**** BEGIN LOGGING AT Tue Jul 20 02:59:56 2010 Jul 20 03:09:49 ty Jul 20 03:11:33 nbd, does hauke's update here https://dev.openwrt.org/changeset/22298/trunk Jul 20 03:11:37 apply to this ticket Jul 20 03:11:46 https://dev.openwrt.org/ticket/7630 Jul 20 03:12:31 it talks about board stuff but i dont know ;p lol Jul 20 05:38:43 anybody here live in munich? /me visiting briefly next month Jul 20 06:35:11 hello Jul 20 06:41:14 hi Jul 20 06:57:45 cshore * r22306 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: Disabled mod-flite for avr32 and ixp4xx as those target are not supported by flite at the present time. Jul 20 08:14:05 cshore * r22307 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: Fixed compilation error on gcc4.2.x (avr32) Jul 20 08:14:08 cshore * r22308 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: Removed accidentally committed warning suppresion Jul 20 09:15:24 who maintains the website? Jul 20 09:15:45 id like to use http://www.mediawiki.org/wiki/MediaWiki for wiki. Jul 20 09:15:57 its more user friendly and editor friendly Jul 20 09:16:39 * realopty will do the migration of the wiki mediaWiKi Jul 20 09:16:49 to* Jul 20 09:17:14 mediawiki is definitly better yeh.. Jul 20 09:17:58 indeed. Jul 20 09:19:59 even has Third-party extensions Jul 20 09:20:36 like this one Jul 20 09:20:36 http://www.mediawiki.org/wiki/Extension:DocumentApproval Jul 20 09:20:43 thats a good one :) Jul 20 09:20:52 openwrt needs an offical documentation Jul 20 09:21:00 kerneis * r22309 /packages/ipv6/ahcpd/ (4 files in 3 dirs): Jul 20 09:21:00 [packages] ahcpd: apply unreleased upstream patch Jul 20 09:21:01 - Tue Jul 20 00:25:27 CEST 2010 Juliusz Chroboczek Jul 20 09:21:01 * Fix typo -- ignore autoconf prefixes for plen > 64. Jul 20 09:21:01 - Use uci_get_state instead of invoking uci directly. Jul 20 09:21:01 Signed-off-by: Gabriel Kerneis Jul 20 09:21:02 kerneis * r22310 /packages/net/babel/ (Makefile files/babeld.init): Jul 20 09:21:02 [packages] babel: use uci_get_state Jul 20 09:21:03 Signed-off-by: Gabriel Kerneis Jul 20 09:23:00 xMff, do you maintain the website? Jul 20 09:26:33 realopty: Bartman007 and Kaloz do Jul 20 09:27:57 Bartman007, Kaloz, what do you think about using mediawiki as the wiki for openwrt? Jul 20 09:30:20 mediawiki is pain. Jul 20 09:31:23 nah shouldnt be too bad Jul 20 09:32:50 been there. done that. not maintainable if not paid work. Jul 20 09:33:37 bad to intregrate Jul 20 09:33:47 horrifying syntax Jul 20 09:34:02 i could go on.. but i'd rather like to stop ;) Jul 20 09:34:36 means.. i can understand why it is not a mediawiki. its dokuwiki now. dev.openwrt.org is trac which also has a wiki. Jul 20 09:34:58 yes, the decision has been made. you are a little late realopty ;-) Jul 20 10:23:36 lol ok Jul 20 10:23:43 * realopty keeps updating wiki Jul 20 10:37:32 mediawiki, no way Jul 20 11:11:29 kaloz * r22311 /trunk/toolchain/uClibc/ (5 files in 5 dirs): the storm target is gone for a long time now Jul 20 11:24:59 realopty: can you mark the one-sentence pages as stub? Jul 20 11:25:04 in the wiki Jul 20 11:27:05 xMff, how so? Jul 20 11:27:11 {stub} ? Jul 20 11:27:12 :\ Jul 20 11:27:20 :!: This page is a stub! ~~~-- Jul 20 11:27:27 yeah ok Jul 20 11:27:36 just getting a start on somethings :/ Jul 20 11:27:47 like a FAQ heheh Jul 20 11:27:48 and I just give tips Jul 20 11:27:58 yes thank you Jul 20 11:29:02 with old pages you have to be careful, for example doc/howto/uninstall refers solely to bcm hardware Jul 20 11:29:42 you should add least add that users are adviced to check the device