**** BEGIN LOGGING AT Fri Aug 13 02:59:57 2010 Aug 13 03:32:43 agb * r22616 /packages/net/nufw/Makefile: [packages] net/nufw: run of autogen.sh needed with recent versions of libtool Aug 13 03:59:03 build #65 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/65 **** ENDING LOGGING AT Fri Aug 13 04:13:39 2010 **** BEGIN LOGGING AT Fri Aug 13 04:14:14 2010 Aug 13 05:44:46 stintel, ping. Aug 13 07:24:04 RealOpty: pong, didn't see him online yet Aug 13 08:00:27 cshore * r22617 /packages/net/freeswitch/ (Makefile files/etc.uci/ files/uci/): [packages] net/freeswitch: Remove shell-script based UCI config. It breaks on restart, and it will be some time before the Lua based one is complete. Aug 13 08:21:33 hello Aug 13 08:27:06 morining Aug 13 09:13:50 <_trine> does openwrt have mysql Aug 13 09:24:36 yes Aug 13 09:24:48 hmm Aug 13 09:24:49 no Aug 13 09:24:52 strange Aug 13 09:25:04 <_trine> I could not find it Aug 13 09:25:14 lots of plugins for it but no package Aug 13 09:25:21 <_trine> exactly Aug 13 09:25:42 strange Aug 13 09:26:54 <_trine> the reason for looking for it is that I have just got cherokee web server to compile and now I am in the process of getting wordpress set up but there is no mysql that I could find Aug 13 09:27:42 <_trine> I'm doing this for my dockstar Aug 13 09:33:03 haha Aug 13 09:33:14 i have asterisk next on my list Aug 13 09:33:23 now that dvbt works and i can record and multicast Aug 13 09:33:37 <_trine> I have asterisk working too Aug 13 09:33:54 i also made moutnd and mpd work Aug 13 09:34:37 <_trine> we need to get mysql working though Aug 13 10:41:37 <_trine> bl Aug 13 10:41:42 <_trine> blogic, Aug 13 10:41:46 <_trine> https://dev.openwrt.org/browser/packages/libs/mysql/Makefile?rev=5386 Aug 13 10:52:05 <_trine> hmm I dont think any of the sources are working for mysql Aug 13 10:52:14 <_trine> in the makefile Aug 13 11:02:28 [florian], ping Aug 13 11:07:46 _trine, blogic: someone submitted mysql-server to -devel a few days ago Aug 13 11:08:24 <_trine> thanks loswillios Aug 13 11:09:11 <_trine> I was just looking at the mysql web site and am surprised registration is required before you can download Aug 13 11:10:00 <_trine> loswillios, whats happening to that mysql submission you have mentioned ,, do you know? Aug 13 11:11:55 if the package is ok, it'll be committed I guess Aug 13 11:12:24 <_trine> who's looking at it Aug 13 11:14:24 no idea Aug 13 11:14:35 <_trine> ok thanks Aug 13 11:17:12 {Nico} apparently ;) Aug 13 11:17:30 <_trine> ah thanks Aug 13 11:18:42 <_trine> https://lists.openwrt.org/pipermail/openwrt-devel/2010-August/007714.html Aug 13 11:19:22 <[florian]> Wipster: pong Aug 13 11:25:27 [florian], hi looking at cpmac.c again to try and force the link up in adjust_link I have made the check always true, turn the carrier on before the queues start, and disable the not oldlink but still no joy. Do you know of any other sections that could be affecting it? I cant spot one Aug 13 11:25:58 <[florian]> Wipster: there is not anything else as far as I can tell Aug 13 12:32:22 _trine: registration isn't required, downloading without is just a bit hidden. you have to click on "ยป No thanks, just take me to the downloads!" at the bottom Aug 13 12:41:01 <_trine> KanjiMonster, thank you Aug 13 13:13:51 acoul * r22618 /packages/net/kismet/Makefile: net/kismet: update kismet to Kismet-2010-07-R1 Aug 13 13:21:26 acoul * r22619 /packages/net/iodine/Makefile: net/iodine: update iodine to 0.6.0-rc1 Aug 13 15:13:27 acoul * r22620 /packages/net/nmap/ (3 files in 2 dirs): net/nmap: update nmap to 5.35DC1 Aug 13 16:04:48 jow * r22621 /trunk/target/linux/generic/patches-2.6.34/200-netfilter_xtables_align.patch: Aug 13 16:04:48 [kernel] fixup 2.6.34 kernel headers to properly export XT_ALIGN() to userspace Aug 13 16:04:48 Ref: http://lkml.org/lkml/2010/4/13/118 Aug 13 16:05:36 florian * r22622 /packages/net/faifa/Makefile: [package] update faifa to r64 Aug 13 16:14:09 mb * r22623 /trunk/target/linux/omap24xx/patches-2.6.35/820-backlight-fixes.patch: n810: Make sysfs backlight controls work Aug 13 17:15:19 mb * r22624 /packages/Xorg/xorg/driver/xf86-video-omap/patches/001-misc-fixes.patch: