**** BEGIN LOGGING AT Thu Jul 30 02:59:57 2009 Jul 30 07:44:33 good morning Jul 30 07:53:06 <[florian]> hey Jul 30 08:18:49 florian * r17046 /packages/net/mac-to-devinfo/files/ (7 files): [package] add copyright headers to mac-to-devinfo script (#5593) Jul 30 08:45:02 phaidros: ping Jul 30 09:08:01 lars * r17047 /trunk/target/linux/s3c24xx/files-2.6.30/drivers/mfd/glamo/glamo-mci.c: [s3c24xx] glamo-mmc: Limit clock rate. Jul 30 09:13:27 hi should I open a bug for that http://pastebin.com/d484c547f << FTBFS on hotplug2 with trunk Jul 30 09:13:31 ? Jul 30 09:14:35 <[florian]> yes please Jul 30 09:20:18 done https://dev.openwrt.org/ticket/5596 Jul 30 09:22:16 glp: pong Jul 30 09:22:22 <[florian]> bigon: thanks Jul 30 09:30:22 phaidros: I had an error message from nuwiki when I tried to move/rename a page? Jul 30 09:32:00 glp: what exactly was the error Jul 30 09:32:07 and, did it work afterwards? Jul 30 09:39:19 phaidros: I'll try to reproduce it, can't remember the exact message - but it happened several times in a row Jul 30 09:39:59 ok Jul 30 09:41:17 phaidros: I wanted to change the path to the meeting log on this page: http://nuwiki.openwrt.org/meta/ircmeeting/2009-07-23 Jul 30 09:41:30 and now the meeting log page seems to have gone missing Jul 30 09:42:54 hm, seems there is no meetinglog page o.O Jul 30 09:44:20 phaidros: which I don't really understand, in as much as I added that page on monday? Jul 30 09:46:50 glp: the page is there. Jul 30 09:46:58 I was just cnfused, coz was looking under meta .. Jul 30 09:47:12 of course it is still under wiki: where one originally put it Jul 30 09:47:48 to really *move* a page, you have to do that from the admin interface. Jul 30 09:48:26 http://nuwiki.openwrt.org/wiki/meetinglog230709 Jul 30 09:48:55 there it is. but i wonder, the 'wiki' namespace shouldn't even exist .. do you remember anything about that? Jul 30 09:49:03 not sure anymore :) Jul 30 09:49:35 phaidros: yes the wiki namespace was my fault, and I tried to change it via the admin interface Jul 30 09:49:43 which then gave me the error message Jul 30 09:52:42 ok, so I try that now, moving from wiki/ to meta/ircmeeting/ , right/ Jul 30 09:52:44 ? Jul 30 09:52:58 phaidros: try it :) Jul 30 09:53:51 arning: preg_replace() [function.preg-replace]: Compilation failed: unmatched parentheses at offset 37 in /var/www/servers/beta.openwrt.org/pages/lib/plugins/pagemove/admin.php on line 414 Jul 30 09:53:55 ther eit is :0 Jul 30 09:54:02 so, I hunt that down Jul 30 09:55:01 phaidros: cool :) Jul 30 10:55:50 glp: I was able to reproduce the error again and again, and finally move the page by hand. Jul 30 10:56:15 that is no real solution, but I emailed the developer of the pagemove plugin already, let's see if he has an idea. Jul 30 10:56:39 I have to say, the error is not obvious there and I cannot understand why it is happening in that part of the code :/ Jul 30 10:59:23 phaidros: me neither Jul 30 11:07:19 bah, it changes the link, but not the name of the file/page Jul 30 11:09:33 I used the stone age variant: edit, ctrl-a, ctrl-c, make new page, edit, ctrl-v, save :) Jul 30 11:19:37 heh, but thats what the pagemove plugin is for .. should be :) Jul 30 11:56:13 phaidros: http://nuwiki.openwrt.org/inbox/howto.build.openwrt Jul 30 11:58:19 glp: nice :) Jul 30 12:04:05 lars * r17048 /trunk/target/linux/s3c24xx/files-2.6.30/ (10 files in 7 dirs): Jul 30 12:04:05 [s3c24xx] glamo: Implement gpiolib for the glamo. Get rid of glamo-spi-gpio and Jul 30 12:04:05 use the generic