**** BEGIN LOGGING AT Wed Feb 09 02:59:57 2011 Feb 09 03:11:35 build #73 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/73 Feb 09 04:04:14 build #82 of brcm47xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/82 Feb 09 09:33:25 hi. anyone around able to compile irssi mipsel with ipv6? Feb 09 09:38:14 isn't that compiled with v6 by default? Feb 09 09:38:17 sorry. im dumb. Feb 09 09:38:21 :) Feb 09 09:38:42 i found my error. Feb 09 09:40:57 i don't know how to spell freenode.net lol Feb 09 09:42:49 lol at ::dead:beef:cafe Feb 09 09:58:24 hello Feb 09 11:05:11 gmorning Feb 09 12:54:28 mb * r25429 /trunk/target/linux/omap24xx/patches-2.6.37/900-n810-battery-management.patch: n810bm: Add charger detection Feb 09 12:54:56 mb * r25430 /packages/Xorg/app/pwrtray/Makefile: pwrtray: Update to latest n810bm API Feb 09 13:09:54 florian * r25431 /trunk/target/linux/brcm63xx/base-files/etc/defconfig/ (SP500W/network SPW500V/ SPW500V/network): Feb 09 13:09:54 [brcm63xx] fix T-Com Speedport W 500V default config Feb 09 13:09:54 The device is called SPW500V, not SP500W. Feb 09 13:09:54 Signed-off-by: Jonas Gorski Feb 09 13:16:12 florian * r25432 /packages/utils/coreutils/Makefile: Feb 09 13:16:12 [package] update coreutils to 8.8 Feb 09 13:16:12 Signed-off-by: Alexander Khryukin Feb 09 13:27:18 build #8 of lantiq is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/lantiq/builds/8 Feb 09 13:55:33 Does anyone know how to print a more specif error message for: "ERROR: please fix package/pippo/Makefile" while building a new package? It should be possible to print the wrong Makefile line number or the missing keyword which cause the error? Feb 09 13:58:35 TOPDIR=`pwd` make -C package/pippo/ DUMP=1 V=99 Feb 09 14:04:30 ok, thanks Feb 09 15:35:35 anyone know if the older wrt54g gen broadcoms are mips32r2? Feb 09 15:36:27 (as opposed to r1) Feb 09 15:41:31 darned connection, noticed recently the bcm4716 even has ASE DSP if I read it right Feb 09 15:44:29 maligor_: they're r1 Feb 09 15:44:39 oh, ok, thanks Feb 09 15:50:10 I wonder why broadcom doesn't tell on their website, all they say on their public docs is: MIPS32 XXXMHz Feb 09 15:50:54 even for the newer chips Feb 09 16:01:36 could anyone point me to the right direction on how to solve a package compilation problem on backfire? It seems to be a make problem since the compilation of my package does not start, here the log: http://openwrt.pastebin.com/AHs8CL0L Feb 09 16:06:29 its missing the kernel .config as far as i can see. have you build the kernel already? Feb 09 16:31:13 the first time I have done a make, then I enabled the module with make menuconfig, and then I compiled the module. The problem is that the directory linux-2.6.32.10 is missing too Feb 09 16:34:14 the old Makefile (the one distributed with ipfw20100319 works fine on kamikaze_8.09.1 kernel version 2.4.35) Feb 09 16:34:26 but gives the same error on backfire Feb 09 16:39:52 the kernel-dir should be there, but on a recent backfire it is linux-2.6.32.27 Feb 09 17:02:36 build #73 of ppc44x is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ppc44x/builds/73 Feb 09 18:26:18 build #78 of cobalt is complete: Failure [failed shell compile_6] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/78 Feb 09 18:54:23 nbd: ping Feb 09 18:55:08 hcg_: pong Feb 09 18:56:45 Are there any tricks I should note for building OpenWRT on a MAC? Feb 09 18:57:05 I have not found any good documentation Feb 09 18:57:19 no, it should tell you what's missing through prereq checks Feb 09 18:57:32 the main thing you need is a case sensitive rootfs (or build in a disk image) Feb 09 18:58:37 a bunch of macports... coreutils, sha1md5sum etc. Feb 09 18:59:23 yes, it is complaining about not having a case-sensitive filesystem - I do not know much about OSX - is it possible to