**** BEGIN LOGGING AT Tue Sep 08 02:59:57 2009
Sep 08 05:54:02 <dkostousov> I wrote some module with uci-config and init-script. How I can execute the init-script after editing uci-config via LuCI?
Sep 08 06:50:27 <dkostousov> How I can restart a init-script after changing config via LuCI?
Sep 08 06:52:39 <frogonwheels> dkostousov:  /etc/init.d/<script> restart    (or reload if it supports it)
Sep 08 06:53:38 <frogonwheels> .. or go through luci's version of the same - if you have that module installed.
Sep 08 07:03:21 <CIA-36> juhosg * r17539 /trunk/target/linux/ar71xx/ (config-2.6.28 config-2.6.30 config-2.6.31): ar71xx: the adm6996 phy driver does not work with the RouterStation, disable it until we find a better solution
Sep 08 07:25:21 <CIA-36> juhosg * r17540 /trunk/ (Config.in include/kernel-defaults.mk): add menuconfig option for enabling debugfs support in the kernel
Sep 08 07:32:09 <CIA-36> juhosg * r17541 /trunk/package/mac80211/ (Config.in Config.in.ath9k Config.in.b43 Makefile): package/mac80211: add menuconfig option for enabling debug support via debugfs
Sep 08 07:51:39 <Hopscotch> good morning
Sep 08 09:25:52 <roh> ola
Sep 08 09:27:42 <roh> hm.. is there a port to the 'congstar dsl-box zwei' yet?
Sep 08 09:29:01 <blogic> roh: whats the soc inside ?
Sep 08 09:31:00 <roh> broadcom
Sep 08 09:31:08 <blogic> which ?
Sep 08 09:31:12 <blogic> 63xx ?
Sep 08 09:31:17 <blogic> or
Sep 08 09:31:19 <roh> 6338
Sep 08 09:31:23 <blogic> does it have a built in dsl modem ?
Sep 08 09:31:24 <blogic> ok
Sep 08 09:31:26 <roh> it is one
Sep 08 09:31:33 <roh> only 1x ethernet
Sep 08 09:31:35 <blogic> apart from ethernet and dsl it works :)
Sep 08 09:31:41 <roh> argh.
Sep 08 09:31:46 <roh> i need the ethernet part.
Sep 08 09:31:48 <blogic> ethernet should work
Sep 08 09:31:58 <roh> one can get them for 4Euro at pollin
Sep 08 09:31:58 <blogic> but for some reason the switch does not come up
Sep 08 09:32:03 <blogic> haha
Sep 08 09:32:31 <roh> http://www.pollin.de/shop/shop.php?cf=detail.php&pg=NQ==&a=NTM4ODgyOTk=
Sep 08 09:32:34 <roh> no switch in there
Sep 08 09:32:38 <blogic> yes
Sep 08 09:32:40 <blogic> inside the soc
Sep 08 09:32:44 <roh> only the soc, a psu and some external parts for the dsl
Sep 08 09:32:48 <blogic> sure
Sep 08 09:32:52 <blogic> it still wont work :)
Sep 08 09:33:05 <blogic> ethernet bug is known but we have not found the solution yet
Sep 08 09:33:13 <roh> it has jtag on the bottom as pads, and serial on top on a 4x pinheader (guessing here, justgot it and made photos)
Sep 08 09:33:20 <blogic> as for dsl, there is a rewrite but borkedcom refuses to let the suthor GPL it
Sep 08 09:33:27 <blogic> hehe
Sep 08 09:33:32 <blogic> sorry
Sep 08 09:33:42 <blogic> it'll be a while until 6338 is fully supported
Sep 08 09:33:48 <roh> blogic huh? did the author write it for brmc?
Sep 08 09:33:55 <blogic> no
Sep 08 09:34:00 <roh> then so what.
Sep 08 09:34:03 <blogic> well
Sep 08 09:34:08 <blogic> there is a thing called NDA
Sep 08 09:34:19 <blogic> i assume you have signed a few in your life :)
Sep 08 09:34:34 <roh> well.. stupid if one signs one which doesnt allow using knowledge to write gpl.
Sep 08 09:34:41 <blogic> not really
Sep 08 09:34:50 <blogic> he was allowed to release everyting apart from dsl
Sep 08 09:34:56 <blogic> and his company signed, not him
Sep 08 09:34:57 <roh> yes. i've singned some too. but they mostly attend to company interna and stuff like that.