spi-gpio driver instead. Jul 30 12:12:13 lars * r17049 /trunk/target/linux/s3c24xx/files-2.6.30/sound/soc/s3c24xx/gta02_wm8753.c: Jul 30 12:12:13 [s3c24xx] revert changes to gta02_wm8753.c in previous commit. They should not Jul 30 12:12:13 have been commited. Jul 30 13:05:53 florian * r17050 /trunk/include/package-defaults.mk: [package] move the global ipv6 option to the very bottom of the options passed to configure scripts, so that we can still override it in package makefiles (#5592, #5586) Jul 30 13:58:50 lars * r17051 /trunk/target/linux/s3c24xx/files-2.6.30/drivers/mfd/glamo/ (glamo-core.c glamo-fb.c glamo-mci.c): Jul 30 13:58:50 [s3c24xx] glamo: The cell can rely on being a child of the core. Use it for Jul 30 13:58:50 passing platform data. Jul 30 15:08:08 lars * r17052 /trunk/target/linux/s3c24xx/ (10 files in 4 dirs): Jul 30 15:08:08 [s3c24xx] Get rid of gpio shadowing. Jul 30 15:08:08 I fail really to see why this would be necessary here. So get rid of it and see if Jul 30 15:08:08 anything breaks... Jul 30 15:40:38 lars * r17053 /trunk/target/linux/s3c24xx/ (3 files in 3 dirs): [s3c24xx] Replace gta02 led driver with generic gpio led driver. Jul 30 17:15:24 jushog, you around, I've got some progress to report for ya, and, one simple question Jul 30 17:15:44 on the wrt160nl, i've got it up and running with uboot netboot stuff no problem Jul 30 17:15:51 everything working except the ethernet right now Jul 30 17:16:07 this morning, i've been working on formatting up bin files that the router with default stuff will accept Jul 30 17:16:14 linksys gpl code builds them like this Jul 30 17:16:36 it takes the kernel uimage, and joins it with the root file system, into a trx, just like the build for broadcom Jul 30 17:16:40 then it adds the patterns etc Jul 30 17:17:01 i've been adding that into the makefiles in the 160nl profile, and now have it building images the router will accept Jul 30 17:17:44 once thats flashed in, it boots, and i'm _just now_ starting to sort out the flash map issues for getting the squashfs mapped correctly so it can actually find it's rootfs Jul 30 17:18:18 when building for initramfs, the build scripts do not run thru the profiles and build device specific images Jul 30 17:18:40 I think I'm gonna hack in some stuff in that part too, so it will spit out an initramfs image that uboot is happy to flash in and boot from the flash Jul 30 17:19:38 the part now that i'm not really clear on, and need to hunt down, and this will help immensly if you have a quick pointer to the correct spot Jul 30 17:20:00 by packaging up the uimage into a trx, then use the addpattern, I've got it to boot the kernel from flash Jul 30 17:20:22 but, i dont have the board set properly, so it's recognizing generic rather than 160nl Jul 30 17:20:29 which i normally pass on the uboot command line Jul 30 17:20:34 when netbooting Jul 30 17:21:24 is there an easy way to wire in a board specific kernel command line in the profile section of the image makefiles ? Jul 30 17:21:44 that's the part I haven't quite figured out yet, still unravelling a lot of how this build env really works Jul 30 17:27:00 can you even include a command line in the flash image? I know that some targets had kernel patches that altered the command line in the kernel code Jul 30 17:27:33 i'm just hunting that part down, and, I think I may have found it, will take a little while to validate it Jul 30 17:28:22 but I barely know how u-boot works, just used the first time ever two days ago Jul 30 