specific wiki page for uinstall instructions Jul 20 11:29:51 most contain the proper way to go back to factory Jul 20 11:31:01 maybe we should redo doc/start in tables instead of nested lists Jul 20 11:32:00 yeah that would look better Jul 20 11:32:54 xMff, where can i find out the ;syntax; of docuwiki? Jul 20 11:33:06 http://wiki.openwrt.org/meta/syntax Jul 20 11:33:27 oh nice Jul 20 11:37:42 xMff, you give me a base to work off with the doc/start page and ill go to work Jul 20 11:49:17 xMff, your working on doc/start yes? Jul 20 11:49:24 yes Jul 20 11:49:29 kk Jul 20 11:50:40 xMff, can we use use php in the wiki :O ? Jul 20 11:50:46 like is it enabled? Jul 20 11:50:48 hopefully not Jul 20 11:51:08 too bad Jul 20 11:51:15 that would be the most gaping security hole I've ever seen Jul 20 11:51:19 * realopty loves php Jul 20 11:52:31 its not enabled Jul 20 11:53:44 but html is Jul 20 11:53:49 yay Jul 20 12:08:57 xMff, know what would make the wiki look more sexy Jul 20 12:09:13 if the whole website had a theme that wasnt depressing Jul 20 12:09:30 id say a them based off 'backfire' Jul 20 12:09:38 its a technical reference, no yellow press tabloid Jul 20 12:09:40 some cool flames or something idk Jul 20 12:10:25 its still dark and disgusting Jul 20 12:11:45 well in the first place its CI Jul 20 12:11:56 if the mainpage changes, we can change the wiki too Jul 20 12:12:13 the mainpage looks dirrent from the wiki Jul 20 12:12:24 the mainpage is blue and the wiki is like teal Jul 20 12:12:57 and the mainpage still points alot to oldwiki :( Jul 20 12:13:05 the main page css changes with daytime Jul 20 12:13:13 oh Jul 20 12:13:15 from "teal" to blue and back Jul 20 12:13:19 never noticed that Jul 20 12:13:39 * realopty never looks at wiki during day Jul 20 12:13:50 black sounds sexy Jul 20 12:13:50 for me both page and wiki have the exact same color scheme Jul 20 12:13:53 right now Jul 20 12:14:01 hmmm odd Jul 20 12:14:06 only font sizes differ slightly Jul 20 12:14:31 try changing your system time to 1pm or so Jul 20 12:14:39 kk Jul 20 12:15:17 wiki didnt change Jul 20 12:15:27 but the homepage changed to teal and matches now :/ Jul 20 12:15:50 thats what I mean Jul 20 12:16:16 i understud that it changed colors with time. but the wiki dosnt change for me Jul 20 12:16:26 yes, thats correct as well Jul 20 12:16:33 I need to hack that in at some point Jul 20 12:16:33 oh ok then lol Jul 20 12:16:51 or hack in user themes ;p Jul 20 12:20:36 * realopty wants to see your changes to doc/start Jul 20 12:20:58 I'm halfway through Jul 20 12:41:05 xMff, u still editing? Jul 20 12:41:12 yes Jul 20 12:41:23 oh ok. u lost your lock though Jul 20 12:51:40 xMff, im adding some icons to the wiki. how and where should i put a link to the Creative Commons Attribution/Share-Alike License? Jul 20 12:51:57 uhm Jul 20 12:52:43 maybe create a page meta/icons, reference each icon there (in a list or table) and put the license stuff below Jul 20 12:53:00 this way editors can also easily copy the icon code off that page Jul 20 12:53:02 kk good idea Jul 20 12:53:18 well its kinda already easy to insert an icon. Jul 20 12:53:26 but having a page to see it will be nice too Jul 20 12:53:28 then you can reference the icons page in meta/syntax Jul 20 12:53:30 all of them * Jul 20 12:53:41 or even include it: http://www.dokuwiki.org/plugin:include Jul 20 12:53:52 examples for includes are on the start page Jul 20 12:54:11 indeed Jul 20 12:57:16 doc/start is done# Jul 20 13:02:15 realopty: you going to be working on docs? Jul 20 13:02:32 * cshore thinks have