**** BEGIN LOGGING AT Fri Jan 21 02:59:57 2011 Jan 21 04:14:18 build #53 of ep93xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/53 Jan 21 04:45:46 build #68 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/68 **** BEGIN LOGGING AT Fri Jan 21 07:52:17 2011 Jan 21 07:56:35 Hauke: ping Jan 21 07:57:04 philipp64|laptop: pong Jan 21 07:57:19 hi. got a minute to review my patches? Jan 21 07:57:26 just 2... they're fairly trivial. Jan 21 08:20:44 ok. maybe another time. Jan 21 08:51:52 philipp64|laptop: at the moment I do not have time to do that Jan 21 12:55:34 more trac spam, #8691 Jan 21 12:56:24 swalker: deleted, thanks Jan 21 17:52:31 kaloz * r25057 /trunk/target/linux/generic/ (67 files in 2 dirs): [target/generic]: preliminary 2.6.38 support Jan 21 17:54:53 nbd * r25058 /trunk/package/mac80211/patches/550-ath9k_fix_ps_wakeup_issues.patch: ath9k: fix some ps wakeup/restore issues that led to crashes and other weird behavior Jan 21 18:09:06 nbd * r25059 /branches/backfire/package/mac80211/patches/570-ath9k_fix_ps_wakeup_issues.patch: ath9k: backport ps wakeup/restore fixes from r25058 Jan 21 18:32:11 nbd * r25060 /trunk/package/mac80211/Makefile: ath9k: kmod-ath9k has no extra menuconfig options, remove MENU:=1 (#8695) Jan 21 18:32:58 nbd * r25061 /branches/backfire/package/mac80211/Makefile: ath9k: backport menuconfig change from r25060 Jan 21 18:37:15 nbd: hey Felix, can you have a look at my patches? thanks. Jan 21 19:07:10 i've successfully build an image for my orion WNR854T (netgear) Jan 21 19:07:25 its possible to change only kernel ? Jan 21 19:07:45 i've applied a small patch to try led support Jan 21 19:08:14 you can make a ramdisk image and boot it by hand to try out Jan 21 19:08:41 the snapshot build doesnt support kexec... Jan 21 19:08:55 hmm the ramdisk ? Jan 21 19:08:59 no Jan 21 19:09:04 do oyu have serial on uboot ? Jan 21 19:09:06 with serial ? Jan 21 19:09:10 oh Jan 21 19:09:11 yes Jan 21 19:09:15 that is how i do it Jan 21 19:09:22 hmm i've made it with putty yea Jan 21 19:09:34 yesterday (firewall was buggy) Jan 21 19:09:35 silly putty? Jan 21 19:09:40 tftp ${mem_addr} foo_uimage; bootm ${mem_addr} Jan 21 19:09:50 :) Jan 21 19:10:02 printenv shows you the environmanet Jan 21 19:10:03 silly ? Jan 21 19:10:13 there will be vars, look for the one that seems to be in ram Jan 21 19:10:35 you mean "set" Jan 21 19:10:37 ? Jan 21 19:10:50 euh... Jan 21 19:10:55 print Jan 21 19:11:11 tftpboot 6000000 vmlinuz Jan 21 19:11:14 cat /proc/iomem Jan 21 19:11:35 00000000-01ffffff : System RAM Jan 21 19:11:41 00021000-00278fff : Kernel text Jan 21 19:11:45 oh Jan 21 19:11:45 0027a000-0029bf33 : Kernel data* Jan 21 19:12:16 f1012000-f10120ff : serial8250.0 Jan 21 19:12:16 f1012000-f101201f : serial Jan 21 19:13:25 00000000-03ffffff : System RAM Jan 21 19:13:29 be jealous Jan 21 19:13:50 32M is good :) Jan 21 19:14:01 i will soon have a 3500L Jan 21 19:14:13 with 64M if i remember Jan 21 19:14:18 64M Jan 21 19:14:28 and usb :) Jan 21 19:14:37 to have some more place to log traffic Jan 21 19:14:50 i paid $15 for this. it has 64M and hopefully i'll get USB working someday heh Jan 21 19:15:19 on the WRT854 there is maybe hidden pins for usb dunno Jan 21 19:15:25 so anyways. You have new kernel only and want to test it. Did what blogic say make sense? Jan 21 19:15:26 the SoC supports 2 ports Jan 21 19:15:34 WNR Jan 21 19:15:36 erf Jan 21 19:16:06 i installed the trunk version yesterday Jan 21 19:16:19 so i think my kernel could be compatible with rootfs Jan 21 19:16:35 i used git to get openwrt sources Jan 21 19:16:56 and i have wnr854t-uImage.jffs2 Jan 21 19:17:02 wnr854t-zImage Jan 21 19:17:09 and root.squashfs Jan 21 19:17:35 i could try a webupgrade... Jan 21 19:17:50 but i dunno how to make the webupgrade image Jan 21 19:18:26 is there a tool to execute a kernel ? like mboot ? Jan 21 19:18:50 i could download the image in /tmp Jan 21 19:18:53 (memory) Jan 21 19:19:15 why not commit to flash the whole new image Jan 21 19:19:24 no way to recover? Jan 21 19:19:29 in fact yea Jan 21 19:19:30 i dunno Jan 21 19:19:38 if i can if there is a problem :p Jan 21 19:19:53 You have a serial console access to the u-boot loader? Jan 21 19:19:58 my serial interface is working nice Jan 21 19:20:02 yea Jan 21 19:20:10 i just need to reopen the box ;p Jan 21 19:20:18 You can boot the kernel from RAM using TFTP Jan 21 19:20:35 tftp from the host right ? Jan 21 19:20:44 the router is the tftp client. Jan 21 19:20:56 to tftpd from uboot or from openwrt ? Jan 21 19:21:00 u-boot won't incorporate a tftp server :( Jan 21 19:21:20 ok... my first wifi router (WRT54G had that :p) Jan 21 19:21:26 on startup Jan 21 19:22:18 ok, will try the tfp upload :) Jan 21 19:22:21 yah its nicer Jan 21 19:23:47 hmm.. need to find where i enable tftpd :) Jan 21 19:30:24 what host Jan 21 19:31:11 neutronscott@iBook:~$ grep tftp /etc/inetd.conf Jan 21 19:31:11 tftp dgram udp wait nobody /usr/sbin/tcpd /usr/sbin/in.tftpd /srv/tftp Jan 21 19:33:14 tftpd-hpa iputils-tftpd atftpd Jan 21 19:33:30 hmm which one is working good for that :) Jan 21 19:34:45 neutronscott, $15 for a device with 64M? what is it? Jan 21 19:35:37 did you sold yourself the chip ? :p Jan 21 19:35:45 soldered Jan 21 19:36:08 I got 2 devices with 64mb ram for essentially free :P Jan 21 19:36:28 I should poke them again sometime Jan 21 19:36:59 too many ideas and too little time Jan 21 19:37:34 wrtu54g-tm Jan 21 19:38:27 was used for t-mobile @home service. they stopped offering it. many of the overstock sellers on internet have it for $15 Jan 21 19:38:59 heh Jan 21 19:39:14 i paid my WNR854T 45EUR 2 years ago Jan 21 19:39:18 too bad only 1 work on kernel Jan 21 19:39:24 me. and i suck. Jan 21 19:39:25 FXO ports on it? Jan 21 19:39:32 yeah. Jan 21 19:39:48 5Gbe Wifi N and 32Mb Jan 21 19:39:54 sounds like a interesting device for cheap Jan 21 19:39:55 500Mhz Jan 21 19:40:13 and mini pci =) Jan 21 19:40:18 Epsylon3, I bet it only has 100BaseT tho? Jan 21 19:40:32 no Gbe = 1000 Jan 21 19:40:38 oh, nice Jan 21 19:40:39 Gigabit Ethernet Jan 21 19:40:50 i <3 my router =\ Jan 21 19:41:19 there is no usb port... Jan 21 19:41:23 I'll probably get some gige router too once I get money from my new job Jan 21 19:41:40 the right one is the WRT3500L Jan 21 19:41:45 8mb flash, 64mb ram. $15. usb port added. heh Jan 21 19:41:50 64M USB for 70$ Jan 21 19:42:17 neutronscott, what soc does it have? Jan 21 19:42:28 one day maybe i learn to use FXO Jan 21 19:42:33 doubt it Jan 21 19:42:38 adm8668 Jan 21 19:43:39 a 200mhz mip4k? Jan 21 19:43:39 http://www.myopenrouter.com/category/40/WNR3500L-Open-Source-Router Jan 21 19:43:51 yeah 4kc Jan 21 19:44:13 still.. decent for $15 :P Jan 21 19:44:19 even if the cpu is probably relatively slow Jan 21 19:44:46 it can do better Jan 21 19:44:56 seems to have optional pci and ide interfaces too Jan 21 19:45:03 i am unsure if my irq handler is at fault. but interrupts messing me up Jan 21 19:45:12 the ethernet driver is NAPI so it's nice Jan 21 19:45:31 the wifi is rt61pci Jan 21 19:45:44 some wifi cards have a parameter to limit the irq per seconds Jan 21 19:46:00 it does. Jan 21 19:46:25 rt ? realtek ? Jan 21 19:46:32 but maybe it should just poll. you know traffic is always gonna be there. Jan 21 19:46:37 Ralink Jan 21 19:46:48 oh Jan 21 19:47:18 oh yea... my tv has that :p Jan 21 19:47:27 Epsylon3: if you're buying a new router, don't buy the WNR3500L Jan 21 19:47:36 too late... why ? Jan 21 19:47:45 oops Jan 21 19:47:58 broadcom... Jan 21 19:48:04 oh Jan 21 19:48:26 any routers have FXO ports that can work in asterisks Jan 21 19:48:28 there is DDWRT on that Jan 21 19:48:36 too Jan 21 19:48:38 yeah, but it only works with proprietary drivers Jan 21 19:48:47 WNDR3700 is a better choice Jan 21 19:48:59 yea... but ive seen there are 2 models Jan 21 19:49:04 and one is not compatible Jan 21 19:49:07 or (was) Jan 21 19:49:09 neutronscott: to make voip calls ? Jan 21 19:49:11 WNDR3700 works well with openwrt Jan 21 19:49:15 mm.. what did WZR-HP-G300NH have Jan 21 19:49:19 <_trine> http://cgi.ebay.co.uk/New!!