**** BEGIN LOGGING AT Wed Aug 05 02:59:56 2009 Aug 05 03:51:35 ok. we _seem_ to be building a x86 64bit target. cross fingers and hope Aug 05 04:21:02 * kupesoft nudges for ticket #5626 fix Aug 05 04:35:37 xMff: ping Aug 05 04:57:45 I have installed openwrt from latest trunk. Aug 05 04:59:29 Now opening http://192.168.1.1/luci/admin/system (in LuCI Administration->System->System fails with error http://lua.pastebin.com/mbd07852 Aug 05 05:03:52 dkostousov: you may have to wait a few hours for xMff to get some sleep. Aug 05 05:04:40 looks like he went to bed approx 3am his time. Aug 05 05:05:10 There are a few other who've delved into lua - though don't know who. Aug 05 05:09:39 Is xMff openwrt's main developer? Aug 05 05:12:20 no- but Luci is his baby. Aug 05 09:49:06 AndyIL: success? Aug 05 09:51:58 about? Aug 05 09:52:43 bcm6332 Aug 05 09:53:21 boot fine! spi driver dont working! Aug 05 09:53:59 this board have spi flash! ;-( Aug 05 09:54:38 spi driver hang on load Aug 05 09:54:53 AndyIL: you no success fixing? Aug 05 09:55:36 how? Aug 05 09:55:41 i don't know Aug 05 09:57:34 For whom worked SPI on bcm63xx Aug 05 09:58:11 i don't think it works, bcm6338 has no spi? Aug 05 10:03:20 6338 has spi! Aug 05 10:03:51 <[florian]> bcm6338 has spi, but does not use it for flash or switch Aug 05 10:06:14 <[florian]> You can add necessary structures for support SPI flash? For check? Aug 05 10:06:39 <[florian]> AndyIL: yes, I will do later Aug 05 10:08:43 <[florian]> Can be now. For you it is 10 minutes. Aug 05 10:10:26 <[florian]> AndyIL: no, I am at work Aug 05 10:11:20 <[florian]> AndyIL: and it is not 10 minutes to add flash support :) Aug 05 10:12:22 <[florian]> for your 20! not more! Aug 05 10:12:34 <[florian]> lol Aug 05 10:13:49 :) Aug 05 10:41:18 florian * r17128 /trunk/target/linux/generic-2.6/config-2.6.27: [kernel] add missing configuration symbol CONFIG_CRYPTO_SALSA20_586 in 2.6.27 config Aug 05 11:18:08 jow * r17129 /packages/net/mini_snmpd/ (Makefile patches/104-ipv6-support.patch): Aug 05 11:18:08 [PATCH] mini_snmpd: support for IPv6 Aug 05 11:18:08 The following patch adds IPv6 support to mini-snmpd in the packages repository. It is then able to handle requests from IPv4 and IPv6 clients. Aug 05 11:18:08 Signed-off-by: Linus L?ssing Aug 05 12:14:01 Hi there, I always got " not syncing: Attempted to kill init" Aug 05 12:14:23 I add some output in arch/mips/kernel/syscall.c Aug 05 12:14:23 the do_execve (/etc/preinit) return 0 Aug 05 12:14:37 but always kernel panic. I am beginner at kernel hacking. so just give me some tips. Aug 05 12:14:39 very thanks. Aug 05 12:32:16 xMff ping Aug 05 12:32:44 . Aug 05 12:33:24 seems it's possible to run multiple instances of dropbear in trunk now. can you add this to luci? Aug 05 12:35:20 have you tried running multiple instances and whether it actually works? Aug 05 12:35:36 no Aug 05 12:36:18 then I'll delay that until I had a chance to test Aug 05 12:37:04 Dogge_: what's the advantage of that Aug 05 12:37:16 (just curious) Aug 05 12:37:20 xMff okie Aug 05 12:37:56 I try to configure openwrt on d-link dir330. Now I have a problem with configuring wan interface of the router. The wan interface should have static ip address. I set up ipaddr to eth1. But ping fails through eth1. Aug 05 12:37:56 dmesg from DIR-330 openwrt trunk r17127 http://pastebin.com/m3828aea8 Aug 05 12:38:02 network configuration http://pastebin.com/m28902bb0 Aug 05 12:38:03 danage: easy way to run ssh on multiple ports without having to make firewall rules Aug 05 12:38:10 'robocfg show' command output http://pastebin.com/m69e25fd2 Aug 05 12:38:17 I have found some difference between config and robocfg out: /etc/config/network defines vlan0 but robocfg shows vlan0 and vlan1 Aug 05 12:38:21 danage: or make one internal with pwd auth and one external with key-only auth