a doc person would be good Jul 20 13:02:33 gonna try to do as much as i can Jul 20 13:02:51 realopty: awesome Jul 20 13:03:49 can any one tell me if these icons are GNU or CC -sharealike 2.5 ? Jul 20 13:03:51 http://commons.wikimedia.org/wiki/Tango_icons Jul 20 13:04:19 cshore, looks like xMff has been doing the documents for a while now Jul 20 13:04:48 * realopty thinks documents will take more then two ppl Jul 20 13:04:54 yeah, but he also does lots of other stuff....we need more ppl Jul 20 13:05:12 I sometime do stuff, usually for things I've written Jul 20 13:05:50 for sure Jul 20 13:05:58 xMff, doc/start looks whole lot better Jul 20 13:07:22 like 100X better Jul 20 13:08:22 lol i just found the nfs server setup page cause its a better format now Jul 20 13:08:26 didnt notice it before Jul 20 13:21:03 kaloz * r22312 /trunk/scripts/ (config.guess config.sub): upgrade the config.* scripts Jul 20 13:22:21 xMff, on doc/start Jul 20 13:22:32 shouldnt the UCI reference Jul 20 13:22:38 point to doc/uci ? Jul 20 13:22:49 insted of techref/uci Jul 20 13:23:02 no. it will contain api descriptions and tuff at some point Jul 20 13:23:06 stuff Jul 20 13:23:11 ok making sure Jul 20 13:42:14 realopty: doc/howto/clientmode is a duplicate Jul 20 13:42:53 of? Jul 20 13:43:25 the same entry in recipes Jul 20 13:43:48 same for creatingpackages, its part of devel Jul 20 13:43:53 hmm doot point Jul 20 13:43:56 my bad lol Jul 20 13:43:58 good Jul 20 13:45:01 The "HOWTOs" section should cover ordinary user stuff, "Devel" should contain stuff that involves compiling, makefiles etc. and "Recipes" is for global setups that involve more than just installing a package Jul 20 13:45:09 i was looking at http://wiki.openwrt.org/doc/start?do=index then clicked on howto lol Jul 20 13:45:20 * realopty understands Jul 20 13:46:27 how do u revert changes ? Jul 20 13:46:52 just edit again Jul 20 13:46:57 yeah lol Jul 20 13:48:35 better Jul 20 13:48:41 indeed Jul 20 13:49:14 At some point we should splitt installation off from howtos and make it a new level 2 heading Jul 20 13:49:25 before config Jul 20 13:50:28 and we need to create some common text blocks which can be included in other pages Jul 20 13:50:52 how we do that Jul 20 13:51:43 sth. like meta/include/warning meta/include/needsrepair etc. with some nicely formatted html or wiki markup Jul 20 13:51:54 like the orange box here: http://en.wikipedia.org/wiki/Spencer_Rochfort Jul 20 13:52:54 yes very nice Jul 20 13:54:43 another example; lots of broadcom device articles share the same tftp instructions Jul 20 13:55:04 you can put those instruction on a separate page somewhere and just include it in each article Jul 20 13:55:14 this way, text fixes are applied to all pages immediately too Jul 20 13:55:24 give me an example Jul 20 13:55:39 of how to include i mean Jul 20 13:55:40 loll Jul 20 13:55:47 * realopty is tired Jul 20 13:55:55 http://wiki.openwrt.org/start?do=edit Jul 20 13:56:02 the {{section>...}} tags Jul 20 13:56:25 oh nice nice Jul 20 13:58:51 its better to spent time on writing conclusive infos once than having it only halfway in multiple places :) Jul 20 13:59:55 don't forget that brcm63xx (broadcom) is different than brcm47xx (broadcom) Jul 20 14:00:16 brcm47xx is usually what's just called broadcom in OpenWRT Jul 20 14:00:18 give me an example of a page that could use the doc:howto:tftp included plz Jul 20 14:00:25 brcm63xx is new Jul 20 14:00:49 realopy: should be in the table of hardware...look for anything that is brcm47xx Jul 20 14:00:50 I'd just do tftp_upload (bcm47xx) and