**** BEGIN LOGGING AT Mon Nov 19 02:59:59 2012
Nov 19 10:56:39 <OpenWrt-Commits> kaloz r34254 trunk/target/linux/generic/patches-3.7/820-usb_add_usb_find_device_by_name.patch * [generic/3.7]: fix 820-usb_add_usb_find_device_by_name.patch
Nov 19 12:38:55 <OpenWrt-Commits> jogo r34255 trunk/target/linux/ brcm63xx/patches-3.6/260-move_bcm963xx_tag.patch generic/patches-3.6/260-move_bcm963xx_tag.patch * Revert "target specific patches have nothing to do in generic.."
Nov 19 14:31:01 <KanjiMonster> swalker: ping
Nov 19 14:37:51 <OpenWrt-Commits> florian r34256 trunk/package/devel/gdb/Makefile * [package] gdb: add missing zlib dependency
Nov 19 14:38:03 <OpenWrt-Commits> florian r34257 packages/utils/strace/ patches patches/001-eglibc_2_16_process_vm_readv.patch * [package] strace: backport an upstream build fix for eglibc >= 2.15
Nov 19 14:38:32 <OpenWrt-Commits> florian r34258 trunk/package/devel/ gdb/patches/001-gdb-pr14523-mips-signal-number.patch gdb/patches * [package] gdb: fix debugging of MIPS(el) applications
Nov 19 15:31:12 <OpenWrt-Commits> florian r34259 packages/utils/smartmontools/Makefile * [package] smartmontools: avoid using space in BUILD_INFO.
Nov 19 16:23:33 <Chocks> moo
Nov 19 16:27:45 <Chocks> I most everything - all but kernel really - from a SVN from about 6 months ago.  And hotplug buttons no longer work.  hotplug2 package itself doesn't seem to have changed since.  Where should I be looking for the problem?
Nov 19 16:44:19 <BasicXP> Good time of the day, jow_laptop. May I please ask you to sync the translations for LuCI in SVN and Pootle? Thank you!
Nov 19 16:44:54 <[florian]> Chocks: what's your target?
Nov 19 16:45:00 <Chocks> rt3883
Nov 19 16:45:04 <Chocks> ramips
Nov 19 16:45:33 <Chocks> I most = I updated most
Nov 19 16:46:04 <Chocks> so all the kernel modules etc are the same.  It's gotta be something user space
Nov 19 16:46:08 <[florian]> did you check the button is still properly registered, module loaded, sysfs entries there etc ...
Nov 19 16:46:26 <Chocks> I will double check. but it's the same kernel
Nov 19 16:48:52 <[florian]> so it's definitively user-space related yes
Nov 19 16:49:10 <Chocks> just flipped back to older build to double check
Nov 19 16:50:16 <Chocks> hm
Nov 19 16:50:33 <Chocks> gpio_keys_polled hasn't made any /sys entries
Nov 19 16:50:43 <Chocks> oh, sorry. wrong grep
Nov 19 16:50:45 <Chocks> all the same
Nov 19 17:23:23 <Chocks> so anyway
Nov 19 17:23:24 <Chocks> I have no idea
Nov 19 17:23:32 <Chocks> gpio-keys-polled is still in evidence
Nov 19 17:32:59 <DonkeyHotei> gpio-keys-polled doesn't seem to do anything for me
Nov 19 17:33:07 <OpenWrt-Commits> florian r34260 trunk/toolchain/binutils/patches/2.22/500-arm_ld_assert_fix.patch
Nov 19 17:33:07 <OpenWrt-Commits> [toolchain] binutils/2.22: backport an upstream fix for an assertion in the ARM backend.
Nov 19 18:10:17 <Chocks> there were some minor changes in hotplug rules in the base-files package, but still nothing
Nov 19 18:14:03 <DonkeyHotei> [florian]: no luck with either fixed phy or gpio phy? i'm just making a cursory check, because i've given up on getting an ar7 device ready for production use anytime soon and got a lantiq one, which needs work too, and blogic has been MIA for almost a week
Nov 19 18:14:22 * Chocks sobs for all
Nov 19 18:16:14 <DonkeyHotei> also, is lantiq voip hardware the only voip hardware supported in owrt?
Nov 19 18:33:29 <mirko> DonkeyHotei: depends on how you define "support" I guess - what do you wanna do?
Nov 19 18:38:29 <DonkeyHotei> mirko: have it work?
Nov 19 18:38:44 * Chocks pokes mirko
Nov 19 18:38:50 <Chocks> something eglibc related perhaps
Nov 19 18:42:50 <mirko> DonkeyHotei: I always had just lantiq voip hardware running - but if you wanna e.g. use asterisk, take a look at the list of channel drivers
Nov 19 18:42:53 <mirko> Chocks: ?
