**** BEGIN LOGGING AT Sat Sep 17 02:59:57 2011 Sep 17 03:03:38 mazilo * r28248 /packages/net/freeswitch/ (2 files in 2 dirs): bumped to latest git version with a support for applications/mod_blacklist Sep 17 07:03:48 nbd * r28249 /trunk/package/mac80211/patches/571-ath9k_reset_debug.patch: ath9k: fix compile without debugfs (#10117) Sep 17 07:41:18 Chocky: ping Sep 17 08:35:05 nbd * r28250 /trunk/package/mac80211/patches/551-ath9k_fix_keymiss_handling.patch: ath9k: fix a regression in handling of MAC key miss events (should improve connection stability) Sep 17 08:35:09 nbd * r28251 /trunk/package/mac80211/patches/522-mac80211_reduce_txqueuelen.patch: mac80211: reduce tx queue length of wireless interfaces - improves latency under load Sep 17 08:42:21 nbd * r28252 /trunk/package/mac80211/patches/551-ath9k_fix_keymiss_handling.patch: ath9k: do not strip MMIC for key miss frames Sep 17 10:57:32 Chocky: committed some stuff that might help with your stability issues Sep 17 16:03:27 nbd * r28253 /trunk/package/mac80211/patches/ (5 files): ath9k: add some cleanup patches and rework tx power handling Sep 17 16:18:04 Chocky: re-ping? Sep 17 17:47:02 nbd: will try, thanks Sep 17 17:56:12 Chocky: please try latest first Sep 17 17:56:22 and if that doesn't work, go back to the old version that i mentioend Sep 17 17:57:39 aye Sep 17 18:09:40 rebooting to new version Sep 17 18:22:54 trying now. wish me luck Sep 17 18:34:36 nbd: trunk fail Sep 17 18:37:06 Chocky: fail how? Sep 17 18:38:02 I don't know exactly, but connectivity stopped after ~15 mins. unlike that old one, which ran for nearly a day before I tried this Sep 17 18:38:36 please re-test to verify if it really got worse Sep 17 18:39:32 did reconnecting to the AP work or did you have to restart wifi? Sep 17 18:39:54 I had to power cycle; I'm in the middle of something else Sep 17 18:40:04 but I'll run until it fails again if you want more info Sep 17 18:41:17 yeah, i'd like the output of cd /sys/kernel/debug/ieee80211/phy0/ath9k; cat wiphy interrupt xmit Sep 17 18:41:26 after it has failed Sep 17 19:40:26 ok Sep 17 19:41:32 nbd: http://pastebin.com/LM6BeBku not sure how long it went down for, my connections stayed up. Sep 17 19:41:37 only that it did Sep 17 19:42:18 in the previous case, my connections would generally drop Sep 17 19:42:27 IRC, that is. Sep 17 21:15:47 Chocky: i'll have a new patch for testing in a minute Sep 17 21:16:19 Chocky: the issue seems to be unrelated to my recent changes, it may be coincidence that it seems to be worse with the latest build Sep 17 21:18:09 right Sep 17 21:18:19 I'm popping out; back later to try whatever Sep 17 21:19:00 http://nbd.name/590-ath9k_rxeol_handling.patch Sep 17 21:19:05 copy that to package/mac80211/patches/ Sep 17 23:18:00 nbd: running now Sep 17 23:51:39 hey, anyone know if switching from gcc 4.6 to 4.5 will fix the swconfig sefaults on ar71xx? Sep 17 23:52:06 wondering if i should wait for this compile, or just switch to backfire Sep 17 23:55:27 :( Sep 17 23:57:32 most likely Sep 17 23:58:02 afair 4.6 is not the default Sep 17 23:58:24 it isnt Sep 18 00:00:53 build #56 of iop32x is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/56 Sep 18 00:00:54 build #55 of etrax is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/55 Sep 18 00:00:56 build #53 of ep93xx is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/53 Sep 18 00:00:58 build #51 of gemini is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/51 Sep 18 00:02:13 cool we'll see Sep 18 00:02:32 curious to see what this txpower fix does Sep 18 00:02:58 i remember setting a bunch of defs in some ath9k files months ago to fix it Sep 18 00:03:14 