tftp_download (bcm63xx, tftp with serial initiated by cfe) Jul 20 14:02:06 xMff: are there ports tags on the new wiki? Jul 20 14:02:11 like brcm4xx? Jul 20 14:02:18 port tags? Jul 20 14:02:33 CateoryBrcm63xx for instance Jul 20 14:02:39 for searching Jul 20 14:02:41 not yet Jul 20 14:03:05 I did not understand the dokuwiki tag system yet Jul 20 14:03:19 and somebody needs to maintain the tags Jul 20 14:03:22 right Jul 20 14:03:38 too much work for me atm, I rather fous on important docs Jul 20 14:04:08 yes, first things first Jul 20 14:04:11 indeed Jul 20 14:04:15 but um Jul 20 14:04:17 http://wiki.openwrt.org/doc/configuring Jul 20 14:04:26 does openwrt come with luci by default? Jul 20 14:04:31 yes Jul 20 14:04:43 i dont think all devices have a DMZ light though Jul 20 14:04:50 correct Jul 20 14:05:10 the mindset for a lot of older docs is that there is only bcm devices Jul 20 14:05:24 thats because whiterussian was bcm only years ago Jul 20 14:05:31 nvm that whole first sentence confuses me Jul 20 14:05:34 and they where all similar Jul 20 14:05:37 of the page http://wiki.openwrt.org/doc/configuring Jul 20 14:06:08 it means; if there is a DMZ LED, OpenWrt will light it up during boot and turn it off if it finished booting Jul 20 14:06:36 its an indicator to see whether something boots at all and when its finished Jul 20 14:09:38 for brcm47xx devices with a controllable power led, the device turns the led on, openwrt turns the led off while booting, and on once booted the led is on again. For failsafe the led flashes Jul 20 14:13:14 cshore * r22313 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: Add back mistakenly removed warning suppression for print/scanf - needed for avr32 Jul 20 14:30:46 LOL Jul 20 14:30:47 http://wiki.openwrt.org/oldwiki/jonathankollasch/waystobrickyourbox Jul 20 14:33:38 xMff, http://wiki.openwrt.org/toh/start looks broke in my browser Jul 20 14:33:43 firefox :/ Jul 20 14:34:00 the ToC on the right is just huge Jul 20 14:34:04 get a bigger screen or use ctrl- Jul 20 14:34:15 got a 32" screen lol Jul 20 14:35:05 ah, he means the tables run throught he ToC Jul 20 14:35:17 yes Jul 20 14:35:25 * realopty needs lots of sleep Jul 20 14:36:02 could the hardware be made to start below the ToC Jul 20 14:36:04 ? Jul 20 14:36:22 I've noticed other pages have this problem too Jul 20 14:36:24 if there was more content Jul 20 14:36:25 yes Jul 20 14:36:28 24" here and it fits. at 1920x1200 Jul 20 14:36:33 hmm Jul 20 14:36:42 xMff: vertically, not horizontally Jul 20 14:37:04 oh Jul 20 14:37:08 hmm Jul 20 14:37:16 I have a 24" firefox and the tables (e..g Abicom Internatiional run through the ToC) Jul 20 14:37:16 my res is 1366*768 Jul 20 14:37:21 cshore: the tables are just too wide Jul 20 14:37:42 xMff: could the tables be made to start below the ToC? Jul 20 14:37:56 cshore: even if you take out the toc, they run out horizontally if the screenspace is too narrow Jul 20 14:37:58 xMff: there would be some white space Jul 20 14:38:10 xMff: but it's an improvement Jul 20 14:39:10 xMff: maybe the tables should be done on two lines each entry Jul 20 14:40:54 in most cases it'd help to just make the table headings shorter Jul 20 14:41:02 ok, when I use fullscreen I have enough room, but I think the ToC should push the hardware tables below it, so that the overlap doesn't happen Jul 20 14:41:32 at least the screen scrolls if they're too wide, you can't correct for the overlap, because it's in the CSS Jul 20 14:41:41 *but Jul 20 14:42:08 the right side is just bulky Jul 20 14:42:21 would be nice if the search