Nov 19 18:43:09 <Chocks> mirko: udev/hotplug2/button breakage in latest user space.  apparently
Nov 19 18:45:14 <mirko> Chocks: Context? Ticket? Elaboration?
Nov 19 18:46:01 <Chocks> from earlier:    I svn updated most everything - all but kernel really - from a SVN from about 6 months ago.  And hotplug buttons no longer work.  hotplug2 package itself doesn't seem to have changed since.  Where should I be looking for the problem?
Nov 19 18:47:34 <Chocks> I see that hotplug sits on a socket to get events.  Not sure where that socket comes from though
Nov 19 18:48:25 <mirko> Chocks: well, I don't know what exactly happened the last 6 months - please investigate the rules in /etc/hotplug.d/, bisect the changes, open a ticket
Nov 19 18:48:44 <Chocks> I did
Nov 19 18:48:49 <Chocks> but it's not clear those are to blame
Nov 19 18:48:57 <Chocks> I'm just looking for ideas where to even look
Nov 19 18:51:46 <Chocks> unfortunately, I don't have the luxury to wait for someone else  to find the solution
Nov 19 18:55:19 <DonkeyHotei> mirko: what directory do i look in for the channel drivers?
Nov 19 18:55:21 <Chocks> I reverted to the exact previous kernel binary to make sure; it's not that
Nov 19 19:02:27 <KanjiMonster> Chocks: you can always try to bisect it if you have a known working version
Nov 19 19:04:02 <Chocks> that's going to take forever
Nov 19 19:04:14 <Chocks> there are eglibc changes and other stuff which requires a full rebuild
Nov 19 19:04:41 <Chocks> I'm only using a small number of packages. there's only so many things which could be wrong
Nov 19 19:13:45 <Chocks> moo
Nov 19 19:17:35 <mirko> DonkeyHotei: please take a look into the asterisk project
Nov 19 19:18:24 <mirko> Chocks: Well, we don't know either
Nov 19 19:18:46 <mirko> Either you collect enough information to make it easy for us helping you debugging the issue, or you do it yourself
Nov 19 19:18:58 <Chocks> back up.
Nov 19 19:18:59 <Chocks> I
Nov 19 19:19:04 <Chocks> I'm not a noob.  you should know that
Nov 19 19:19:09 <mirko> strasidlo: You#re the guy who committed the asterisk-patch, right?
Nov 19 19:19:12 <Chocks> which information do you want?
Nov 19 19:19:27 <strasidlo> mirko, yes
Nov 19 19:19:29 <Chocks> it's not the hotplug configuration files.  I said that earlier
Nov 19 19:19:41 <Chocks> and hotplug itself has not changed
Nov 19 19:20:04 <mirko> Chocks: well, right now I#ve the feeling you want me to find the error in your setup - I obviously can't do that
Nov 19 19:20:14 <Chocks> then you've definitely misread
Nov 19 19:20:23 <Chocks> :o
Nov 19 19:20:26 <mirko> Also: since I don't know what's your problem, I can't give you a solution, nor any hint on how to debug
Nov 19 19:20:27 <[florian]> Chocks: just start a bisection, seriously
Nov 19 19:20:55 <mirko> strasidlo: What hardware did you test the asterisk stuff with?
Nov 19 19:21:02 <Chocks> florian: I'm not going to spend hours building many versions of openwrt going back months.  seriously
Nov 19 19:21:08 <mirko> strasidlo:  and since it's an RFC: any particular you're not certain about?
Nov 19 19:21:14 <Chocks> what I *am* asking is how hotplug should be working in openwrt
Nov 19 19:21:15 <[florian]> Chocks: then you are on your own
Nov 19 19:21:19 <Chocks> no
Nov 19 19:21:26 <Chocks> why do we have to be so difficult in here?
Nov 19 19:21:29 <[florian]> hotplug listens to the uevent socket provided by the kernel
Nov 19 19:21:30 <strasidlo> mirko, I am testing it with tp-link 1043nd and x86 generic (on virtual machine)
Nov 19 19:21:34 <[florian]> pretty much like on your desktop
Nov 19 19:21:51 <Chocks> is that socket configured in user space anywhere?
Nov 19 19:22:01 <[florian]> it's built into the kernel
Nov 19 19:22:12 <[florian]> and unless specifically disabled, which I don't think is the case
Nov 19 19:22:16 <[florian]> it's always tehre
Nov 19 19:22:40 <Chocks> so. it's the same kernel, same kernel modules.  same hotplug.
Nov 19 19:22:47 <strasidlo> mirko: I also asked my "voip" team to test it thoroughly if it segfaults and goes well, also they'll do some performance testing
Nov 19 19:22:54 <[florian]> is your board actually supported by mainline OpenWrt?