Aug 05 12:38:32 ah yes, makes sense Aug 05 12:39:33 it was always almost possible to run multiple instances by tweaking two lines in the init script, but as a sideeffect of a recent patch it could run ootb now Aug 05 12:40:24 cool Aug 05 12:40:35 something else: Bartman007 needs a change inthe dokuwiki config files Aug 05 12:40:44 you have access right? Aug 05 12:40:48 yes Aug 05 12:40:52 <[florian]> xiangfu: still here? Aug 05 12:41:00 i forgot what it was, xMff did he contact you yet? Aug 05 12:41:04 <[florian]> xiangfu: which toolchain are you building with? Aug 05 12:41:10 yes Aug 05 12:41:10 danage: no Aug 05 12:41:56 it's about toh and some plugin me thinks, sounds like he's almost done Aug 05 12:42:39 [florian]: what you mean? I use "staging_dir/toolchain-mipsel_gcc-4.1.2_uClibc-0.9.30.1/usr/bin" Aug 05 12:43:00 <[florian]> xiangfu: I meant the version of the compiler and uclibc version Aug 05 12:43:13 <[florian]> xiangfu: did you change anything in the kernel sources? Aug 05 12:43:24 <[florian]> xiangfu: which touches the executables handling Aug 05 12:43:52 [florian]: no Aug 05 12:44:07 * moonflux hates autoconf. understanding a configure script failing makes my brain hurt Aug 05 12:44:31 <[florian]> xiangfu: ok, can you try to get a core dump of what is happening and run gdb on it?* Aug 05 12:45:04 I try to configure eth0.1 as WAN interface but nothing changes Aug 05 12:45:33 [florian]: I will try. never try to get a core dump. I need learn. and some time. sorry. Aug 05 12:51:46 xMff: also, i won't make the wiki meeting here tomorrow Aug 05 12:52:30 danage: okay, log will be published anyway as usual Aug 05 12:52:36 cool Aug 05 12:53:00 <[florian]> xiangfu: core dump should be written on the medium Aug 05 12:53:04 <[florian]> xiangfu: i.e your sd card Aug 05 12:56:19 lars * r17130 /trunk/target/linux/s3c24xx/files-2.6.30/arch/arm/mach-s3c2442/mach-gta02.c: [s3c24xx] gta02: Workaround hardware bug on rev 5 and earlier Aug 05 13:03:59 [florian]: there is no core file after panic. Aug 05 13:05:23 <[florian]> xiangfu: ok, try to start /bin/sh instead of /etc/preinit Aug 05 13:05:33 <[florian]> xiangfu: which kernel version is this? Aug 05 13:06:06 <[florian]> xiangfu: also, please make sure that you have CONFIG_SOFT_FLOAT in your openwrt .config Aug 05 13:11:30 [florian] : CONFIG_SOFT_FLOAT=y Aug 05 13:11:43 [florian]: kernel version is 2.6.28-10 Aug 05 13:11:45 <[florian]> xiangfu: ok Aug 05 13:11:55 [florian]: just try the /bin/sh. same error. Aug 05 13:11:57 <[florian]> xiangfu: well, can you send me your openwrt .config ? Aug 05 13:14:06 [florian]: wget http://www.openmobilefree.net/other/file/openwrt.config Aug 05 13:15:12 <[florian]> xiangfu: thanks I see nothing wrong in your config Aug 05 13:28:08 [florian]: if "init=/bin/sh" there is no "kernel Panic *** kill init", just stop. no console prompt . Aug 05 13:46:08 <[florian]> xiangfu: ok, tyr Aug 05 13:46:18 <[florian]> try init=/bin/busybox help Aug 05 13:48:23 florian * r17131 /trunk/target/linux/rdc/ (8 files in 2 dirs): [rdc] add experimental support for 2.6.30 Aug 05 13:49:27 "Kernel panic ** kill init!" appear again. Aug 05 13:50:16 <[florian]> xiangfu: ok Aug 05 13:50:29 <[florian]> xiangfu: try the same rootfs with a 2.6.24 kernel Aug 05 13:52:50 [florian]: stop at "Warning: unable to open an initial console." Aug 05 13:53:33 <[florian]> xiangfu: which uart does this device use ? 