was in the top of the banner on the right Jul 20 14:42:42 and 'actions' is at the bottom Jul 20 14:42:44 or something Jul 20 14:42:45 idk : Jul 20 14:42:45 \ Jul 20 14:43:09 really the problem is tables don't scroll if they hit the edge of a another object Jul 20 14:43:20 but that's an HTML problem Jul 20 14:43:39 or CSS/HTML/UA Jul 20 14:43:47 * realopty needs caffeine Jul 20 14:43:59 cant go to sleep till my son wakes up :/ Jul 20 14:44:17 hold on Jul 20 14:50:48 reload Jul 20 14:50:52 should be better now Jul 20 14:51:24 you just have to put
before the first table to terminate the ToCs right float Jul 20 14:51:32 fixes the problem :) Jul 20 14:51:45 it looks awful here now Jul 20 14:51:53 lol Jul 20 14:51:56 pwnt Jul 20 14:52:14 xMff, what looks awful? Jul 20 14:52:22 the gigantic gap Jul 20 14:52:31 LOL Jul 20 14:52:36 <_iTroll> hello guys, im building a fresh kamikaze. Just checked out kamikaze r22313 and did a "$ ./scripts/feeds update; $ ./scripts/feeds install -a; $ make defconfig; $ make menuconfig " and getting a build error on the first package, sed. Am I missing something stupid here? Jul 20 14:52:54 xMff: I think it's better than the run-through on the more usual screen size Jul 20 14:52:57 <_iTroll> error is: make[4]: Entering directory `/home/itroll/packages/8.09/build_dir/host/sed-4.1.2' Jul 20 14:53:03 yeah thats why i said it should be filled with content Jul 20 14:53:04 <_iTroll> make[4]: *** No targets specified and no makefile found. Stop. Jul 20 14:53:37 <_iTroll> as if ./configure wasn't run on sed Jul 20 14:54:13 or put the toc on the right where search is, and then put search above it, and then sink actions down to the very bottom Jul 20 14:54:50 i think that would look more clean :/ Jul 20 14:55:02 maybe even toc on the Left :) Jul 20 14:55:07 evil Jul 20 14:55:32 xMff: do the actions and such float, or are they fixed? Jul 20 14:55:33 would look more pleasing and open up some space Jul 20 14:55:53 realopty: unless I'm mistaken the problem is the variable size of the ToC Jul 20 14:55:58 horizontally Jul 20 14:56:24 its jsut the braindeadness of floating content in css :) Jul 20 14:56:40 hehe Jul 20 14:56:42 yeah, that's what I mean Jul 20 14:57:33 seriously though look at the TOH page as a reference Jul 20 14:57:38 I was wondering if it could be unfloated in the widget column Jul 20 14:57:40 all that wasted space on the right side :/ Jul 20 14:58:09 could you guys post a link to the page you are working on - so i can click on it before leaving from work :D Jul 20 14:58:22 http://wiki.openwrt.org/toh/start Jul 20 14:58:24 INSTRUCTIONS Jul 20 14:58:24 1. Flash the SimpleShare with v1.07 of the original firmware that has telnet enabled. Jul 20 14:58:24 2. Turn on boot_wait in the nvram Jul 20 14:58:24 nvram set boot_wait=on Jul 20 14:58:24 nvram set wait_time=5 Jul 20 14:58:25 nvram commit Jul 20 14:58:26 3. Unplug the Simpleshare Jul 20 14:58:31 OOPS Jul 20 14:58:33 :/ sorrt Jul 20 14:58:59 hrhr Jul 20 15:00:18 ok ... i just was curiuous aboout what you are doing ... i'll go home now ... yea 31 degrees outside ... Jul 20 15:00:24 ciao Jul 20 15:00:30 peace Jul 20 15:06:24 hmmm Jul 20 15:11:10 xMff, dont work? Jul 20 15:11:26 realopty: ? Jul 20 15:11:40 Syntax Highlighting Jul 20 15:12:04 it does Jul 20 15:12:22 hmm ok whats the statement to /end a code :\ Jul 20 15:12:25 http://wiki.openwrt.org/meta/syntax#syntax.highlighting Jul 20 15:12:32 lol im reading that Jul 20 15:12:42 ... Jul 20 15:12:43 just says how to start it Jul 20 15:12:45 tried that Jul 20 15:12:48 it dont work :/ Jul 20 15:12:59 