**** BEGIN LOGGING AT Thu Sep 17 02:59:57 2009 Sep 17 03:30:58 hi.. Sep 17 03:31:14 do u think superchannel is possible with atheros card with ath5k on superchannel ? Sep 17 08:43:13 juhosg * r17600 /trunk/target/linux/adm5120/patches-2.6.30/ (001-adm5120.patch 005-adm5120_usb.patch): adm5120: refresh 2.6.30 patches Sep 17 08:43:16 juhosg * r17601 /trunk/target/linux/adm5120/patches-2.6.30/810-adm5120sw-netdev_ops.patch: adm5120: use netdev_ops on kernels >= 2.6.30 (based on a patch by Alexandros C. Couloumbis, closes #5846) Sep 17 09:38:35 Hi, there is a device known as Procare eShare ES-8068 with 200 MHz cpu with 64mb ram and 8mb rom selling for around $50. It has no wifi but two usb ports. There is a toolchain and 2.6 series kernel for it. Sep 17 09:41:35 arm cpu Sep 17 09:43:10 $42 shipped that is Sep 17 09:49:35 and complete datasheet Sep 17 10:03:31 aet: what's inside? Sep 17 10:04:09 str9104 aka CNS1104 Sep 17 10:04:26 not atheros or anything else for the matter Sep 17 10:06:53 there is an os for it already but imo having openwrt would be alot easier in the long run Sep 17 10:07:18 hi guys Sep 17 10:10:49 you have any documentation about porting openwrt to new platforms? Sep 17 10:15:14 aet: http://kamikaze.openwrt.org/docs/openwrt.html#x1-540002.3 Sep 17 10:15:54 also, http://nuwiki.openwrt.org/doc/techref/buildroot Sep 17 10:18:09 loswillios: thnaks Sep 17 11:22:35 kaloz * r17602 /trunk/ (25 files in 25 dirs): update to latest 2.6.27 and 2.6.30 kernels Sep 17 11:54:50 nbd * r17603 /branches/8.09/package/madwifi/files/lib/wifi/madwifi.sh: madwifi: merge uapsd setting change from r17510 to fix powersave with some mobile devices Sep 17 11:56:33 N Sep 17 12:34:35 <_trine> if I use this :- Strip unnecessary functions from libraries;; from;; Global build settings ─my compile always fails at make[3] package/preconfig Sep 17 12:36:48 yes, the feature still has some bugs in some constellations Sep 17 12:36:58 please run make V=99 to see what error message it prints Sep 17 12:37:14 <_trine> ok Sep 17 12:40:02 nbd: prolly threads :) Sep 17 12:42:09 OutBackDingo: did you get any further with that? Sep 17 12:42:15 <_trine> nbd, http://pastebin.com/m57127382 Sep 17 12:42:36 whee, segmentation fault Sep 17 12:42:37 that's a new one Sep 17 12:42:38 ;) Sep 17 12:43:48 nbd: nope, i know what the error is, wheres it needs to be fixed, but required someones brains like urs or xMffs to change / commit it Sep 17 12:43:56 heh yeaah tis is a new one Sep 17 12:44:45 <_trine> without using the strip option it all compiles Sep 17 12:45:11 OutBackDingo: ok, so where is the error? Sep 17 12:45:28 _trine: yes, because it segfaults in a program that is only used for the stripping option Sep 17 12:46:14 <_trine> nbd, I did use the other stripping option for kernel stuff and that's ok Sep 17 12:46:14 nbd: mklibs.py is what runs the routine Sep 17 12:46:37 needs to be modified to allow removal of -lc for pthreads Sep 17 12:46:59 which should resolve my problem Sep 17 12:47:14 so where does the -lc come from in the first place? Sep 17 12:47:32 one should propably implement some kind of config file where overrides for specific libs can be defined Sep 17 12:48:14 I believe that pthread will not be the last lib with issues Sep 17 12:48:18 might not be necessary Sep 17 12:48:36 considering that pthread is part of the core libraries and does pretty weird stuff, it's normal to give it special treatment Sep 17 12:48:36 xMff: itll be my last issue i belive Sep 17 12:48:49 libc and libpthread both define the same kinds of symbols for a reason Sep 17 12:48:51 true too Sep 17 12:48:58 libpthread implements wrappers that contain mutexes Sep 17 12:49:40 right, and when the strip function runs, it realizes its in both pthreads and libc and fails Sep 17 12:49:49 the removal of -lc did work Sep 17 12:50:17 however, looking at mklibs.py id do more damage breaking pythin then getting it fixed Sep 17 12:51:08 in my case my whole tree will compile fine, until you add olsr, which uses pthreads Sep 17 12:52:09 where is the mklibs.py located in the buildroot? Sep 17 12:52:27 staging_dir/host/bin Sep 17 12:53:08 ty Sep 17 12:53:10 \build_dir/host/mklibs/src Sep 17 15:09:36 hi there, I'm trying to port the ifplugd to OpenWRT. I made nice Makefile and got it compiling and running. but ifplugd is always detecting all the vlans of the ethernet switch as "link beat detected" Sep 17 15:10:07 do I have to activate something in the kernel or would the switch driver have to be changed for that? Sep 17 15:14:58 T_X: this is normal Sep 17 15:15:21 T_X: you'd made it aware of the broadcom specific switch api to make it detect link changes on single ports Sep 17 15:15:33 *you'd need to make Sep 17 15:24:22 xMff: ah, okay. atheros with its ip175c switch chip though Sep 17 15:25:15 T_X: in this case you'd need to use the swconfig api :) Sep 17 15:25:19 so the driver has to explicitly bounce this SIOCETHTOOL ioctl