**** BEGIN LOGGING AT Tue Jul 15 02:59:56 2008 Jul 15 05:04:35 thepeople * r11835 /trunk/package/wpa_supplicant/Makefile: fix md5sum Jul 15 05:43:44 thepeople: r11834 should fix https://dev.openwrt.org/cgi-bin/trac.fcgi/ticket/2356, no? Jul 15 05:45:22 swalker: yep, I will close it Jul 15 09:07:15 blogic * r11836 /trunk/target/linux/ifxmips/files/include/asm-mips/ifxmips/ (ifxmips.h ifxmips_irq.h): adds more irq defines to ifxmips Jul 15 10:22:30 florian * r11837 /trunk/target/linux/rb532/files/drivers/block/rb500/ata.c: Fix the CF serial number byteswapping (#3133) Jul 15 11:48:36 moin moin Jul 15 11:48:56 oh wrong channel - ups Jul 15 11:48:57 hello Jul 15 11:49:04 hello Jul 15 13:50:09 hi guys Jul 15 13:50:16 hello Jul 15 14:00:03 hi Nix Jul 15 14:00:09 hello Jul 15 17:03:27 florian * r11838 /trunk/package/broadcom-wl/files/lib/wifi/broadcom.sh: Support for setting monitor mode (#1306) Jul 15 17:11:23 florian * r11839 /trunk/target/linux/generic-2.6/patches-2.6.25/ (2 files): Add ocf to 2.6.25 for ubsec ssb integration Jul 15 17:11:43 florian * r11840 /trunk/target/linux/generic-2.6/config-2.6.25: Do not build ocf modules by default Jul 15 17:12:51 florian * r11841 /trunk/package/ubsec_ssb/ (20 files in 3 dirs): Add ubsec ssb from Daniel Mueller (#2417) Jul 15 17:16:03 florian * r11842 /trunk/target/linux/brcm47xx/profiles/WGT634U.mk: Include ubsec ssb in the WGT634U profile Jul 15 17:17:06 florian * r11843 /trunk/target/linux/rb532/ (6 files in 4 dirs): Move files and patches to files-2.6.23 and patches-2.6.23 Jul 15 17:18:38 florian * r11844 /trunk/target/linux/rb532/ (59 files in 18 dirs): Add basic 2.6.24 support for rb532, korina napi code has to be adapted to work Jul 15 17:56:49 yay! Jul 15 17:57:35 [florian]: does ubsec work with openvpn yet? Jul 15 18:10:44 ah, seems openssl patches for that are missing Jul 15 18:12:23 if tha tubsec driver uses cryptoapi it should wrok Jul 15 18:12:38 still no idea why he adds it to .25 instead of bringing b0rkedcom to .26 Jul 15 18:19:33 looking at the README, it doesn't seem to support cryptoapi Jul 15 18:21:46 it's seems some hacked-up-something from the ocf driver Jul 15 18:22:18 it also contains lots of BSDisms that should be removed Jul 15 18:23:28 [florian]: where does this come from? Jul 15 18:23:46 http://www.danm.de/files/src/bcm5365p I guess Jul 15 18:24:56 [florian]: is this tested, working? why isn't it a patch against ocf? Jul 15 18:29:40 [florian]: http://www.danm.de/files/src/bcm5365p/REPORTED_DEVICES Jul 15 18:29:41 .... Jul 15 18:30:02 heh ;-) Jul 15 18:30:43 * NAiL has a WL500gP Jul 15 18:31:18 does somebody here have a WGT634u? Jul 15 18:31:23 yes Jul 15 18:31:48 loswillios: the readme says it's an ocf driver, so.. Jul 15 18:32:39 hmm Jul 15 18:32:53 can ocf accelerate openssh? :D Jul 15 18:32:58 sure Jul 15 18:33:20 ocf "looks promising" from what google presents Jul 15 18:33:21 loswillios: i think i know how to fix the irq routing issue Jul 15 18:33:22 ocf can accelerate openssl through cryptodev -> anything using openssl can be accelerated Jul 15 18:33:34 openvpn uses openssl, right? Jul 15 18:33:40 NAiL: yup Jul 15 18:33:43 great :D Jul 15 18:34:02 NAiL: I have no problems with new support, simple the way it was added Jul 15 18:35:06 nbd: I wasn't aware of any routing issues, didn't find a ticket either Jul 15 18:35:19 Kaloz: grabbed from a random source and thrown into the kernel? Jul 15 18:36:01 NAiL: grabben froma source without testing, added a different patch to an older kernel, added to a board profile it doesn't work on... Jul 15 18:36:11 loswillios: http://www.danm.de/files/src/bcm5365p/REPORTED_DEVICES Jul 15 18:36:31 ah, cool! Jul 15 18:38:26 Kaloz: ah. Well, I happen to be setting up openvpn on a wl500gP and definitely wouldn't mind testing it Jul 15 18:38:58 (which ofcourse should've been done before commit, but it's there now) Jul 15 18:38:59 NAiL: wanna clean it up, too? ;) Jul 15 18:39:10 Kaloz: are you insane? :-P Jul 15 18:39:53 well, actually it can't be that hard.. question is ig b0rkedcom boots .26 or not Jul 15 18:41:30 nbd: hey Jul 15 18:41:35 luke-jr: hey Jul 15 18:41:40 nbd: speaking of irq… I am having a lot of trouble with it XD Jul 15 18:41:49 stuck in loop calibration Jul 15 18:42:05 what kind of device? Jul 15 18:42:23 Scientific-Atlanta WebSTAR DPC2100, bcm3349 Jul 15 18:42:33 cable modem Jul 15 18:42:38 being stuck in calibration loop is usually a sign of a) misdetected timer irq line or b) bug in irq routing Jul 15 18:42:52 no irqs are firing at all atm Jul 15 18:43:19 then review the irq setup/dispatch code Jul 15 18:43:31 can't calibrate the delay loop without irq Jul 15 18:43:54 I don't understand it ☹ Jul 15 18:43:55 are you trying this with the bcm63xx port in openwrt? Jul 15 18:43:55 hm Jul 15 18:43:58 yes Jul 15 18:44:10 I just set bit 7 of the int mask to 1 and it seems to be working Jul 15 18:44:39 so the question becomes, why the usual code isn't enabling it Jul 15 18:45:18 hi Jul 15 18:45:24 hi Jul 15 18:46:09 whoa, some of that code is fugly Jul 15 18:46:30 nbd: I saw your name on it; you didn't write it? :o Jul 15 18:46:40 'sleep 10' finishes in 3 seconds :o Jul 15 18:46:41 my name? on 63xx? Jul 15 18:46:45 just a quick question regarding commit 11842 (enabling of ubsec sbb for wgt634u): shouldn't it be *disabled* for the wgt and enabled for the asus wl500gp? because according to http://www.danm.de/files/src/bcm5365p/REPORTED_DEVICES it doesn't work on the wgt634u ... Jul 15 18:46:52 yeah, 63xx irq.c Jul 15 18:47:13 NAiL: wanna test .26 on that crap? :) Jul 15 18:47:15 are we looking at the same file? Jul 15 18:47:37 oh, i overlooked the patch Jul 15 18:47:42 * Copyright © 2006,2007 Felix Fietkau Jul 15 18:47:43 KanjiMonster: florian fscked it up Jul 15 18:47:57 Kaloz: Not quite sure how to exactly Jul 15 18:47:59 interesting Jul 15 18:48:04 i wonder how my name got in there Jul 15 18:48:08 lol Jul 15 18:48:20 NAiL: maybe I rest a bit and leave the stuff i was doing and give you a binary? :) Jul 15 18:48:21 maybe c&p from other files? Jul 15 18:48:24 nbd: florian Jul 15 18:48:25 :P Jul 15 18:48:55 Kaloz: uhm, well, sure :) Jul 15 18:49:13 nbd: the new irq.c look a bit better? Jul 15 18:49:13 luke-jr: ok, one issue that i see is that this code overrides the behavior of int 0-7 Jul 15 18:49:22 luke-jr: 0-7 should be handled by generic irq code Jul 15 18:49:32 and all the external interrupts should be >=8 Jul 15 18:49:33 hm Jul 15 18:49:42 so you should check on some other mips targets Jul 15 18:49:53 atheros, routerboard, whatever Jul 15 18:50:16 the behavior of the mips cpu irq lines is basically the same on every piece of mips hw that i've played with Jul 15 18:50:28 all the board specific stuff cascades from the mips cpu irq lines Jul 15 18:50:44 Kaloz: yeah, the commit log never forgets ;) just thought I mention it in case nobody notices it and you get tickets about it not working or so ... Jul 15 18:51:00 KanjiMonster: the whole commit is b0rked :) Jul 15 18:51:01 gotta go now Jul 15 18:51:02 cya later Jul 15 18:51:05 ttyl Jul 15 18:51:08 thanks for the tips Jul 15 18:51:12 you're welcome Jul 15 19:05:53 nbd: just use svn blame to know who's responsible for that line ;) Jul 15 20:31:18 KanjiMonster: svn blame only works for stuff actually in svn Jul 15 20:37:00 this is in svn. r11178 is the commit in question. Jul 15 20:46:18 hi Jul 15 20:46:24 hey Jul 15 20:46:57 hi Jul 15 20:47:04 hello Jul 15 20:58:31 * NAiL pokes Kaloz Jul 15 21:51:58 Bartman007: the current irq.c is not in svn Jul 15 21:52:04 except as a patch file Jul 15 22:32:56 luke-jr: but using svn blame on the patch gives you the commit id where the lines in question were changed. and the change log on that commit explains that the interrupt handing was pulled from AR7, which is where the copyright comes from :-) Jul 15 22:33:30 i c **** ENDING LOGGING AT Wed Jul 16 02:59:56 2008