what did you try? Jul 20 15:13:02 http://wiki.openwrt.org/toh/simpletech/simpleshare Jul 20 15:13:44 don't indent with two spaces, that will make subsequent text on the same line preformatted Jul 20 15:14:17 kk Jul 20 15:14:27 and for list points everything must be on the same line. You can use \\ to force a line break Jul 20 15:14:39 oooo ty ty Jul 20 15:14:59 * realopty didnt know how to overcome that Jul 20 15:15:11 but the did i do something wrong? Jul 20 15:15:25 looked okay Jul 20 15:15:49 http://wiki.openwrt.org/doc/techref/initscripts its used here too Jul 20 15:16:10 oh Jul 20 15:16:18 ^ insted of < Jul 20 15:16:25 no Jul 20 15:16:26 no Jul 20 15:16:30 hmmmmm Jul 20 15:16:35 has to have a | ? Jul 20 15:16:51 no | is for making tables Jul 20 15:17:00 they use it to draw a box around Jul 20 15:17:24 ah its cause the spaces Jul 20 15:17:37 build #69 of ifxmips is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ifxmips/builds/69 Jul 20 15:22:10 brb, need to get an appointment before its 6pm Jul 20 15:22:41 * realopty needs food Jul 20 16:20:19 obsy * r22314 /packages/net/transmission/Makefile: [packages] transmission: update to 2.02 Jul 20 16:23:50 cshore * r22315 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: Removed dependency on alsa-lib; FreeSWITCH links it's own copy, not the system library Jul 20 16:23:52 cshore * r22316 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: Fixed/enabled mod_fax compilation Jul 20 16:30:31 cshore, failsafe only works somethims Jul 20 16:30:33 es Jul 20 16:30:53 i think it has to do with hotplug.d not diag module :/ Jul 20 16:30:55 are you sure, or is it just tricky to time it right? Jul 20 16:31:14 worksforme Jul 20 16:31:19 perfectly Jul 20 16:31:30 maybe just tricky timing Jul 20 16:31:38 on both brcm63xx and brcm47xx Jul 20 16:31:41 for my device :/ Jul 20 16:31:53 i have to press and release a few times Jul 20 16:31:56 there's only two seconds to get the button Jul 20 16:32:03 if i hold it down it wont go into failsaife Jul 20 16:32:08 that's right Jul 20 16:32:23 you have to get the button just after preinit starts Jul 20 16:32:50 you should listen for the UDP broadcast message indicating to press reset for failsafe Jul 20 16:32:59 oh, the method is to press and release not hold down? Jul 20 16:33:04 yes Jul 20 16:33:12 oh okay then Jul 20 16:33:39 just keep pressing and release for a few seconds once the led indicates that preinit is starting Jul 20 16:33:46 *releasing Jul 20 16:33:56 okies Jul 20 16:34:06 * realopty is glad it does work Jul 20 16:34:27 i had to reflash the other day cause it didnt work and i messed up network config LOL Jul 20 16:59:13 Hello all Jul 20 16:59:42 I need some advice to submit a patch Jul 20 17:47:35 if the gpio for the reset button is high by default and low when button is pressed does this means it's polarity is reverse ? Jul 20 17:49:34 I think so Jul 20 17:49:51 ok Jul 20 17:59:28 cshore * r22317 /packages/net/freeswitch/Makefile: [packages] net/freeswitch: bump package version number from previos changes Jul 20 18:48:14 build #64 of orion is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/64 Jul 20 19:41:20 hauke * r22318 /trunk/target/linux/brcm47xx/ (12 files in 2 dirs): Jul 20 19:41:20 brcm47xx: fix for boardflags by xMff Jul 20 19:41:20 and checkpatch.pl fixes Jul 20 20:44:38 <[Fate]> nbd: ping Jul 20 20:44:41 pong Jul 20 20:45:05 <[Fate]> these defaults in changeset 22303, where did you take these from? Jul 20 20:45:14 atheros sources Jul 20 20:45:23 for the ar724x driver Jul 20 20:45:52 <[Fate]> the one