8250 / ttyS0 ? Aug 05 13:53:41 ttyS0 Aug 05 13:54:07 <[florian]> ok, then I really think there is something fucked up with your toolchain Aug 05 13:54:25 oh Aug 05 13:55:43 <[florian]> xiangfu: ok, try that please: http://openwrt.pastebin.com/m40ef1799 Aug 05 13:56:18 ok. Aug 05 14:09:36 1. change MIPS_FPU_EMU to y. 2. rm build_dir/linux-xburst/linux-2.6.28.10/ 3. make V=99 -j16 Aug 05 14:09:58 [florian]: still the Kernel panic. :-(. Aug 05 14:12:00 <[florian]> xiangfu: then it's definitively related to your toolchain or the kernel does not have support for ELF executables Aug 05 14:15:06 [florian]: very thanks. Aug 05 14:15:27 <[florian]> xiangfu: you are welcome Aug 05 14:15:30 [florian]: that is hard for me to look into the toolchain source code. :-) Aug 05 14:15:40 <[florian]> xiangfu: it worked with 2.6.24 did not it? Aug 05 14:16:04 <[florian]> xiangfu: was there any patch required for the toolchain in the pi-project code source? Aug 05 14:17:47 <[florian]> xiangfu: actually are we sure the mips32 core in the jz47x0 core is fully mips32 compliant? Aug 05 14:17:49 [florian]: before I merge with upstream. the openwrt toolchains compile the 2.6.24 kernel and rootfs. it's work. Aug 05 14:19:32 <[florian]> xiangfu: I still think it's related to soft-float, can you try this in addition to the patch wich re-enabled kernel FPU emulator: http://openwrt.pastebin.com/m20e99a5 Aug 05 14:19:36 <[florian]> xiangfu: make distclean Aug 05 14:19:50 <[florian]> rm .config; make menuconfig again Aug 05 14:21:06 ok. it's will need ~ 1 hour Aug 05 14:34:37 I have a very simple patch for package/kernel/modules Aug 05 14:34:40 http://openwrt.pastebin.ca/1519228 Aug 05 14:35:13 this will add more kernel modules for a usb ethernet (dm9601) and for motorola phone serial Aug 05 14:35:26 could somebody with write access dump that into trunk ? Aug 05 14:35:38 <[florian]> groz: ok I will handle this in a couple of minutes Aug 05 14:35:44 just adds a couple more kernel modules to make menuconfig Aug 05 14:35:48 no problem florian Aug 05 14:35:58 just that I'm gonna be gone starting tomorrow Aug 05 14:36:06 and I have a few little things worth committing before then Aug 05 14:36:23 <[florian]> how long will you vanish this time :p ? Aug 05 14:36:31 i'm on vacation for 3 weeks Aug 05 14:36:50 and at least half of that 3 weeks, completely out of touch, no cell coverage Aug 05 14:37:01 groz: me too, where are you going? Aug 05 14:37:18 starting off headed to cypress hills interprovincial park in saskatchewan Aug 05 14:37:21 there for a week Aug 05 14:37:27 then from there, headed to southern bc Aug 05 14:37:36 gonna be camped out at Mt Kobau for a wekk Aug 05 14:37:37 week Aug 05 14:37:50 and, probably a week on the road getting to/from it all Aug 05 14:38:10 ahhh ok america Aug 05 14:38:15 NO Aug 05 14:38:17 Canada Aug 05 14:38:22 which is america Aug 05 14:38:23 haha, big difference Aug 05 14:38:34 i didn't say us & a Aug 05 14:38:35 where are you ? Aug 05 14:38:36 :) Aug 05 14:38:42 * danage europe, going to toscany Aug 05 14:38:47 hehe Aug 05 14:38:56 toscany, that's on our 'to do someday' list Aug 05 14:39:02 i'll report back Aug 05 14:39:21 chris and I have another little trip that we want to do first tho Aug 05 14:39:30 it's supposed to be REALLY nice, i'm adding some natural spas/thermal springs on our to-do list as i'm typing Aug 05 14:39:35 starting here in canada, want to drive in a small motorhome, to chile Aug 05 14:39:50 now that's AMERICA Aug 05 14:39:55 uh huh Aug 05 14:40:31 we have started planning out the logistics, and, visa requirements etc Aug 05 14:40:58 then just have to figure out when she can take 6 months off work, and we can go do it Aug 05 15:09:23 florian * r17132 /trunk/package/kernel/modules/usb.mk: [package] add usb-serial-motorola-phone, thanks groz Aug 05 15:09:50 thanks florian Aug 05 15:10:31 florian * r17133 /trunk/package/kernel/modules/usb.mk: [package] add usb-net-dm9601-ether, thanks groz Aug 05 15:10:34 <[florian]> yw Aug 05 15:11:04 both of those are quite trivial, but, very useful for _me_, so likely useful for somebody else too Aug 05 15:12:08 <[florian]> that's why I usually apply such patches quickly since the review is easy Aug 05 15:12:37 yah, the stuff I got coming later today, not