-Linksys-WRTU54G-TM-Wireless-Router-54Mbps-_W0QQitemZ110632279619QQcmdZViewItem?rvr_id=201288372767&rvr_id=201288372767&cguid=8c1ec17012d0a47a2475e305ffc5ba4a Jan 21 19:49:19 blogic: yes Jan 21 19:49:24 sort of Jan 21 19:49:32 the lantiq socs have almost working voip Jan 21 19:49:41 been tempted to get the buffalo myself Jan 21 19:49:51 the 3500L is 20€ cheaper too :p Jan 21 19:50:36 guess it was some atheros Jan 21 19:50:55 and.... Jan 21 19:51:07 the 3500L use a 12V 1.5A transformer Jan 21 19:51:17 the 3700 use a 12V 2.5A Jan 21 19:51:23 the 3700 is beasty Jan 21 19:51:26 :p Jan 21 19:51:31 it's more powerful than my laptop Jan 21 19:51:38 :D Jan 21 19:51:56 30 watts? Jan 21 19:52:18 that's what the AMD Zacate E-350 drains when playing bluray movies Jan 21 19:52:28 including system Jan 21 19:52:53 playing bluray == routing gigapackets Jan 21 19:52:58 i broke many router transformers Jan 21 19:53:08 my dlink DI 624 Jan 21 19:53:18 my linksys wrt 54g Jan 21 19:53:34 you mean from drawing too much current? Jan 21 19:53:46 both of them was 2.5A Jan 21 19:53:50 tried sticking a screwdriver into the center and wrapping it all in conductive material? Jan 21 19:53:54 5V if i remember Jan 21 19:54:32 ok, one year or two after but ;p Jan 21 19:55:08 i wonder if anybody would be interrested in in doing an openwrt port to the realtek multimedia soc Jan 21 19:55:55 i see them all the time in all the cheap hdd/usb media players.. and there are multiple sdk around (also thanks to gpl-violations.org) Jan 21 19:56:15 ^^ Jan 21 19:56:22 cheap stuff tends to be rather boring tho Jan 21 19:56:28 i dont really have the time to do it myself but i could help collecting whats needed Jan 21 19:56:32 like the first NAS I got Jan 21 19:56:34 dreadful POS Jan 21 19:56:37 all LCD tvs use linux :) Jan 21 19:56:49 Epsylon3, not all Jan 21 19:57:10 hmm yea ... all recent Jan 21 19:57:19 they decode h264 etc in fullhd with only a few watt and (most of the time) no active cooling. thats quite nice. only the default sw sucks Jan 21 19:57:30 hello Jan 21 19:57:32 in all brandings Jan 21 19:57:41 and I hear my co-workers curse at some tv's they're trying to port adobe flash to, heh Jan 21 19:58:07 never understood why Jan 21 19:58:08 :) soon android on tv :p Jan 21 19:58:09 lol Jan 21 19:58:19 to open security holes Jan 21 19:58:20 take the store down :p Jan 21 19:58:35 I just hope they never make me work on adobe flash codebase Jan 21 19:58:35 i think google already has some agenda in that direction. google-tv Jan 21 19:59:05 seems like a huge nuisance Jan 21 19:59:06 yea :) with cam sensor :) Jan 21 19:59:12 it's 400MB packed Jan 21 19:59:53 survey your neighbor.. with the wifi tv :pp Jan 21 20:00:21 I doubt he'd mind if I grabbed one, he has two Jan 21 20:00:34 the place is full of weird devices Jan 21 20:03:00 that I can't name any of ofcourse, they'd grab my salaries for that :P Jan 21 20:24:36 xMff: ping Jan 21 20:25:53 ... flashing... Jan 21 20:26:15 any committers around? Jan 21 20:26:22 var/git/openwrt/openwrt/bin/orion Jan 21 20:28:47 looking for someone to review my patches... Jan 21 20:29:35 Linux WNR854T 2.6.36.2 #4 Fri Jan 21 17:38:48 CET 2011 armv5tel GNU/Linux Jan 21 20:29:39 wooo :) Jan 21 20:30:01 and i have a green led !!! Jan 21 20:30:30 nice for a first day :) Jan 21 20:30:36 dual wan :) Jan 21 20:31:03 in fact i've two external IP here (Numericable Fiber 100Mb+30Mb) Jan 21 20:31:56 the second one on the TV decoder Jan 21 20:31:57 root@WNR854T:~# ls /sys/class/leds/ Jan 21 20:31:57 wnr854t:blink:power wnr854t:green:power Jan 21 20:32:49 blinking is working :) sweet Jan 21 20:33:06 ohhh and you can set speed with brightness :) Jan 21 20:33:25 wha Jan 21 20:34:27 so... now the wifi led :) Jan 21 20:34:37 maybe adm5120's irq code is better to use for adm8668 then what i came up with ... seems same kind of controller Jan 21 20:34:47 i've already made the fix in sources :) Jan 21 20:37:56 the https://dev.openwrt.org/ticket/8160 is working... Jan 21 20:38:02 on trunk Jan 21 20:38:06 its a patch Jan 21 20:38:23 opened 3months ago Jan 21 20:40:02 not in trunk yet still eh Jan 21 20:40:17 ? Jan 21 20:40:50 you mean linux kernel trunk ? Jan 21 20:41:31 openwrt/trunk Jan 21 20:42:24 neutronscott, only if you can reflash it outside the kernel heh Jan 21 20:42:40 yea, no it is not Jan 21 20:42:50 i had to apply the patch to the trunk Jan 21 20:43:01 openwrt Jan 21 20:43:09 its a patch with make a patch lol Jan 21 20:43:12 :) Jan 21 20:44:36 you can't flash it? Jan 21 20:47:14 ^^ i'm coding on a 32" terminal :) Jan 21 20:47:31 almost matrix with compiling with V=99 Jan 21 20:47:37 when Jan 21 20:50:39 fls( Jan 21 20:51:29 magical Jan 21 20:52:02 wwanna hear something funny Jan 21 20:52:19 my machine which i compile on... 73bogomips :( Jan 21 20:52:45 mine use 20Watts Jan 21 20:52:53 Atom 330 Jan 21 20:52:56 2 cores Jan 21 20:53:12 Giada Cube N Jan 21 20:53:19 what's that cost Jan 21 20:53:21 nvidia ION platform Jan 21 20:53:32 200 USD with the HD and ram Jan 21 20:53:41 its tiny Jan 21 20:53:49 sounds it. whats the clockrate. Jan 21 20:54:05 heuu let me see dont remember :p Jan 21 20:54:19 1600.057 Jan 21 20:54:36 2 cores and HyperThreading Jan 21 20:54:40 so 4 cores Jan 21 20:54:47 4 threads i mean Jan 21 20:54:55 thought HT never helped anything Jan 21 20:55:15 flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc arch_perfmon pebs bts aperfmperf pni dtes64 monitor ds_cpl tm2 ssse3 cx16 xtpr pdcm movbe lahf_lm dts Jan 21 20:55:15 bogomips : 3199.97 Jan 21 20:55:18 i built small system like this. back with first mini-itx board. so today it sucks.. heh Jan 21 20:55:36 cpu : 7447A, altivec supported Jan 21 20:55:36 clock : 1420.000000MHz Jan 21 20:55:46 I have bogomips : 4810.80 :P Jan 21 20:55:55 my cpu is faster at doing nothing Jan 21 20:56:06 :p Jan 21 20:56:12 hah yes bogomips are bogus Jan 21 20:56:13 no its fine really Jan 21 20:56:26 the worst thing is the 3D Card Jan 21 20:56:32 which use most of power Jan 21 20:56:33 i am certain my 73bogomip G4 is little better than my 200bogomip mips 4kc Jan 21 20:56:42 CPU is something like 5Watts Jan 21 20:56:43 pretty sure ion+atom use more than 20watts tho Jan 21 20:57:18 the power is 12V 5A yea Jan 21 20:57:37 but when i use it (Gnome loaded) its 20W Jan 21 20:57:43 this cube is $180 on newegg without ram+hdd Jan 21 20:57:57 I'll probably buy a panda board or a E-350 soon as a secondary Jan 21 20:58:05 euh let me check, i ve the Black one Jan 21 20:58:10 the panda board is tempting but rather expensive Jan 21 20:58:17 not the one like a router Jan 21 20:58:47 http://pandaboard.org/ Jan 21 20:58:52 yea Giada Cube N3 Jan 21 20:59:35 i dont like the style Jan 21 21:00:05 http://www.xbmcboxes.com/wordpress/wp-content/uploads/2010/11/giada-cube-n3-back.jpg Jan 21 21:00:08 rear side Jan 21 21:00:33 there is also a SD card usb and audio on the