in the dd-wrt one or did you get a different version from atheros directly? Jul 20 20:45:54 <[Fate]> *svn Jul 20 20:46:09 from atheros directly Jul 20 20:46:22 ubnt uses the same values Jul 20 20:46:35 <[Fate]> let me guess, under nda? =) Jul 20 20:46:55 <[Fate]> I tried porting the Tx lockup workaround from the version in the dd-wrt svn to openwrt, but to no avail Jul 20 20:46:56 well, i have an nda, but i'm allowed to use the code Jul 20 20:47:03 <[Fate]> maybe your init vals help Jul 20 20:47:09 on which hw? Jul 20 20:47:19 <[Fate]> ar7240 (tp-link wa-901nd) Jul 20 20:47:37 i think that one uses the ap91 ethernet registration code, which uses different values Jul 20 20:47:42 so my commit doesn't change anything for that Jul 20 20:47:47 but you could try putting in the values that i added Jul 20 20:47:55 or just leaving out the overrides in the ap91 eth init code Jul 20 20:48:07 <[Fate]> it isn't in trunk yet, my patch's still waiting approval ;) Jul 20 20:48:12 <[Fate]> and I don't use ap91 Jul 20 20:48:18 ah, ok Jul 20 20:48:34 did you use any other values or did you just leave them unset? Jul 20 20:48:46 <[Fate]> I used the vals from the dd-wrt svn's version Jul 20 20:49:27 i have some difficulty to modifiy nvram.c from arch/mips/nvram.c, where do i have to modify it ? Jul 20 20:49:30 cfg1 and cfg2 look seem to be the same though Jul 20 20:49:35 if directly it give me an error while patching Jul 20 20:49:38 did you look at the ag7100 or the ag7240 driver? Jul 20 20:50:07 <[Fate]> ag7240. I used 0x0fff0000, 0x00001fff and 0x008001ff for fifo1, fifo2 and fifo3 Jul 20 20:50:22 <[Fate]> I tried some others, but didn't seem to make any difference Jul 20 20:50:39 where did you find those? i just looked at the driver and i see this: Jul 20 20:50:39 ag7240_reg_wr(mac, AG7240_MAC_FIFO_CFG_1, 0x10ffff); Jul 20 20:50:39 ag7240_reg_wr(mac, AG7240_MAC_FIFO_CFG_2, 0x015500aa); Jul 20 20:52:08 <[Fate]> correction, you're right, had the wrong src version open Jul 20 20:52:16 <[Fate]> the ones I mentioned were the "others" Jul 20 20:52:22 ok Jul 20 20:52:38 yes, wrong fifo settings could be the cause for stability issues Jul 20 20:53:36 <[Fate]> I'm not sure. seeing that they have the workaround in in spite of the values you committed now Jul 20 20:53:47 the workaround is for a specific issue Jul 20 20:53:55 that happens mainly when the cable is unplugged and plugged backi n Jul 20 20:53:59 <[Fate]> nope Jul 20 20:54:05 doing that repeatedly can corrupt the internal fifo state Jul 20 20:54:10 <[Fate]> I can trigger it very easily by running iperf Jul 20 20:54:13 <[Fate]> #6754 Jul 20 20:54:21 with your wrong values, right? Jul 20 20:54:27 <[Fate]> nope Jul 20 20:54:31 <[Fate]> with any values Jul 20 20:54:37 even with the ubnt ones? Jul 20 20:54:53 <[Fate]> where can I view those? Jul 20 20:54:59 they're in mach-ubnt.c Jul 20 20:55:20 what i was telling you is the reason why the workarounds were added in the first place Jul 20 20:55:28 i know this, because i have access to the atheros bug tracker Jul 20 20:56:23 <[Fate]> to work around the state when the TX dma engine locks up, right? Jul 20 20:57:21 yes Jul 20 20:57:22 nbd * r22319 /trunk/toolchain/ (9 files in 5 dirs): gcc: split up the build process into three distinct stages (minimal, initial, final), to clean up the dependency handling nastiness and to improve support for rebuilding parts of the toolchain Jul 20 20:57:52 <[Fate]> I'm pretty sure I face that independently of the fifo init vals, because I tried the atheros ones, but let me revalidate Jul 20 20:58:33 can you get