quite so simple :) Aug 05 15:13:09 but I'll send that all to juhosg, since he appears to be the keeper of ar71xx Aug 05 15:13:39 juhosg, are you awake this morning, and got a minute ? Aug 05 15:20:31 so I'm puzzling on a bit of an issue right now, I need to glue 2 files together, and deal with sector alignment Aug 05 15:20:51 but, I need the final result to be 'off by 32' to make space for the addpattern Aug 05 15:20:55 dd if=$(KDIR)/vmlinux-$(2).uImage bs=64k conv=sync; \ Aug 05 15:21:17 essentially, I want the resulting file from this one, to be 32 bytes shy of filling the last block Aug 05 15:21:38 I can kludge it by padding out with 65504 bytes Aug 05 15:21:48 but that ends up essentially wasting a flash sector Aug 05 15:23:48 any thoughts ? Aug 05 15:25:23 <[florian]> cannot you create an empty file of 32bytes and append it at the end of uImage, so that you do not pad out to 64KB ? Aug 05 15:25:52 well, problem is, addpattern is going to add 32 bytes at the beginning in the next step Aug 05 15:26:01 so, my rootfs will be 'off by 32' in flash Aug 05 15:26:13 so i'm trying to make this image 32 bytes short Aug 05 15:26:17 not 32 bytes long Aug 05 15:26:41 dd if=$(KDIR)/vmlinux-$(2).uImage bs=64k conv=sync; \ Aug 05 15:26:41 dd if=$(KDIR)/root.$(1) bs=64k; \ Aug 05 15:26:47 that's the full 'generate image' set Aug 05 15:26:53 <[florian]> oh ok Aug 05 15:26:58 where it's appending rootfs to kernel Aug 05 15:27:07 and that rootfs has jffs2 markers on the end Aug 05 15:27:15 i'm trying to make those markers align on the sector boundary Aug 05 15:27:20 now, one way that'll work Aug 05 15:27:36 is stick 65504 bytes between them Aug 05 15:27:46 so after addpattern runs Aug 05 15:27:52 the whole mess is again sector aligned Aug 05 15:27:55 for the rootfs Aug 05 15:28:26 this is an artifact of the cybertan stuff Aug 05 15:29:06 it all works just fine if i manually erase the rootfs_data after flashing Aug 05 15:29:15 but I'm just trying to make the 'magic' work now Aug 05 16:10:10 groz: it is not possible to append the rootfs after you had run the addpattern stuff? Aug 05 16:10:18 trx -o foobar.trx -h foobar.uImage Aug 05 16:10:18 addpattern -i foobar.trx -o foobar.tmp Aug 05 16:10:18 ( \ dd if=$(KDIR)/fooobar.tmp bs=64k conv=sync; \ dd if=$(KDIR)/root.$(1) bs=64k; \ Aug 05 16:10:22 ) Aug 05 16:10:25 hmm, why didn't I think of that Aug 05 16:10:26 haha Aug 05 16:10:34 i'll try it that way Aug 05 16:10:58 but wonder if that'll mess up the crc checks, only take a few minutes to check Aug 05 16:11:28 but i did just build up a bit of a kludge, where i'm padding out with the correct size between kernel and rootfs Aug 05 16:11:38 just flashing it now, it all looks good in a hexdump Aug 05 16:11:58 i think the _ideal_ solution will be to change the padding when adding jffs2 mark to the rootfs Aug 05 16:12:25 but i am running out of time before vacation Aug 05 16:12:35 so, wanted to get something that works, even if not ideal Aug 05 16:12:38 and, this one just worked Aug 05 16:12:48 I flashed it in, and it's correctly building the jffs for mini-fo Aug 05 16:12:56 but it does waste a sector Aug 05 16:13:19 it's also parsing the uboot header now to find stuff, and ignoring the trx altogether Aug 05 16:14:11 dealing with two issues here in reality, Aug 05 16:14:23 a) correctly find squashfs at boot when doing dynamic partitions Aug 05 16:14:41 b) mangle the bin file so that stuff is aligned after addpattern Aug 05 16:17:07 if the crc check fails we can try something else, but it is worth a try before we start thinking Aug 05 16:17:20 yes, i'm just doing the makefile changes now Aug 05 16:17:28 ok Aug 05 16:17:32 but, if this doesn't work, I do have now something that _does work_ Aug 05 16:17:49 and since this is an 8 meg flash device, for now, I'll accept wasting one sector to make it 'just work' Aug 