Sep 08 09:35:03 <blogic> me too
Sep 08 09:35:10 <blogic> i never sign a source code related nda
Sep 08 09:35:20 <blogic> and i sign no contratcs that force me to write non gpl code
Sep 08 09:35:27 <blogic> but that is personal preference i assume
Sep 08 09:35:32 <roh> means i cannot remember any nda which prohibits me from writing gpl code. only ones which prohibit me from leaking some documents
Sep 08 09:36:02 <roh> but a well written gpl driver with comments is more worth than the usual buggy documentation where one needs to find out most bugs himself
Sep 08 09:36:20 <blogic> :)
Sep 08 09:36:28 <roh> well.. so i got 5 adsl2+ modems spare now ;)
Sep 08 09:36:38 <roh> not really a loss.
Sep 08 09:36:58 <roh> hoped to use them as 'ethernet to microcontroller modems'
Sep 08 09:37:46 <blogic> hehe
Sep 08 09:37:57 <blogic> pollin has the netio thing
Sep 08 09:38:03 <blogic> whihc works just fine
Sep 08 09:38:12 <blogic> if you just need uC 2 eth
Sep 08 09:38:20 <roh> sucks
Sep 08 09:38:33 <roh> the stack is crap and the parts can't handle bandwith
Sep 08 09:38:51 <roh> i want it to be usable in segments where one could face a few mbit of multicast traffic.
Sep 08 09:38:52 <blogic> the stack is crap ?!
Sep 08 09:38:57 <blogic> uip ?
Sep 08 09:39:03 <blogic> ah ok
Sep 08 09:39:09 <blogic> hmm
Sep 08 09:39:27 <roh> the ip stack in avr. just not enough ram/cpu to do things properly. and high-level protocolls like mdns or so you cant do at all
Sep 08 09:39:35 <blogic> yes
Sep 08 09:39:46 <roh> i mean.. one could prolly use one avr per osi-layer ;)
Sep 08 09:39:58 <blogic> and you need dsl ?
Sep 08 09:40:13 <blogic> or just a ethernet port ?
Sep 08 09:41:04 <roh> only eth for now.
Sep 08 09:41:19 <roh> if i need a dsl modem i think i can life with the oem which is already on.
Sep 08 09:41:49 <roh> after all there is nothing to config there. routing and terminating pppoe needs to be done on a linux where i got root anyhow
Sep 08 09:59:02 <dkostousov> frogonwheels: how the LuCI restart network after configuration was changed? I can't find that code in the LuCI src.
Sep 08 11:45:18 <frogonwheels> dkostousov: I'm not sure. it could be through uci triggers or something.
Sep 08 13:12:39 <acoul> ping mb_
Sep 08 13:12:59 <nbd> acoul: he's not in this channel
Sep 08 13:13:39 <nbd> btw. i'm almost done with a modified version of the bfs patch which adds compile time scheduler selection
Sep 08 13:13:44 <nbd> for 2.6.30 for now
Sep 08 13:13:50 <nbd> i'll port it to .31 as well if this works
Sep 08 13:14:06 <acoul> I have a BFS patched brcm47xx/2.6.31-rc9 image for him to test
Sep 08 13:14:09 <nbd> i managed to remove most of the parts where this patch touches generic code
Sep 08 13:14:37 <nbd> cool
Sep 08 13:14:55 <acoul> I don't have the hardware here to test
Sep 08 13:14:59 <nbd> can you commit the brcm47xx 2.6.31 patches and kernel config (without bfs)?
Sep 08 13:15:12 <nbd> but not switch to it by default
Sep 08 13:15:14 <acoul> yes ofcourse
Sep 08 13:15:19 <nbd> thanks
Sep 08 13:19:21 <acoul> I will commit the brcm47xx/2.6.31 as soon as I test it (next 5 hours)
Sep 08 13:19:45 <acoul> if you need the patches sooner let me know
Sep 08 13:19:58 <nbd> no, i don't need them at the moment
Sep 08 13:19:59 <nbd> take your time
Sep 08 14:37:37 <nbd> mb__: hi
Sep 08 14:38:06 <nbd> i reworked the bfs patch so that it touches much fewer places in the kernel and it's now possible to switch between bfs and cfs via kconfig
Sep 08 14:38:59 <mb__> very cool, nbd
Sep 08 14:39:10 <nbd> i think i'll commit it soon
Sep 08 14:39:14 <mb__> Did you also run some tests with it?