update omapfb patches Aug 13 17:47:46 hello Aug 13 19:30:01 jow * r22625 /trunk/target/linux/ar71xx/base-files/etc/uci-defaults/vlan-migration: [ar71xx] add uci-defaults script to migrate vlan 0 to vlan 1 after sysupgrade on devices using the RTL8366s switch Aug 13 19:31:18 jow * r22626 /branches/backfire/target/linux/ar71xx/base-files/etc/uci-defaults/vlan-migration: [backfire] merge r22625 Aug 13 19:46:51 dingo * r22627 /packages/net/wing/ (10 files in 6 dirs): Aug 13 19:46:51 [PATCHTEAM] - WING Wireless Mesh Routing Protocol Aug 13 19:46:51 Signed-off-by: Roberto Riggio Aug 13 19:53:36 http://pastebin.org/480145 I don't understand this. I'm trying to build this host package. python-dev is installed on my system and /usr/include/python2.6/Python.h exists. Yet the compiler is unable to find it even if -I is defined on the commandline. How come? Aug 13 19:55:24 and if I copy&paste that command into a shell it does actually work and compile the .o correctly. Aug 13 19:55:44 I really don't get this. What does bias gcc's include path in buildroot? Aug 13 19:58:25 oh, nevermind. I need to go to bed. Aug 13 19:58:39 It does work correctly. I'M just stupid Aug 13 20:01:10 I picked the wrong xterm with the wrong ssh session in it ;) Aug 13 21:06:10 looking forward to attaching the my serial cable this weekend, see what's actually going on be so much easier Aug 13 21:34:13 nico * r22628 /packages/libs/libtorrent/patches/122-fix-uintptr_t-use.patch: [packages] libtorrent: fix build failures on octeon & ppc Aug 13 21:35:44 hi all Aug 13 21:36:28 can anyone tell me how do I make a change in a specific target kernel-config and it gets all the way up to the build dir? Aug 13 21:40:56 <_trine> lfcorreia, I think you have to make changes ~/trunck/target/linux/config-default for example Aug 13 21:41:27 <_trine> lfcorreia, I think you have to make changes ~/trunck/target/linux/target/config-default for example Aug 13 21:41:32 _trine, hum... ok, et me check Aug 13 21:42:05 <_trine> and you can make the kernel as well Aug 13 21:42:15 that's the file I've been changing to no avail Aug 13 21:42:33 since the kernel is not a package, how do I build it separately? Aug 13 21:42:58 <_trine> make kernel_menuconfig Aug 13 21:43:19 well, again I'm explaining my problem Aug 13 21:43:34 someone decided that the ramips platform didn't need USB Aug 13 21:43:53 and therefore, you can't select it ANYWHERE Aug 13 21:44:11 <_trine> oh well Aug 13 21:44:13 so please, preety please tell me how to enable it Aug 13 21:45:01 because editing target/linux/ramips/rt305x/config-2.6.34 is useless Aug 13 21:45:09 it does nothing Aug 13 21:45:32 _trine, so sorry if I'm being annoying, but this has been weeks now, on and off Aug 13 21:45:48 guess I'll just send an email Aug 13 21:47:29 <_trine> I thought I had told you Aug 13 21:48:01 yes, make kernel_menuconfig, go to device drivers - usb support, Aug 13 21:48:16 then nothing happens, the menu option doen't even show up Aug 13 21:48:37 <_trine> does here Aug 13 21:48:48 my svn tree is updated Aug 13 21:48:57 I've even clean it up twice Aug 13 21:49:11 lfcorreia: that could be something in upstream. I recalled similar issues for the malta target, it had PCI disabled Aug 13 21:49:24 let me look it up Aug 13 21:49:29 thanks Aug 13 21:50:47 hm, there it was enough to override the CONFIG_PCI descision made by Linux' arch defconfig Aug 13 21:51:33 xMff, i've looked into other mips based boards and found nothing relevant Aug 13 21:51:52 let me look again Aug 13 21:52:43 for example, ifxmips does have USB enabled Aug 13 21:52:52 so it's not mips upstream related Aug 13 21:53:05 * lfcorreia is now comparing config files Aug 13 21:54:03 did you add the "usb" feature to target/ramips/Makefile ? Aug 13 21:54:23 by default its only squashfs Aug 13 21:54:35 should be something like FEATURES:=squashfs usb Aug 13 21:54:42 let me check Aug 13 21:55:29 xMff, well, ifxmips doesn't need usb there, but i'll try it on Aug 13 22:00:38 lfcorreia: what symbols can't be selected? Aug 13 22:00:58 USB, the whole lot Aug 13 22:01:05 when running kernel_menuconfig here for ramips it