05 16:18:59 we can optimize it later Aug 05 16:19:28 well, if we are happy with 'can optimize later', then, what I have now, will work, but, i have a philosophy question for you Aug 05 16:19:41 the parser patch for doing dynamic partitions out of the headeers Aug 05 16:19:50 currently requires a make kernel_menuconfig option Aug 05 16:20:02 should it be left that way, or, changed to 'just compile in' Aug 05 16:20:19 i'm thinking, this can probably be expanded to make the dynamic partitions work on all devices Aug 05 16:20:30 cuz it's only taking lengths out of the uboot header now Aug 05 16:20:56 only reason it's still trx specific, it's using the NL16 signature to identify 'where to find' the uboot stuff Aug 05 16:21:22 i think the exact same code would work on stuff without the trx header in there Aug 05 16:21:32 if it's assuming the uboot header at start of flash Aug 05 16:21:45 so, re-work this just a bit so it starts using 'beginning of flash' Aug 05 16:21:55 then detecting the NL16 offsets to find uboot header Aug 05 16:22:12 and once uboot pointer is set up, the exact same code for dynamic partitions will work on any device Aug 05 16:22:36 the only real difference between this one and the others Aug 05 16:22:41 the NL16 signature is wrt160nl specific Aug 05 16:22:45 is on this one, uboot header is offset 60 bytes in Aug 05 16:22:53 yes, so, works like this _should work_ Aug 05 16:23:02 uboot_header = offset 0 Aug 05 16:23:15 if (NL16 detect) uboot header = offset 60 Aug 05 16:23:23 now from here, parse it all based on uboot pointer Aug 05 16:23:29 and there are boards with RedBoot and MyLoader as well Aug 05 16:23:43 ahh, and they have different signatures at offset 0 ? Aug 05 16:24:25 * groz trying to make this more generic, ie, once we've located uboot header in flash, from there, we can do dynamic rootfs partition on any of them Aug 05 16:24:27 they don't use uImage Aug 05 16:24:31 ohhh Aug 05 16:24:31 ok Aug 05 16:24:36 then ignore me :) Aug 05 16:25:02 i was mistakenly thinking, they all used a uimage in a wrapper Aug 05 16:25:22 or some variation of that Aug 05 16:25:34 nope Aug 05 16:26:10 ok, seeing all those other utils in there in the makefile Aug 05 16:26:25 was just assuming, different vendors, different variations on the trx concept cybertan used Aug 05 16:27:08 trx is used only on the wrt160nl Aug 05 16:27:56 ok, would you consider this 'safe' as a detect Aug 05 16:27:59 master->read(master, 4 * master->erasesize, sizeof(buf), &len, buf); Aug 05 16:27:59 if(strncmp(buf, "NL16", 4) == 0) { Aug 05 16:27:59 printk(KERN_INFO "TRX on WRT160NL detected\n"); Aug 05 16:29:42 yes, this should be safe Aug 05 16:30:04 the thought I have here, if that is 'safe', this could be added in so that it's always there Aug 05 16:30:12 eliminating the need for make kernel_menuconfig Aug 05 16:30:20 for the nl build Aug 05 16:33:28 or possibly, can it work to add a kernel config option via the profile ? Aug 05 16:33:40 similar to how the profile brings along default packages ? Aug 05 16:33:50 why would you do that? Aug 05 16:34:20 cuz the way one of the patches is now, it's a wrt160nl specific moduel that needs to be 'compiled in' Aug 05 16:34:28 what module? Aug 05 16:34:36 the trx detection/ Aug 05 16:34:41 we can enable it in the generic ar71xx config Aug 05 16:34:41 parser for partitions and trx yes Aug 05 16:34:52 ok, i wasn't going to do it in generic Aug 05 16:34:56 that's small so overhead is minimal Aug 05 16:35:04 cuz the other devices are smaller, and its more 'kernel' size Aug 05 16:35:17 yes, it is very small, and if it's ok in the generic Aug 05 16:35:20 then point is moot Aug 05 16:35:46 ok, i'll be back in 5, then I'll clean this just a bit Aug 05 16:36:04 i will have the last patches needed for a 'just works' wrt160nl build shortly Aug 05 16:36:39 these two