left Jan 21 21:00:45 and 2 other usb 2 on the front Jan 21 21:00:54 oh Jan 21 21:00:58 this one doesnt have wifi Jan 21 21:01:07 bad picture :p Jan 21 21:01:25 Foxconn NT-330I-A-B-NA-A Jan 21 21:01:27 looks neat Jan 21 21:02:27 this with a SSD 60GB could be great :) Jan 21 21:02:44 in fact i love the CPU Jan 21 21:02:52 but i dont like the GPU (ION) Jan 21 21:02:59 55°c Jan 21 21:03:08 cpu doesnt need fan Jan 21 21:03:15 only the gpu Jan 21 21:05:47 i dont needgood gpu Jan 21 21:05:50 i dont game Jan 21 21:05:52 i dont watch movie Jan 21 21:05:57 i just need compile Jan 21 21:05:59 :D Jan 21 21:06:38 then don't get a ion device Jan 21 21:06:54 ion adds to the price quite a chunk Jan 21 21:09:41 is only $200. Jan 21 21:09:55 where is a dual core atom without gpu? Jan 21 21:10:08 mini Jan 21 21:10:11 the alternative to ion is intel chipsets Jan 21 21:10:23 they just have the intel gpu integrated Jan 21 21:15:28 there is a new intel SoC Jan 21 21:15:42 here (in france) we have a new fiber box with that Jan 21 21:17:04 http://www.natchers.com/wp-content/uploads/2010/12/Freebox_Revolution_03.jpg Jan 21 21:17:25 hmmm Jan 21 21:17:32 its wrote Atom 1.2ghz Jan 21 21:17:34 but... Jan 21 21:18:02 yea its a Intel Atom CE4100. Jan 21 21:18:37 http://www.natchers.com/actualite-2010/13976/free-lance-la-freebox-revolution/ Jan 21 21:18:42 (sorry french) Jan 21 21:19:08 fiber + Dsl Jan 21 21:19:13 Gigabit, NAS Jan 21 21:19:16 3D Jan 21 21:19:31 esata etc... Jan 21 21:20:13 ah and blueray :) Jan 21 21:20:27 25W :) Jan 21 21:20:43 12V 2A Jan 21 21:21:06 and DVBT also Jan 21 21:23:00 and sound on the box lol Jan 21 21:23:53 la revolution! Jan 21 21:24:10 yea, the remote has a 3Axis accelerometer Jan 21 21:24:17 why are some libraries, i.e. libpt.a, libopal.a, etc., are not installed under staging_dir/target-ARCH/usr/lib directory? Jan 21 21:24:17 and giro... Jan 21 21:24:37 Epsylon3, sounds like a mess Jan 21 21:25:17 rather like the ... "mobile phone" I've seen Jan 21 21:25:28 it's 3x the size of a normal one Jan 21 21:25:36 but it has "everything" Jan 21 21:25:37 7" ? Jan 21 21:25:47 dunno.. Big Jan 21 21:25:58 ^^ its an ipad ? :p Jan 21 21:26:10 no... it looks like a candybar mobile phone scaled up Jan 21 21:26:19 brick Jan 21 21:26:34 friggen amusing device Jan 21 21:26:56 but it has virtually everything you could imagine of a ""mobile"" device Jan 21 21:27:05 ok hmmm... i need to make a patch of a patch hmmm Jan 21 21:27:35 is there some commands to speedup in git repository ? Jan 21 21:27:46 huh? Jan 21 21:27:59 ive seen in menuconfig Jan 21 21:28:06 external linux git Jan 21 21:28:16 git only supports cloning full trees Jan 21 21:28:23 is there something to get patches from the linux tree Jan 21 21:28:23 and working ontop of those Jan 21 21:28:34 well... yes Jan 21 21:28:42 i mean... Jan 21 21:28:53 i want to change source of my platform Jan 21 21:28:56 to add missing leds Jan 21 21:29:19 to get the patch of what i've modified in kernel tree Jan 21 21:29:23 well, since you're working on openwrt, you can add it to the openwrt platform aptches Jan 21 21:29:26 err.. patches Jan 21 21:29:38 yea, ive seen that Jan 21 21:29:44 but, the kernel tree Jan 21 21:29:52 where is it from ? Jan 21 21:29:56 those were in target iirc Jan 21 21:30:17 it was on Svn ? or its the make which have downloaded it Jan 21 21:30:21 on git Jan 21 21:30:35 say WHAT Jan 21 21:30:47 the "make" process has downloaded some sources Jan 21 21:30:57 but that take 3 hours to compile :p Jan 21 21:31:43 in fact i can see what is changed on openwrt Jan 21 21:31:56 but that are patches on patch files Jan 21 21:32:02 so i cant modify them Jan 21 21:32:06 there are line references Jan 21 21:32:38 you want to modify kernel in your build_dir Jan 21 21:32:48 yea Jan 21 21:33:03 git/openwrt/openwrt/build_dir/linux-orion_generic/linux-2.6.36.2/arch/arm/mach-orion5x Jan 21 21:33:09 !! Jan 21 21:33:35 hmm right i think i know Jan 21 21:33:43 euh Jan 21 21:33:57 you have modified it and want to make a patch for target/linux/orion/patches/ ? Jan 21 21:34:03 yes Jan 21 21:34:18 should have made a copy of it unpatched first :) Jan 21 21:34:32 i'm seeing that Jan 21 21:34:39 what happens at the second "make" Jan 21 21:34:44 i just use diff -ru Jan 21 21:34:55 there is no new patches applied ? Jan 21 21:35:02 it will not patch again Jan 21 21:35:06 ooohh Jan 21 21:35:08 ok :) Jan 21 21:35:12 right i understand Jan 21 21:35:23 because i tried to change .config Jan 21 21:35:33 and was replaced at two levels :p Jan 21 21:35:44 you can do make kernel_menuconfig Jan 21 21:35:44 CRYPTO_HASH was missing to build Jan 21 21:35:52 :) Jan 21 21:36:15 there is also CRYPTO_HASH2 now Jan 21 21:36:23 dunno exactly what it is Jan 21 21:36:49 i think its for the Marvell hardware CRC Jan 21 21:37:14 is there some levels between 99 and nothing ? Jan 21 21:37:23 to see normal kernel generation Jan 21 21:37:33 not the matrix screen :)) Jan 21 21:37:42 i'm sure they are 98 levels between 99 and 0 Jan 21 21:37:54 but i am unsure which are defined if any :) Jan 21 21:38:10 yea ;) its why i ask :) Jan 21 21:39:06 so... i can create a git on kernel tere Jan 21 21:39:07 tree Jan 21 21:39:24 i look at include/verbose.mk is only 99 or not Jan 21 21:39:27 to make patches easily :) Jan 21 21:40:00 * neutronscott shrugs. Jan 21 21:40:17 after Windows.... you have "Screens" Jan 21 21:40:23 i never learn git Jan 21 21:40:29 i've a terminal per LCD :p Jan 21 21:41:03 git init Jan 21 21:41:11 git commit -a Jan 21 21:41:16 and finished Jan 21 21:41:19 build #66 of orion is complete: Failure [failed shell] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/66 Jan 21 21:41:22 i use svn Jan 21 21:41:26 after that... git diff Jan 21 21:41:44 git doesnt need server Jan 21 21:42:00 you commit locally Jan 21 21:42:52 that'd be nice. since i have no access Jan 21 21:43:11 or git status Jan 21 21:43:13 like svn Jan 21 21:43:25 the big difference Jan 21 21:43:34 in git you have only one ".git" folder Jan 21 21:43:38 at the root folder Jan 21 21:43:47 so you can copy tree Jan 21 21:44:00 without all the .svn dirs Jan 21 21:44:11 both are usefull Jan 21 21:44:29 but git can have both Jan 21 21:44:36 not svn Jan 21 21:44:51 git-svn package Jan 21 21:45:28 and branches are not in different directories Jan 21 21:45:41 there are packed in .git database Jan 21 21:45:49 the diffs in fact Jan 21 21:48:38 ok make[1] world .... better Jan 21 21:55:15 KAMIKAZE (bleeding edge, r25056) ------------------ Jan 21 21:57:02 wow trunk modules are working too :) Jan 21 21:57:50 luci is the best web interface ? Jan 21 21:59:48 Epsylon3: yes it is the best Jan 21 21:59:58 :) Jan 21 22:00:13 i'm thinking about including it in my image directly Jan 21 22:01:13 ouch no segfault :p Jan 21 22:08:21 luci > xwrt? Jan 21 22:13:12 depends Jan 21 22:16:55 ive not used it. just installed it. let's see what it do Jan 21 22:17:15 hmm. a blank screen Jan 21 22:26:06 /usr/bin/lua: /usr/lib/lua/luci/sys.lua:33: loop or previous error loading module 'luci.model.uci' Jan 21 22:26:20 what did you install? Jan 21 22:26:26 I mean which package Jan 21 22:27:14 opkg install luci Jan 21 22:27:21 must be a depend issue. Jan 21 22:27:37 http://downloads.openwrt.org/snapshots/trunk/brcm47xx/packages/Packages.gz Jan 21 22:28:11 odd Jan 21 22:31:00 i dont understand lua. Jan 21 22:31:14 but it has a stack trace Jan 21 22:31:15 maybe it's a umask issue Jan 21 22:32:12 try this: lua -lluci.model.uci -e true Jan 21 22:32:37 