**** BEGIN LOGGING AT Sun Aug 02 02:59:57 2009 Aug 02 09:23:49 Not that I'm _needing_ ipv6 yet - but any idea why iptables6 or whatever it's called fails to be built (in x86) Aug 02 09:23:51 ? Aug 02 09:24:31 <[florian]> it's now fixed as of 17076 Aug 02 09:24:56 oh thanks. is that updating package/iptables ? Aug 02 09:25:38 [florian]: is that only _just_ fixed (since 12 hours ago?) Aug 02 09:26:15 <[florian]> svn update and the problem disappears :) Aug 02 09:26:17 (there's a new version of the Makefile, so I'm guessing so... trying now) Aug 02 09:26:19 cool Aug 02 09:27:40 juhosg * r17092 /trunk/target/linux/ar71xx/base-files/lib/upgrade/platform.sh: [ar71xx] fix sysupgrade Aug 02 09:30:00 [florian]: I'm making a new target for a x86 'kontron' board. Aug 02 09:30:20 [florian]: I've got the board - just awaiting CPU/Case so that I can actually try the image :) Aug 02 09:30:42 <[florian]> frogonwheels: oh ok,cool :), what CPU do you use? Aug 02 09:53:44 <[florian]> w/win 4 Aug 02 09:55:00 hi [florian] Aug 02 09:55:22 <[florian]> hii danage Aug 02 10:33:04 Hi All! Aug 02 10:33:09 <[florian]> hi AndyIL Aug 02 10:33:15 have some questions! Aug 02 10:33:23 <[florian]> ok :D Aug 02 10:33:59 i have bcm6332 router and dont have original firmware Aug 02 10:34:26 possible dump original firmware ? Aug 02 10:34:27 <[florian]> 6332 ? really ? Aug 02 10:34:32 Yes! Aug 02 10:34:53 <[florian]> I did not see any of them yet, can we buy some from russia? Aug 02 10:35:24 <[florian]> you could use JTAG to dump the flash Aug 02 10:36:08 router dont have Jtag and have SPI flash ;-) Aug 02 10:37:00 <[florian]> oh, that becomes interesting Aug 02 10:37:07 <[florian]> are you sure there is no jtag? Aug 02 10:37:08 possible make special kernel with dump function ? Aug 02 10:37:10 <[florian]> even on test points? Aug 02 10:37:27 <[florian]> you can build a kernel which has a MTD partition of the whole flash Aug 02 10:37:49 <[florian]> AndyIL: where can I buy such an unit? Aug 02 10:38:27 I very well understand in JTAG Aug 02 10:39:36 <[florian]> AndyIL: the flash is mapped at a physical address space, the EJTAG part of the CPU allows you dump the whole flash and you do not need to drive the flash pins Aug 02 10:40:37 They are sold by operator ADSL Aug 02 10:40:51 <[florian]> AndyIL: ah ok :/ Aug 02 10:41:28 How it to make? Aug 02 10:43:26 <[florian]> AndyIL: use JTAG in DMA mode Aug 02 10:44:02 PCB dont have JTAG pin Aug 02 10:45:26 <[florian]> are you sure? even on test points? Aug 02 10:46:17 YES! Aug 02 10:46:17 <[florian]> if you have no JTAG you need to recompile a firmware for your router which registers the whole flash Aug 02 10:46:30 How it to make? Aug 02 10:46:44 <[florian]> no idea, do you have any sources for it? Aug 02 10:47:05 possible build openwrt kernel! Aug 02 10:47:49 <[florian]> it will not boot Aug 02 10:48:03 <[florian]> the kernel does not know about bcm6332 Aug 02 10:48:15 <[florian]> and there are loads of differences Aug 02 10:48:46 i make bootable kernel Aug 02 10:50:09 6338 and 6332 almost identical Aug 02 10:50:36 <[florian]> ok, make a ramdisk version of OpenWrt (make menconfig -> Target images -> [*] ramdisk) Aug 02 10:50:49 OO! Aug 02 10:50:59 OK! Aug 02 10:51:03 <[florian]> then use CFE to dowload the kernel with TFTP Aug 02 10:51:51 It is clear Aug 02 10:52:22 <[florian]> SPI flash is not yet supported by OpenWrt ! Aug 02 10:52:55 Here in it and a question! Aug 02 10:53:11 <[florian]> this is not a question Aug 02 10:53:18 mtd supporten SPI flash Aug 02 10:53:42 <[florian]> yes, but I have not tested driving SPI flash with bcm63xx_spi Aug 02 10:54:17 And we will check up! Aug 02 10:54:30 <[florian]> ok Aug 02 10:54:58 <[florian]> AndyIL: you also need to use physmap-flash and not bcm963xx-flash to access the whole flash Aug 02 10:55:05 [florian]: could you flashimage fw_sinus_1054dsl_v2_1.bin for me real quick Aug 02 10:55:13 <[florian]> danage: ok Aug 02 10:55:15 that will also tell us whether i've reconfigured tftpd properly Aug 02 10:55:27 <[florian]> danage: connection closed whenever I type enter Aug 02 10:55:49 meh Aug 02 10:55:59 <[florian]> how? Aug 02 10:56:06 oh yes, of course, i know why... i need to plug the cable in... sec Aug 02 10:56:11 DUH :) Aug 02 10:57:06 <[florian]> CFE this router have dm command Aug 02 10:57:19 <[florian]> dump memory! Aug 02 10:57:37 <[florian]> in serial console Aug 02 10:57:46 <[florian]> AndyIL: perfect ! Aug 02 10:57:53 <[florian]> AndyIL: you need a little script then ;) Aug 02 10:58:07 <[florian]> only on screen! Aug 02 10:58:32 my terminal add cr lf Aug 02 10:58:48 <[florian]> AndyIL: http://openwrt.pastebin.com/m1a441e11 Aug 02 10:58:51 <[florian]> AndyIL: remove cr lf Aug 02 11:00:38 possible make direct reader from serial and convert it in bynary? Aug 02 11:01:42 Everything is possible ;) Aug 02 11:01:58 <[florian]> AndyIL: yes Aug 02 11:02:37 florian * r17093 /trunk/target/linux/generic-2.6/patches-2.6.30/ (3 files): [kernel] refresh 2.6.30 patches Aug 02 11:02:57 CFE> dm 0xbfc00000 2097152 Aug 02 11:02:57 bfc00000: 10 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ Aug 02 11:02:57 bfc00010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ Aug 02 11:02:57 bfc00020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ Aug 02 11:03:37 incorrect! Aug 02 11:04:04 <[florian]> try dm 0x1fc00000 Aug 02 11:04:35 florian * r17094 /trunk/target/linux/ar7/config-default: [ar7] resync kernel configuration Aug 02 11:04:35 <[florian]> no it`s correct boot address! Aug 02 11:04:57 first 4 bytes incorrect Aug 02 11:05:07 <[florian]> AndyIL: 0x1fc00000 also is the correct address this is the physical address of the flash Aug 02 11:05:17 <[florian]> 0xbfc00000 is the virtual address of the flash Aug 02 11:06:05 http://openwrt.pastebin.com/m2769bcf5 Aug 02 11:06:18 OK!! Aug 02 11:06:22 try Aug 02 11:06:59 <[florian]> AndyIL: it does not boot, I have told you :) it is similar but different :p Aug 02 11:07:32 wow it crashed the entire computer Aug 02 11:08:08 <[florian]> CFE reboot Aug 02 11:09:54 [florian]: can you try reconnecting to both? it should work Aug 02 11:10:18 <[florian]> danage: ok I am reconnected to ser2net Aug 02 11:10:25 <[florian]> needed make image work ONLY in readonly mode! Aug 02 11:10:31 good. i already flashed the original firmware. no crashes Aug 02 11:10:40 <[florian]> danage: ok Aug 02 11:10:46 you should get a login prompt Aug 02 11:10:50 <[florian]> danage: I will see what is so special about the original firmware Aug 02 11:11:06 i don't know the login though, let me see Aug 02 11:12:10 it should be admin and 0000 Aug 02 11:12:26 you can start busybox with sh Aug 02 11:12:50 <[florian]> ah ah Aug 02 11:12:51 <[florian]> awesome Aug 02 11:13:49 working? let me know in case ftp doesn't come up Aug 02 11:13:51 i'll be around Aug 02 11:14:01 [florian]: I've ordered an intel mobile duo 2-core CPU Aug 02 11:14:03 <[florian]> fyi, ctrl-d makes it escape the stupid cli Aug 02 11:14:38 [florian]: I'm hoping it arrives tomorrow :) Aug 02 11:15:53 [florian]: and don't hack my pc kthx ;) Aug 02 11:16:26 <[florian]> danage: eh eh Aug 02 11:17:01 <[florian]> possible make image work ONLY in readonly mode! Aug 02 11:19:48 <[florian]> AndyIL: yes Aug 02 11:20:05 kaloz * r17095 /trunk/target/linux/ (ppc40x/config-2.6.30 ppc44x/config-default): enable the watchdog driver for ppc4xx Aug 02 11:21:16 <[florian]> who? Aug 02 11:21:40 <[florian]> 2.6.30.4 not boot! Aug 02 11:22:16 [florian]: I'm looking forward to seeing this board work. it's mini-itx with 3x gigabit lan, 4x serial, compact flash on board (seems it uses IDE to acces it) VGA , sound on board etc. Aug 02 11:22:40 perfect for a router-on-steroids type situation :) Aug 02 11:23:52 <[florian]> AndyIL: 2.6.27 does not boot on your device either Aug 02 11:23:58 I'm just hoping I've got all the openwrt options sorted so that it will boot with everything working. Aug 02 11:24:03 <[florian]> AndyIL: you need to patch openwrt for your bcm6332, they are too different Aug 02 11:24:06 <[florian]> http://openwrt.pastebin.com/m64b6b110 Aug 02 11:24:19 Kontron actively supports linux (they tell you what drivers to use), which is great :) Aug 02 11:24:37 <[florian]> AndyIL: ah ok Aug 02 11:24:37 <[florian]> about bcm6338 and 2.6.27 -> 2.6.30 Aug 02 11:24:52 <[florian]> AndyIL: this is *experimental* : Aug 02 11:25:06 <[florian]> whatch problem? Aug 02 11:25:30 <[florian]> AndyIL: I do not know yet why Aug 02 11:25:56 <[florian]> boot log look? Aug 02 11:26:19 <[florian]> AndyIL: I had the same thing on bcm6338 too Aug 02 11:26:37 <[florian]> what? Aug 02 11:26:46 <[florian]> AndyIL: same problem here Aug 02 11:26:49 <[florian]> AndyIL: no solution yet Aug 02 11:27:18 [florian]: seems like you're fighting on several fronts here :) Aug 02 11:27:28 <[florian]> ok! Aug 02 11:30:33 <[florian]> incorect parsing partition offset? Aug 02 11:31:45 <[florian]> AndyIL: I do not think so Aug 02 11:32:09 bcm963xx_flash: Partition 1 is kernel offset 4c023000 and length 0 (2.6.30) Aug 02 11:32:23 bcm963xx_flash: Partition 1 is kernel offset 10100 and length bff00 (2.6.27) Aug 02 11:32:29 <[florian]> oh, so you are right ! Aug 02 11:32:35 <[florian]> I did not compare them Aug 02 11:33:02 needed small fix ;-) Aug 02 11:34:15 <[florian]> AndyIL: yes :p Aug 02 11:34:41 <[florian]> where ? Aug 02 11:35:20 <[florian]> AndyIL: I think in target/linux/brcm63xx/files/drivers/mtd/maps/bcm963xx-flash.c Aug 02 11:38:23 <[florian]> I understand. You will make? Aug 02 11:38:35 <[florian]> AndyIL: try please Aug 02 11:38:39 <[florian]> AndyIL: I will not do it today Aug 02 11:39:33 <[florian]> have any progress about 6345? Aug 02 11:40:14 <[florian]> AndyIL: no sorry Aug 02 11:57:25 hi small question: how to obtion the bootlog from an wl500gP? Aug 02 11:57:34 obtain. Aug 02 11:58:00 serial cable Aug 02 11:59:50 shit. Aug 02 12:00:47 I'm trying linux 2.6 on my wl500gP and it hangs while booting into normal mode. failsafe works so far. Aug 02 12:05:35 I want to get some informations, what's actually going wrong. Aug 02 12:07:32 serial console is the most reliable but maybe netconsole works for you too Aug 02 12:07:50 assuming that at least the kernel boots properly Aug 02 12:08:44 it does, I can boot into failsafe. Aug 02 12:09:22 thy give netconsole a try Aug 02 12:09:28 *then Aug 02 12:14:30 I have to make another image, don't I? Aug 02 12:20:02 I think yes Aug 02 12:21:12 <[florian]> about partition address i`s printf problem in bcm963xx-flash.c line 264 & 329 %x --> %llx Aug 02 12:23:23 where to adjust kernel parameters to enable netconsole? Aug 02 12:24:26 Gottox: I think in make kernel_menuconfig Aug 02 12:24:37 ah Aug 02 12:24:41 the bootcommands Aug 02 12:25:18 got to enable netconsole in the kernel too. Aug 02 12:28:17 or is there some way to boot into normal mode from failsafe without rebooting? Aug 02 12:29:23 <[florian]> AndyIL: can you check with cat /proc/mtd that it is the same as expected? Aug 02 12:30:46 insmod: cannot insert 'b43.ko': unknown symbolin module, or unknown parameter Aug 02 12:31:58 Gottox: yes, you could start the init scripts one by one Aug 02 12:32:10 after mount_root Aug 02 12:37:53 florian * r17096 /trunk/package/ar7-atm/patches/170-bus_id_removal.patch: [ar7] compile fixes in ar7-atm for new kernels Aug 02 12:46:06 <[florian]> http://pastebin.com/m4dc0e639 Aug 02 12:46:42 <[florian]> AndyIL: endianness problem ! Aug 02 12:46:52 <[florian]> AndyIL: it should be something like cpu_to_le32(SQUASHFS_MAGIC) Aug 02 12:47:42 <[florian]> printk(KERN_ALERT "<-magic-> 0x%08x\n",le32_to_cpu(sb.s_magic)); Aug 02 12:48:10 <[florian]> AndyIL: strange Aug 02 12:49:29 <[florian]> in flash image signature 0x73717368 possible mksquashfs build incorrect image? Aug 02 12:49:54 <[florian]> AndyIL: it is possible, yes Aug 02 12:50:04 <[florian]> AndyIL: oh yes, we us mksquashfs4 now Aug 02 12:50:17 <[florian]> ;-) Aug 02 12:50:42 <[florian]> fix prinf problem? Aug 02 12:51:13 <[florian]> AndyIL: I have to test Aug 02 12:53:07 <[florian]> ok Aug 02 12:53:56 xMff: strange problem with openwrt. after booting normal, I can't ping the router, but if I call the initscripts one by one from failsafe it works (except wifi) Aug 02 12:55:10 <[florian]>possible add automatic switching use mksqsh or mksqsh4 for new kernel? in imagebild script? Aug 02 12:55:36 <[florian]> AndyIL: is not that done already? Aug 02 12:57:17 <[florian]> AndyIL: you used squashfs3 or 4? Aug 02 12:57:37 3 Aug 02 12:57:59 3 use automatic Aug 02 12:57:59 <[florian]> ok Aug 02 12:58:02 <[florian]> try with 4 Aug 02 12:58:06 ok! Aug 02 13:01:20 florian * r17097 /trunk/target/linux/ar7/ (14 files in 3 dirs): [ar7] add experimental support for 2.6.30 Aug 02 13:02:10 nice Aug 02 13:02:30 <[florian]> got to test this thoroughly this afternoon :po Aug 02 13:04:22 :) Aug 02 13:05:09 <[florian]> i can`t find where call call mksquashfs Aug 02 13:10:02 AndyIL: did you run make oldconfig after updating your tree to 2.6.30? Aug 02 13:10:43 no! Aug 02 13:10:49 do it Aug 02 13:10:53 OK! Aug 02 13:27:45 juhosg * r17098 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/irq.c: [ar71xx] fix MISC IRQ handling on the AR7240 Aug 02 13:29:57 anyone arround whos running an sd card on a fonera with kamikaze 8.09? Aug 02 13:46:42 <[florian]> AndyIL: with an oldconfig it works for me Aug 02 13:50:41 florian * r17099 /trunk/target/linux/brcm63xx/files/drivers/mtd/maps/bcm963xx-flash.c: [brcm63xx] fix printing of offsets Aug 02 14:05:05 [florian] thanks (printf) Aug 02 14:05:30 nbd thanks (make oldconfig) Aug 02 16:56:58 [florian]: someone posted in the bcm63xx thread on the forums that there would be no DSL driver for 2.6 - neither blob nor sourcewise? Aug 02 17:33:49 ping juhosg Aug 02 17:35:04 jushog, regarding ethernet mac addresses on the wrt160nl, i changed options in the nvram makefile, then built nvram for this platform (it was excluded for this platform initially) Aug 02 17:35:19 the nvram utility worked fine to extract the ethernet mac addresses from nvram Aug 02 17:35:53 now trying to decipher just where in the start up scripts I should place this, so the mac gets set during the initial bring-up of ethernet Aug 02 17:36:32 i need to get it to do ifconfig ethx hw ether $(nvram get lan_hwaddr) along the way Aug 02 17:56:49 <[florian]> danage: that's completely false information Aug 02 17:58:00 <[florian]> danage: I do not even know where the guy heard about that, and if he has any informations they are wrong :) Aug 02 18:00:12 florian ping Aug 02 18:00:35 in the build for base stuff, the nvram utility is dependant on broadcom targets Aug 02 18:00:36 <[florian]> groz: pong Aug 02 18:00:49 <[florian]> groz: that's right Aug 02 18:00:49 turns out, really needs to be dependant on anything thru