**** BEGIN LOGGING AT Mon Jan 23 02:59:57 2012 Jan 23 03:28:40 is there a way to detect wrt54g version from openwrt commandline? Jan 23 03:28:50 * russell-- is looking for a v1.0 Jan 23 03:29:08 (needs a couple rp-tnc pigtails) Jan 23 03:39:34 build #113 of ppc44x is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/113 Jan 23 07:02:15 build #116 of rb532 is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/116 Jan 23 07:02:16 build #111 of uml is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/111 Jan 23 07:02:19 build #112 of x86 is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/112 Jan 23 07:02:21 build #108 of ar7 is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/108 Jan 23 07:52:18 build #106 of avr32 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/106 Jan 23 08:53:48 juhosg * r29869 /trunk/target/linux/ar71xx/files-2.6.39/arch/mips/ar71xx/mach-eap7660d.c: ar71xx: fix compiler warning if PCI is not enabled Jan 23 08:53:49 juhosg * r29870 /trunk/target/linux/ar71xx/ (2 files in 2 dirs): ar71xx: 3.2: fix build errors if PCI is not enabled Jan 23 08:59:24 thank you Gabor Jan 23 10:24:59 juhosg * r29871 /trunk/target/linux/ramips/ (6 files in 5 dirs): (log message trimmed) Jan 23 10:24:59 ramips: scripts updates Jan 23 10:24:59 In this patch: Jan 23 10:24:59 * rename Argus leds to avoid underscores Jan 23 10:24:59 * rename Belkin F5D8235 v1 leds from f5d8234 to f5d8235 Jan 23 10:25:00 * remove Belkin F5D8235 v1 status led defined as storage led (it was Jan 23 10:25:00 defined as usb led earlier, just in wrong place) - it should have Jan 23 12:28:24 swalker * r29872 /packages/net/dmapd/ (Makefile files/dmapd.init patches/001-dmapd_conf.patch): [packages] dmapd: update to 0.0.46 (#10734, thanks flyn), create DPAP directory, refresh patches Jan 23 12:33:04 swalker * r29873 /packages/libs/boost/ (Makefile patches/100-do-not-use-librt.patch): [packages] boost: fix compile failure (#10535, thanks zhenglei), refresh patches Jan 23 13:56:39 jow * r29874 /packages/ (4 files in 4 dirs): Jan 23 13:56:39 [packages] rrdtool, rrdtool-1.0.x: backport upstream commit r2069 Jan 23 13:56:39 Explicitely cast float cookie value to double, fixes erroneous Jan 23 13:56:39 "RRD was created on another architecture" errors on x86. Jan 23 13:56:39 Patches provided by Thomas Heil Jan 23 15:54:00 jogo * r29875 /packages/net/znc/ (5 files in 2 dirs): packages: znc: update to 0.204 Jan 23 16:59:26 moo Jan 23 16:59:30 Chocks: :) Jan 23 16:59:56 Chocks: the patchset is located where? sorry in case i already asked you that (propably more than once..) Jan 23 17:00:40 mirko: btw. i now have a phone for porting openwrt to Jan 23 17:00:52 when i get around to doing more android/openwrt work Jan 23 17:01:17 nbd: a phone is a start ;) Jan 23 17:01:26 nbd: when you're gonna start hacking some day, just gimme a call Jan 23 17:01:34 sure Jan 23 17:01:47 the nice thing about this phone is that much of the SoC code is already mainline Jan 23 17:01:56 nico recently claimed the bug, but hasn't done anything, I don't think. trac not responding right now. it's ticket 9483 Jan 23 17:01:58 nbd: which one? Jan 23 17:02:01 i just found a source for ifxhcd Jan 23 17:02:02 (exynos4) Jan 23 17:02:06 nbd: just don't make the huge mess that OpenMoko did. Jan 23 17:02:08 its the lantiq internal usb driver Jan 23 17:02:16 the phone is samsung galaxy s2 Jan 23 17:02:17 you couldn't have more mess/branches/UIs if you tried Jan 23 17:02:25 Chocks: he won't and i won't let that happen as well ;) Jan 23 17:02:42 I presume you are familiar with what they did ;-) Jan 23 17:02:54 yep, we are Jan 23 17:03:03 Chocks: 17:58 < mirko> Chocks: the patchset is located where? sorry in case i already asked you that (propably more than once..) Jan 23 17:03:25 * Chocks contributed to that project, but my stuff got stomped all over. too much focus on not actually having the phone work Jan 