lua: error loading module 'uci' from file '/usr/lib/lua/uci.so': File not found Jan 21 22:32:58 oh :) Jan 21 22:33:06 liblua-uci installed ? Jan 21 22:33:11 mmhmm Jan 21 22:33:48 uci.so: ELF 32-bit LSB shared object, MIPS, MIPS32 version 1 (SYSV), dynamically linked, corrupted section header size Jan 21 22:34:10 ? Jan 21 22:34:12 well /usr/lib/uci.so != /usr/lib/lua/uci.so Jan 21 22:34:23 one is the library, the other the lua binding Jan 21 22:34:51 i have no /usr/lib/uci.so Jan 21 22:35:06 thats quite uncommon Jan 21 22:35:30 liblua.so.5.1.4usr/lib/ Jan 21 22:36:02 yeah, well ok. but do you have a /usr/lib/*lua/*uci.so ? Jan 21 22:36:25 # echo /usr/lib/*lua/*uci.so Jan 21 22:36:25 /usr/lib/lua/uci.so Jan 21 22:36:53 ok. lua -luci Jan 21 22:36:58 still throwing errors? Jan 21 22:37:06 maybe i shall just compile from source :) Jan 21 22:37:44 I think its a uci 2010-09-28 vs 2011-01-19 issue Jan 21 22:38:40 luci - 0.10+svn6778-1 Jan 21 22:39:38 no, I mean this https://dev.openwrt.org/changeset/25048 Jan 21 22:40:39 libuci - 2010-09-28.2-2 Jan 21 22:40:39 libuci-lua - 2011-01-19.1-1 Jan 21 22:40:52 yes Jan 21 22:40:54 could also be the binaries weren't made for my target? same cpu type Jan 21 22:41:03 heh Jan 21 22:41:32 no, libuci-lua looks for uci.so.2011-01-19 and you have still uci.so.2010-09-28 Jan 21 22:41:55 you could just symlink it, the abi should be unchanged Jan 21 22:42:48 Upgrading libuci on root from 2010-09-28.2-2 to 2011-01-19.1-1... Jan 21 22:42:58 root@OpenWrt:/www/cgi-bin# ./luci Jan 21 22:42:58 /usr/lib/lua/luci/dispatcher.lua:445: Fatal: Indexcache is not sane! Jan 21 22:43:03 success! Jan 21 22:43:06 heh Jan 21 22:43:13 heh Jan 21 22:43:20 rm -r /tmp/luci-* Jan 21 22:44:10 2 days ago Jan 21 22:44:16 yeah guess i need to svn up Jan 21 22:45:33 thats nice. ask me to login AND tell me no password is set! :D Jan 21 22:46:59 well, stick to the procedures, bureaucracy and all Jan 21 22:47:01 this is much different than luci i remember Jan 21 22:48:02 guess my cpu is very slow.. or i've kernel issues.. Jan 21 22:48:14 whats your device? Jan 21 22:48:38 I use a dir-300 as reference for slowness here Jan 21 22:48:57 i'm the guy trying to get wrtu54g-tm working good :D Jan 21 22:49:23 ah I recall Jan 21 22:49:41 florian did inital commit and gave me a to-do list.. Jan 21 22:50:00 i got stuck on wifi issue, but maybe it is kernel timing issue.. Jan 21 22:50:25 it never felt this slow on 2.4 kernel :( Jan 21 22:51:29 are you able to commit my two patches on list? :) Jan 21 22:51:53 which ones? Jan 21 22:52:01 actually, nevermind the clean-up patch. i'm reworking includes directory after this Jan 21 22:52:44 said something like 'adm8668 target: fix firstboot with new mtd driver' Jan 21 22:58:04 is OK idea? i couldnt find how other machines handle this. Jan 21 22:58:14 er, other u-boot platforms. Jan 21 23:00:34 xMff: what is luci2 Jan 21 23:00:35 ? Jan 21 23:01:17 https://forum.openwrt.org/viewtopic.php?id=25988 Jan 21 23:02:22 blogic: ^ Jan 21 23:03:35 ? Jan 21 23:04:15 you know better what luci2 is Jan 21 23:06:20 luci2 is a collection of libraries and tools Jan 21 23:06:25 we made them as a prototype Jan 21 23:09:49 neutronscott: I cannot judge the code Jan 21 23:10:09 neutronscott: I think you should wait for florian or nbd to ack it Jan 21 23:10:55 * nbd doesn't have time for reviewing platform code at the moment Jan 21 23:10:58 * nbd is working on ubus Jan 21 23:12:25 it's better than the last and no one uses it yet but me so *shrugs* :D Jan 21 23:19:19 wow... the big segfault :p Jan 21 23:19:50 i had to remove the wifi card to connect and change opkg repository to my box :p Jan 21 23:20:15 and cant build without download ! Jan 21 23:23:10 so... second firmware soon... and maybe wifi led :) Jan 21 23:24:52 ^^ just realized the leds are blinking... too fast to be seen Jan 21 23:25:01 on activity Jan 21 23:27:23 i think all packages with kernel dependencies should be separated in a sub package in repository Jan 21 23:27:35 i mean kmodules Jan 21 23:28:10 because when you build your own kernel... (for platform leds) Jan 21 23:28:43 you have segfaults with repo modules Jan 21 23:29:01 some are loaded automatically by dependencies Jan 21 23:33:52 hmm ccache could be useful in toolchaine :/ Jan 21 23:34:04 its slooow Jan 21 23:40:44 how'd u get mixed modules tho Jan 21 23:40:53 not replacing filesystem? Jan 21 23:41:00 tftpboot kernel? Jan 21 23:42:02 nfsroot :) Jan 21 23:42:31 i need to do that again. i'm sure my flash has seen an abnormal amount of erase cycles.. Jan 21 23:44:35 yea :p Jan 21 23:44:46 and no sdcard Jan 21 23:45:25 to compile external packages... Jan 21 23:45:28 like luci Jan 21 23:45:45 i need to make links ? Jan 21 23:45:46 i should add usb support to u-boot Jan 21 23:46:07 ln -s ../../git_packages/net/multiwan Jan 21 23:46:36 yea :) i made that for the SmartQ Jan 21 23:46:40 a mini tabled Jan 21 23:46:49 qi and uboot Jan 21 23:46:58 tablet Jan 21 23:48:00 qi? Jan 21 23:48:19 http://tanguy.wdscript.fr/smartq5 Jan 21 23:48:25 euh for qi... Jan 21 23:49:17 https://github.com/tpruvot/smartq-magicsd/tree/qi-smartq Jan 21 23:50:39 https://github.com/tpruvot/smartq-magicsd/blob/2.6.24-rt/qi/README Jan 21 23:50:56 Qi (named by Alan Cox on Openmoko kernel list) is a minimal bootloader that Jan 21 23:50:56 "breathes life" into Linux. Its goal is to stay close to the minimum needed Jan 21 23:50:56 to "load" and then "boot" Linux -- no boot menus, additional peripheral init Jan 21 23:50:56 or private states. Jan 21 23:51:11 qu'est-ce que c'est Jan 21 23:51:21 a truncated uboot Jan 21 23:51:23 :p Jan 21 23:51:33 very small Jan 21 23:51:44 for arm Jan 21 23:51:48 cannot linux itself load itself Jan 21 23:52:16 i dont know arm.. but MIPS it does all the init needed it seems Jan 21 23:52:38 its almost the same Jan 21 23:52:42 i bet if i burned it to rom it'd load heh Jan 21 23:52:59 its a bios in fact ;) Jan 21 23:53:51 which allocate memory, set cpu registers Jan 21 23:53:54 gpios Jan 21 23:54:06 and jump to a ram address Jan 21 23:54:13 on ext2 Jan 21 23:54:15 or Jan 21 23:54:45 its led based... Jan 21 23:54:50 good for routers Jan 21 23:59:24 CPU: 0% usr 0% sys 0% nic 16% idle 0% io 0% irq 82% sirq Jan 22 00:12:27 erf, get a "Not Found" in luci Jan 22 00:12:42 i just want to install my image : p Jan 22 00:13:35 sysupgrade / mtd Jan 22 00:15:14 hmm yea Jan 22 00:15:41 -c Jan 22 00:17:05 mtd ? Jan 22 00:17:14 dont understand this command Jan 22 00:17:35 Saving config files... Jan 22 00:17:35 Switching to ramdisk... Jan 22 00:17:35 mount: mounting mini_fo:/overlay on /mnt failed: Function not implemented Jan 22 00:17:35 Performing system upgrade... Jan 22 00:17:35 Unlocking image ... Jan 22 00:17:35 Writing from to image ... [e] Jan 22 00:20:27 oops Jan 22 00:22:42 ^^ :) thanks Jan 22 00:22:58 ok, this time i've the wifi directly :) Jan 22 00:25:13 ok hmm added the wan led in fact :p Jan 22 00:37:27 fine :) Jan 22 00:37:34 that works :) Jan 22 00:37:42 Hauke: you around? Jan 22 00:38:04 philipp64|laptop: yes Jan 22 00:38:40 I'm wondering how much of the Wifi (ath9k) hardware out there today would support multiple SSID's. Jan 22 00:39:34 I'd like to see OpenWRT have the option of configuring a "regular" internal network (or bridged to the wired internal network), and a "guest" network with its own set of policies and firewall rules. Jan 22 00:39:45 all ath9k hardware supports multiple SSIDs Jan 22 00:39:46 is that feasible? Jan 22 00:40:09 you