stuff? Sep 17 15:25:27 no Sep 17 15:25:31 hmm Sep 17 15:25:31 other ioctls Sep 17 15:25:37 iirc Sep 17 15:25:46 or generic netlink in the case of swconfig Sep 17 15:26:23 Collecting package info: done Sep 17 15:26:23 ERROR: please fix target/linux/cavium/Makefile Sep 17 15:27:41 i copied target/linux/orion/ to target/linux/cavium and changed BOARD to cavium in Makefile Sep 17 15:28:33 i removed tmp but that did not help either Sep 17 15:36:49 aet: orion uses subtargets Sep 17 15:37:14 and taht si? Sep 17 15:37:50 subtargets split up a target into different configurations Sep 17 15:38:17 if you only copy the makefile, you need to remove the subtargets variable and make sure that BOARDNAME and Target/Description are set Sep 17 15:45:13 ok Sep 17 15:46:52 is there some file where i could find these CFLAGS in prebuilt toolchain? Sep 17 15:49:38 what do you mean? Sep 17 15:50:45 i have a prebuilt uclibc toolchain that i do not want to use but instead let openwrt build a new one Sep 17 15:51:33 make dirclean gets rid of all the toolchain stuff that it built, and it also implies make clean Sep 17 15:52:23 i mean -march and -mtune Sep 17 15:53:08 the CFLAGS are stored in .config Sep 17 15:53:25 and the defaults are set based on the arch and can be changed in the target makefile Sep 17 15:57:32 is there some specific way to make config-default? Sep 17 16:20:03 okay, it is building Sep 17 17:09:47 xMff: u want a ticket on the mklibs ? Sep 17 17:09:52 or leave it alone Sep 17 17:12:56 create a ticket so that the details are available Sep 17 17:20:15 {standard input}:39: Error: selected processor does not support `bx r6' Sep 17 17:21:41 xMff: FYI https://dev.openwrt.org/ticket/5851 Sep 17 17:21:57 though u prolly just got mail LOL Sep 17 17:23:16 whaaa kernel upgrade... hrmm Sep 17 17:24:12 _trine: question, what version of gcc you building with ? Sep 17 17:49:30 <_trine> OutBackDingo: 4.1.2 Sep 17 17:49:46 <_trine> as far as I recall Sep 17 17:50:04 _trine: can u try it with 4.4.1 with stripped enabled Sep 17 17:50:22 mine did a full stock trunk build there Sep 17 17:50:28 ath2k Sep 17 17:50:45 <_trine> OutBackDingo: I can but not today Sep 17 17:50:51 ahh ok Sep 17 17:50:56 whats ur target Sep 17 17:51:02 <_trine> ar71xxx Sep 17 17:51:18 ill try one this evening Sep 17 17:51:21 <_trine> ok Sep 17 17:51:35 <_trine> I'm just in the middle of letter writing Sep 17 17:56:10 hello guys... I'm trying to build an image for wrt54gl to use sd card, so I select kmod-broadcom-mmc and kmod-fs-ext2 kmod-fs-ext3 and kmod-fs-vfat, the build done ok but after flash router do not start Sep 17 17:56:28 I have not serial port to check what happend Sep 17 17:56:41 anyone have any idea? Sep 17 17:57:30 bump it into failsafe and exute thr initscripts manually Sep 17 17:58:14 kmod-broadcomm-mmc It's ok, or i need copy mmc.o from other site Sep 17 17:59:18 ok xMff, to get failsafe I need push reset button after plug power, right? Sep 17 17:59:40 when the dmz led starts glowing, hit reset until dmz starts blinking Sep 17 17:59:44 will bypass any init Sep 17 18:00:06 then telnet to 192.168.1.1 and start executing the scripts one by one until you find the offending one Sep 17 18:00:30 xMff, oh, not DMZ never start Sep 17 18:01:01 then you probably flashed a broken image Sep 17 18:01:14 or a too big one or whatever Sep 17 18:01:46 ok, will build and try againg thanks xMff Sep 17 18:33:31 xMff: thanks, but look like my buildroot is bad, so I will de make clean Sep 17 21:24:20 excuseme xMff now I have installed an optimized mmc Sep 17 21:24:32 xMff: but i have an error Sep 17 21:24:53 when mmc start Sep 17 21:25:26 [WARN] mmc_init: impossible to get card indentification info for reason code: 02 Sep 17 21:25:27 [ERROR] mmc_init: got an error when trying to get card configuration registers: 01 Sep 17 21:25:27 [ERROR] mmc_check_media: change detected but was not able to initialize new card: ffffffff Sep 17 21:26:32 xMff: this sd is a Kingston with 1G Sep 17 21:27:50 mmc driver support sd card with 2, 4, ...??? Sep 17 21:40:58 excuseme xMff now I have installed an optimized mmc Sep 17 21:41:06 xMff: but i have an error Sep 17 21:41:15 when mmc start Sep 17 21:41:30 [WARN] mmc_init: impossible to get card indentification info for reason code: 02 Sep 17 21:41:50 [ERROR] mmc_init: got an error when trying to get card configuration registers: 01 Sep 17 21:42:08 [ERROR] mmc_check_media: change detected but was not able to initialize new card: ffffffff Sep 17 21:42:43 xMff: any Idea? Sep 17 22:16:33 lars * r17604 /trunk/target/linux/generic-2.6/patches-2.6.31/998-openwrt_lzma_options.patch: Fix lzma cmd args. Sep 17 23:33:55 how do i actually prevent wiresless-tools package from being built? Sep 17 23:40:26 (by disabling all wireless drivers of course) Sep 18 00:38:12 so i now got a firmware but no idea how to make it work with bootp Sep 18 01:24:14 it boots **** ENDING LOGGING AT Fri Sep 18 02:59:56 2009