23 17:03:37 "trac not responding right now. it's ticket 9483" Jan 23 17:03:47 Chocks: ah, that belonged to me.. Jan 23 17:04:08 504tastic Jan 23 17:04:20 github seems to be down as well - at least for me Jan 23 17:04:53 nbd: is the goal general hacking, or some particular aim? Jan 23 17:05:15 at some point i want to have a phone that runs a software stack that's under my control Jan 23 17:05:19 with openwrt as base system Jan 23 17:05:21 Chocks: building own images with a working toolchain might be a goal :> Jan 23 17:05:27 and i want it to be usable ;) Jan 23 17:05:47 so, the goal of OpenMoko, but with sane hardware and saneish software base Jan 23 17:06:14 well, openmoko was based around the idea of building free hardware Jan 23 17:06:29 i don't want to be involved with the hardware stuff and just focus on the software part Jan 23 17:06:38 OM wanted to build it's own software stack and UI Jan 23 17:06:40 yes, I think that idea is a fail, at least for the foreseeable future. leave that to people who can do it well Jan 23 17:06:46 that's what already present with android Jan 23 17:07:24 but you still need dialer and what not to talk to hardware you don't find on normal OpenWrt hardware Jan 23 17:07:26 i think there's a chance that OM played a similar role with establishing demand for free software in phones as OLPC did for netbooks Jan 23 17:07:26 i appreciate what OM did.. many mistakes i'm not gonna do twice Jan 23 17:07:29 that's not even ironic Jan 23 17:07:32 i'm really glad Jan 23 17:07:42 so OM may have failed, but it was still important Jan 23 17:07:42 what about UI and apps? Jan 23 17:07:51 nbd: sure Jan 23 17:07:58 they showed up what's not gonna work - and it wasn't totally foreseeable in many ways Jan 23 17:08:16 well, imho they did get quite far Jan 23 17:08:25 they sold a product Jan 23 17:08:29 not just one Jan 23 17:08:29 I think the main problem from OM was weak project management. no obvious direction Jan 23 17:08:35 they just had more different ideas than people implementing them ;) Jan 23 17:09:01 the sound on my phone never worked properly, then worked even less properly after I got the fix Jan 23 17:09:09 Chocks: which phone? Jan 23 17:09:11 Ax? Jan 23 17:09:32 OM's phone, whatever it was called Jan 23 17:09:41 Chocks: there are quite some revisions Jan 23 17:10:11 I was very early in the large batch they did when they really started selling them Jan 23 17:10:43 Chocks: i see.. i guess it was a A5 - miles away from the last revision (which btw didn't ever got sold :P) Jan 23 17:10:54 that sounds likely Jan 23 17:11:46 hm, they are still at it Jan 23 17:12:33 only 666 EUR. ugh Jan 23 17:12:51 that price sounds evil Jan 23 17:12:51 anyway Jan 23 17:12:53 ;) Jan 23 17:13:06 there's a needed fix to sysupgrade for e/glibc Jan 23 17:13:21 what eglibc version are you guys using? Jan 23 17:13:36 just building 2.14 now to try. been using 2.13 for months Jan 23 17:13:57 i think having 3 versions of it are a bit much, it's probably time to fix and focus on one Jan 23 17:14:04 probably the one piece of openwrt I haven't had problems with Jan 23 17:14:29 although you can get into weird build states if you don't select certain options in eglibc Jan 23 17:15:08 Chocks: indeed Jan 23 17:15:48 and I'm using stuff like POSIX messaging, which you need to turn on there and in the kernel Jan 23 17:30:51 moo Jan 23 18:23:33 [florian]: ping Jan 23 18:30:26 nbd: what's involved in getting sysupgrade to run on x86? I'm almost ready to finance the work myself... Jan 23 18:30:58 well, it needs a good plan and somebody to implement it Jan 23 18:32:21 ok, who's most knowledgeable about the necessary parts? Jan 23 18:38:29 you mean sysupgrade with ext? filesystems? sysupgrade with squashfs image works fine already on x86 Jan 23 18:38:58 there are two main issues (I think I summarized it here a while ago) Jan 23 18:39:08 there is no ext* equivalent for