can create multiple wifi interfaces (up to 4) and use them in whatever way you want Jan 22 00:40:18 with different network settings, different firewall settings, etc. Jan 22 00:40:28 yea :) ive made that Jan 22 00:40:39 in fact 2 (on same channel) Jan 22 00:40:48 its one of my 2 wan Jan 22 00:40:59 so we'd just need to munge the configuration scripts to handle seeing multiple instances of wifi-device? Jan 22 00:41:18 Epyslon3: got patches you could submit? Jan 22 00:41:25 ive a ath5k Jan 22 00:41:32 no patches needed Jan 22 00:41:46 just yea Jan 22 00:41:50 that would work too... should work for any phy that supports mac80211, right? Jan 22 00:41:51 dont set it as dhcp Jan 22 00:42:02 else that change default gateway Jan 22 00:42:03 that's a bit of a limitation. Jan 22 00:42:20 with multiwan package Jan 22 00:42:25 that can help Jan 22 00:42:32 (i think) Jan 22 00:42:37 i've made it manually Jan 22 00:42:48 to do the same on another router later Jan 22 00:42:55 on the other side Jan 22 00:43:15 maybe with WDS Jan 22 00:43:37 if you want my config Jan 22 00:44:41 config 'wifi-iface' Jan 22 00:44:41 option 'device' 'wifi0' Jan 22 00:44:41 option 'mode' 'ap' Jan 22 00:44:41 option 'ssid' 'MimoGigabit' Jan 22 00:44:41 option 'encryption' 'psk-mixed' Jan 22 00:44:41 option 'key' 'xxxxxxxxxxx' Jan 22 00:44:41 option 'network' 'lan' Jan 22 00:44:42 config 'wifi-iface' Jan 22 00:44:42 option 'device' 'wifi0' Jan 22 00:44:43 option 'ssid' 'CABLE' Jan 22 00:44:43 option 'mode' 'sta' Jan 22 00:44:44 option 'network' 'salon' Jan 22 00:44:44 option 'encryption' 'psk' Jan 22 00:44:45 option 'key' 'yyyyyyyyyyy' Jan 22 00:45:38 philipp64 : ohh yea Jan 22 00:45:42 i use madwifi Jan 22 00:45:56 ath5k is buggy here Jan 22 00:46:02 to much modules Jan 22 00:47:14 --- Jan 22 00:47:18 config 'interface' 'salon' Jan 22 00:47:18 option 'proto' 'static' Jan 22 00:47:18 option 'ipaddr' '192.168.0.2' Jan 22 00:47:18 option 'netmask' '255.255.255.0' Jan 22 00:47:18 option 'broadcast' '192.168.0.255' Jan 22 00:47:19 option 'peerdns' '0' Jan 22 00:47:19 config 'route' Jan 22 00:47:20 option 'target' '192.168.0.1' Jan 22 00:47:20 option 'netmask' '255.255.255.0' Jan 22 00:47:21 option 'gateway' '192.168.0.1' Jan 22 00:47:21 option 'metric' '2' Jan 22 00:47:22 option 'interface' 'salon' Jan 22 00:47:40 pastebin Jan 22 00:47:46 yea sorry Jan 22 00:48:05 dunno for the metric, but 3 is slow **** ENDING LOGGING AT Sat Jan 22 00:48:05 2011 **** BEGIN LOGGING AT Sat Jan 22 00:52:36 2011 Jan 22 00:52:58 my default gateway is set on "wan" Jan 22 00:53:16 option 'defaultroute' '1' Jan 22 00:53:26 I was referring to: Jan 22 00:53:26 (01:38:13) Epsylon3: dont set it as dhcp Jan 22 00:54:30 (01:38:24) Epsylon3: else that change default gateway Jan 22 00:54:30 ive dhcp on both routeurs Jan 22 00:54:30 servers Jan 22 00:54:30 and yea, only my wan is dhcpp Jan 22 00:54:51 oh Jan 22 00:55:11 so the metric is for dhcp ? Jan 22 00:55:23 in fact i dont really understand metrics Jan 22 00:55:26 no in trunk you can have metric for interfaces Jan 22 00:55:46 it wil lget applied to any defaultroute created, be it due to dhcp or pppd Jan 22 00:56:07 so you could have two interfaces set to dhcp and make one use a higher metric Jan 22 00:56:23 then you have two default routes but with different weight, so they don't conflict Jan 22 00:56:56 ok, the less is the good one ? Jan 22 00:57:05 yes Jan 22 00:57:12 :) ok i understand Jan 22 00:57:14 smaller = more prio Jan 22 00:57:57 so, i will make a patch tomorrow for the wnr854t leds... Jan 22 00:58:07 oh, and this reset button.. need to check :) Jan 22 00:59:03 test Jan 22 00:59:37 Destination Gateway Genmask Flags Metric Ref Use Iface Jan 22 00:59:37 192.168.1.0 * 255.255.255.0 U 0 0 0 br-lan Jan 22 00:59:37 192.168.0.0 * 255.255.255.0 U 0 0 0 ath0 Jan 22 00:59:37 85.170.20.0 * 255.255.252.0 U 0 0 0 wan Jan 22 00:59:37 default 85-170-20-1.rev 0.0.0.0 UG 0 0 0 wan Jan 22 00:59:48 metric are ignored Jan 22 00:59:58 maybe a busybox bug Jan 22 00:59:58 how did you set them? Jan 22 01:00:07 option 'metric' '2' Jan 22 01:00:13 in wan? Jan 22 01:00:22 oh.... on a static route Jan 22 01:00:33 :/ Jan 22 01:00:54 the one to the second wifi Jan 22 01:01:42 yea ok, so i need to set it on interface ? Jan 22 01:01:51 yes Jan 22 01:02:57 route: netmask and route address conflict Jan 22 01:03:04 :/ Jan 22 01:04:13 not related to the metric i think Jan 22 01:04:25 wth u doin! Jan 22 01:05:14 i break my stable config :p Jan 22 01:05:30 (fully tested since.... yesterday :p) Jan 22 01:07:17 build #61 of ps3 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/61 Jan 22 01:09:09 reminds me. wonder what nifty things came of that ps3 key leak Jan 22 01:10:00 next week i will connect with serial to my isp netgear CBVG835G =) Jan 22 01:10:08 834G Jan 22 01:10:24 100Mb cable modem **** BEGIN LOGGING AT Sat Jan 22 04:42:16 2011 Jan 22 11:05:42 what is the param : list 'ht_capab' 'DSSS_CCK-40' Jan 22 11:05:52 in wifi-device **** ENDING LOGGING AT Sat Jan 22 11:17:40 2011 **** BEGIN LOGGING AT Sat Jan 22 11:22:23 2011 Jan 22 11:28:30 Kernel Execute-In-Place from ROM Jan 22 11:28:34 hmm Jan 22 12:05:13 ergh, cant find datasheet of the Marvell 88F5181 Jan 22 12:35:55 Epsylon3: unfortunately I don't know how much they differ, but you can find datasheets for the 5182 and the 5281 - they might be similar enough for most purposes Jan 22 12:37:13 yea, i was searching the 5180 too Jan 22 12:37:59 found :) http://www.embeddedarm.com/documentation/third-party/MV88F5182-datasheet.pdf Jan 22 12:38:05 5182 Jan 22 13:05:30 ok, third firmware :) i've enabled kexec support :) Jan 22 13:13:02 :) Jan 22 13:55:57 hi. i just subscribed to openwrt-devel but i can't send and emails to it, i only get back: Jan 22 13:56:10 header: "the results of your email command" Jan 22 13:56:21 - Results: Ignoring non-text/plain MIME parts Jan 22 13:56:52 any hints what i do wrong there? its a normal text message with a patch in text format attached Jan 22 13:59:07 maybe some administrative keyword in the subject Jan 22 14:00:00 i don't think so, header is "Madwifi country code patch" Jan 22 14:02:28 soma you got the activate mail and clicked the link in it ? Jan 22 14:02:33 ah, i know what the problem is, stupid me :) Jan 22 14:02:46 sent to openwrt-devel-request. sorry Jan 22 14:02:51 :D Jan 22 14:02:53 heh Jan 22 14:03:39 now it works :D Jan 22 14:05:37 soma: there's a bit of whitespace noise in it, this shouldn't be part of a patch unless it's a patch to fix code style ;) Jan 22 14:07:20 ok, if people think that patch is worth including then i'll fix that Jan 22 14:11:32 I think it's a good idea as it reduces confusion, but I'm not the one responsible for the madwifi package ;) Jan 22 14:12:56 also makes automatic setups easier. no need to find out first if we are using madwifi or brcm (i use that patch already for my imagebuilder since a while) Jan 22 14:13:11 : connect on list website Jan 22 14:13:19 there is an option for Mime or Text Jan 22 14:14:55 soma: it isn't just broadcom; mac80211 drivers use them too Jan 22 14:15:15 they use two-letter-style, right? Jan 22 14:15:30 Epsylon3: thats only for receiving digests as far as i can see Jan 22 14:16:02 iso-something country codes Jan 22 14:16:43 ah, its all three iso ;) Jan 22 14:18:04 iso3166 it seems, and thats two-letter, yes Jan 22 14:19:13 yes Jan 22 14:19:19 you need to use iso3166 Jan 22 14:19:53 the 11D ies use the iso3166 2 char id i think Jan 22 14:19:59 or i am