Sep 08 14:39:18 <nbd> yes
Sep 08 14:39:19 * mb__ is excited
Sep 08 14:39:27 <nbd> on atheros wisoc it doesn't make a difference when routing from ethernet to ethernet
Sep 08 14:39:42 <nbd> but maybe that's because the bus is the bottleneck there, not the cpu
Sep 08 14:39:52 <mb__> yeah
Sep 08 14:40:22 <nbd> atheros' implementation of AHB seems really crappy
Sep 08 14:40:46 <nbd> at least on the old ar2315
Sep 08 14:44:06 <nbd> here's my current patch: http://nbd.name/bfs-209nbd.patch
Sep 08 14:59:12 <mb__> nbd: ignore diablo. He's a troll
Sep 08 14:59:28 <nbd> ok ;)
Sep 08 15:01:22 <mb__> Hm, what package is "vmstat" in?
Sep 08 15:01:44 <nbd> procps
Sep 08 15:01:54 <nbd> according to apt-cache search&show
Sep 08 15:02:01 <mb__> cool, thanks
Sep 08 15:02:18 <mb__> Do we have that in openwrt?
Sep 08 15:02:25 <nbd> not yet
Sep 08 15:02:39 <mb__> hm..
Sep 08 15:02:58 <nbd> oh, wait
Sep 08 15:02:59 <nbd> we do have that
Sep 08 15:03:20 <nbd> and it includes vmstat
Sep 08 15:03:24 <nbd> it's in the packages feed
Sep 08 15:03:58 <mb__> cool
Sep 08 15:07:39 <mb__> Very hard to find out what's going on... Especially when ingo keeps posting unrelated bullshit of how he fails to reproduce it.
Sep 08 15:07:50 <nbd> yeah
Sep 08 15:08:15 <nbd> i think pluggable schedulers aren't such a bad idea at all
Sep 08 15:08:16 <mb__> I think benh might have a point there
Sep 08 15:08:46 <mb__> Yeah, I don't understand why mainline doesn't want it. The one-fits-all simply does not scale
Sep 08 15:09:40 <nbd> i don't think tlb misses are the cause for this. accesses to KSEG0 should never cause a tlb miss
Sep 08 15:09:54 <nbd> since the first wired tlb entry is the one for KSEG0, iirc
Sep 08 15:13:37 <mb__> hm
Sep 08 15:16:30 <mb__> whoopsy, this oopses the box
Sep 08 15:16:58 <nbd> ?
Sep 08 15:17:50 <mb__> running vmstat and iperf threw an oops
Sep 08 15:17:57 <mb__> Doesn't do it again...
Sep 08 15:20:32 <nbd> the routing throughput has less fluctuation with BFS than with CFS
Sep 08 15:20:47 <mb__> cool
Sep 08 15:20:49 <nbd> and it seems at least a little bit faster
Sep 08 15:21:15 <nbd> around 1Mb/s difference
Sep 08 15:21:15 <mb__> yeah I would not expect much difference for routing, as it's mostly tasklet based
Sep 08 15:21:25 <nbd> (BFS is around 37Mb/s
Sep 08 15:24:08 <CIA-36> florian * r17542 /trunk/target/linux/brcm63xx/files/arch/mips/bcm63xx/boards/board_bcm963xx.c: [brcm63xx] prevent gpio-buttons registration failure
Sep 08 15:27:57 <nbd> hm, i need to run the test with CFS again, found a typo in my patch that might have altered its behavior ;)
Sep 08 15:31:33 <mb__> SQUASHFS error: Unable to read inode 0x1257
Sep 08 15:31:37 <mb__> lol? What's that?
Sep 08 15:32:48 <mb__> reflashing doesn't help
Sep 08 15:34:35 <nbd> that's odd
Sep 08 15:34:44 <nbd> does it happen on the same inode all the time?
