zone_wan_REJECT -j ACCEPT
Nov 13 02:20:38 <LetoTo1> support for having a subnet from your isp is not present, and these rules make it very hard
Nov 13 02:20:55 <LetoTo1> eg I have a ppp0 1.2.3.4 on which my ISP tunnels 2.3.4.0/28
Nov 13 02:21:25 <LetoTo1> it requires a lot of rule changes to ensure those IP's get forwarded, and non-masq'ed to 1.2.3.4
Nov 13 02:21:48 <LetoTo1> also, the gui is missing the option to configure additional IP addresses on br-lan (I need to put a 2.3.4.x address on it)
Nov 13 02:22:05 <LetoTo1> oh, and the GUI "save" does not work, but "save and apply" works
Nov 13 02:22:22 <LetoTo1> that's my bug reports on 8.09_RC1 for today. thanks :)
Nov 13 02:23:14 <LetoTo1> oh. and no openswan packge :P but I'll fix that myself :P
Nov 13 02:36:23 <LetoTo1> kernel OCF broke too. cannot select it in make menuconfig
**** ENDING LOGGING AT Thu Nov 13 02:59:57 2008