pretty sure that is the case Jan 22 14:20:56 but documenation still says the option "country" is only valid for brcm and madwifi Jan 22 14:21:29 ah ok, in trunk there is also mac80211 Jan 22 14:43:34 What is the OpenWRT's variable for target OS? Jan 22 15:19:35 nunojpg * r25062 /packages/net/nmap/Makefile: [packages] nmap: update to version 5.36TEST4 Jan 22 15:29:06 make finishes even if the image goes over the flash size and the device images are not created Jan 22 15:29:20 I think earlier this was giving a error Jan 22 15:29:25 is this the intended behaviour? Jan 22 15:43:17 . Jan 22 18:19:05 axel * r25063 /packages/net/bmxd/Makefile: [packages] bmxd: move to git repository at bmx6.net Jan 22 18:28:27 could somebody apply the patch of issus 8160 Jan 22 18:28:30 issue Jan 22 18:28:58 ... i want to submit another one after that on same files Jan 22 19:01:26 any committers around? Jan 22 19:05:57 why ? Jan 22 19:08:01 re Jan 22 19:08:56 was hoping to get something reviewed and committed... Jan 22 19:09:07 ok, didnt find the wifi led gpio, i think it is maybe connected to the mini pci card... but wan and power leds are working (dual color) Jan 22 19:11:39 on the wnr854T Jan 22 19:12:05 blogic: you're a committer? Jan 22 19:14:20 i have commit access if that is what you mean Jan 22 19:14:27 Epyslon3: dual color? fancy! :-) Jan 22 19:14:51 blogic: yes. Just resent to patches. they're fairly straightforward. Jan 22 19:14:58 *two Jan 22 19:15:33 what is geos ` Jan 22 19:15:34 ? Jan 22 19:16:47 and can i change secret to password :) Jan 22 19:17:00 set network.lan.ipaddr="192.168.0.1" Jan 22 19:17:12 that should be 1.1 imho to make it consistent with the docs Jan 22 19:17:23 as in flash, telnet 192.168.1.1 Jan 22 19:17:31 that is whatr most of the docu uses Jan 22 19:17:45 http://www.traverse.com.au/productview.php?product_id=117 Jan 22 19:18:05 oh, sure, change it to 1.1. Jan 22 19:19:15 In a nutshell, Geos is an SBC with Geode 800 MHz processor, mini-PCI slot, 2 ADSL2+ coprocessors, CF slot, 2 Gbit Ethernet ports, USB, GPIO, etc. Jan 22 19:19:53 Oh, also IDE... and serial console. Jan 22 19:20:26 ah nice Jan 22 19:20:33 yeah, gimme a few minutes Jan 22 19:20:36 scratch that, they aren't Gig ethernet... Just Fast ethernet. Jan 22 19:22:08 they're about $420 USD (with case, p/s, .... but no Wireless, CF or DRAM). Jan 22 19:22:39 looks nice Jan 22 19:22:53 if your isp can do bonding ... Jan 22 19:23:13 and that is the solos dsl stuff ? Jan 22 19:23:25 yeah Jan 22 19:23:26 that reminds me.... let's see... do I select mppp? Jan 22 19:23:36 yeah, it's the Solos processor. Jan 22 19:23:47 mpp for gpio ? Jan 22 19:24:09 oddly, there aren't a lot of ADSL chipsets out there... probably because everyone has gone to SoC + Ethernet for a "bridge in a box" design. Jan 22 19:24:19 no, multilink ppp. Jan 22 19:24:20 there is lantiq and brcm essentially now Jan 22 19:24:23 for bonding. Jan 22 19:24:51 and then a few small ones that do standalone modem chips Jan 22 19:25:15 the lantiq dsl driver is already open and works in openwrt Jan 22 19:25:19 brcm does not however Jan 22 19:25:43 and we should statrt a vci/vpi index in the wiki Jan 22 19:25:53 so people know what to use in different countires Jan 22 19:28:22 does anyone make PCI cards besides Traverse? Jan 22 19:28:36 I have a couple of really old 3Com cards... Jan 22 19:33:19 oh, while I remember, linux-atm-2.5.2 has been out about a month... Jan 22 19:33:30 digging through the patches to see which ones go away. Jan 22 19:35:07 . Jan 22 19:35:30 atm ? its a realtime kernel ? Jan 22 19:36:06 asynchronous transfer mode Jan 22 19:36:09 dsl etc. Jan 22 19:36:49 yea i know ATM :) Jan 22 19:36:49 hmmm.... need to get it from sourceforge, which means munging the URL... Jan 22 19:36:54 but linux-atm Jan 22 19:37:20 is any modern network interface synchronous? Jan 22 19:37:22 its a linux stack to implement the atm protocol Jan 22 19:37:55 actually, I haven't a clue about ethernet, guess it might be Jan 22 19:37:58 oh ok Jan 22 19:38:02 a module Jan 22 19:38:09 yes Jan 22 19:38:44 it's because ive seen it near the linux-2.6.32 folder ... somewhere Jan 22 19:40:27 rmmm i dunno how to link the gpio_key to reset function :p Jan 22 19:40:36 gpio key is working Jan 22 19:40:48 in /sys/kernel/debug/gpio Jan 22 19:42:09 GPIOs 0-31, orion_gpio: Jan 22 19:42:09 gpio-0 (wnr854t:green:power ) out lo Jan 22 19:42:09 gpio-1 (Reset Button ) in hi irq-33 edge-both Jan 22 19:42:09 gpio-2 (wnr854t:blink:power ) out hi Jan 22 19:42:09 gpio-3 (wnr854t:amber:wan ) out lo Jan 22 19:43:00 is there a macro for demunging SF URL's? Jan 22 19:43:10 http://sourceforge.net/projects/linux-atm/files/linux-atm/2.5.2/linux-atm-2.5.2.tar.gz/download Jan 22 19:43:24 wget Jan 22 19:43:26 works Jan 22 19:43:33 note that the file name is not the actual `basename` of the URL. Jan 22 19:43:38 yeah Jan 22 19:43:45 @SF/linux-atm Jan 22 19:52:38 patches need tweaking... Jan 22 19:52:52 for the linux-atm 2.5.2 version bump, I mean. Jan 22 19:53:00 blogic: how do my two patches look otherwise? Jan 22 19:53:10 (other than using the .0 net, I mean.) Jan 22 19:53:58 xMff: would it make sense to include a "salon" network in the config, but disable the wifi-iface? Jan 22 19:56:36 and why do we patch linux-atm to not link with the flex library? Jan 22 19:59:05 because it is apparently not needed judging from the patch Jan 22 20:00:57 how do I update the patch if the line numbers change, etc? Jan 22 20:01:52 yeah, the flex default implementation of yywrap is "return 1" so it makes a whole lot of sense to not link it Jan 22 20:03:14 make package/refresh V=99 Jan 22 20:03:27 make package/foo/refresh V=99 Jan 22 20:05:51 philipp64|laptop: doing a svn chekout now to make a test tree Jan 22 20:06:33 in the same way that we have files that get included into the final image with "root/files", would it make sense to have a .config.user that gets merged with everything? what I mean is, I always like to have "iftop" and "iproute2" on all of my images, so it would be nice to have something that gets sourced before "make menuconfig" gets run and flips things on... Jan 22 20:06:55 did you just commit? Jan 22 20:07:09 it would appear here Jan 22 20:07:45 eh? Jan 22 20:08:14 sorry, what would appear where... what are we talking about? Jan 22 20:08:27 CIA-77 posts all commits to this channel Jan 22 20:11:08 oh, blogic is testing my patch. got it. Jan 22 20:11:59 I sent Karl and Guy and others the link to the 2 linux-atm patches in Openwrt to see if they want to cherry-pick anything for 2.5.3. Jan 22 20:37:38 ok, so no one likes the .config.user idea... Jan 22 20:38:54 there is something like that but I forgot how it is used Jan 22 20:41:25 ah right Jan 22 20:41:30 $HOME/.openwrt/defconfig Jan 22 20:41:39 this one is always sourced Jan 22 20:47:21 ah, cool. Jan 22 20:50:12 except that there is no iproute2 package. sigh. Jan 22 20:50:38 oh, yes there is. weird. why isn't it showing up in my .config... Jan 22 20:50:48 it is called "ip" Jan 22 20:50:53 as symbol Jan 22 20:51:16 CONFIG_PACKAGE_ip=y Jan 22 20:54:43 say, we're building a very old version of br2684... we should be building the one in the linux-atm tarball. Jan 22 20:56:44 is there an easy way to modify br2674 to be part of the linux-atm package? Jan 22 21:02:29 arf :/ pci bus is not working :/ Jan 22 21:04:44 mirko * r25064 /packages/utils/gpiotoggling/ (. Makefile): [packages/gpiotoggling] add package which allows toggling GPIO ranges in kernel