mtd jffs2write Jan 23 18:39:11 you did :) Jan 23 18:39:27 and there is no way to tell in advance whether the to-be-flashed image is ext* or squash Jan 23 18:40:14 2 is solvable by using the proposed combined extended image format Jan 23 18:40:35 1 no idea, I suppose one could lift some code from mke2fs or so Jan 23 19:23:38 which x86 platforms use flash hardware that's not SD/other abstraction? Jan 23 19:45:13 build #106 of ixp4xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/106 Jan 23 20:04:23 blogic: are you about? Jan 23 20:06:12 this nand stuff id driving me up the wall :( Jan 23 20:07:07 well Jan 23 20:07:10 what can i say Jan 23 20:07:19 have you tried turning it on and off again ? Jan 23 20:07:23 :-D Jan 23 20:07:34 :) i've tried everything in my book Jan 23 20:07:44 i honestly have no clue Jan 23 20:07:49 i will have time soonish Jan 23 20:07:55 we could try to get me a unit Jan 23 20:08:34 i'm contemplating rebuilding u-boot with diags for comparison. At least I know I can build a u-boot where nand works. Jan 23 20:09:23 where are you again? Jan 23 20:13:17 did you see my mail ref PCI? Jan 23 20:15:20 no Jan 23 20:15:30 what is the problem with PCI ? Jan 23 20:17:31 have a read of mail from yesterday. comments from nand driver from openrg source indicate they mask all PCI requests whilst nand is selected :( Jan 23 20:19:01 hmm Jan 23 20:19:13 show me the openrg source that does that please Jan 23 20:20:50 Chocks: SD (secure digital) abstraction? Flash on an x86 is usually an IDE emulation. Jan 23 20:21:58 or did you mean /dev/sdX (scsi) emulation? Jan 23 20:22:03 https://sourceforge.net/projects/hh2b4ever/files/Original%20Open%20Source/ - i think you pulled this before. in danube_nand for the disable/enable. one moment for where the comment is, i'll find it... Jan 23 20:23:31 pastebin it please ;) Jan 23 20:24:11 danube_nand.c: http://pastebin.com/ypWFw6pk Jan 23 20:26:34 pci.c: http://pastebin.com/FjLhyuaF - not where the comment is, but interesting there is mention of EBU..... Jan 23 20:28:39 reset.c http://pastebin.com/Be8wEt84 - still not found the comment, but note the disable of something in pci before nand reset.... Jan 23 20:29:42 build #104 of kirkwood is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/104 Jan 23 20:31:58 ahh finally, command in u-boot memsetup.S: http://pastebin.com/cPksNMKd Jan 23 20:35:03 hi guys. i'm trying to pivot /overlay with usb flash drive, but it just gets mounted to /tmp/overlay-disabled. how to debug that? Log: http://pastebin.com/71p2d4zP Jan 23 20:35:27 mbs_: do you have a jffs2 at all? Jan 23 20:35:54 build #106 of xburst is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/106 Jan 23 20:36:10 in kernel: yes. as partition (rootfs_data): no Jan 23 20:37:07 btsimonh: do you have internal pics of your router? Jan 23 20:37:18 i tried for ages and could not get it to work :( (or get clear answers). I have no flash at the moment, so thought I could use USB, but did not manage it. Jan 23 20:37:43 Delboy: here: http://wiki.openwrt.org/toh/bt/homehub_v2b#gpos Jan 23 20:39:22 ok, do you have power on usb port by setting that gpio 13? Jan 23 20:40:02 blogic: no, DISABLE_CFRAME is NOT defined, so another red herring on the comment. BUT PCI_CR_PCI seems set for a reason unknown to the developer. Jan 23 20:41:21 Delboy_: yes, USB power was on 13; different to sx source I copied. Jan 23 20:42:00 i had it working with a build like 3 weeks ago, but i don't even get any error message from blockmount to see what's wrong Jan 23 20:42:46 phillip: I'm well aware of that. my question is, what are the x86 platforms that use jffs2 etc? Jan 23 20:42:58 all Jan 23 20:43:30 there are squashfs+jffs2 images that run on top of block2mtd and there are ext* images Jan 23 21:29:21 mirko: build fail Jan 23 21:30:08 http://pastebin.com/dKc4VK7g Jan 23 21:56:46 Delboy_: you don't have NAND , just NOR, correct? Jan 