**** BEGIN LOGGING AT Sun Oct 09 02:59:57 2011 Oct 09 05:37:38 mwalle: it's rather the other way round Oct 09 06:08:38 obsy * r28389 /packages/net/transmission/Makefile: [packages] transmission: update to 2.40 (final) Oct 09 14:14:46 nico * r28390 /trunk/package/gdb/Makefile: package/gdb: disable expat properly, remove conditional dependency on libexpat, bump release number Oct 09 14:23:45 nico * r28391 /branches/backfire/package/gdb/Makefile: [backfire] package/gdb: backport [28390] (closes: #8688) Oct 09 14:41:35 * mystica555_ decides to idle here too now :) Oct 09 14:41:55 Which USB-JTAG cable should I get for working with broadcom devices ? Oct 09 15:48:43 mystica555_: good choice, its quiter here Oct 09 15:48:59 ;) i figure, i ask enough highly technical questions... Oct 09 15:49:04 vs 'how do i port forward' Oct 09 15:51:18 nbd * r28392 /trunk/package/mac80211/ (51 files in 2 dirs): mac80211: update to wireless-testing 2011-10-05 + pending patches Oct 09 15:51:45 and stuff like that is nice to see too :) Oct 09 16:07:53 nbd * r28393 /trunk/package/ (libubox/Makefile ubus/Makefile): update libubox and ubus to the latest version Oct 09 16:09:20 nbd: question about netifd: will allow us to run a custom script when a STA connects or disconnects? Oct 09 16:10:25 regarding the rtl8366 switch and swconfig, is there any particular reason why the other features in this patch were not applied (*yet): https://dev.openwrt.org/ticket/7977 Oct 09 16:10:32 (only jumboframes for now) Oct 09 16:10:45 solca: I'd say that not really within the scope of netifd Oct 09 16:10:58 solca: such functionality belongs into hostapd Oct 09 16:11:29 eventually i'd like to have a separate daemon that can call scripts based on ubus events Oct 09 16:11:38 that daemon can then also handle interface up/down events Oct 09 16:11:41 and other interesting things Oct 09 16:11:52 yeah, nl80211 should provide some suitable events to notice sta connects afair Oct 09 16:12:12 hostapd events are better than nl80211 events Oct 09 16:12:20 because nl80211 doesn't know when auth is done Oct 09 16:12:22 or even successful Oct 09 16:12:26 actually, it does Oct 09 16:12:28 true, nl80211 lacks the wpa awareness Oct 09 16:12:29 but it doesn't send events for that Oct 09 16:14:12 mystica555_: lack of time maybe Oct 09 16:14:48 ok; i was just wondering if there may have been something with the patch that wouldnt actually work. that looks like rather useful functionality :) Oct 09 16:14:57 have you tested it? Oct 09 16:15:12 yeah, at least the link speed setting is needed by some people who need to force 10 mbit to their cable modem Oct 09 16:15:14 no, i havent yet, i just saw that when looking for info on that switch Oct 09 16:16:11 patches look ok though Oct 09 16:16:16 so i guess they can be committed Oct 09 16:16:23 * nbd goes back to hacking build system stuff Oct 09 16:16:27 :) Oct 09 16:17:00 until i looked at that, i wasnt quite aware that the switch in these routers was quite so nice Oct 09 16:24:19 btw. i'm currently adding something that makes it easier to work with source trees that are managed in git Oct 09 16:24:40 if you enable a kconfig option, you can simply symlink the .git tree of a workdir to package//git-src Oct 09 16:24:51 and it'll do a checkout from that git tree in the prepare step Oct 09 16:25:46 sweet Oct 09 16:26:30 interesting; may want to mention that to paulfertser on the other chan ;) Oct 09 16:26:55 it doesn't help with his use case yet Oct 09 16:26:58 oh k Oct 09 16:27:01 but maybe i'll extend it Oct 09 16:27:09 right now i need it for other things Oct 09 16:27:14 e.g. for netifd testing Oct 09 16:36:54 ah, nice. it works Oct 09 16:37:14 nbd * r28394 /trunk/ (Config.in include/package.mk): build system: add a feature that allows you to pull sources from a git tree instead of the usual tarball Oct 09 16:37:22 it also leaves the .git symlink in the package build dir Oct 09 16:37:28 so you can commit from there Oct 09 16:57:16 nbd * r28395 /trunk/include/package.mk: fix a small logic error in package download Oct 09 17:05:52 i wish i had thought of that git-src thing earlier Oct 09 17:05:57 makes things so much easier ;) Oct 09 18:18:15 [florian]: it's broken, unfortunately. here's the bootlog: http://openwrt.pastebin.ca/2088450 Oct 09 18:19:14 i _can_ bring up the console, but first it's very slow, then it stops reacting after about a minute or so Oct 09 18:27:12 failsafe works fine, so i am suspecting the ethernet driver or the wlan driver Oct 09 18:31:48 danage: I remember florian having the the same symptom once, and it was the jffs2 garbage collector thread Oct 09 18:32:13 do you know what i can do KanjiMonster Oct 09 18:33:10 danage: no, sorry; I don't remember if [florian] managed to fix it or not (and if, how) - you have to ask him ;) Oct 09 18:35:10 ok thanks. i guess i'll wait until he turns up Oct 09 18:35:47 it does look like a memory leak or something of the sort. at one point, some sort of watchdog was invoked Oct 09 18:39:57 maybe not a memory leak but a just not enough memory; I saw in your bootlog the device has only 8 MB; maybe there's an active loop trying to allocate until it succeeds Oct 09 18:40:56 what's the current memory footprint? i did compile all sorts of stuff into the image, luci for example Oct 09 19:04:16 danage: definitely try removing luci and uhttpd Oct 09 19:04:40 KanjiMonster: if i must... meh, i want luci :) Oct 09 19:05:40 8 megs of ram is...claustrophobic Oct 09 19:07:34 it's vintage-tastic! Oct 09 19:21:25 XD Oct 09 20:35:58 updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/index.html Oct 09 21:33:27 KanjiMonster: same problem Oct 09 21:38:08 hrm Oct 09 21:38:28 its gotta be the driver Oct 09 21:38:50 [florian] said something about wlan not working. i'll try without. Oct 09 23:22:18 build #61 of rdc is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/rdc/builds/61 Oct 09 23:57:56 nico * r28396 /packages/lang/distribute/Makefile: [packages] distribute: fix MD5SUM (closes: #10154) Oct 10 00:55:11 hm the zyxel NBG5715 is listed on openwrt's donors page; is this actually available as a compilable target now? Oct 10 00:55:14 its not in the TOH Oct 10 00:56:43 build #92 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/92 Oct 10 00:57:27 or if not, what does having hardware donated actually mean? Oct 10 01:11:53 build #90 of ubicom32 is complete: Failure [failed compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/90 Oct 10 01:14:42 nor can i find any reference to it grepping through my source tree Oct 10 02:46:16 nico * r28397 /branches/packages_10.03.1/multimedia/ffmpeg/ (Makefile patches/ patches/901-fix_libswscale_on_x86.patch): [backfire/packages] ffmpeg: fix build failure on x86 (closes: #8840), install libswscale headers (closes: #9447) **** ENDING LOGGING AT Mon Oct 10 02:59:57 2011