**** BEGIN LOGGING AT Mon Nov 28 02:59:57 2011 Nov 28 03:01:09 jow_laptop: loswillios: 'git branch -l' didn't show branches on packages Nov 28 03:01:36 jow_laptop: re kernel versions: some things (usb) are fixed for me in >3.0 Nov 28 03:02:14 but you are right, every release is getting bloated and more things are broken than fixed Nov 28 03:02:37 for me the best release was 2.6.32 but sadly trunk drop it for a bad release like 2.6.39 Nov 28 03:02:50 IMHO Nov 28 03:04:36 * solca_ forgot git is not the official version control system :) Nov 28 05:31:53 jow * r29346 /packages/net/samba3/patches/110-compile_fixes.patch: [packages] samba3: kill pointless AC_TRY_RUN() Nov 28 05:32:20 I can't think of any appropraite punishment for inventing stuff like AC_TRY_RUN Nov 28 06:11:01 jow * r29347 /packages/net/samba3/ (Makefile files/samba-nmbd.init files/smb.conf.template): [packages] samba3: add missing samba-nmbd.init, remove interface setting from config template Nov 28 06:57:58 build #103 of ramips is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/103 Nov 28 08:08:40 build #105 of cobalt is complete: Exception [exception compile_5 shell_14 compile_12] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/105 Nov 28 08:08:41 build #103 of pxcab is complete: Exception [exception compile_4 shell_14 compile_12] Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/103 Nov 28 08:08:44 build #77 of rdc is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/rdc/builds/77 Nov 28 08:08:46 build #107 of atheros is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/107 Nov 28 09:46:55 build #97 of ppc44x is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/97 Nov 28 09:59:32 blogic: ping Nov 28 10:20:38 Delboy_: hey Nov 28 10:22:17 blogic: hi Nov 28 10:23:33 well I wanted to ask you what wlan chip is on arv45xx but i found it on forum ar2417 :) Nov 28 10:23:57 yes Nov 28 10:24:49 btw I was just looking at that ath5k eeprom stuff i already posted but Nov 28 10:25:49 it looks like i did this wrong in that platform data patch if (pdata && pdata->eeprom_data && pdata->eeprom_data[0] == AR5K_EEPROM_MAGIC_VALUE) Nov 28 10:25:56 blogic * r29348 /trunk/target/linux/lantiq/Makefile: lantiq: bump to 3.1.3 Nov 28 10:26:07 ok Nov 28 10:26:08 nice :D 3.1.3 Nov 28 10:26:13 yes :D Nov 28 10:26:31 http://phrozen.org/lantiq.dts Nov 28 10:26:35 DT \o/ Nov 28 10:26:47 been busy with this for over a week now Nov 28 10:26:56 anyhow ... whats wrong with that line you pasted ? Nov 28 10:27:15 all the time I thought that those platform data eeproms start with AR5K_EEPROM_MAGIC_VALUE Nov 28 10:27:25 aa5555aa ? Nov 28 10:27:31 yeah Nov 28 10:27:34 yes Nov 28 10:27:39 they do imho ?! Nov 28 10:27:49 i can dig out a board later on to check Nov 28 10:28:55 yes but I see in madwifi patch KSEG1ADDR(lantiq_madwifi_eep_addr + addrsel))); Nov 28 10:29:25 addrsel is the offset for the ebu Nov 28 10:29:26 and lantiq_madwifi_eep_addr is at 0xb07f0400 Nov 28 10:29:34 * blogic looks Nov 28 10:31:21 and in arv45xx is #define ARV45XX_BRN_ATH         0x3f0478 Nov 28 10:31:33 blogic: nice Nov 28 10:31:46 Delboy_: i will ask nbd about this Nov 28 10:31:50 the madwifi patch works Nov 28 10:31:58 i am using it right now on my dsl gateway Nov 28 10:32:01 hmmm Nov 28 10:32:16 let me wrap up some other stuff first and then i will get some hw to play with Nov 28 10:32:25 ok Nov 28 10:32:26 i am also getting a board this week that has ath9k eep in flash Nov 28 10:32:46 so i can finally fix luca olivettis patch for him