Sep 08 15:35:29 <mb__> yeah
Sep 08 15:35:34 <mb__> I'm recompiling now
Sep 08 15:44:11 <mb__> still the same
Sep 08 15:45:00 <nbd> maybe a part of your flash chip has gone bad
Sep 08 15:45:31 <mb__> yeah
Sep 08 15:45:33 <mb__> :(
Sep 08 16:08:34 <mb__> hargh. And now the powerbook doesn't boot into X. I should stop all activity for today. THis is not my day
Sep 08 16:14:57 <mb__> PMU battery failed, as it seems. That causes the clock to reset and X to not start. Very weird.
Sep 08 16:42:14 <acoul> http://thread.gmane.org/gmane.linux.kernel/886319
Sep 08 16:43:39 <nbd> mb__: btw. on atheros wisoc bfs only gives 3 mbit more than cfs in an iperf test
Sep 08 16:43:52 <nbd> still an improvement
Sep 08 16:44:42 <mb__> nice
Sep 08 17:10:56 <acoul> nbd - mb__ what do you think of this: http://ck.kolivas.org/patches/bfs/bfs210-test2.patch
Sep 08 17:19:43 <maligor> I wouldn't call 3mbit "only"
Sep 08 17:21:26 <mb__> acoul: I'd say it's a no-op for us
Sep 08 17:22:40 <netprince> :q
Sep 08 17:22:46 <netprince> crap
Sep 08 17:24:20 <acoul> I will test BFS on an old wrap, rb532 and rb133
Sep 08 17:45:01 <acoul> mb__ are you talking about this: http://en.wikipedia.org/wiki/Noop_scheduler :-)
Sep 08 18:43:04 <fofware> hello every body
Sep 08 18:44:40 <fofware> how I can build an image whit specific network and firewall setting?
Sep 08 18:48:06 <fofware> I did modiffy network file in ...basefiles/etc/config/network and target.../firewall-1/ipkg/firewall/etc/config/firewall to do that
Sep 08 18:49:00 <fofware> but changes are not in new images, how I can do that?
Sep 08 18:49:12 <fofware> but changes are not in new images, how can i do that?
Sep 08 19:02:24 <loswillios> fofware: files/etc/config/firewall
Sep 08 19:15:38 <fofware> thanks loswillios, yeah was more easy, only change package in trunk, thanks
Sep 08 19:16:17 <CIA-36> juhosg * r17543 /trunk/target/linux/ar71xx/files/arch/mips/ (3 files in 2 dirs): ar71xx: add preliminary support for the RB-411U (thanks to Feher Zoltan for testing)
Sep 08 20:50:37 * fofware is away: Estoy ocupado
Sep 08 22:07:15 <cirdan> just fyi luci deps are all messed up right now in trunk
Sep 08 22:07:32 <cirdan> luci-admin-full is missing a ton of things needed to get it running
Sep 08 22:07:42 <cirdan> tons
Sep 09 00:04:28 <frogonwheels> cirdan: you sure you added them all from the feeds?
Sep 09 00:05:05 <frogonwheels> cirdan: scripts/feeds -p luci -a
Sep 09 00:11:17 <cirdan> yeah
Sep 09 00:11:24 <cirdan> it didnt complain about missing deps
Sep 09 00:11:37 <cirdan> i installed them manually until it worked
Sep 09 00:16:57 <CIA-36> nbd * r17544 /trunk/package/base-files/files/ (3 files in 3 dirs): Revert "fix nonhotpluggable networks not being brought up. Closes #2781" (r17527) until we've solved the race conditions it creates
Sep 09 00:17:15 <CIA-36> nbd * r17545 /trunk/package/madwifi/patches/ (21 files): madwifi: fix several noderef and ibss merge issues that led to memory leaks and high system load
Sep 09 00:30:38 <CIA-36> nbd * r17546 /trunk/package/madwifi/patches/441-fix_ibss_node_handling.patch: madwifi: replace an implicit memcpy with an explicit one (cosmetic fix)
Sep 09 00:32:52 <CIA-36> nbd * r17547 /branches/8.09/package/madwifi/patches/ (20 files): madwifi: merge r17545, r17546 from trunk
Sep 09 00:40:18 <swalker> cirdan: file a ticket in their trac?
Sep 09 00:41:25 <cirdan> not yet
Sep 09 00:41:46 <cirdan> wasnt sure if they hung out here, or anything just trying to document my progress
Sep 09 00:41:59 <cirdan> right now the firewall is frustrating me
**** ENDING LOGGING AT Wed Sep 09 02:59:56 2009