**** BEGIN LOGGING AT Wed Mar 30 02:59:57 2011 Mar 30 03:55:12 hmmm.. I'm seeing: http://fpaste.org/CTip/ ... but it looks like libusb-1.0 is being built before opkg has? Mar 30 03:55:35 aren't there dependencies to cause all of the host stuff to be built before any of the target stuff? Mar 30 04:13:59 and the makefile isn't modifying PKG_BUILD_DEPENDS... Mar 30 06:28:30 if there's a committer about, can I please get http://patchwork.midlink.org/patch/819/ committed? Mar 30 06:31:56 anyone know if Openwrt works with hotplug2 disabled and udev turned on??? Mar 30 07:11:52 xMff: ping Mar 30 07:36:15 build #89 of ar71xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ar71xx/builds/89 Mar 30 08:26:07 build #106 of cobalt is complete: Failure [failed shell compile_6] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/106 Mar 30 08:27:56 philipp64|laptop: what do you need udev for? Mar 30 08:46:52 morning Mar 30 09:31:27 acinonyx * r26360 /packages/net/quagga/ (13 files in 2 dirs): [packages] quagga: Update to v0.99.18 Mar 30 09:31:38 acinonyx * r26361 /packages/net/quagga/Makefile: [packages] quagga: Change to PKG_FIXUP:=autoreconf Mar 30 09:31:39 acinonyx * r26362 /packages/net/quagga/patches/ (12 files): [packages] quagga: Renumber patches Mar 30 09:44:17 hcg * r26363 /trunk/target/linux/omap35xx/gumstix/ (3 files in 2 dirs): [omap35xx] Add uboot-envtools to gumstix builds Mar 30 10:48:41 nbd * r26364 /trunk/package/qos-scripts/files/etc/config/qos: qos-scripts: remove the layer7 based classifiers from the default configuration - they are unreliable and prone to memory leaks Mar 30 10:49:39 nbd * r26365 /branches/backfire/package/qos-scripts/files/etc/config/qos: qos-scripts: remove the layer7 based classifiers from the default configuration - they are unreliable and prone to memory leaks (backport of r26364) Mar 30 10:50:04 got_milk: ping Mar 30 11:10:32 cshore: ping Mar 30 11:41:24 hcg * r26366 /trunk/target/linux/omap35xx/ (3 files in 2 dirs): [omap35xx] Update filesystem selection Mar 30 11:42:49 nbd * r26367 /trunk/package/hostapd/patches/730-nl80211_enable_qosdata.patch: nl80211: send EAP frames as QoS data frames for QoS aware clients, should fix EAP rekeying timeout issues with intel clients (#8830) Mar 30 11:42:53 nbd * r26368 /trunk/package/mac80211/patches/560-ath9k_fix_reported_signal_strength.patch: ath9k: fix reported signal strength Mar 30 12:29:14 hi Mar 30 12:29:30 i ported the kernel patches to 2.6.37.6 and compiled the kernel Mar 30 12:29:33 but i get this Mar 30 12:29:36 http://nopaste.info/183cacc9e8.html Mar 30 12:29:39 what am i missing? Mar 30 12:34:20 ahh i found the error Mar 30 12:37:31 nbd * r26369 /trunk/package/mac80211/Makefile: mac80211: add all drivers to PKG_CONFIG_DEPENDS Mar 30 12:42:04 nbd * r26370 /branches/backfire/package/ (108 files in 10 dirs): backport mac80211, hostapd, iw, crda from trunk as of r26369 Mar 30 13:35:29 hi Mar 30 13:35:49 i ported the kernel patches to 2.6.37.6 and compiled the kernel - but i get this http://nopaste.info/183cacc9e8.html - what am i missing? Mar 30 13:36:11 thoight it works - the patch i used is allready on the ml Mar 30 13:44:28 build #36 of lantiq is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/lantiq/builds/36 Mar 30 13:58:19 build #106 of orion is complete: Failure [failed shell_9] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/106 Mar 30 14:08:34 flyn * r26371 /packages/ (libs/libdmapsharing/Makefile net/dmapd/Makefile): Update dependencies for dmapd and libdmapsharing Mar 30 14:13:20 build #110 of brcm47xx is complete: Failure [failed shell_9] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/110 Mar 30 14:25:33 nbd: pong (from ~4 hours ago) Mar 30 14:55:28 build #109 of ubicom32 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/109 Mar 30 15:06:30 got_milk: did you ever run into any issues with intel clients that happened at rekeying time? Mar 30 15:11:31 nbd * r26372 /trunk/package/kernel/modules/crypto.mk: kernel: add missing dependency for ocf on kmod-crypto-hash Mar 30 15:18:28 nbd: I'm thinking that was causing the connection drops Mar 30 15:18:46 I applied the patch you posted in ticket 8830 and the dropping stopped Mar 30 15:18:51 cool Mar 30 15:18:59 there's still latency spikes but no actual drops Mar 30 15:19:12 how