23 22:01:53 btsimonh: yes Jan 23 22:02:01 i have the same board so i know this Jan 23 22:02:42 :) - i could not find nand in danube u-boot sources in openwrt, but i guess most don't need it. Jan 23 22:16:08 u-boot is not built for most platforms Jan 23 22:26:42 philip: over here Jan 23 22:27:01 I'm not sure what you're asking. Jan 23 22:27:35 on which platforms would you be using, e.g, the mtd tool and jffs2 images? Jan 23 22:27:39 which x86 Jan 23 22:27:46 most x86 platforms have standard BIOS support, so their bootstrap loader (Coreboot+SeaBIOS, Award, etc.) can you whatever you want them to. Jan 23 22:28:00 um Jan 23 22:28:13 I use Geos, Alix2, Fish River Island, Soekris... Jan 23 22:28:14 what does that have to do with anything Jan 23 22:28:33 the last Soekris stuff I use had CF cards, which was via the IDE interface Jan 23 22:28:43 right, which is what I said earlier. Jan 23 22:28:51 the CF's emulate IDE. Jan 23 22:28:56 I know that. Jan 23 22:29:00 and I said I know that Jan 23 22:29:07 but it's nothing to do with what I'm asking Jan 23 22:29:41 openwrt in most cases on x86 would be using ext2/3 etc, right? Jan 23 22:30:37 well, you're assuming I know enough to know whether ext2 or ext3 wins me anything over mtd and jffs2. Jan 23 22:30:42 frankly, I have no idea. Jan 23 22:30:42 no Jan 23 22:31:11 I'm asking if there's any platforms on x86 where the filesystem isn't over USB/ATA/IDE, etc, etc Jan 23 22:31:20 and is access via flash, as per most routers Jan 23 22:31:31 for the root-overlay, yes, it would be using ext3. Jan 23 22:31:50 i like jffs2 even for x86 devices Jan 23 22:31:55 because it has one major advantange Jan 23 22:32:02 the filesystem has no problem with sudden power outage Jan 23 22:32:08 it's been designed to handle that well Jan 23 22:32:11 the only one I know of is the FRI (Fish River Island) which does SD NAND or motherboard NAND. Jan 23 22:32:41 ok. I was wondering about the point that jow raised earlier about figuring out which one to upgrade - ext vs jffs2 Jan 23 22:32:53 jffs2 can often even deal with interrupted writes properly Jan 23 22:32:57 because it's a log-structured filesystem Jan 23 22:33:05 nbd: jffs2 on top of what? Jan 23 22:33:09 block2mtd Jan 23 22:33:24 on top of any block device Jan 23 22:33:28 umm.... upgrade how? I'm assuming that sysupgrade just overwrites the root RO image with a new one.. yes? Jan 23 22:33:34 yes, ok. I was thinking something else. Jan 23 22:33:41 umm...yourself Jan 23 22:33:47 I'm simply asking questions Jan 23 22:33:58 sysupgrade overwrites the RO image and appends jffs2 data containing the config Jan 23 22:34:14 it can only do this append process for squashfs+jffs2 or for jffs2 Jan 23 22:34:22 ummm is me thinking aloud trying to remember and wondering if I should trust my recollection. Jan 23 22:35:00 so it doesn't support multiple partitions in the flash then? Jan 23 22:35:15 huh? Jan 23 22:35:29 it's possible, if you have enough ram, to download new image, mount it locally, pivot_root, reformat rootfs, copy new system to it Jan 23 22:35:39 because otherwise you could just have the ext3 start on /dev/sda2 and have /dev/sda1 be the space for the squashfs. Jan 23 22:35:47 I wrote a system like that for a soekris board Jan 23 22:36:03 Chocks: right... ram is usually not an issue on x86 platforms. Jan 23 22:36:11 I've yet to see a box with less than 128MB. Jan 23 22:36:24 philipp64|laptop: the nice thing about the way sysupgrade works on non-x86 stuff is that it works even when the split between RO and RW parts changes Jan 23 22:36:28 and my largest squashfs image so far has been about 57MB. Jan 23 22:37:03 nbd: that sounds nice, but I've got without *any* way to do an in-place upgrade that I'm desperate. Jan 23 22:37:16 I frankly wouldn't care if it wasted 50% of my CF's capacity. Jan 23 22:37:23 *gone Jan 23 22:37:33 so sysupgrade doesn't work with squashfs images on x86 