17:28:46 I'm trying to end up with a bin file that I can load into stock factory firmware Jul 30 17:28:56 and it'll come up correctly, ie, without changes to the on board uboot Jul 30 17:29:25 then you propably have to change the commandline in the kernel source Jul 30 17:30:14 but I have no real clue, so don't trust this comment :) Jul 30 19:10:16 BusyBox v1.13.4 (2009-07-29 13:06:33 PDT) built-in shell (ash) Jul 30 19:10:16 Enter 'help' for a list of built-in commands. Jul 30 19:10:16 Jul 30 19:10:16 _______ ________ __ Jul 30 19:10:16 | |.-----.-----.-----.| | | |.----.| |_ Jul 30 19:10:17 | - || _ | -__| || | | || _|| _| Jul 30 19:10:19 |_______|| __|_____|__|__||________||__| |____| Jul 30 19:10:21 |__| W I R E L E S S F R E E D O M Jul 30 19:10:23 KAMIKAZE (bleeding edge, r16658) ------------------- Jul 30 19:10:25 * 10 oz Vodka Shake well with ice and strain Jul 30 19:10:27 * 10 oz Triple sec mixture into 10 shot glasses. Jul 30 19:10:29 * 10 oz lime juice Salute! Jul 30 19:10:31 --------------------------------------------------- Jul 30 19:10:33 root@OpenWrt:/# Jul 30 19:10:35 wrt160nl Jul 30 19:10:37 booted from an initramfs image, written to it's flash Jul 30 19:10:45 root@OpenWrt:/# ifconfig -a Jul 30 19:11:39 hehe, ok, that paste was to big Jul 30 19:12:36 argh Jul 30 19:12:45 i have running squashfs patch for wrt160nl here Jul 30 19:12:49 including trx parser Jul 30 19:13:09 lol, and i've been working on this one for a couple days now Jul 30 19:13:13 an working ethernet Jul 30 19:13:18 OHHHHHHHHHHH Jul 30 19:13:19 i did so last week... Jul 30 19:13:28 i've not got ethernet working yet Jul 30 19:13:35 email-address? Jul 30 19:15:08 gerryr@rozeware.bc.ca Jul 30 19:39:17 groz: ping Jul 30 19:41:09 hi Jul 30 19:41:20 do you want to finish my rtl8306 driver? Jul 30 19:41:36 all that's missing is making it detect the chip and pass over the mii bus Jul 30 19:41:48 i'd love to Jul 30 19:41:57 but, my schedule has me able to work on this stuff for 2 more days Jul 30 19:41:57 the hardware configuration bits are tested and working on a different platform Jul 30 19:42:04 then i'm likely gone for better part of 3 weeks Jul 30 19:42:15 if the stuff I just got from hopscotch 'drops in and works' Jul 30 19:42:30 then I'll have a function build with ethernet, and correctly generating bin files Jul 30 19:42:40 so, all we need at that point, is full suppport on the 8306 Jul 30 19:42:45 groz: http://nbd.name/rtl8306.c Jul 30 19:42:47 and this platform will be close to soup Jul 30 19:42:56 this driver hooks into my swconfig stuff and it should attach through the phy layer Jul 30 19:43:12 ok, what I'll do, is first apply these other patches I just got Jul 30 19:43:17 get that building and flashing properly Jul 30 19:43:24 then dive into this switch driver Jul 30 19:43:26 ok Jul 30 19:43:47 i got a deal on a stack of 5 boards here the other day Jul 30 19:43:49 i'm still waiting for an wrt160nl, so i can't test the driver at the moment Jul 30 19:44:01 well, i can test it Jul 30 19:44:06 and, if _you_ want to test it Jul 30 19:44:19 i can set up a 160nl with serial console Jul 30 19:44:20 well, i also don't have the time to deal with this right now Jul 30 19:44:30 give you access to everything, including a login to a working build env Jul 30 19:44:31 too many projects running simultaneously Jul 30 19:44:36 haha, all of us Jul 30 19:44:44 me too, but, this is a fun distraction for a few days Jul 30 19:44:47 ;) Jul 30 19:44:50 * groz got a few weeks off from the airplanes Jul 30 