space, controllable via sysfs Jan 22 21:04:58 userspace ? Jan 22 21:05:02 :) Jan 22 21:05:15 that use /dev/gpio ? Jan 22 21:05:28 because i dont have /dev/gpio Jan 22 21:06:50 Epsylon3: /sys/gpiotoggling Jan 22 21:08:06 will take a look thanks Jan 22 21:09:15 it's hacked however works let you measure maximumm toggle speed of a range of gpios Jan 22 21:09:52 wrote it to gather information which gpio's are which Jan 22 21:11:37 Epsylon3: root@OpenWrt:~# ls /sys/gpiotoggling/ Jan 22 21:11:38 count fire interval loop start stop verbosity Jan 22 21:12:16 CIA-77 : where is this package ? Jan 22 21:12:20 i dont have it in git Jan 22 21:12:30 Epsylon3: then pull it Jan 22 21:12:38 oh :) a new one :) Jan 22 21:12:52 yea, got it Jan 22 21:13:05 that's why it says "add" in the commit-msg Jan 22 21:13:26 just to clarify, CIA-77 is a bot and this package has nothing to do with the pci bus Jan 22 21:13:50 xMff: which pci bus? even more confusion - yay ;) Jan 22 21:14:32 wow patches 2.6.38 Jan 22 21:14:39 lot of files since yesterday Jan 22 21:16:00 mini pci Jan 22 21:16:28 hmm something messed up i think in my kernel config Jan 22 21:16:42 how can i force a new kernel tree ? Jan 22 21:16:50 to download and patch again Jan 22 21:17:29 make target/linux/clean Jan 22 21:18:05 thanks Jan 22 21:18:36 so the git is on the 2.6.38 now ? Jan 22 21:19:24 it has support for .38 Jan 22 21:19:41 which does not mean that all targets use it yet Jan 22 21:19:48 :) ok Jan 22 21:20:04 how's the x86 support? Jan 22 21:20:10 I'm running .37 right now. Jan 22 21:20:51 dunno... in fact, i'm on 2.6.35 on my linux box :p Jan 22 21:21:00 ubuntu 10.10 Jan 22 21:21:56 i'm setting up an external kernel git for openwrt Jan 22 21:22:04 will be easier to track changes Jan 22 21:39:50 got another patch coming... Jan 22 21:40:33 ok Jan 22 21:40:42 image just finished building Jan 22 21:40:57 let me also merge that pacth and then you can try to boot the image and tell me if it works Jan 22 21:43:53 already booted all 3 patches... they work. Jan 22 21:44:10 my next patch is a merger of br2684ctl and linux-atm. Jan 22 21:45:33 ok Jan 22 21:45:38 uff Jan 22 21:50:04 kmod-gpiotoggling - 2.6.36.2+b57866441932eabc75b6c4252103121f606dc820-1 Jan 22 21:50:09 mouarf Jan 22 21:50:22 the git commit ? Jan 22 21:50:46 hmm... I moved br2684ctl into linux-atm... now how do I force an update of the package database where it looks for stuff? Jan 22 21:52:10 make package/index Jan 22 21:52:53 or make :) Jan 22 21:52:55 2 times Jan 22 21:53:04 make oldconfig / make Jan 22 21:53:31 ok, is there some doc ? Jan 22 21:53:39 ls /sys/gpiotoggling/ Jan 22 21:53:39 count fire interval loop start stop verbosity Jan 22 21:58:01 euh.... Jan 22 21:58:14 maybe the fire should reset old state Jan 22 21:58:22 after some seconds... Jan 22 21:58:41 you can break your gpio pins Jan 22 21:59:43 hmmm.. reboot time Jan 22 22:08:01 rmmm i really hope i dont have broke my pci line :/ Jan 22 22:08:28 there are gpios on pci int Jan 22 22:08:50 will try with the trunk one... Jan 22 22:16:48 ouf :) no its a kernel problem Jan 22 22:17:22 will clean my build dir Jan 22 22:18:49 blogic: how did the build go? Jan 22 22:19:45 gime a sac Jan 22 22:20:22 i think i will hack my isp modem :) http://tpruvot.free.fr/nc100/ Jan 22 22:20:52 its a cbvg834g Jan 22 22:20:57 netgear Jan 22 22:21:06 100Mb cable modem Jan 22 22:21:31 rebooting w/ patch #4 now. Jan 22 22:21:32 there is a serial connector, same kind as the DG834G Jan 22 22:24:46 http://www.netgear.com/service-provider/products/routers-and-gateways/cable-gateways/CBVG834G.aspx Jan 22 22:29:48 linux-2.6.36.2 Jan 22 22:29:58 what is the corresponding tag ? Jan 22 22:30:17 i see only the rc tags Jan 22 22:30:22 its rc2 ? Jan 22 22:31:24 no Jan 22 22:32:06 blogic: I just rebooted with the modified image. works. Jan 22 22:32:26 I'll wait until the linux-atm update gets committed, since patch #4 is layered on it. Jan 22 22:34:16 grrr... "svn diff" doesn't show moves... Jan 22 22:34:54 yea, i hate that Jan 22 22:35:07 moves in the kernel tree Jan 22 22:35:19 like includes :p Jan 22 22:35:19 ? Jan 22 22:35:22 it does Jan 22 22:35:24 svn move Jan 22 22:35:40 hmm Jan 22 22:35:49 there is maybe a flag for the diff Jan 22 22:36:01 because diff make a --- then a +++ Jan 22 22:37:03 nope Jan 22 22:37:20 git had invented new diff metadata to represent moves and patch gained support for it Jan 22 22:37:30 I think svn does not emmit such extra context Jan 22 22:38:21 argh Jan 22 22:38:30 cant find the 2.6.32-2 ref :p Jan 22 22:38:44 36.2 Jan 22 22:39:01 its maybe 37-rc2 ? Jan 22 22:39:02 openwrt has > 20 targets, if you speak of kernel versions you should also mention the platform Jan 22 22:39:17 for orion Jan 22 22:39:35 what do you mean with "the 2.6.32-2 ref" Jan 22 22:39:50 the one used in openwrt Jan 22 22:40:18 cat /proc/version Jan 22 22:40:18 Linux version 2.6.36.2 (buildslave@dd.das-iro.de) (gcc version 4.3.3 (GCC) ) #1 Fri Jan 21 22:09:19 CET 2011 Jan 22 22:43:07 thats in trunk: https://dev.openwrt.org/browser/trunk/target/linux/orion/Makefile Jan 22 22:43:47 yea... Jan 22 22:43:59 blogic * r25065 /trunk/target/linux/x86/geos/ (7 files in 4 dirs): Jan 22 22:43:59 [geos] Jan 22 22:43:59 * adds sane default uci environment Jan 22 22:43:59 * adds sane default package selection Jan 22 22:43:59 contributed by Philip Prindeville Jan 22 22:44:04 so its the 2.6.36 Jan 22 22:44:36 blogic: yippee! Jan 22 22:44:55 jow * r25066 /branches/backfire/package/base-files/files/lib/upgrade/keep.d/base-files-essential: [backfire] merge r24742 (#8700) Jan 22 22:45:11 blogic * r25067 /trunk/package/linux-atm/ (Makefile patches/200-no_libfl.patch): Jan 22 22:45:11 [linux-atm] Jan 22 22:45:11 * bump to 2.5.2 Jan 22 22:45:18 done Jan 22 22:45:58 there's no way to set the grub console baud rate from a target.mk file, is there? Jan 22 22:47:17 grub ? on the router ? Jan 22 22:47:25 yup. Jan 22 22:47:40 :) Jan 22 22:48:27 philipp64|laptop: thats done in Config.in Jan 22 22:48:34 philipp64|laptop: just amend it for geos Jan 22 22:52:45 Use external kernel tree is a string (CONFIG_EXTERNAL_KERNEL_TREE) Jan 22 22:52:51 euh Jan 22 22:53:10 same as CONFIG_KERNEL_GIT_LOCAL_REPOSITORY ? Jan 22 22:53:47 xMff: it's in there now... target/linux/x86/image/Config.in line 25... but it's not taking. Jan 22 22:54:19 it only has effect on virgin configsd Jan 22 22:54:37 once the corresponding symbol is set (from a loaded config) it is not overridden anymore Jan 22 22:55:19 actually, isn't the order wrong? Jan 22 22:56:36 just found the CCACHE option :) Jan 22 23:01:42 unknown hw_mode 'turboG' Jan 22 23:02:22 yeah, such a mode was never defined Jan 22 23:03:12 philipp64|laptop: looks like you forgot the initscript Jan 22 23:03:15 philipp64|laptop: fwiw, a lot of "old" are being built Jan 22 23:05:17 xMff: how's that forgot? Jan 22 23:05:29 deleted but not added again Jan 22 23:05:36 Atheros HAL provided by OpenWrt, DD-WRT and MakSat Technologies Jan 22 23:05:36 wifi0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps Jan 22 23:05:36 wifi0: turboG rates: 6Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps Jan 22 23:05:58 hmm no 108Mbps Jan 22 23:05:59 like the comment at the top says, it doesn't show up in the diff. Jan 22 23:06:13 ah right, svn Jan 22 23:07:09 hence the earlier question. Jan 22 23:08:13 any pppd guru around? Jan 22 23:08:17 philipp64|laptop: why does mving it up 1 line fix it ? Jan 22 23:08:34 i mean the console patch Jan 22 