convert an existing rootfs into a case-sensitive one, or should I make a virtual FS for OpenWRT? Feb 09 18:59:38 it can only be chosen at install time Feb 09 18:59:43 so you need to build in a disk image Feb 09 19:00:59 Ah, OK, thanks for that. I will google that and then get onto the macports stuff. Feb 09 19:01:14 Another quick one - do you use glibc at all? Feb 09 19:01:33 no Feb 09 19:01:45 i don't see much reason for it either, iirc eglibc is binary compatible Feb 09 19:01:49 I am having issues building glibc 2.4 for an ARM platform Feb 09 19:01:50 and it's smaller and more tested than glibc Feb 09 19:02:08 at least on embedded hw Feb 09 19:02:33 I am wanting to try out the Oracle Embedded JVM, and that is linked against glibc 2.4 - is eglibc that compatible? Feb 09 19:03:09 not sure, but probably worth a try Feb 09 19:03:13 Of course the Embedded JVM is only available in binary form Feb 09 19:03:22 yeah, i figured Feb 09 19:03:51 iirc is one of eglibc's goals to be binary compatible to glibc Feb 09 19:04:56 I will probably wind up porting OpenJDK eventually once I find the time to do it Feb 09 19:39:20 hcg_: parallels? Feb 09 19:39:58 I have Virtualbox, but there are performance issues - I would prefer to have it native Feb 09 19:57:24 nbd: hi Feb 09 20:10:42 mb * r25433 /trunk/target/linux/omap24xx/patches-2.6.37/900-n810-battery-management.patch: n810bm: Set minimum voltage threshold to 3200 Feb 09 20:12:21 tornado: hi Feb 09 20:13:32 nbd: Problem solved like a charm. you're a great man. like always :) Feb 09 20:13:39 nice Feb 09 20:14:22 nbd: Thank you very much for your great news :) Feb 09 20:27:01 nbd: for some strange reason the packages fileutils, awk and find refuse to install using macports - any ideas? Feb 09 20:28:56 hcg_: what about coreutils? Feb 09 20:29:08 and which awk did you try to install? Feb 09 20:29:15 gawk is the right one Feb 09 20:29:52 Ah, thanks for the tips Feb 09 20:37:16 florian * r25434 /packages/utils/coreutils/Makefile: [package] coreutils: rework Makefile. Feb 09 20:37:19 florian * r25435 /packages/libs/jpeg/Makefile: Feb 09 20:37:19 [package] libjpeg: fix parallel build Feb 09 20:37:19 For parallel builds, make targets that build need to be separated from targets that install ;-) Feb 09 20:37:19 Signed-off-by: Joe Roback Feb 09 20:37:23 florian * r25436 /packages/libs/libusb/patches/004-no_unused.patch: (log message trimmed) Feb 09 20:37:23 [package] libusb: prevent building unused stuff Feb 09 20:37:23 This patch prevents the processing of the 'doc' and 'tests' Feb 09 20:37:23 subdirectories of the libusb package. Since the result of building these Feb 09 20:37:23 directories is never included in the ipkg package, omitting them saves Feb 09 20:37:24 compilation time and avoids possible errors if building the docs fails Feb 09 20:37:24 because of some openjade error (as it happened in my case). Feb 09 20:37:29 florian * r25437 /packages/libs/opencdk/ (Makefile patches/200-verify-clearsigned.patch): Feb 09 20:37:29 [package] opencdk: fix verification of clearsigned documents Feb 09 20:37:29 This patch fixes a bug in the upstream library that prevents the verification of Feb 09 20:37:29 clearsigned documents with \r\n line endings. Obviously, it should go under Feb 09 20:37:29 feeds/packages/libs/opencdk/patches/ Feb 09 20:37:29 Signed-off-by: Daniel A. Nagy Feb 09 20:37:30 florian * r25438 /packages/libs/opencdk/Makefile: [package] opencdk: change default mirror Feb 09 20:53:35 soma: I compiled against trunk (rev. 25432) and the compilation goes fine. (the old one was backfire_10.03), thank you very much Feb 09 23:54:17 I doubt this is gpl compliant: broadcom patches the atm bridge driver (which is gpl'd) so it directly calls functions of the proprietiary atm driver (on bcm63xx) Feb 10 00:16:33 Is a Netgear WNR3500L still