19:45:17 either way, if you don't get around to finishing this, it's not a big deal Jul 30 19:45:17 only one flight left to do on friday, to 'finish' training on my replacement Jul 30 19:45:21 at some point i will find the time Jul 30 19:45:23 then, 4 weeks free of real work Jul 30 19:45:37 3 of which we plan to spend out in a campsite in the middle of nowhere Jul 30 19:45:57 sounds like fun Jul 30 19:46:06 i hope it will be, but, as part of the fun Jul 30 19:46:08 * nbd is going spend some time on a campsite soon as well Jul 30 19:46:21 i'm gonna have a couple routers along, and, probably put a full build env onto my netbook Jul 30 19:46:38 painful slow, but, one way to while away days when the weather is not so nice Jul 30 19:46:51 i expect to have decent connectivity on the campsite that i'm going to Jul 30 19:46:54 http://har2009.org Jul 30 19:46:55 ;) Jul 30 19:47:07 haha, well, i'm bringning along my own connectivity Jul 30 19:47:15 usb 3g modem in the wrt160nl Jul 30 19:47:19 ;) Jul 30 19:47:26 got that working yesterday from the uboot netboot Jul 30 19:47:38 but, for at least one week, we will be in a campsite that has Jul 30 19:47:47 no cellular coverag Jul 30 19:47:58 we will be completely out of touch with the world Jul 30 19:48:16 sometimes that's not a bad thing Jul 30 19:48:30 no, we are both really looking forward to that week Jul 30 19:48:37 great rainbow trout fishing there Jul 30 19:48:44 cool Jul 30 19:48:48 mmm, I love rainbow trout. Jul 30 19:48:53 the lake is good for swimming, maybe a little cold it's glacier fed Jul 30 19:49:17 i think chris will spend the whole week photographing the wildlife Jul 30 19:49:32 * groz gonna spend the week murdering fish Jul 30 19:51:21 nbd, i just looked at that webpage Jul 30 19:51:31 you need to take ur vacation where there ARE NO PUTERS Jul 30 19:52:03 no geeks, just pretty girls on the beach in skimpy / no bikinis Jul 30 19:52:17 yeah, i'll do that as well, but another time ... Jul 30 19:52:22 haha Jul 30 19:54:37 http://www.mksp.ca/pages/history/2008/2008photos.php Jul 30 19:54:46 thats where we will be for one week of the trip Jul 30 20:01:09 ok, now i submitted my patches to the devel-list Jul 30 20:01:19 i hope, at least part of it will be accepted... Jul 30 20:02:26 Hopscotch: patches look good, but i'll wait for juhosg or kaloz to apply them, since they maintain the target Jul 30 20:03:13 cool Jul 30 20:03:22 the detection is really ugly though :/ Jul 30 20:04:00 yeah, i was wondering about that Jul 30 20:04:05 does the search space really need to be that big? Jul 30 20:04:06 here's what I did Jul 30 20:04:19 i modified the build env, so, when building for 160nl Jul 30 20:04:34 i'm patching in the board=WRT160NL kernel parameter Jul 30 20:04:49 similar to how some of the other targets in ar7xx do it in the board profile Jul 30 20:05:09 nbd: i'm looking for the "WRT160NL u-boot version: 1.0.0" greeting Jul 30 20:05:39 nbd: i guess, one could narrow it down to the u-boot const segment Jul 30 20:05:55 does uboot not pass any magic into the fw_args? Jul 30 20:06:03 nope, nothing as far as i could see Jul 30 20:06:12 hang on Jul 30 20:06:15 i'll get into uboot Jul 30 20:06:21 have you looked at the source? Jul 30 20:06:27 hmmm Jul 30 20:06:31 not too much Jul 30 20:06:32 i noticed that they did provide them in a newer version of the gpl tarball Jul 30 20:06:48 gave up after i saw that all the interesting parts are missing Jul 30 20:06:50 ... Jul 30 20:07:13 bootargs=console=ttyS0,115200 root=31:04 rootfstype=squashfs