big are the latency spikes? Mar 30 15:19:23 and how frequent? Mar 30 15:19:40 ~200ms for about 3-4 seconds, every 3-5 minutes Mar 30 15:20:10 pinging from sta to ap or from ap to sta? Mar 30 15:20:33 sta -> ap Mar 30 15:20:38 ok Mar 30 15:20:49 and does this have a significant throughput impact when running e.g. iperf? Mar 30 15:21:09 maybe it's just doing some kind of background scanning or something like that Mar 30 15:21:12 it destroys the throughput Mar 30 15:21:15 ok Mar 30 15:21:30 I thought something like that too Mar 30 15:21:40 is it something that happens with other APs too? Mar 30 15:21:45 or is it still an openwrt specific issue? Mar 30 15:21:55 definitely openwrt-specific Mar 30 15:21:58 ok Mar 30 15:22:03 can't reproduce on other hardware or on stock Mar 30 15:22:05 firmware Mar 30 15:22:07 ok Mar 30 15:22:30 then i think the most likely cause is aggregation session teardown/renegotiation Mar 30 15:23:09 intel clients really like to do this for some reason Mar 30 15:23:23 and i don't entirely trust sequence number handling in the ath9k aggregation path yet Mar 30 15:24:04 it would make sense Mar 30 15:24:10 (to me, at least) Mar 30 15:24:52 please edit package/mac80211/Makefile and add -DCONFIG_MAC80211_HT_DEBUG to the BUILDFLAGS variable Mar 30 15:25:04 that should produce useful information about that kind of issues in the kernel log Mar 30 15:25:16 will do Mar 30 15:25:16 then you can see if that corresponds to the latency issues Mar 30 15:26:37 build #114 of at91 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/114 Mar 30 15:40:40 florian * r26373 /trunk/package/kernel/modules/i2c.mk: [octeon] package i2c master driver Mar 30 15:40:42 florian * r26374 /trunk/package/kernel/modules/usb.mk: [package] fix USB modules build on octeon Mar 30 16:01:40 ping nbd Mar 30 16:01:57 there is still one error left after your commit for ocf .... Mar 30 16:01:58 ERROR: "crypto_alloc_ablkcipher" [crypto/ocf/cryptosoft.ko] undefined! Mar 30 16:02:47 oh, right Mar 30 16:02:50 i'll take care of it Mar 30 16:02:51 thanks Mar 30 16:03:55 nbd * r26375 /trunk/package/kernel/modules/crypto.mk: kernel: ocf also requires kmod-crypto-mananger (which depends on kmod-crypto-hash) Mar 30 16:04:00 tripolar: there, should work now Mar 30 16:05:20 nbd: i will test but it should as cryptomanager includes - CONFIG_CRYPTO_BLKCIPHER Mar 30 16:05:28 yes Mar 30 16:11:15 nbd works Mar 30 16:11:17 nbd: built and flashed latest trunk with mac80211 debugging, I'll do some testing and report back in a bit Mar 30 16:11:28 tripolar: great Mar 30 16:11:30 got_milk: ok, thanks Mar 30 16:15:38 loswillios * r26376 /packages/net/openvpn/Makefile: [packages] openvpn: update to 2.1.4, thanks obsy Mar 30 16:16:39 nbd could you please apply my 2.6.37.6 patch - i sent a new version to the ml Mar 30 16:17:36 not right now, i'm about to leave Mar 30 16:32:24 nbd: you still around? Mar 30 16:32:59 got_milk: he'll be back in an hour Mar 30 16:35:21 ah, thanks blogic Mar 30 16:47:43 mb * r26377 /trunk/target/linux/generic/patches-2.6.38/980-vm_exports.patch: Mar 30 16:47:43 2.6.38 vm_exports.patch: Fix merge error introduced in r26332 Mar 30 16:47:43 Thanks to acoul for noticing. Mar 30 16:53:41 mb * r26378 /trunk/target/linux/generic/patches-2.6.38/980-vm_exports.patch: shmem_set_file: Remove ima_counts_get() Mar 30 17:23:08 build #84 of au1000 is complete: Exception [exception] Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/84 Mar 30 17:23:10 build #112 of s3c24xx is complete: Exception [exception] Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/112 Mar 30 17:23:13 build #113 of atheros is complete: Exception [exception] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/113 Mar 30 17:23:16 build #111 of brcm47xx is complete: Exception [exception] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/111 Mar 30 17:29:18 build #92 of gemini is complete: Exception [exception] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/92 Mar 30 18:15:53 build #80 of rdc is complete: Exception [exception] Build details are at http://tksite.gotdns.org:8010/builders/rdc/builds/80 Mar 30 20:15:01 florian * r26379 /packages/admin/pmacct/ (4 files in 2 dirs): Mar 30 20:15:01 [package] pmacct: add uacctd Mar 30 20:15:01 This patch add uacctd, an ulog based accounting deamon