me a binary of the original driver from tp-link? Jul 20 20:58:41 for this particular device Jul 20 20:58:48 <[Fate]> yes Jul 20 20:58:54 i could check the disassembly and see what fifo values they use Jul 20 20:59:23 <[Fate]> email? upload somewhere or..? Jul 20 20:59:36 upload Jul 20 21:00:26 <[Fate]> sec Jul 20 21:01:33 <[Fate]> http://www.hollants.com/ag7240_mod.ko.gz Jul 20 21:05:11 that one uses the same values that i pasted above Jul 20 21:05:43 at least for cfg1, cfg2 Jul 20 21:05:54 <[Fate]> yes, but their firmware survives my iperf torture test. but let me just recheck once again with trunk. Jul 20 21:43:43 <[Fate]> nbd: guess what =) Jul 20 21:44:13 <[Fate]> I don't know why, but the vals do seem to make a serious difference Jul 20 21:53:31 build #58 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/58 Jul 20 21:54:24 realopty: I made some includes you can stick to the top of pages... http://wiki.openwrt.org/meta/infoboxes Use that instead of the "this is outdated" etc. texts Jul 20 22:01:44 [Fate]: :) Jul 20 22:02:04 <[Fate]> peak iperf performance of 101mbit/s Jul 20 22:02:05 <[Fate]> stable Jul 20 22:02:09 <[Fate]> will update #6754 Jul 20 22:13:49 <[Fate]> nbd: do you have any idea what those vals actually do? Jul 20 22:13:56 <[Fate]> they're only partially documented in the srcs Jul 20 22:17:19 [Fate]: do you by any chance have any other ar724x based devices? Jul 20 22:17:54 <[Fate]> nbd: unfortunately no. just another one of the same type. Jul 20 22:18:26 ok Jul 20 22:18:59 i'd like to know if the tl-wr741nd will work with the same values or if it needs the ones in the ap91 eth code Jul 20 22:19:20 <[Fate]> could you have a look at #7528, safe for my init values there which by now is obsolete? Jul 20 22:23:38 looks good, but is the dsa stuff really necessary? did you try without it? Jul 20 22:24:40 <[Fate]> afair: yes. also, i couldn't explain to myself how it would work without it. Jul 20 22:24:55 <[Fate]> they seem to use the switch chip for vlan stuff in the original firmware Jul 20 22:25:05 <[Fate]> afaict we can't do that yet Jul 20 22:25:53 florian * r22320 /trunk/package/switch/src/ (switch-adm.c switch-core.c switch-robo.c): [package] remove whitespaces in kmod-switch sources Jul 20 22:25:53 florian * r22321 /trunk/package/kernel/modules/crypto.mk: [package] fix kmod-crypto-hw-padlock dependency against kmod-crypto-aes (#7643) Jul 20 22:26:37 florian * r22322 /packages/libs/curl/Makefile: [package] fix inverted libcurl/curl descriptions (#7648) Jul 20 22:28:15 so did it work without or not? Jul 20 22:28:26 the switch is built into the cpu Jul 20 22:28:42 if it's initialized in a way to not add any headers, then it should work without the dsa driver Jul 20 22:29:00 and that's definitely preferable, since dsa stuff is bad for performance Jul 20 22:29:33 <[Fate]> testing was some weeks ago, that's why. let me check how easily I can disable the dsa stuff for a test Jul 20 22:30:06 <[Fate]> it certainly isn't done with commenting out in Kconfig **** ENDING LOGGING AT Tue Jul 20 22:30:09 2010 **** BEGIN LOGGING AT Tue Jul 20 23:10:15 2010 Jul 21 00:32:07 xMff, cool Jul 21 00:35:11 xMff, very nice actually. Jul 21 00:44:53 xMff, ping Jul 21 00:47:02 here Jul 21 00:47:34 should there be a link to the ToH in doc :/ Jul 21 01:10:06 xMff, do we need a link to the ToH in doc/start ? Jul 21 01:10:13 no Jul 21 01:10:18 o ok :/ Jul 21 01:10:19 its linked on the mainpage already Jul 21 01:10:31 yeah Jul 21 02:05:22 who does the kernel mtd stuff on brcm47xx ? **** ENDING LOGGING AT Wed Jul 21 02:59:57 2010