init=/sbin/init mtdparts=ar7100-nor0:256k(u-boot),7808k(linux),64k(nvram),64k(ART),64k(filesystem) bootver=1.1.5 Jul 30 20:07:57 there is nothing in the uboot environment that jumps out as 160NL specific Jul 30 20:08:33 yes, i looked really hard into the running u-boot Jul 30 20:08:50 only idea i had was to use the mac address Jul 30 20:08:58 linksys uses ranges for the different models Jul 30 20:09:11 yah, and they can very arbitrarily change later Jul 30 20:09:16 yep Jul 30 20:10:04 that's kinda why i liked my method, i patch the kernel command line just before building the bin file Jul 30 20:10:13 and put the wrt160nl directly into the command line Jul 30 20:10:46 results in special images, but i can live with that :) Jul 30 20:10:47 $(STAGING_DIR_HOST)/bin/patch-cmdline $(KDIR)/vmlinux-$(2) '$(3)' Jul 30 20:11:13 that puts the board=WRT160NL directly inot the kernel startup Jul 30 20:11:30 so, if you have selected 160NL in make menuconfig Jul 30 20:11:34 it'll be there in the bin file Jul 30 20:11:38 ok, that is some force Jul 30 20:11:45 seems to do the job Jul 30 20:11:51 yes, but, it means never chasing hardware changes Jul 30 20:11:58 the kick my hack out... didn't like it too much anyways :) Jul 30 20:11:58 if you know you are building for an nl Jul 30 20:12:02 you do it at build time Jul 30 20:12:13 Hopscotch: as far as i can see, it does hand over a pointer to a board info struct to the kernel Jul 30 20:12:22 well, first lemme see how this all works, i'm just getting ready to boot with your other two patches Jul 30 20:12:26 i haven't applied that one yet Jul 30 20:12:43 nbd: this also doesn't contain anything useful Jul 30 20:13:22 nbd: i dumped that stuff in the prom.c file Jul 30 20:13:53 Hopscotch: afaics it *should* contain the ethernet mac address Jul 30 20:14:03 nbd: yes, that it does Jul 30 20:14:09 that's at least 3 bytes of magic Jul 30 20:14:10 ;) Jul 30 20:14:14 hehe Jul 30 20:14:21 just compare the OUI Jul 30 20:14:44 only works as long as linksys doesn't start making other ar71xx/ar91xx based devices Jul 30 20:14:54 they will probably be able to run with similar code Jul 30 20:15:27 and if not, we'll deal with the individual differences later Jul 30 20:15:28 that's a bit of a bet Jul 30 20:15:34 ok :) Jul 30 20:15:58 we could use this first test to make sure that it's linksys, then do other linksys specific tests to check the exact model name Jul 30 20:16:17 that reduces the probability of fallout on other kinds of devices Jul 30 20:16:32 very true Jul 30 20:16:37 as long as we make sure that the board info pointer doesn't point to a place where the device could crash Jul 30 20:16:47 but that should be easy Jul 30 20:20:22 ok, booted an initramfs from uboot with your patches applied hoptscotch Jul 30 20:20:34 ethernet is up, eth1 dhcp from the network fine Jul 30 20:26:36 ok, the trx parser will need a little more smarts too Jul 30 20:26:50 if we boot with a squash root, it needs to create one more partition Jul 30 20:26:54 which is 'all unused flash' Jul 30 20:27:01 right Jul 30 20:27:06 so that we can format that with jffs, and then do the mini-fo trick Jul 30 20:27:09 i knew, something was missing :) Jul 30 20:27:20 another option is Jul 30 20:27:40 it already makes a partition called 'firmware', which is essentially kernel + filesystem + empty Jul 30 20:27:52 could change it to just mark that one as starting after the squash Jul 30 20:28:04 instead of starting where kernel starts Jul 30 20:28:17 or rename it to rootfs and use the