shows USB_SUPPORT as enabled Aug 13 22:01:26 you mean usb symbols in normal (openwrt) menuconf or kernel menuconf? Aug 13 22:01:26 yes, but can you select anything in the menu? Aug 13 22:01:32 both Aug 13 22:02:01 <{Nico}> lfcorreia: the kernel config needs CONFIG_USB_ARCH_HAS_HCD or CONFIG_USB_EHCI_HCD Aug 13 22:02:46 lfcorreia: I can enter both without any changes Aug 13 22:02:52 CONFIG_USB_ARCH_HAS_HCD Aug 13 22:04:54 for example: Kernel Modules -> USB Support -> kmod-usb-net Aug 13 22:05:15 its using linux 2.6.34.1 Aug 13 22:06:30 xMff, then it's probably me, if it matters at all, i'm on Fedora 13 x64 Aug 13 22:06:55 lfcorreia: what is displayed if you / and search for usb? Aug 13 22:07:01 that should dump the dependencies Aug 13 22:07:13 *you use Aug 13 22:07:21 xMff, please explain Aug 13 22:07:29 <{Nico}> maybe trash the tmp directory Aug 13 22:07:44 in both menuconfig and kernel_menuconfig you can use / (slash) to search for symbols Aug 13 22:07:52 like vi or less Aug 13 22:08:01 ah, ok Aug 13 22:08:11 search for usb there Aug 13 22:08:25 and check what results are shown Aug 13 22:08:45 can you rule out that package/kernel/modules/usb.mk is broken for some reason? Aug 13 22:08:55 e.g. local changes and forgot an endef or so Aug 13 22:09:03 RealOpty: according to my buddy you need kernel <8xxx and the xbox must not be repaired after august 2009 Aug 13 22:09:26 xMff, i've just cleaned out my tree, will start again from the beggining, hold on for a while Aug 13 22:09:34 sure Aug 13 22:33:49 xMff: the rt305x subtarget unsets the CONFIG_USB_ARCH_HAS_EHCI/HCD/OHCI symbols for the kernel config Aug 13 22:34:15 lfcorreia: ^ Aug 13 22:34:32 KanjiMonster, and how can this be fixed? Aug 13 22:35:46 lfcorreia: remove these three lines from ramips/rt305x/config-2.6.34, and force a reparsing by removing the tmp directory in the buildroot dir - we already had this discussion a few days ago ;) Aug 13 22:36:39 KanjiMonster, yes, we had that discussion, should I propose a patch? because even with removing those lines, usbcore and others aren't being built Aug 13 22:36:59 ah, okay, I didn't know that they aren't built Aug 13 22:39:53 KanjiMonster, I would help fix it, but there's a ton of magic behind the scenes that I still don't know about :( Aug 13 22:41:38 lfcorreia: perhaps try talking to blogic about this - I don't know much about the ralink platform except from looking at the target definitions. And from there it should work as well for the rt305x target as for the rt288x target when removing these three lines Aug 13 22:45:18 hi all Aug 13 22:46:39 I just wanted to report that 10.03.1-rc1 upgrade worked out of the box for me (with keeping settings - first time on my DIR-825). Aug 13 22:47:45 but the reg domain issue really is annoying ... I'm from Germany and just because of this I have to compile my own OpenWRT firmware for DIR-825 to be able to use channels 12,13,14 Aug 13 22:47:57 atleast channel 12-13 are legal in germany Aug 13 22:48:36 if possible please consider changing defaults such that the regdomain can actually be set in OpenWRT out of the box Aug 13 22:54:27 hmm, upgrading replaced my fstab file by a symlink in /etc/config/fstab pointing to /etc/fstab (I guess) Aug 13 22:56:29 joda_bot: regdomain is a legal, not a technical issue atm Aug 13 22:58:24 :-( Aug 13 22:58:36 legal problem in which country? :) Aug 13 22:58:56 is there a way to change the eeprom regdomain? Aug 13 22:59:36 as rom is probably "read only memory" ... Aug 13 22:59:39 :-( Aug 13 23:00:05 xMff: would it possible to provide a alternative kernel module? Aug 13 23:00:11 xMff: I don't know how much work that would be, but how about creating images with forced regdomains for those devices which have a known wrong one (e.g. one for germany, one for us, one for japan, etc)? This would be somewhat of a middle step of allowing to set/force it, but only through flashing a new image Aug 13 23:01:21 KanjiMonster: probably Aug 13 23:01:23 joda_bot: iirc at least on some atheros cards its possible, but I guess this definitely voids your warranty ;) Aug 13 23:01:59 KanjiMonster: show me how :D Aug 13 23:02:57 also is