combined with what juhosg has already committed will turn it into a 'just works' platform Aug 05 16:39:20 groz: ok, the only question: what we can do with the SPI driver? Aug 05 16:42:19 well, i haven't looked at it, or touched it Aug 05 16:42:25 and things are just working Aug 05 16:42:44 are there issues I haven't bumped into yet ? Aug 05 16:43:23 one of the reasons I'd like to get what I have over to you sometime today is this, tomorrow I leave Aug 05 16:43:33 and i'll be out of the loop for 3 weeks Aug 05 16:44:15 and i've reached my first interim goal, got it up and running, working as an access point in the van, with the 3g modem Aug 05 16:44:24 internet to go in the camper Aug 05 16:44:27 for our trip Aug 05 16:46:52 iirc, you had to revert this change: https://dev.openwrt.org/changeset/16767/trunk/target/linux/ar71xx/files/drivers/spi/ar71xx_spi.c Aug 05 16:47:06 oh right, I forgot about that Aug 05 16:47:08 ooops Aug 05 16:47:39 and i'm sure, reverting that will likely break something else wont it ? Aug 05 16:48:30 i guess I should take a peek in that part, see if there is some global var available on which to check Aug 05 16:48:52 and then wrap that particular call inside an 'if (some condition) Aug 05 16:49:35 lemme finish checking your other idea on trx wrapping first Aug 05 16:51:40 actually, what I got now is working good, optimizing for that extra sector can wait Aug 05 16:57:14 you there Bartman007? Aug 05 17:04:33 groz: the spi_driver should be working. if it does not, it has a bug then, and i would like to fix it before users are starting to flash openwrt into their wrt160nl Aug 05 17:07:07 yah, i agree ju, i had forgotten that one, it's been a very busy few days here, and I'm constantly slipping back to play with this as I can Aug 05 17:07:08 groz: can you try this patch: http://openwrt.pastebin.ca/1519376 ? Aug 05 17:07:13 i'm just looking at the spi now Aug 05 17:07:36 yep, i can try that, only take a few minutes Aug 05 17:07:50 ok Aug 05 17:08:31 now with the trx stuff, should I get you a modified version of christiens patch Aug 05 17:08:50 or, should I look at incorporating it into the files stuff in the ar711 tree Aug 05 17:08:54 in the mtd part Aug 05 17:09:31 modified variant of christians original patch is somewhat easier for me Aug 05 17:11:16 why is the new fuse24 (kernel module for linux 2.4 kernels) package using only fuse 2.5.3? Aug 05 17:12:09 because not many people care about updating packages for linux 2.4 Aug 05 17:13:39 oh, and before i forget, juhosg, I'm gonna get you a full flash dump from a virgin wrt160nl sometime today, and I'll zip the whole thing up into a spot you can download at your lesiure Aug 05 17:14:25 nbd: I'm considering either updating this package, or possibly getting a new (preferably cheap) device that has usb, wifi, and 2.6 kernel support. Got any suggestions? Aug 05 17:14:39 what kind of device do you have? Aug 05 17:15:12 groz: the modified patch would be perfect, and thank you in advance for the dump Aug 05 17:15:32 This is my device: http://lithostech.com/openwrt-wifi-radio-part-1 asus wl-520gu Aug 05 17:16:12 dunno which devices are cheap these days Aug 05 17:18:34 steve, define 'chea' Aug 05 17:18:36 cheap Aug 05 17:19:00 cheap as humanly possible... far less than 100USD preferably Aug 05 17:19:14 ok, then my suggestion is no good Aug 05 17:26:38 hello guys!!!, I need some teach me something Aug 05 17:27:09 groz: what were you going to suggest? Aug 05 17:27:10 I'm trying to write a patch Aug 05 17:27:27 and everthing go well Aug 05 17:27:47 juhosg * r17134 /trunk/package/iw/Makefile: [package] iw: update to 0.9.15 Aug 05 17:27:54 i've been working on cleaning up patch sets for WRT160nl steve, and it meets all of those requirements, I want it specifically for the usb port with my usb attach 3g modem Aug 05 17:28:26 