its like the hardcoded stuff overrides any crda values Sep 18 00:03:28 i dunno who keeps breaking it upstream :( they should be yelled at Sep 18 00:03:49 crda lists regulatory limits, the driver caps those to hardware limits which are defined in the eeprom Sep 18 00:04:08 something besides the country code? Sep 18 00:04:16 in the eeprom Sep 18 00:04:18 like 0x0 Sep 18 00:05:15 yes Sep 18 00:05:20 oh Sep 18 00:05:43 buffalo capped them at 21dbm? for the wzr-hp-g300nh Sep 18 00:05:53 lame, i have done more than that with it Sep 18 00:06:03 i think i had it at 25dbm Sep 18 00:06:09 a year ago Sep 18 00:06:09 did you measure it? Sep 18 00:06:16 no Sep 18 00:06:25 so some tool told you it did 25 Sep 18 00:06:42 iwconfig on my wrt54gl says its transmitting with 32dbm :) Sep 18 00:06:47 heh Sep 18 00:07:03 but for these mikrotik r52n, they should do 25dbm Sep 18 00:07:07 i think 27? Sep 18 00:07:12 capped at 23 Sep 18 00:16:31 nbd: still went down. need the same again? Sep 18 00:25:55 the FCC documents list a peak power of 26.33 dBm (429mW) for the wzr-hp-g300nh Sep 18 00:27:26 the average maximum is ~23dBm across tested channels Sep 18 00:28:37 https://fjallfoss.fcc.gov/eas/GetApplicationAttachment.html?id=1172739 Sep 18 00:28:42 after reading those measurement tests, an upper limit of 23dbm does not look unreasonable to me Sep 18 00:30:17 page 99 Sep 18 00:31:07 thanks for looking into it Sep 18 00:32:02 what caused a lot of confusion is that in earlier openwrt versions, the driver did report the regulatory, not the hardware limit Sep 18 00:33:37 that led people to believe that there's 30 dbm possible which are now somehow "locked" Sep 18 00:35:04 the R52N seems to be rated for 26dbm Sep 18 00:37:54 https://fjallfoss.fcc.gov/eas/GetApplicationAttachment.html?id=1109564 page 37 Sep 18 00:41:09 it wont allow me to view those direct links Sep 18 00:41:21 ah Sep 18 00:41:37 i see it rated at 350mw or so (25dbm at least) in adverts though Sep 18 00:41:51 i dont think the r52n is made anymore, for some time Sep 18 00:42:00 for me government agency >> adverts :) Sep 18 00:42:08 totally Sep 18 00:42:28 can you open this? https://fjallfoss.fcc.gov/oetcf/eas/reports/ViewExhibitReport.cfm?mode=Exhibits&RequestTimeout=500&calledFromFrame=N&application_id=413699&fcc_id=%27TV7R52N%27 Sep 18 00:44:20 yes Sep 18 00:44:25 "test report unii I" Sep 18 00:44:49 wzr-hp-g300nh is here, in case you're interested: https://fjallfoss.fcc.gov/oetcf/eas/reports/ViewExhibitReport.cfm?mode=Exhibits&RequestTimeout=500&calledFromFrame=N&application_id=306828&fcc_id=%27FDI-09101560-0%27 Sep 18 00:46:49 should maybe add those to the wiki but I am too lazy atm Sep 18 00:47:47 yes i see the results Sep 18 00:48:44 its showing like 25dbm on 802.11n Sep 18 00:49:15 closer to 30dbm at 20mhz Sep 18 00:49:29 where does 21dbm come from then? Sep 18 00:49:45 you're saying thats what the eeprom says? Sep 18 00:50:02 there ought to be a way to override that, then :-) Sep 18 00:50:19 * m4t thinks its gimped Sep 18 00:51:51 the wzr ? Sep 18 00:53:02 most likely, yes Sep 18 00:53:21 yea the wzr Sep 18 00:53:24 the r52n too Sep 18 00:53:39 * m4t <3 2.4ghz radiation Sep 18 00:54:14 I think its all just snake-oil without an external measurement Sep 18 00:55:28 gcc 4.5 fixed swconfig it seems, on ar71xx trunk Sep 18 00:55:37 i left binutils at default too, 2.20.1 i think Sep 18 00:59:01 for the wzr look at the chain 0 .. chain 2 columns Sep 18 01:49:32 yikes, trunk definitely didnt fix my xbox360 woes on the wzr-hp-g300nh Sep 18 01:49:39 i saw 10,000+ms latency :( Sep 18 01:50:07 it was only 1500ms or so in backfire ;) **** ENDING LOGGING AT Sun Sep 18 02:59:57 2011