right now? Jan 23 22:37:55 I tried to get it to work about 2 years ago and it bricked my system. Jan 23 22:37:58 in the case I was on about, there was more than enough room on the CF for a "factory" file system. something like I'm trying to do now on ar71xx Jan 23 22:38:10 although grub did tricky bits there Jan 23 22:38:49 i'd like to see someone propose a sane system for in-place upgrades that can deal with changes in the layout Jan 23 22:38:57 actually, I had an Airlink router some years ago that was x86 Jan 23 22:39:04 but it was u-boot etc Jan 23 22:39:11 nbd: send me an email with a set of steps to test (just to make sure I'm not botching it) and I'll send you the logs and report on success. Jan 23 22:39:35 but as I remember trying to do the sysupgrade as described on the wiki, it failed miserably. Jan 23 22:40:02 steps to try: flash a squashfs image, boot it, try running sysupgrade on another squashfs image Jan 23 22:40:14 nothing special Jan 23 22:40:21 I'll donate a box (and Alix 2D13 with 256MB and 2G flash) to anyone who commits to fixing this. Jan 23 22:40:42 so if lack of hardware is the blocking issue, I'm happy to unblock that. Jan 23 22:40:43 i don't really have time to come up with a good x86 in-place upgrade system Jan 23 22:40:53 but i'm willing to help other people develop such a thing Jan 23 22:41:01 not by writing code, but by working out the idea Jan 23 22:41:42 yeah, Daniel was talking about what was involved for about 6 months and then went silent about 3 months ago... so I'm guessing he decided not to do it. Jan 23 22:42:07 he seemed convinced that the coding was the easy part... Jan 23 22:42:22 working out what to do and handling the corner cases seemed like the issue to him. Jan 23 22:43:04 btw, on FRI the the NAND was 2 or 4 GB as I remember. Jan 23 22:43:38 I just don't understand why x86 requires a different solution from the other boxes. Jan 23 22:43:44 *architectures. Jan 23 22:44:06 mostly because x86 uses block devices with partition tables Jan 23 22:44:16 and because x86 mostly uses grub etc Jan 23 22:44:39 i wish i had time to replace that crap with something sane Jan 23 22:44:41 right... but.... you only really need grub if you support booting multiple os's right? Jan 23 22:44:53 I mean you could use syslinux and kexec if you wanted... Jan 23 22:45:01 you need a bootloader of some kind. Jan 23 22:45:12 syslinux isn't a bootloader? Jan 23 22:45:23 but they are rather different from u-boot etc in nature, due to BIOS Jan 23 22:45:34 x86 is a special child Jan 23 22:46:51 ok... I would have thought that having BIOS provide low-level functionality would make things easier, not harder. Jan 23 22:47:02 I guess I'm just flaunting my ignorance. Jan 23 22:48:01 heh, BIOS making things easier Jan 23 22:48:03 nbd: what is the install procedure right now for openwrt on x86? Jan 23 22:48:12 just a tar.gz? Jan 23 22:48:21 Chocks: dd an image to a cf card or sometjing Jan 23 22:48:28 oh right Jan 23 22:48:35 build can spit out ext2 etc image Jan 23 22:48:39 yeah Jan 23 22:48:47 took me a while to make that work under os x ;) Jan 23 22:48:51 grub was the hard bit Jan 23 22:49:11 nbd: I don't understand Jan 23 22:49:19 what don't you understand? Jan 23 22:49:32 the OS X/grub issue Jan 23 22:50:11 parts of grub are built with the host toolchain, parts with the cross toolchain Jan 23 22:50:20 right, comedy Jan 23 22:50:22 because there's a grub program that needs to run to control the installation Jan 23 22:50:25 not something I've ever built myself Jan 23 22:50:29 which actually calls *target* code to do the wirk Jan 23 22:50:30 work Jan 23 22:50:31 ;) Jan 23 22:51:05 even on some linux distros it was tricky Jan 23 22:51:34 at some point i figured out that some kernel security things (i think PaX) were also causing it to fail Jan 23 22:51:53 and i had to write my own partition table generator Jan 23 22:51:57 because the existing