requested for testing? (currently on list of requested hardware) Feb 10 00:26:28 jmccrohan: I suppose yes, otherwise it would've been removed Feb 10 00:33:40 on the hardware request for Ralink RT288x/RT305x... does it matter if the device is not on that list? (i can't give it up, but i can test builds of for this hardware_ Feb 10 00:34:24 I have been in contact Peter on the myopenrouter forums with a view to getting Netgear to donate a unit to OpenWrt. Feb 10 00:37:51 netgear's handling of the wnr3500l was ... weird Feb 10 00:38:14 they market it as open source router and ship it with a somewhat defunct openwrt Feb 10 00:38:21 having to use ddwrt is pissing me off bigtime Feb 10 00:38:24 thats not nice Feb 10 00:38:52 i think i'll sell it and get a WNR3700v2 when they filter out to retailers. Feb 10 00:38:57 ubiquiti did the same crap with the rspro, also shipping a crippled openwrt Feb 10 00:39:11 its like pc makers who ship with freedos to circumvent licensing Feb 10 00:39:14 at least ubiquiti sells hardware that can be supported with 100% free software Feb 10 00:39:35 marketing the wnr3500l as open, but not wndr3700 was a really crappy choice Feb 10 00:39:39 oh, and since i forgot to mention the exact device... it's the airlink 101 ar670w Feb 10 00:40:33 To get it clear, what exactly is not supported by OpenWrt on WNR3500L? Switch or wl? Feb 10 00:40:38 nbd: I think they rather meant "router for 3rd party firmware" and since tomato and dd-wrt are very popular they had to stick to bcm - imo Feb 10 00:40:55 jmccrohan: wireless only works with proprietary drivers Feb 10 00:41:05 so there's much less interest in supporting devices like that in openwrt Feb 10 00:41:31 as opposed to wndr3700, where everything works properly with 100% free software Feb 10 00:41:42 Thats what i thought. What is the current state of b43? Feb 10 00:41:50 tomato does not work at all on the wndr3700 and dd-wrt is still struggling with madwifi on it :) Feb 10 00:42:14 xMff: internally dd-wrt is using ath9k already Feb 10 00:42:17 ;) Feb 10 00:42:30 jmccrohan: works for some old chips, but not for the 11n ones Feb 10 00:42:31 yeah, I heard Feb 10 00:43:15 So if I am willing to taint my kernel for the timebeing, should the WNR3500L be supported by trunk? Feb 10 00:43:23 though there's been a lot of N-Phy activity in b43 lately Feb 10 00:43:26 jmccrohan: no idea, haven't really looked into it Feb 10 00:43:32 I suppose they scrape the useful bits from brcm80211 Feb 10 00:43:53 yeah, they managed to get legacy stuff working on some n-phy chips already Feb 10 00:44:05 still far away from having a usable 11n driver though Feb 10 00:45:08 anyways, i did spot dd-wrt has the rt2880 in it's tree (kernel at least), so it shouldn't take too much to integrate, possibly. Feb 10 00:45:15 jmccrohan: I'd follow the rt-n16 development a bit, once that is ironed out the wndr3500l will be easy to get running Feb 10 00:45:47 dandel: depends, the blob they use is most likely not portable to any recent kernel Feb 10 00:45:58 dandel: openwrt already works on several rt2880/rt3052 devices Feb 10 00:46:01 dandel: and it has a horrible interface (ini files) Feb 10 00:46:22 I don't know which blob it uses, but the kernel is 2.6.x based. Feb 10 00:46:25 and the wifi driver is getting closer to being usable, at least that's what i hear Feb 10 00:46:32 i mean the open source wifi driver Feb 10 00:46:51 xMff: I already am. :) Essentially the same device. Feb 10 00:47:23 dandel: trying to integrate the code from dd-wrt would be a waste of time Feb 10 00:47:36 jmccrohan: right now we lack any support for "modern" broadcom socs (afaik) Feb 10 00:47:42 nbd, I'm only talking about kernel. Feb 10 00:48:10 dandel: even there. it's better to try the openwrt kernel code and fix up the remaining issues as they show up Feb 10 