and get him to stop ranting Nov 28 10:32:55 :D Nov 28 10:32:59 seriously Nov 28 10:33:07 he mailed me that the patch dont work Nov 28 10:33:27 and everytime anyone on the ml says anything about wifi/lantiq he shouts "my patch is rotting" .... Nov 28 10:33:35 its very childish Nov 28 10:33:43 loswillios: yes :D Nov 28 10:36:04 blogic: should I update that ath5k pci ids patch, or wait until you fix luca's patches to detect pci id properly? Nov 28 10:36:28 you mean add the comments ? Nov 28 10:36:33 yep Nov 28 10:37:11 what chipset do you have ? Nov 28 10:37:17 i can fix that when i merge the patch Nov 28 10:37:24 the id that was already there was a 2417 Nov 28 10:37:30 btu whats the new one ? Nov 28 10:38:03 sx673 has ar2414 and sx762 has ar2413 Nov 28 10:38:43 both 2413 and 2414 are detected as ff16 Nov 28 10:39:28 ok Nov 28 10:39:39 i will add that when i tested/megred the patch Nov 28 10:39:47 the patch does wot for the sx76x though ? Nov 28 10:39:51 the patch does work for the sx76x though ? Nov 28 10:39:56 * blogic cant type today Nov 28 10:40:22 ahhh .... you use the arteep.bin thing Nov 28 10:40:26 i will test it :D Nov 28 10:40:36 yeah that array in .h Nov 28 10:47:03 Delboy_: more trolling from luca :D Nov 28 10:47:10 i rest my case Nov 28 12:22:47 jow_laptop: not sure you got my last night feedback here, may i bother you again? seems like after a last night's trunk flashing, after the router starts and is configured for pppoe auth and pppoe is up, some sites wont work, prolly mss problems. if i restart the firewall via /etc/init.d/firewall it works Nov 28 12:23:19 so far on dlink dir825 Nov 28 12:23:25 ok Nov 28 12:23:46 no idea. diff the iptables rulesets next time Nov 28 12:23:47 ill test the tplinks tonight and get back if i got relevant info Nov 28 12:23:55 yes, i will tonight Nov 28 12:23:56 the ones in -t raw Nov 28 12:24:07 sorry, -t mangle Nov 28 12:24:12 ok, mangle Nov 28 12:49:18 blogic * r29349 /trunk/target/linux/omap24xx/Makefile: omap24xx: bump to 3.1.3 Nov 28 12:54:25 jow * r29350 /packages/utils/zoneinfo/Makefile: [packages] zoneinfo: update to 2011n (#10317) Nov 28 12:55:23 {Nico}: hey, how's radvd 1.8.3 testing? ive been using it for few days without problems :) Nov 28 13:00:21 jow_laptop: what's the most pain-free way of having two separate compile configs? Nov 28 13:00:42 scripts/env Nov 28 13:00:48 my currenct config targets ar71xx... say I wanted to do a new config for ARM without nuking the existing one Nov 28 13:01:28 ooh, lovely Nov 28 13:02:18 jow_laptop: is scripts/env able to handle kernel config too..? Nov 28 13:02:22 no Nov 28 13:02:40 kernel config seems broken to me Nov 28 13:02:56 as in, the menuconfig interface actually corrupts itself as you traverse it Nov 28 13:03:47 although... what if I cd to /build_dir// and do a make xconfig from there? Nov 28 15:51:53 whoo boy, I've gone back to try and update a patch to include some .config changes, and I can't even begin to remember how I did all this quilt stuff Nov 28 15:52:07 guess it's going to take a few more times before this gets comfortable and familiar :) Nov 28 15:52:38 Delboy_: pong Nov 28 17:01:16 nbd: ping Nov 28 17:01:25 pong Nov 28 17:01:29 hi Nov 28 17:01:32 hi Nov 28 17:01:48 did blogic asked you about platform data eeprom? Nov 28 17:01:53 no Nov 28 17:03:18 no Nov 28 17:03:21 ;) Nov 28 17:03:31 :) Nov 28 17:03:32 Delboy_: i was plannin to do so later Nov 28 17:03:45 i noticed you pinged me a few times Nov 28 17:03:48 yeah but i pinged nbd about it y-day :D