tools sucked Jan 23 22:52:01 what kind of nonsense was the grub tool doing? Jan 23 22:52:25 well it has to make stage 1 find stage 2 Jan 23 22:52:28 and stage 2 find stage 3 Jan 23 22:52:37 so it does some things that depend on the geometry Jan 23 23:48:25 anyone awake :) Jan 23 23:48:46 i'm wondering what is behind leds definition: http://wiki.openwrt.org/doc/uci/system#examples Jan 23 23:49:23 ie which proces/daemon takes care of the leds configuration written in /etc/config/system Jan 23 23:54:47 Chocks and nbd: just tell me where to mail the alix2 system ... :-) Jan 23 23:55:31 luka: LEDs are mostly driven during start up. they are handled by the kernel Jan 23 23:55:37 yes Jan 23 23:55:41 just found out Jan 23 23:55:42 package/base-files/files/etc/init.d/led Jan 23 23:55:58 Chocks: zcat openwrt-x86-generic-combined-squashfs.img.gz | dd of=/dev/sdc (or something like that) works Jan 23 23:55:59 just wrote some code here to mess with them Jan 23 23:56:23 Chocks: you know which kernel part is responisble for this Jan 23 23:56:36 russell: yes, have used squash on x86 in the past. Jan 23 23:56:53 each platform registers LEDs, but I don't know where the core is Jan 23 23:57:24 ok, thx Jan 24 00:00:31 btw. i'd like to get some more testing feedback on netifd Jan 24 00:00:41 * russell-- just did a half dozen of CF images last night for some old net45xx's Jan 24 00:00:42 as soon as i've implemented the missing bits, i want to enable it by default Jan 24 00:00:47 my build is not bringing up br-lan Jan 24 00:00:48 wah Jan 24 00:00:49 wah Jan 24 00:00:50 wah Jan 24 00:01:38 nbd: i have tested it with my current config and it did not work with it Jan 24 00:01:48 i can paste you the config if you want Jan 24 00:01:48 what kind of config? Jan 24 00:01:50 and which version? Jan 24 00:01:59 i recently updated netifd Jan 24 00:02:10 yes i think it was before the update Jan 24 00:02:13 ok Jan 24 00:02:15 where can we read more about it? Jan 24 00:02:18 let me paste you the config Jan 24 00:02:32 russell--: not much to read yet Jan 24 00:02:35 russell--: there is a wiki page somebody posted Jan 24 00:02:41 let me give you a linke Jan 24 00:02:43 there is? interesting Jan 24 00:02:46 please! Jan 24 00:03:02 i'd like to understand what i'm testing ;-) Jan 24 00:03:06 it's going to replace all those ugly network scripts Jan 24 00:03:13 nice Jan 24 00:03:20 https://forum.openwrt.org/viewtopic.php?id=34220 Jan 24 00:03:28 takes you to: http://wiki.openwrt.org/doc/techref/netifd Jan 24 00:04:52 nice to see somebody turning my words from an irc discussion into a wiki page Jan 24 00:05:02 nbd: here is my config http://pastebin.com/r5WkA92c Jan 24 00:05:24 and what didn't work Jan 24 00:05:24 i'm using lantiq target and switch config is for adm6996i chip Jan 24 00:05:28 how did it fail? Jan 24 00:05:55 well i was debuggind some other issue and i did not look into it Jan 24 00:06:10 i can test again tomorrow Jan 24 00:06:17 that would be good Jan 24 00:06:21 please ping me again when you do Jan 24 00:06:27 ok Jan 24 00:06:34 but you think that config should work? Jan 24 00:06:47 i don't see anything wrong with that config Jan 24 00:06:56 well it's working now Jan 24 00:07:06 so when you ran netifd, was lan working Jan 24 00:07:11 or did it break all interfaces? Jan 24 00:07:48 one sec Jan 24 00:09:12 http://pastebin.com/vRsBtr9a Jan 24 00:09:22 note the option nat Jan 24 00:09:38 the problem i was debugging other day was Jan 24 00:10:03 that i could not get arp responsed from the router back to my pc when there was this option Jan 24 00:10:21 weird Jan 24 00:10:26 because of that i could not get dhcp or ping the router Jan 24 00:10:27 yes Jan 24 00:10:40 firewall script issue then Jan 24 00:10:43 and i had this working about month ago Jan 24 00:11:11 i looked at the svn changes and i did not find anything that could be the problem Jan 24 00:11:31 anyway when i