cybertan Aug 02 18:00:58 <[florian]> groz: right again Aug 02 18:01:01 works fine on my wrt160nl whic is ar711 Aug 02 18:01:03 haha, ok Aug 02 18:01:13 and, i built it in, changed the depends to 'depnds nothing' Aug 02 18:01:16 so i could add it in Aug 02 18:01:26 and, i can no get correct mac addresses for this thing easily Aug 02 18:01:29 just an fyi Aug 02 18:01:52 not sure exactly how that dependancy should be changed to be 'technically correct' Aug 02 18:02:03 <[florian]> well the simple thing is probably to make it depend on target that support the nvram format Aug 02 18:02:48 plenty of little nuances so far but, got the board up and running correctly , now figured this out on getting macs right Aug 02 18:02:58 im not far off of having it 'all correct' now Aug 02 18:03:12 but there is a philisophical question to ask regarding format of bins Aug 02 18:03:21 for now, since the uboot on this thing uses trx Aug 02 18:03:32 i followed the brcm example of making a multi-part trx Aug 02 18:03:53 it could be done tho, building the entire setup as a raw uboot image Aug 02 18:04:04 then just put that inside a trx wrapper in a single part trx Aug 02 18:04:32 <[florian]> yes you could do that Aug 02 18:04:46 well, just wondering, and i know you are likely the wrong person to ask Aug 02 18:05:00 in the long run, which is a 'better' way for the project as a whole Aug 02 18:05:23 linksys specific image building, or, generic image inside a linksys wrapper ? Aug 02 18:05:58 <[florian]> I would say that generic image inside a linksys wrapper sounds like a good long term solution, but that's thinking about it a couple of seconds ;) Aug 02 18:06:18 :) Aug 02 18:06:34 i know, i've been pondering for a day now, i have linksys specific build working, and working well Aug 02 18:06:43 squash complete with the split Aug 02 18:06:46 and jffs Aug 02 18:06:57 but, wondering if there is an overriding reason to not do it this way Aug 02 18:07:07 i've got it parsing trx header to find the partitions Aug 02 18:07:18 and make the line between kernel and squash that way Aug 02 18:07:39 that involves a cybertan specific option in the kernel config Aug 02 18:07:55 which can be safely compiled in elsewhere, and, without the correct header on the flash partition Aug 02 18:07:59 it'll ignore it Aug 02 18:08:19 but i've been out of the 'philosophy' loop for a while Aug 02 20:19:28 hello openwrt developer Aug 02 20:22:38 please how i can check if my DLink device is supported? Aug 02 20:22:48 DLink 2054U ADSL router Aug 02 20:28:15 the Makefile for s3fs hardcodes the compiler reference... like this: all: g++ -ggdb -Wall .... does that mean I have to keep a patchfile in /package/s3fs/patches in order to get it to compile? Aug 02 21:10:22 [florian]: i'm about to shut this box down. let me know when you need access, ok? Aug 02 23:24:34 ping Hopscotch Aug 03 02:56:39 is this channel good for providing help in porting a package to openwrt? Aug 03 02:56:56 yep Aug 03 02:57:22 stevecrozz: you may need to wait for answers. tends to be reasonably low volume, so might take hours for an answer Aug 03 02:57:43 frogonwheels: I'm not exactly an embedded systems developer, but I'm trying my best to port s3fs to openwrt by forcing it to compile against uClibc++ Aug 03 02:57:57 I could really use some help from someone who's done this thing before Aug 03 02:58:02 you've got a package/ Makefile ? Aug 03 02:58:17 yeah, i wrote one, but the compile fails all over the place Aug 03 02:58:39 have you got a custom configure section? Aug 03 02:58:58 it's worth putting what you've got into a pastebin Aug 03 02:59:04 I can show you what I've got, sure **** ENDING LOGGING AT Mon Aug 03 02:59:57 2009