Nov 28 17:03:53 haven't been at the computer much Nov 28 17:03:58 basically ... Nov 28 17:04:05 i'm still in indonesia ;) Nov 28 17:04:21 for those boards that have the legacy ath eep inside flash we used the madwifi patch Nov 28 17:04:29 there is now a patch that does the same for ath5k Nov 28 17:04:41 but we are not sure exactly what offset into the eep we should use Nov 28 17:05:13 nbd: its this mail from 2 days ago --> [OpenWrt-Devel] [PATCH v2 1/2] mac80211: read eeprom from platformdata Nov 28 17:06:19 ok, use the offset in a way that works for ath5k, then fix madwifi to work in the same way Nov 28 17:06:55 and if madwifi doesn't work, then that's ok as well Nov 28 17:07:00 i want to get rid of it anyway Nov 28 17:07:33 ok Nov 28 17:07:34 haha Nov 28 17:07:44 too much madness Nov 28 17:07:45 :D Nov 28 17:07:51 last time i tried ath5k it was usable Nov 28 17:07:59 kk Nov 28 17:08:07 (after me doing a round of bugfixes) ;) Nov 28 17:08:15 hrr hrr Nov 28 17:08:22 Delboy_: ok i will simply test your patch tomorrow Nov 28 17:08:28 i dont have the board on me just now Nov 28 17:14:09 any of you kids tried phc on OpenWrt? Nov 28 17:14:17 requires an older version of PHP, which is not overly helpful Nov 28 17:14:51 (PHP compiler) Nov 28 17:15:22 * blogic looks around to see if he can spot the kids Nov 28 17:20:34 nbd, ping Nov 28 17:20:55 OutBackDingo: pong Nov 28 17:21:08 nbd, who manages ssh keys for WRT Nov 28 17:21:14 i need mine updated Nov 28 17:21:24 Kaloz Nov 28 17:21:34 ahhhh ... ok thanks Nov 28 17:32:28 When trying to compile perl, the compilation spits out error messages as shown here (http://pastebin.com/dE4uQac9), see line #1300. Nov 28 17:34:01 mazilo: Nov 28 17:34:10 mazilo: line 1300 has no error Nov 28 17:34:58 blogic: What about line #1301? Nov 28 17:42:46 guess not Nov 28 18:17:12 anyone here have a WNDR3800? Nov 28 18:17:20 I think the flash chip is 32MB, not 16 Nov 28 18:19:44 I've just dumped all 10 MTD partitions from my WNDR3800 stock device Nov 28 18:19:58 mtd2 and mtd9 contain completely separate data Nov 28 18:20:09 mtd9 appears to be squashfs Nov 28 18:20:18 mtd2 I'm not sure about Nov 28 18:20:39 in any case, since they are definitely not the same part of the flash, it would lead me to believe that the total size of the flash is 32MB, not 16 Nov 28 18:22:33 ugh, never mind, they are Nov 28 18:22:37 almost anyway Nov 28 18:22:52 mtd2 has a 64 byte header pre-pended Nov 28 18:23:10 remove it and the two partitions are identical, so much for that then Nov 28 18:23:45 was about to say that mtd aprtitions may intersect Nov 28 18:24:38 hm, still not quite identical Nov 28 18:25:06 once you get to 0xE3FC00 it differs Nov 28 18:25:21 odd arrangement Nov 28 18:53:16 I made some kernel patches ok using quilt, but I'm trying to edit the kernel config now. I tried make clean,prpare QUILT=1, then quilt edit build_dir/linux-atheros/linux-2.6.37.6/.config Nov 28 18:53:23 but that hasn't updated into the right place. Nov 28 18:54:12 is there some trick to updating the default config files? Nov 28 18:54:25 or should I just add the option to the Kconfig file? Nov 28 19:04:50 use kernel_menuconfig make target Nov 28 19:05:09 or kernel_oldconfig Nov 28 19:06:04 no, I'm trying to make an option appear in the default. Nov 28 19:06:20 do I just have a diff containing the patch, and also touching the config.2.6.37 file? Nov 28 19:06:28 so a diff contianing a diff and a normal change? Nov 28 19:07:20 no idea what you are talking about Nov 28 19:07:38 mb__: hi Nov 28 19:07:39 Those two targets are for modification of the platform kernel config Nov 28 19:07:42 did you see my mail ? Nov 28 19:08:05 blogic: The version bump? Nov 28 19:08:14 yes Nov 28 19:08:19 looks good Nov 28 19:08:59 ok Nov 28 19:09:05 just so you know if it blows up it was me Nov 28 19:09:07 In general I don't have a problem with people bumping my stable revisions. It just has to compile. Nov 28 19:09:11 but 3.1.3 boots fine on lantiq Nov 28 19:09:18 it did :D Nov 28 19:09:22 :) Nov 28 19:10:24 whats the n810 support like ? Nov 28 19:10:36 i was looking at used units on ebay Nov 28 19:11:12 the major remaining problem still is charging support. Nov 28 19:12:01 I'll have to get my ass up and reverse that stupid calibration code. :) Nov 28 19:14:02 ok Nov 28 19:14:06 i will order one in this case Nov 28 19:19:10 jow * r29351 /packages/net/samba3/patches/ (10 files): [packages] samba3: add missing patches from previous update commit, refresh series Nov 28 19:20:54 blogic: The missing piece is a 200 instruction ARM snipped that implements some software floating point logic for converting the calibration values. Nov 28 19:21:40 ok Nov 28 19:21:51 blogic: As this heavily screws my brains, my current approach is to put it into a small C program, feed it with values and execute it. Maybe the output pattern is obvious. I bet it's some really trivial calculation. Nov 28 19:22:34 nbd * r29352 /trunk/ (toolchain/Makefile tools/Makefile): build: add a lib64 symlink in staging_dir/host and staging_dir/toolchain* for systems that prefer this as library path (e.g. current SuSE), fixes mpfr and gcc build Nov 28 19:22:48 It's just stupid float (double precision) Nov 28 19:23:01 ok Nov 28 19:23:06 got a link to the code ? Nov 28 19:26:39 blogic: http://pastebin.de/20565 Nov 28 19:27:04 blogic: it's not really a sub, though. As far as I can see it has at least two entry points Nov 28 19:27:17 It takes a float double as argument Nov 28 19:27:47 and it most likely returns an integer in some 1000th scaling. Nov 28 19:52:57 anyone out there using UDT? Nov 29 02:34:35 nbd: ping Nov 29 02:47:05 solca_: pong Nov 29 02:47:15 nbd: hi! Nov 29 02:47:30 hi Nov 29 02:48:14 I'm debugging a regression in hostapd where ap_max_inactivity is not doing any effect Nov 29 02:48:35 and would like to ask you if inactivity has something to do at the driver level (ath9k)? Nov 29 02:48:53 or if is just handled completely in userspace in hostapd? Nov 29 02:49:57 hostapd asks mac80211 for the inactivitity time Nov 29 02:50:00 because I'm checking a prior backfire hostapd version in trunk but it's the same thing Nov 29 02:50:21 probably a mac80211 thing then Nov 29 02:50:30 you should be able to query the inactivity time using iw Nov 29 02:50:36 inactivity is reported right via 'iw' Nov 29 02:50:40 hm Nov 29 02:50:48 but many APs getting million ms Nov 29 02:51:09 and never get dropped by inactivity Nov 29 02:51:09 add a debug statement to the driver op in hostapd that queries it Nov 29 02:51:20 and to the code that compares it with the max inactivity Nov 29 02:52:18 ok, will do it, but now I know that it's not a ath9k thing at least Nov 29 02:52:30 hostapd doesn't know about ath9k Nov 29 02:52:34 it only knows about mac80211 Nov 29 02:52:59 yeah but if iw reports the right inactivity then the problem must lie in hostapd Nov 29 02:53:04 yep Nov 29 02:53:17 ok, thank you nbd! Nov 29 02:53:44 no problem, let me know how it goes Nov 29 02:53:54 ok **** ENDING LOGGING AT Tue Nov 29 02:59:57 2011