removed that line everything worked like expected Jan 24 00:11:55 i will test again tomorrow netifd with my current config Jan 24 00:12:07 anything i should pay speciall attention to? Jan 24 00:12:21 no, just if everything works normally Jan 24 00:12:36 maybe leave it enabled in your builds Jan 24 00:12:45 and let me know if anything breaks Jan 24 00:12:59 if it segfaults for some reason, it'll leave a core file in /tmp Jan 24 00:13:38 btw. are you going to be at fosdem? Jan 24 00:13:49 no... Jan 24 00:13:58 but i will watch your talk Jan 24 00:14:10 any shows in Bay Area any time soon? Jan 24 00:14:17 someone said it will be recorded... Jan 24 00:14:28 * Chocks has never made a fosdem Jan 24 00:14:57 * russell-- recently watched nbd's ccc talk from several years ago Jan 24 00:15:09 that one sucked Jan 24 00:15:12 lol Jan 24 00:15:18 * Chocks blames Wookey Jan 24 00:15:23 which one was it? Jan 24 00:15:37 some old presentation i did on how to use the openwrt build system Jan 24 00:16:05 http://www.youtube.com/watch?v=dCKeVfXlJ-c Jan 24 00:16:12 22c3 Jan 24 00:16:32 kind of ancient at this point Jan 24 00:16:39 yeah Jan 24 00:16:55 the only thing that's the same is that spelling of openwrt! ;-) Jan 24 00:17:48 nbd: nice to connect a face with a name Jan 24 00:18:10 :) Jan 24 00:18:14 * luka12345 likes latex presentations (and latex in general) Jan 24 00:18:43 * russell-- kind of got my fill of latex in the 1990's ;-) Jan 24 00:19:12 i still use beamer-latex Jan 24 00:19:15 it's very convenient Jan 24 00:19:35 https://events.linuxfoundation.org/events/embedded-linux-conference Jan 24 00:19:53 once i learned latex i could never go back Jan 24 00:20:03 every thing i can i do in latex Jan 24 00:20:20 that's pretty near me Jan 24 00:20:36 :) Jan 24 00:20:52 Chocks: it seems to be pretty heavily Yocto focused and android i guess Jan 24 00:20:58 right Jan 24 00:21:00 ugh Jan 24 00:21:43 there are some nice breweries in the area too Jan 24 00:22:23 i hope the recording at fosdem is going to work Jan 24 00:22:50 i think last year (or the year before) they announced it as well, but they didn't record the sessions Jan 24 00:23:32 i want something to replace the boring 22c3 video Jan 24 00:23:35 ;) Jan 24 00:23:38 * Chocks tires of making broken images. that one lacked uci. ugh! Jan 24 00:23:53 nbd: hahaha Jan 24 00:24:23 Chocks: you can check the sanity of your .config with ./scripts/diffconfig.sh Jan 24 00:24:35 yeah, it wasn't that Jan 24 00:24:42 it was hacking I did to make a minimal image Jan 24 00:24:52 the default packages bring in too much stuff Jan 24 00:25:26 and I don't really need ppp, which appears to be a core dependency Jan 24 00:25:43 it's not a core dependency, just enabled by defaulty Jan 24 00:26:10 something drags it in. perhaps luci Jan 24 00:26:15 could be Jan 24 00:26:21 I'm not building it, but I get dependency errors Jan 24 00:27:52 no, I don't think it's luci Jan 24 00:27:55 maybe firewall Jan 24 00:29:01 hm Jan 24 00:38:51 hmm, compiling for wr740n gives me three images, v1, v3 and v4 Jan 24 00:38:59 lucky Jan 24 00:39:27 mine is v2.3, used to work fine with the v3 output image but last time i compiled and flashed was in november last year Jan 24 00:41:02 damn usb to serial ca-42s take so long to arrive from china Jan 24 00:41:14 yeah Jan 24 00:41:20 you can find US ones tho, if you are lucky Jan 24 00:41:33 i dont buy from ebay so much these years Jan 24 00:41:50 aye Jan 24 00:42:15 i got some from local gsm phone shops, turned out to be some sort of pl2303 cheap immitations that report the prolific at usb listing but actually wont work Jan 24 00:42:46 unlucky Jan 24 00:42:55 i guess i was lucky enough to get the fake ones.. Jan 24 00:43:35 pretty cheap tho, like 3 usd or so Jan 24 00:44:10 i have this one here, it's pretty reliable: http://www.avrfreaks.net/index.php?func=viewItem&item_id=745&module=Freaks%20Tools Jan 24 00:44:17