00:48:28 the kernel code that's in dd-wrt (which originates from ralink) is horrible Feb 10 00:48:58 the only reason i am even using dd-wrt is I can't find the base kernel source anymore (removed from site) Feb 10 00:49:14 nbd: i don't "get" dd-wrt. it makes little sense. either have it full open or fully closed Feb 10 00:49:46 stuff like obfusticating the webui is just stupid Feb 10 00:49:55 nbd, where's the openwrt kernel code to rt2880 then. Feb 10 00:50:13 dandel: in the openwrt 'ramips' target Feb 10 00:50:45 jmccrohan: users seem to love it Feb 10 00:51:00 different mindset I suppose Feb 10 00:52:12 heh. http://tech.slashdot.org/story/11/02/09/2124210/Cisco-Linksys-Routers-Still-Dont-Support-IPv6#comments - first post Feb 10 00:52:15 nbd, I see, thanks... anything i should be looking for then? Feb 10 00:52:49 xMff: lol Feb 10 00:53:27 dandel: no idea what state the code is in Feb 10 00:54:13 IPv6 support in DD-WRT is horrible. Pasting scripts into text fields in the WebUI. Feb 10 00:54:28 they're currently fixing it Feb 10 00:54:30 jmccrohan: "it's intuitive and easy" Feb 10 00:55:17 Speaking of IPv6, anyone have it natively yet? Feb 10 00:55:23 people don't mind doing weird unintuitive stuff, as long as there's a howto that tells them exactly what to do Feb 10 00:56:14 nope :( not even 6to4 works reliable here, must have a bad server nearby Feb 10 00:56:27 hmm... I'll haft to be careful then... I don't have the jtag tools, and information on how to hardmod jtag onto the board. :/ Feb 10 00:56:57 dandel: it's unlikely that you will need jtag to recover Feb 10 00:57:18 dandel: most devices come with a boot loader that you can use to reflash (if you've connected to the serial console) Feb 10 00:57:53 nbd, it takes knowing how to connect the serial console to the device. Feb 10 00:58:17 xMff: I only experimented with it recently. local SixXS PoP works pretty well. UPC can't even get "unlimited" correct, i doubt they'll have IPv6 soon. Feb 10 00:58:24 dandel: sure, but if you want to play with kernel code you need that anyway Feb 10 00:59:06 trying to fix a kernel without having serial console output is not a good idea Feb 10 00:59:34 jmccrohan: yeah, same everywhere Feb 10 01:00:46 jmccrohan: not even equinix can provide us with ipv6 yet, and thats not exactly consumer grade Feb 10 01:01:27 nbd, I know, and part of why i want to get openwrt working is because the dd-wrt implementation is horrible. Feb 10 01:01:33 the 11n chip support in b43 is currently limited to the bcm4321 (iirc only with b/g, so no actual 11n). anything newer isn't supported (yet) Feb 10 01:02:05 KanjiMonster: I see Feb 10 01:02:20 I haft to reset the device to change wifi settings... it hangs when loading the wifi settings page, and i haven't figured out what happened yet. Feb 10 01:07:46 nbd, if you want a fun port, look at the airlink+ ar315w (airlink101) Feb 10 01:10:32 xMff: there is some groundwork done for the bcm4322 (and b43 also happily attaches if you enable the n phy support ;), but it currently isn't usable Feb 10 01:13:27 dandel: what would be fun about it? Feb 10 01:13:42 marvell powered device, similar to tew-432brp Feb 10 01:14:20 heh, and you call that fun? Feb 10 01:15:08 It's fun in a sense that it's pretty much an device asking to be unlocked from the poor firmware that is used originally Feb 10 01:16:37 i think i'll stick with the decent hardware platforms instead Feb 10 01:16:51 ;) Feb 10 01:17:04 nbd, what constitutes a decent hardware platform, in your opinion. Feb 10 01:18:15 hardware that works well and where there's either extensive documentation or reasonably readable free software reference code available. Feb 10 01:19:16 also, this thing seems to have *very* little RAM/flash Feb 10 01:20:59 these days i mostly stick to hardware where the chip manufacturers