and i have that running now, just working out the details of getting it all to build correctly and run all the hardware on board Aug 05 17:28:26 but in /target/.... / dir I don't see the Makefile of openwrt and the /files dir Aug 05 17:28:51 looks like i could pick that up for about $100... I'll consider it Aug 05 17:29:00 yah, it's around $100 Aug 05 17:29:12 and it's 'real new', so likely to be issues on it for a while Aug 05 17:29:24 * groz not gonna reccomend it unless you are handy with serial console Aug 05 17:29:43 and comfortable patching things up and flashing from serial console Aug 05 17:29:50 serial headers pretty easy to add to that device? Aug 05 17:29:54 and I need to know how modify Makefile and /files dir from my patch, any one can help me? Aug 05 17:29:57 it comes with serial header on board Aug 05 17:30:03 just open it up, and plug in Aug 05 17:30:22 no soldering required, the header has pins Aug 05 17:30:23 nice Aug 05 17:30:40 * groz like it so far, but Aug 05 17:30:51 i'm also not skeered of bricking them Aug 05 17:31:03 i bought 5 'just in case' Aug 05 17:31:16 and, i got a bit of a deal on a stack of 5 Aug 05 17:31:42 yeah, I'm more interested in developing my own software, rather than spending my time fixing openwrt bugs :) Aug 05 17:32:19 guess that depends what you call a bug, getting drivers working on a new platform is not 'bugs' to me, it's 'moving forward' Aug 05 17:33:43 good point groz, and stevecrozz If openwrt help you why not you help to openwrt? Aug 05 17:34:21 I think I am helping, I just got s3fs compiled for openwrt and created a makefile Aug 05 17:34:48 what kindd of software you putting on the box ? Aug 05 17:35:26 I'm going to build a "cloud player" which would be a music player that uses s3fs for storage Aug 05 17:36:20 basically just mpd / s3fs with a custom hardware interface that I have yet to build Aug 05 17:37:00 xMff: around? Aug 05 17:40:08 Anyone got a ballpark figure on what kind of space a complete trunk compile takes Aug 05 17:40:29 how many platforms you gonna compile for ? Aug 05 17:41:26 juhosg: That spi patch you just sent me seems to have done the trick Aug 05 17:41:39 i nuked the entire kernel tree, applied that patch, built from scratch Aug 05 17:41:51 it's detected the correct flash now Aug 05 17:42:12 now I gotta merge back my trx patch and it should be into 'just works' territory Aug 05 17:43:30 [florian]: ping Aug 05 17:44:43 anybody know about support for a peplink micrel-ks8695 (arm922tid) board? i'm telnet'd into one running linux right now. Aug 05 17:44:57 groz: ok, thanks. i have to test the spi patch on several boards before i commit it Aug 05 17:47:18 that makes sense Aug 05 17:53:19 yeah, it must be working on AR7130, AR7161, AR9130, AR9132 and AR7240 SoCs Aug 05 18:06:32 nbd: ping Aug 05 18:58:42 Kaloz, are you still awake ?? Aug 05 19:36:43 <[Fate]> a general kernel-related question: is the order in which static device drivers get called by the kernel in any way determinable? Aug 05 19:36:56 <[Fate]> eg. when the parport code gets called, when the ide driver gets called etc.? Aug 05 20:01:42 groz: One target, I want to build the entire packagespace for brcm43xx with 2.6.30.4, but I only have like 10gb to do so on Aug 05 20:04:53 kupesoft: it's somewhere in the 10-12 GB range last time I checked. Aug 05 20:05:18 Why is the architecture for brcm43xx now "brcm43xx" instead of mipsel? Aug 05 20:06:01 strange, brcm-2.4 is now "brcm-2.4" instead of mipsel as well Aug 05 20:06:07 the 8.09.1 release trees take up 183GB IIRC, and that's for 18 targets. Aug 05 20:06:40 Why change the arch name? Aug 05 20:06:44 Is that a bug Aug 05 20:07:02 kupesoft: the change was intentional to allow further seperation/optimization of individual targets. Aug 05 20:07:02 Since aug 4th on downloads.openwrt.org/snapshots Aug 05 