Nov 19 19:23:10 <Chocks> no
Nov 19 19:23:23 <Chocks> despite considerable efforts to the contrary by myself
Nov 19 19:24:41 <[florian]> ah, then maybe your button drivers have not been moved to gpio_keys_button?
Nov 19 19:24:48 <[florian]> though ramips seem to provide some helper
Nov 19 19:25:05 <Chocks> please explain further
Nov 19 19:26:06 <[florian]> if you browse through the svn history, you will see that we used to use the gpio-buttons driver, and we now use gpio-keys-polled instead
Nov 19 19:26:30 <strasidlo> mirko: I rewrote the Makefile a little, so there are several changes against asterisk18 - this is the first reason why I am asking. I am not sure about my comment I sent to my first mail. I would also appreciate if someone else would test this package on different architectures - also I want the package to be clean and awesome. :)
Nov 19 19:27:16 <Chocks> yes, I recall that changed, and had to change stuff in my setup.  But I am using gpio-keys-polled before and now.  Anyway, that's a kernel module which hasn't changed.
Nov 19 19:27:39 <[florian]> the registration of the actual data happens in your board-specific file, just like leds
Nov 19 19:27:47 <Chocks> I know
Nov 19 19:27:52 <[florian]> just making sure
Nov 19 19:29:31 <OpenWrt-Commits> acinonyx r34261 trunk/tools/ipkg-utils/patches/200-force_gnu_format.patch * [tools] ipkg-utils: Force gnu format for tar (#12496)
Nov 19 19:30:59 <Chocks> I'm trying to make certain it's not something that changed in kernel due to module rebuild etc
Nov 19 19:31:07 <Chocks> but still smells like user space to me
Nov 19 19:32:46 <Chocks> with strace I can see that nothing is being delivered on the socket hotplug is listening on.  contrary to earlier versions
Nov 19 19:36:34 <[florian]> can you check the gpio mapping/pinmuxing etc...
Nov 19 19:37:58 <OpenWrt-Commits> florian r34262 trunk/toolchain/ (7 files in 3 dirs) * [toolchain] binutils: add support for 2.23.1
Nov 19 19:38:02 <OpenWrt-Commits> florian r34263 trunk/toolchain/gcc/common.mk * [toolchain] gcc-4.7-linaro: update to 2012.11
Nov 19 19:38:09 <OpenWrt-Commits> florian r34264 trunk/toolchain/gcc/common.mk * [toolchain] gcc-4.6-linaro: update to 2012.11
Nov 19 19:38:12 <OpenWrt-Commits> florian r34265 trunk/toolchain/ binutils/Makefile binutils/Config.in binutils/patches/2.21 * [toolchain] binutils: remove 2.21
Nov 19 19:38:31 <Chocks> it's the exact same kernel binary.  how could that have changed?
Nov 19 19:38:59 <[florian]> you power cycled and it used to work because of some previous initialization?
Nov 19 19:39:23 <[florian]> just formulating hypothesis
Nov 19 19:39:42 <Chocks> that seems like an odd suggestion.  This setup/kernel was used on 300 boxes in the field.
Nov 19 19:40:03 <Chocks> maybe it's something else magic that changed in base-files
Nov 19 20:00:41 <DonkeyHotei> mirko: all i see is lantiq
Nov 19 20:41:55 <mirko> DonkeyHotei: so it's just lantiq
Nov 19 20:42:45 <DonkeyHotei> mirko: know any lantiq devices without dsl hardware?
Nov 19 20:43:47 <mirko> DonkeyHotei: the GRX288 chipset from Lantiq doesn't have DSL at all
Nov 19 20:44:17 <mirko> But I don't know about end-user hardware
Nov 19 21:21:56 <Chocks> at this point, I begin to suspect the eglibc change from 2.13 to 2.15
Nov 19 21:29:18 <Chocks> sigh
Nov 19 21:29:22 <Chocks> I will try 2.16
Nov 19 23:12:36 <OpenWrt-Commits> florian r34266 trunk/toolchain/gcc/common.mk * [toolchain] gcc: build with HOST_CFLAGS
Nov 19 23:51:42 <Chocks> is ubusd required in simplistic setups?
Nov 19 23:52:12 <Chocks> that is, which basic services depend upon it?
Nov 20 00:19:05 <swalker> KanjiMonster: pong
Nov 20 00:20:51 <KanjiMonster> swalker: have you seen mcgrof's (Luis Rodriguez) mail? Thought it's something similar to what you are doing with your up-to-date list of packages
Nov 20 00:21:30 <swalker> KanjiMonster: I have and mostly agree
Nov 20 00:22:00 <KanjiMonster> Chocks: netifd does; which is rather essential
Nov 20 00:22:12 <Chocks> right.  I thought that might be the case.
**** ENDING LOGGING AT Tue Nov 20 02:59:57 2012