from pmacct project (http://www.pmacct.net/) Mar 30 20:15:01 It also put all pmacct packages in a submenu called (13 packages) Mar 30 20:15:01 Thanks in advance. Mar 30 20:15:02 Signed-off-by: Etienne CHAMPETIER Mar 30 20:15:03 florian * r26380 /packages/utils/wattsup/ (. Makefile src/ src/Makefile src/wattsup.c): Mar 30 20:15:03 [package] add wattsup Mar 30 20:15:04 This patch adds a command line utility to interface with the WattsUp Mar 30 20:15:04 power meters (https://www.wattsupmeters.com/) Mar 30 20:15:05 Signed-off-by: Roberto Riggio Mar 30 20:15:09 florian * r26381 /packages/utils/flashrom/ (Makefile patches/ patches/001-msr_dev_path.patch): Mar 30 20:15:09 [package] flashrom: use msr path compatible w/ hotplug2 Mar 30 20:15:09 The issue wasn't that flashrom-0.9.3 was broken, but that the incompatible device path put the box into a funky state that couldn't be reverted. Mar 30 20:15:09 This fix causes it to use the hotplug2 path. Mar 30 20:15:09 With this, we don't need to revert to 0.9.1, superseding my early patch to revert... Mar 30 20:15:10 Signed-off-by: Philip Prindeville Mar 30 20:33:34 jow * r26382 /trunk/package/firewall/ (Makefile files/lib/core_interface.sh): [package] firewall: prevent duplicate values in interface state vars Mar 30 20:35:09 jow * r26383 /branches/backfire/package/firewall/ (Makefile files/lib/core_interface.sh): [backfire] merge r26382 Mar 30 20:43:32 nbd: pong Mar 30 20:52:02 [florian]: ping Mar 30 21:02:57 cshore... maybe you know. Is there a way to add a CONFIG_package_xxx_yyy variable to further customize package 'xxx'? I'm thinking that it might be useful to include "atmdiag" and "atmdump" as part of "atm-utils". Mar 30 21:03:16 If we were using Config.in files, this would be trivial... Mar 30 21:04:46 philipp64|laptop: you mean like https://dev.openwrt.org/browser/trunk/package/busybox/Config.in or inline like https://dev.openwrt.org/browser/trunk/package/mac80211/Makefile#L83 ? ;) Mar 30 21:06:11 KanjiMonster: so we hybridize Config.in and the Makefile Package/ stuff? Mar 30 21:06:28 philipp64|laptop: yes Mar 30 21:15:48 ok... looking for another package out there that does something similar to what I want... Mar 30 23:13:26 luka * r26384 /packages/utils/at/ (5 files in 3 dirs): update at package Mar 30 23:15:25 Weird - when building 'newt' it installs into ipkg-x86 but then tries to copy out of ipkg-install !? Mar 30 23:15:55 Ie: I get ................./newt-0.52.12/ipkg-install/usr/bin/whiptail': No such file or directory Mar 30 23:16:14 luka * r26385 /packages/utils/at/patches/110-getloadavg.patch: update at package - forgot to add patch Mar 30 23:17:24 It's been a while since I built.. and I've finally gotten around to building from git so that I can manage my local changes better, and contribute them back Mar 30 23:42:56 can someone mark this as applied? http://patchwork.midlink.org/patch/790/ Mar 30 23:46:28 say, I'm confused... do I need to have CONFIG_PACKAGE_coreutils=y or not? it seems to change daily... Mar 31 00:32:44 what target do i want for a ubiquiti bullet M5? Mar 31 00:35:00 nm, forum says ar71xx Mar 31 02:14:20 nbd: you around? Mar 31 02:14:38 got_milk: yes Mar 31 02:14:56 I assume this will be of use to you: http://pastebin.com/c2dNBc9i Mar 31 02:15:42 no other client I've tested printed anything like that in the log Mar 31 02:15:57 only the intel client did Mar 31 02:17:01 got_milk: ok, so i was right about one thing and wrong about another Mar 31 02:17:20 i was right about the intel client killing the aggregation sessions Mar 31 02:17:26 but it's not a sequence number issue Mar 31 02:17:44 it seems that the intel client lets it re-establish the aggregation session Mar 31 02:17:54 only to kill it again 5 seconds later Mar 31 02:18:03 this looks a lot like a bug in the intel driver Mar 31 02:18:20 but maybe it can be worked around by increasing the time between aggregation session establish attempts Mar 31 02:18:32 i'll try to hack up a patch for that tomorrow Mar 31 02:20:01 cool Mar 31 02:20:21 I'm a tad busy over the next little while but I'll try to make time to test patches, etc Mar 31 02:58:10 dingo * r26386 /packages/net/wing/ (3 files in 3 dirs): [patch-team] update wing to 20110329 - Signed off by Roberto Riggio **** ENDING LOGGING AT Thu Mar 31 02:59:58 2011