23:08:36 did you read the comment? Jan 22 23:09:07 :D Jan 22 23:09:13 it's not intuitive... it surprised me too. Jan 22 23:09:27 this dude here has fun with his line, I already patched pppd to make it survive the oddities but the last thing is interesting (4 diff. service offers and only the second is working) https://forum.openwrt.org/viewtopic.php?pid=126243#p126243 Jan 22 23:09:32 but apparently when it sees a match it stops processing the rest of the 'default' lines. Jan 22 23:10:57 blogic * r25068 /trunk/target/linux/x86/image/Config.in: Jan 22 23:10:57 [geos] Jan 22 23:10:57 * Make sure that Geos baud-rate defaults correctly Jan 22 23:10:57 Changeset 23414 somehow got reordered. The unconditional default trumps later settings. Jan 22 23:10:57 contributed by Philip Prindeville Jan 22 23:11:31 I'm feeling very prolific today.... :-) Jan 22 23:12:17 btw, if there's a /base-files and /files ... which gets done first? Jan 22 23:13:32 oh, I also tested the br2684ctl build... though I don't have RFC-1483 service so I can't test it. Jan 22 23:13:42 that is, deploy it. Jan 22 23:14:02 what command dumps the contents of a package? Jan 22 23:14:11 tar Jan 22 23:14:12 a .ipk that is. Jan 22 23:14:20 its a tar.gz Jan 22 23:14:38 which contains two tar.gz Jan 22 23:14:39 :) Jan 22 23:14:52 assign it to Seven Zip Jan 22 23:16:00 hmmm... compressing compressed files? Jan 22 23:16:59 the file build_dir/target-i386_uClibc-0.9.31/root-x86/usr/lib/opkg/info/br2684ctl.list looks good. Jan 22 23:17:12 hmmm... should I be using 0.9.32? Jan 22 23:18:03 option 'turbo' ? Jan 22 23:20:19 hmmm... building 2.6.38 fails... can't download several files. guess they haven't been uploaded yet. Jan 22 23:21:20 yea the 38 is in rc Jan 22 23:21:22 rc2 Jan 22 23:22:21 there is 6 to 8 rc generally Jan 22 23:22:49 one per week Jan 22 23:22:53 if i remember Jan 22 23:23:02 or 2 weeks Jan 22 23:23:12 ah... so rc2 is probably pretty half-baked then. Jan 22 23:23:43 and rc2 is 15 hours old :p Jan 22 23:24:00 ew. Jan 22 23:24:06 so 2.6.37 is final Jan 22 23:24:12 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=tags Jan 22 23:24:19 blogic: how does the br2684ctl patch look? Jan 22 23:24:57 swalker: rereading your comment... which are you talking about? what's an example? Jan 22 23:25:00 philipp64|laptop: did you check the size of the new version? Jan 22 23:26:25 6k... looks ok... didn't notice what the previous size was. Jan 22 23:26:38 ldd tells me it's sane. Jan 22 23:27:06 the prev .ipk was 5k Jan 22 23:27:07 br2684ctl --help looks sane. Jan 22 23:27:22 ah, you mean the package size. Jan 22 23:27:25 so okay, it didn't grow extremely Jan 22 23:27:28 thought you meant the binary. Jan 22 23:27:43 yeah, its fine Jan 22 23:28:18 package size is 4K Jan 22 23:28:30 philipp64|laptop: i will test it tommorw Jan 22 23:28:40 thanks! Jan 22 23:28:44 its a more intrisuve patch Jan 22 23:28:52 you have RFC-2684 bridging thwn? Jan 22 23:32:36 me ? Jan 22 23:32:38 yes Jan 22 23:32:45 i use it at home on my gateway Jan 22 23:34:38 philipp64|laptop: anything red on https://home.comcast.net/~sdwalker/uscan/uscan.shtml, some "older" than others Jan 23 00:01:17 ah, so you're asking me if I want to do more version bumps then? Jan 23 00:04:57 swalker: next on my todo list was to get the leds-geos module done... then look at cleaning up the cross-compilation support in ppp (deja vu... the last time I hacked ppp was 18 years ago). Jan 23 00:04:57 not specifically, although it's nice to be closer to upstreams, just that "old" packages aren't uncommon for various reasons Jan 23 00:05:31 well, weaning us off 2.6.32 as the baseline for a lot of targets would help!!! Jan 23 00:05:55 libnl definitely needs to be bumped. Jan 23 00:07:20 I wonder... do we actively feed patches upstream? if I could bump libnl to 2.0 and all of the patches (8 files???) went away, that would be very little work. Jan 23 00:08:05 not sure why we have the wrong package source URL though... Jan 23 00:10:48 xMff: I used to be a pppd guru... didn't see your question... what's up? Jan 23 00:11:14 check the linked thread Jan 23 00:22:57 i removed folder build_dir/linux-orion_generic/linux-2.6.36.2 Jan 23 00:23:12 but make doesnt recreate it... Jan 23 00:23:49 hmm i will clone openwrt again :/ Jan 23 00:25:02 an AC is an access concentrator, as far as I know... Jan 23 00:26:22 hmm nevermind Jan 23 00:26:34 bad path for my local git:p Jan 23 00:27:19 philipp64|laptop: I basically wonder whether its really up to pppd to deal with it Jan 23 00:27:26 that's a good thing :) Jan 23 00:27:44 (the link to the git) not a copy Jan 23 00:27:52 philipp64|laptop: this seems like a classical case of garbage in garbage out Jan 23 00:28:29 how so? Jan 23 00:28:54 if the is sends announcements that lead to nowhere something must be wrong Jan 23 00:28:59 *the isp Jan 23 00:29:33 I don't know how ordinary CPE equipment would deal with it, maybe it would be configured to only look for one specific AC Jan 23 00:30:37 earlier he had issues with the AC frames being malformed, that triggered a die() in the packet parsing of the pppoe plugin Jan 23 00:32:13 I'd tell him to start on the ppp mailing list. Jan 23 00:32:38 also, getting the model/make/sw version of his ISP's equipment is really helpful. Jan 23 00:32:53 yeah. I just skipped the wikipedia article on pppoe and it sounded like the user (or the client system) should be able to pick a specific PoP Jan 23 00:33:13 however I cannot find a pppd option to do that Jan 23 00:33:24 any words of caution before I plow into a version bump to libnl 2.0? Jan 23 00:33:33 or should I go to 1.2 instead? Jan 23 00:38:10 i think bbconfig should be added in busybox... Jan 23 00:38:14 also in kernel Jan 23 00:38:34 config module Jan 23 00:38:47 to compare kernel configs... Jan 23 00:46:22 luci is only available in OpenWrt-ImageBuilder ? Jan 23 00:46:44 ./scripts/feeds update; ./scripts feeds install -a -p luci Jan 23 00:46:55 =) Jan 23 00:47:04 faster than google :) Jan 23 00:48:18 wow Jan 23 00:48:29 almost Android market :p Jan 23 00:50:30 philipp64|laptop: "rp_pppoe_ac ..." is most likely what he needs. It can be specified in the pppd options Jan 23 00:50:50 this allows a user to manually select one of multiple ACs Jan 23 00:52:11 on a normal distro the pppoe-discover util would take care of it and present the user with a choice of services Jan 23 00:52:28 for the initial config Jan 23 01:05:28 erf, libxt_layer7.c:355: error: invalid application of 'sizeof' to incomplete type 'struct xt_layer7_info' Jan 23 01:05:44 in iptables-1.4.10 Jan 23 01:09:41 Userspace I/O platform driver Jan 23 01:09:58 could be usefull for a router :p Jan 23 01:10:27 Device Driver/Userspace I/O drivers Jan 23 01:28:00 xMff: sorry, was walking the dog... she's getting slower with age. Jan 23 01:28:48 ^^ Jan 23 01:29:00 its cold outside Jan 23 01:31:50 tryiing a 2.6.37 compilation :) Jan 23 01:33:22 rmmmm Jan 23 01:33:30 argh Jan 23 02:11:15 anyone know why we need so many patches to libnl? what will break if we build off-the-shelf libnl2.0 unpatched? Jan 23 02:18:55 many? its only 8 Jan 23 02:20:32 100 adds a static lib, 110 fixes toolchain issues, 120 adds missing header stuff, the rest should be obvious Jan 23 02:42:06 build #69 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/69 Jan 23 02:55:34 ouch! just replaced my AR5413 with an AR9160 (Ubiquity Networks SR71) and my upload speed just when up 100kb/s. **** ENDING LOGGING AT Sun Jan 23 02:59:57 2011