code that's already in Jul 30 20:28:20 but that leaves problems later, trying to upgrade to a larger kernel Jul 30 20:28:28 at least i saw something that could do the trick Jul 30 20:28:39 yes, i'll dig into that shortly Jul 30 20:28:51 right now, i see ethernet working, and, I need some food Jul 30 20:28:54 it's past lunchtime Jul 30 20:28:55 haha Jul 30 20:29:12 hehe Jul 30 20:29:19 here it's past bedtime for me Jul 30 20:31:52 blinking blue leds are evil Jul 30 20:32:50 what part of the world are you in ? Jul 30 20:33:16 a small piece of electrians tape can fix a blinking blue led Jul 30 20:33:55 as can a solder iron Jul 30 20:34:04 but i don't understand these designers Jul 30 20:34:12 lately every fscking led is blue Jul 30 20:34:19 i always think the police is behind me Jul 30 20:34:27 i'm in europe Jul 30 20:34:49 ok Jul 30 20:34:49 im on north american west coast Jul 30 20:34:55 * groz can see pacific ocean out front window Jul 30 20:35:06 *sigh* Jul 30 20:35:26 travelling the us is on my list Jul 30 20:35:45 not mine Jul 30 20:35:57 * groz in canada, no interest in going south of 49th these days Jul 30 20:36:09 the border goons are to snarley these days Jul 30 20:36:10 ah, canada - even better Jul 30 20:36:40 we were in washington state last week for a couple days, border goons gave us huge hassle on the way down Jul 30 20:37:02 at one point, I thought they were gonna rip the camper apart Jul 30 20:37:06 that's exactly why i didn't go there during the last years Jul 30 20:37:26 the doller was cheap, but their angst is hilarious Jul 30 20:37:26 hehe, i love to tease our visitors from down south Jul 30 20:37:42 'you only need the drivers license to come to canada, but, you need your passport to go home again' Jul 30 20:38:03 oh, that cheap dollar, has me upset Jul 30 20:38:10 * groz gets paid on $us Jul 30 20:38:13 in Jul 30 20:38:17 ouch Jul 30 20:38:27 yah. last summer was brutal Jul 30 20:39:01 from our view going to the us was like going to turkey Jul 30 20:39:03 pricewise Jul 30 21:26:51 any idea how can i compile some code which uses libc functions only and upload them to my router which runs on mips32 on dd-wrt ? Jul 30 21:29:28 florian * r17054 /trunk/target/linux/brcm63xx/ (18 files in 5 dirs): [brcm63xx] add experimental support for 2.6.28.10, tested on bcm6338 Jul 30 21:43:16 is arm in openwrt still big-endian-only? Jul 30 21:44:15 we have the orion target which is little endian Jul 30 21:44:41 him_: why are you asking about dd-wrt in this channel? Jul 30 21:45:05 him_: we already told you not to bug us with it ;) Jul 30 21:46:45 nbd: thanks. just got an arm board, checking debian prefers LE, openembedded prefers LE as well, i remember ixp4xx is BE-only Jul 30 21:47:33 i've been able to compile openwrt for ixp4xx le easily, with minimal changes Jul 30 21:47:46 but the kernel byteswapping stuff wasn't reliable enough Jul 30 21:48:00 i think the only reason why debian prefers LE is they had no actively maintained BE port for a while Jul 30 21:48:11 for a router, BE is definitely better than LE Jul 30 21:49:50 nbd: makes sense. thanks. Jul 30 23:17:55 lars * r17055 /trunk/target/linux/s3c24xx/ (3 files in 3 dirs): [s3c24xx] Post cleanup cleanup Jul 30 23:37:56 lars * r17056 /trunk/target/linux/s3c24xx/files-2.6.30/drivers/mfd/glamo/glamo-core.c: [s3c24xx] It's to late... Jul 31 00:26:11 ping blogic Jul 31 01:23:05 is there any icecast compatible streaming tool (darkice, ezstream ...) available on openwrt? **** ENDING LOGGING AT Fri Jul 31 02:59:57 2009