20:07:09 okay Aug 05 20:07:17 it breaks my tree kind of, I have recompile everything Aug 05 20:07:50 no biggy Aug 05 20:09:17 juhosg, I just emailed you my 160nl patches, so, after you test that spi patch against the other boards Aug 05 20:09:25 this should fall into the 'just works' category Aug 05 20:09:33 squash and jffs roots both tested here Aug 05 20:09:57 root@OpenWrt:/# uname -a Aug 05 20:09:57 Linux OpenWrt 2.6.30.4 #1 Wed Aug 5 00:55:18 EDT 2009 mips unknown Aug 05 20:09:57 root@OpenWrt:/# /usr/sbin/dbus-daemon --system Aug 05 20:09:57 Segmentation fault Aug 05 20:10:01 by flashing tftp thru the uboot console, altho I did do bin files to the linksys firmware via web interface the other day Aug 05 20:14:31 Looks like I'm fixing dbus. Aug 05 20:19:26 or at least trying :( Aug 05 20:23:04 ping Hopscotch Aug 05 20:27:09 hmm, sort of lost with dbus here Aug 05 20:27:55 I probably won't be of much help, but have you tried stracing it to see where it segfaults? Aug 05 20:30:27 hi Aug 05 20:30:33 can i build a ramdisk image after i built a normal one quickly? Aug 05 20:30:41 or do i need to recompile the whole kernel? Aug 05 20:31:56 Bartman007: I just hand chomped basically all the CFLAGS while compiling dbus-daemon and it works Aug 05 20:31:57 hmmmm Aug 05 20:33:24 kupesoft: but dbus isn't compiled with any extra cflags from what I can see, did you remove the defaults? Aug 05 20:33:36 ah Aug 05 20:34:00 just changing it to ramdisk in menuconfig worked fine Aug 05 20:34:15 it went through everything but no recompile Aug 05 20:34:37 freezer_: it should keep all the compiled kernel objects around, so it is a quick make run Aug 05 20:34:53 run? Aug 05 20:35:27 run == iteration, ie: `make' Aug 05 20:35:46 Bartman007: Yeah, Aug 05 20:35:58 No, I looked at the log of dbus compiling Aug 05 20:36:04 Took the line where dbus-daemon is compiled and linked Aug 05 20:36:08 MIPS: machine is Generic AR71xx board Aug 05 20:36:20 cd'd in to the build dir, and ran gcc without any cflags Aug 05 20:36:25 it doesnt detect WRT160NL Aug 05 20:36:52 are you doing a tftp boot from uboot freezer ? Aug 05 20:37:02 freezer_: AFAIK we haven't found any clean way to detect the wrt160nl hardware Aug 05 20:37:27 bartman, the patches I just sent off to juhosg a few minutes ago Aug 05 20:37:30 have that in them Aug 05 20:37:42 i set board=WRT160NL in uboot Aug 05 20:37:43 groz: oh, no more hardcoded BOARD value? Aug 05 20:37:43 he needs to test some of the spi stuff on other boards before it's all committed Aug 05 20:37:49 yay :) Aug 05 20:37:57 haha well Aug 05 20:38:17 i have to go back and double check, it may still be wired into the kernel command line stuff Aug 05 20:38:21 but, shouldn't be needed Aug 05 20:38:30 detecting now based on the NL16 header on the flash Aug 05 20:38:42 Bartman007: http://dpaste.com/75713/ Aug 05 20:38:46 ^^ anyone else who knows dbus Aug 05 20:39:03 groz: ethernet still not working? Aug 05 20:39:10 I wonder what argument is doing this... Aug 05 20:39:13 everything working freezer Aug 05 20:39:23 the patches for ethernet were committed a couple days ago Aug 05 20:39:36 i built yesterday Aug 05 20:39:43 did i forget them in menuconfig? Aug 05 20:39:46 mine is running here, as an access point, ethernet attached to my local lan Aug 05 20:40:00 and it's got internet connection thru a usb attach 3g modem Aug 05 20:40:01 or is it because it doesnt detect it as an WRT160NL Aug 05 20:40:03 I just arbitrarily chomped the some of those cflags, hmm Aug 05 20:40:12 s/the some/some/ Aug 05 20:40:14 it's because it's not detecting wrt160nl Aug 05 20:40:37 how do i need to boot? Aug 05 20:40:40 setenv bootargs 'board=WRT160NL'; tftpboot 0x81000000 openwrt-ar71xx-uImage-gzip.bin; bootm Aug 05 20:40:53 ok gonna check Aug 05 20:40:59 i just cut and paste that into the uboot console Aug 05 20:41:04