there any way to see if HT mode is active? Aug 13 23:03:18 only in hostapd debug atm I believe Aug 13 23:03:20 I can't really tell from my read outs if it finally would work :-) and my notebook is just 11g Aug 13 23:03:41 or maybe with iw something Aug 13 23:04:02 tried those, I can't tell from that just that the wireless is HT capable Aug 13 23:04:28 I see Aug 13 23:04:32 rdate seems broken on 10.03.1-rc1? Aug 13 23:04:39 well if you find something I'd be interested as well Aug 13 23:04:52 k Aug 13 23:05:01 for the gui stuff Aug 13 23:09:42 joda_bot: there is an howto for older madwifi compatible cards, so unless we are talking about an old card, it won't help you Aug 13 23:28:12 KanjiMonster: no help that is then.. Aug 13 23:28:19 :-P Aug 13 23:28:46 ;) Aug 14 00:45:26 jow * r22629 /trunk/package/ (4 files in 4 dirs): Aug 14 00:45:26 [package] base-files, uci: properly revert list state on config_load Aug 14 00:45:26 This fixes a bug where lists end up with duplicate values if config_load was Aug 14 00:45:26 invoked multiple times. Aug 14 00:54:35 jow * r22630 /trunk/package/uhttpd/ (10 files in 3 dirs): Aug 14 00:54:35 [package] uhttpd: Aug 14 00:54:35 - more robust handling of network failures on static file serving Aug 14 00:54:35 - support unlimited amount of authentication realms, listener and client sockets Aug 14 00:54:35 - support for interpreters (.php => /usr/bin/php-cgi) Aug 14 01:34:09 jow_laptop: xMff: there is still a bug on 10.03.1-rc1 with /etc/init.d/fstab replacing a symlink with a warning message Aug 14 01:34:38 I never understood that bug Aug 14 01:34:53 in start() Aug 14 01:35:00 this line has to go away: #echo '# WARNING: this is an auto generated file, please use uci to set defined filesystems' > /etc/fstab Aug 14 01:35:18 I just comment it out on every upgrade Aug 14 01:35:19 that does not replace a symlink Aug 14 01:35:26 it does Aug 14 01:35:38 because /etc/fstab points to /tmp/fstab Aug 14 01:35:46 maybe it creates file which prevents a symlink from being created Aug 14 01:35:57 but echo foo > bar does not destry existing symlinks Aug 14 01:35:58 an /tmp/fstab is generated by block-mount / block-extroot Aug 14 01:36:28 then the destruction of the symlink is a bug in mini_fo? Aug 14 01:37:08 eh, well > that should replace any existing file ... did you think of >> ? Aug 14 01:37:55 > is truncate || create Aug 14 01:37:59 not replace Aug 14 01:38:18 sorry, you're right the symlink seems to be ok... Aug 14 01:38:29 something else must have gone wrong Aug 14 01:39:07 what is the expected behaviour? Aug 14 01:39:34 /tmp/fstab gets populated? Aug 14 01:40:54 please try https://dev.openwrt.org/changeset/22148 Aug 14 01:41:03 if you can confirm the fix I'll merge it Aug 14 01:43:10 hm Aug 14 01:43:24 ok, a race condition with locking might really be a good point Aug 14 01:43:26 let me see Aug 14 01:44:00 I can't really confirm anything specific yet, just that mountextroot worked fine most of the times and then failed consistently at times Aug 14 01:50:35 hm definitly something wrong there Aug 14 01:52:05 xMff: there still is the bug that if is_rootfs is true the target of the mount is not ignored Aug 14 01:52:11 it only works with /overlay Aug 14 01:52:20 and without that patch the overlay looked broken Aug 14 01:52:35 atleast / had no free space according to df Aug 14 01:52:45 next reboot now Aug 14 01:54:39 yeah Aug 14 01:54:42 that helped Aug 14 01:55:35 the patch? Aug 14 01:57:02 did not work without the patch you pointed me to, but with the patch and setting the is_rootfs mount's target to /overlay I now got this: http://pastebin.com/PUgVpYkw Aug 14 01:58:19 k Aug 14 01:59:07 I really find it confusing that mtdblock4 and sda1 are mounted on /overlay Aug 14 02:02:50 hmm, what happens if both overlays differ ever so slightly? Aug 14 02:03:06 my harddisk an additional folder in root called "/home" Aug 14 02:03:15 and the jffs2 image doesn't Aug 14 02:03:38 result it's not visible in the / folder Aug 14 02:04:28 /mtd4/nohome Aug 14 02:04:28 /disk/home/users... Aug 14 02:04:28 / <-- shows no home with 2 overlays :D **** ENDING LOGGING AT Sat Aug 14 02:59:57 2010