actually like having decent open source software support and work with the community Feb 10 01:21:08 nbd, I'm trying to decode how much ram it has. Feb 10 01:21:51 it uses EtronTech EM636165TS-6 ram, which from what i can figure out equates to 64MBit x16 sdram tsop (6ns).. and it uses two of these chips. Feb 10 01:22:30 ah, so RAM is not as bad as i thought Feb 10 01:22:32 but the rest is still crap ;) Feb 10 01:23:13 it has 8mb of flash. Feb 10 01:24:17 but it has marvell wireless Feb 10 01:24:30 and the ethernet switch sucks Feb 10 01:25:21 I expect it not to get any better. Feb 10 01:26:37 and who still buys 802.11g routers? Feb 10 01:26:46 anyways, i at least want to get enough tweaks to possibly get ipv6, and wireless bridge. (It's an old router of mine) Feb 10 01:27:06 my recommendation is to not waste any time on that thing Feb 10 01:27:16 the platform is dead anyway Feb 10 01:27:21 nobody will care about it Feb 10 01:27:37 then what about the rt2880 802.11n airlink ar670w of mine? Feb 10 01:28:33 how much experience do you have with kernel and driver hacking? Feb 10 01:30:38 not a lot with driver/kernel hacks... mainly had experience figuring out acpi bugs and triangulating the patch which caused the bug. Feb 10 01:31:25 well, to fix up the platform you'd probably have to familiarize yourself with the code more than just playing with patches and testing them Feb 10 01:32:48 I'll probably start with the ar315 since i don't care if i brick that. Feb 10 01:34:09 anyways, how do i attach jtag to the router? (i see an header on the board which is 20 pins) Feb 10 01:34:41 buy a proper arm ejtag cable, then connect it to the header ;) Feb 10 01:37:18 http://www.amontec.com/jtagkey-tiny.shtml Feb 10 01:38:42 xMff, any cheaper than that? Feb 10 01:38:58 I think it is already one of the cheapest, sorry Feb 10 01:39:10 aside from hand soldering some cables Feb 10 01:39:14 or, possibly the schematics to build a cable for a com port? Feb 10 01:39:43 com port? you mean parallel port, right? Feb 10 01:39:53 No, i mean serial port. Feb 10 01:40:00 you can't do jtag over a serial port Feb 10 01:40:02 I have no parallel port. Feb 10 01:40:26 parallel port stuff works in principle, but is often unreliable Feb 10 01:40:29 ... and slow Feb 10 02:23:49 i'm on a ipv6 kick today. will be attempting to provide dns64 within dnsmasq tonight. :D Feb 10 02:24:00 unless there is already a better solution? Feb 10 02:24:51 How does OpenWrt handle the init sequence? Does a script only start when all scripts with a lower init value complete? Or does it run them in parallel? Feb 10 02:25:14 usually parallel, look at /etc/rc.d/rcS i beleive Feb 10 02:25:39 sequential Feb 10 02:25:42 00..99 Feb 10 02:25:55 parallel stuff is triggered by hotplug Feb 10 02:26:09 But does n+1 depend on n being totally finished? Feb 10 02:26:39 Or does it just fire them off in order? Feb 10 02:26:49 they're fired in order Feb 10 02:27:04 whether a script block or goes into background for its stuff varies Feb 10 02:27:34 but rcS says & Feb 10 02:28:16 rcS just loops over all init scripts Feb 10 02:28:26 rcS itself is in background, the init scripts not Feb 10 02:28:40 if [ "$1" = "S" -a "$foreground" != "1" ]; then run_scripts "$1" "$2" & Feb 10 02:29:01 * neutronscott shrugs Feb 10 02:29:35 yeah and run_scripts() contains a for loop Feb 10 02:29:48 the entire loop runs in background, not each script by itself Feb 10 02:30:09 ah. my baddd Feb 10 02:30:56 well sorry for interjecting and making that answer harder for ya but i also learned ;) Feb 10 02:31:35 xMff, neutronscott: Thanks :) Feb 10 02:48:52 nbd, the firmware i found dates back to linux 2.0.x ><; (on the marvel source, from asus) Feb 10 02:51:07 haha Feb 10 02:55:35 at least it's better than nothing. **** ENDING LOGGING AT Thu Feb 10 02:59:57 2011