**** BEGIN LOGGING AT Thu Feb 10 02:59:57 2011 Feb 10 03:07:07 lol... interesting note in the config files... i wonder what this means... GameBoyAdvance CONFIG_ARCH_GBA" AplioTrio Feb 10 03:21:34 nbd, it might be me, but it appears asus inconspicuously left the kernel out of the files (major hints in the makefiles that it should be using linux 2.4.x Feb 10 03:21:37 !?! Feb 10 03:22:07 neutronscott, i'm looking for the source to the marvell chip which my old airlink ar315w uses. Feb 10 03:29:18 dandel: are you actually surprised that the code is a mess? Feb 10 03:29:53 not really... just a good grin at what is in the mess. Feb 10 03:31:03 from the what i can tell, the tarball from asus should contain... kernel_version=2.4.22-uc (however, i can only find 2.0.39-uc) Feb 10 03:33:39 That's from the file, mvwlan.o which claims the license is "GPL" and has the vendor/device combo of "Marvell"/ "Wireless LAN Driver" (also shows kernel ver) Feb 10 03:38:16 nbd * r25439 /trunk/package/mac80211/patches/532-ath9k_remove_pending_frames_workaround.patch: ath9k: remove an old tx stop workaround which conflicts with the powersave fixes Feb 10 03:38:21 nbd * r25440 /trunk/package/iw/ (9 files in 2 dirs): iw: update to 0.9.22 Feb 10 03:38:24 nbd * r25441 /trunk/package/mac80211/patches/540-mac80211_add_rx_rate.patch: mac80211: add a patch for keeping track of the rx bitrate of remote stations Feb 10 03:38:28 nbd * r25442 /trunk/package/iw/patches/100-rx_rate.patch: iw: add support for showing the rx bitrate Feb 10 03:50:08 nbd, it's possible the old software release was missing a few things (noticed the newest rev is over 40mb larger) Feb 10 03:50:47 wouldn't surprise me Feb 10 03:50:51 jacked up Feb 10 03:50:55 happens all the time Feb 10 03:51:15 i emailed linksys about "GPL" appearing in binaries and eventually got the sources i seeked. Feb 10 03:51:27 3months tho. Feb 10 03:52:21 neutronscott, that's better than what i did when asking for gpl sources from android tablets out of china... got ignored completely. Feb 10 03:53:32 nbd * r25443 /trunk/toolchain/kernel-headers/Makefile: kernel-headers: get rid of the arch/$(ARCH)/include/asm vs include/asm-$(ARCH) copying mess and stick to what the kernel is using Feb 10 03:53:34 I even went as far as requesting the source via an support ticket... which was closed when they opted to take the official stance of no further updates for that device (and their other linux devices at the same time) Feb 10 03:57:39 nbd, yep... Just as i thought, it contained the kernel... now to see if they add any binary blobs. Feb 10 04:03:39 thought the chinese devices were usually just generic builds Feb 10 04:03:45 and a quick u-boot thrown in.. Feb 10 10:07:33 build #68 of avr32 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/avr32/builds/68 Feb 10 15:51:35 I am getting a white screen when accessing LuCI, and cmd says bad precompiled header. Feb 10 15:51:52 This was after an error when trying to install a package. Feb 10 15:52:06 Has LuCI corrupted itself? :-/ Feb 10 16:06:23 jmccrohan: my guess would be that you probably installed some luci 0.10 libs/apps that don't work with 0.9 Feb 10 16:08:41 soma: ah. i thought they were backward compatable. Feb 10 16:10:08 no, i think they aren't Feb 10 16:15:49 ok. how then can i force menuconfig to build 9.0 packages? it was a fresh backfire checkout/package update built yesterday. Feb 10 16:24:37 you can change feeds.conf.default Feb 10 16:33:35 soma: thanks. exactly what i wanted. odd that it was already pointing to /branches/luci-0.10/. I wonder where the 0.9 packages came from. Feb 10 16:34:58 if you plan to completely install that router it would be better to go with 0.10. i was thinking you had a router with older firmware on it and then installed 0.10 packages Feb 10 16:36:12 When I said that, I didn't realise that luci-core was now at 0.10. :) Feb 10 17:15:38 heil * r25444 /packages/net/haproxy/Makefile: Feb 10 17:15:38 Package: haproxy Feb 10 17:15:38 - increase default bufsize to 16384 so that Feb 10 17:15:38 http error 400 is not thrown any longer due to Feb 10 17:15:38 large get requests (reduce this in the global section with (tune.bufsize 8192) Feb 10 17:15:39 to save memory Feb 10 17:15:39 Thx. to Willy Feb 10 17:16:07 weird, the bcom4716's really do support MIPS ASE DSP R2 Feb 10 17:16:15 heil * r25445 /packages/net/haproxy/Makefile: Package: haproxy (forgot to increse Version) Feb 10 17:17:41 shame they tend to have that broadcom wifi also Feb 10 17:34:18 build #63 of etrax is complete: Failure [failed shell compile_3] Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/63 Feb 10 17:36:59 ping nbd Feb 10 17:37:04 pong Feb 10 17:37:17 heil * r25446 /packages/net/haproxy/Makefile: Package: haproxy bump to version 1.4.11 Feb 10 17:37:27 nbd: did you read the mail on the mailinglist from Philip Prindeville? Feb 10 17:37:36 about Enable DSCP QoS for openssh-5.7p1 Feb 10 17:37:59 id say stuff like this souldnt be enabled in the default config Feb 10 17:38:05 or am i wrong? Feb 10 17:38:15 haven't had time to look into this Feb 10 17:38:31 its only the options for the ssh/d config files Feb 10 17:38:50 but id say this is none default stuff and can make trouble with old hardware Feb 10 17:41:57 nbd is qos enabled in the default image? Feb 10 18:04:14 florian * r25447 /trunk/ (2 files in 2 dirs): [kernel] add missing 2.6.37 config symbols. Feb 10 18:04:17 florian * r25448 /trunk/target/linux/malta/Makefile: [malta] update to 2.6.37 Feb 10 18:04:22 florian * r25449 /trunk/target/linux/realview/ (Makefile config-default): [realview] switch to 2.6.37 Feb 10 18:04:23 florian * r25450 /trunk/package/pjsip/Makefile: [package] pjsip: fix linking failure w/ ARM eabi Feb 10 18:36:16 tripolar * r25451 /packages/net/openssh/patches/200-dscp-qos.patch: Feb 10 18:36:16 [openssh] add DSCP QoS options to ssh_config and sshd_config Feb 10 18:36:16 thanks Philip Prindeville Feb 10 19:59:02 tripolar: pfifo_fast is the default in Linux and honours the QoS values in the TOS field Feb 10 19:59:14 IMHO that patch should have not been merged Feb 10 20:06:04 jow * r25452 /trunk/package/base-files/files/usr/share/udhcpc/default.script: [package] base-files: derive lease_acquired dhcp state var from /proc/uptime Feb 10 20:06:43 jow * r25453 /branches/backfire/package/base-files/files/usr/share/udhcpc/default.script: [backfire] merge r25452 Feb 10 21:37:01 I submitted my patches to trac ticket system, but I get no follow up.... Feb 10 21:39:58 that sucks Feb 10 22:32:51 ping KanjiMonster Feb 10 22:33:01 xMff: pong Feb 10 22:33:14 did you see the rt-n16 stuff in the mailinglist? Feb 10 22:33:18 yeah Feb 10 22:33:44 did this guy got in touch with you? Feb 10 22:33:49 sort-of Feb 10 22:34:39 he sent an email to Micheal Buesch and me, then after telling him he should post to a mailing list posted patches to linux wireless Feb 10 22:34:44 so not really much discussion there Feb 10 22:35:33 I see Feb 10 22:35:55 he asks for a different branch due to some code changes, wonder where that was discussed, I didn't follow it Feb 10 22:38:33 as far as I understood a different target, to keep the experimential code away from the working code for older bcm47xx Feb 10 22:38:50 I see Feb 10 22:39:01 might want to wait until hauke is bad Feb 10 22:39:09 erm back Feb 10 22:39:10 lol Feb 10 22:39:51 freudian slip ? :P Feb 10 22:40:23 downside of two screen... reading on the right while typing on the left and thinking about something else Feb 10 22:40:49 didn't to too much lately for the same reasons Michael Buesch is currently complaining, the whole pci(e) code needs to be changed/fixed up, too; and I don't have anything to test with (and neither he has) Feb 10 22:42:29 btw having a separate "new broadcom" target makes sense anyway, since then we can make use of the mips 74k's enhanced instruction set, which should improve performance Feb 10 22:43:02 does that really require a separate target? Feb 10 22:43:08 What about a config option? Feb 10 22:43:18 could be a subtarget Feb 10 22:43:33 since subtargets can override cflags as well Feb 10 22:44:24 yeah, subtarget is fine Feb 10 22:45:21 okay Feb 10 22:45:40 KanjiMonster: As you also figured out already, his code is not complete as-is. So any "see? this is very small and nonintrusive" arguments have to be taken with a grain of salt Feb 10 22:46:40 What bothers me most is the major layer violation he's doing, however. Feb 10 22:46:47 mb__: I think the pci(e) core code will be /the/ location where it will get hairy Feb 10 22:46:57 He's basically repeating the same crap that broadcom did in their SB code. Feb 10 22:47:15 yep Feb 10 22:48:02 So I'm still for an additional "HND" layer with separate busses (AI or SSB) below it. Feb 10 22:53:40 which should probably get a different name, since "HND" stands for "home networking devision" ;) Feb 10 22:57:51 yeah i know :) Feb 10 22:58:13 which reminds me that I should finish cleaning up the serial flash support code for ssb/bcm47xx Feb 10 23:04:15 xMff: btw, the arm machine update breaks 2.6.38-rc4 (got included there) Feb 10 23:05:11 ok. Feb 10 23:05:16 it worked with rc3 Feb 10 23:05:16 ipv6 in client mode seems to not wor propperly Feb 10 23:05:26 not dhcp but stateless Feb 10 23:05:28 what kind of client mode? Feb 10 23:06:34 it works somtimes, i also get a local one. but not a global - the ipv6 however discover the net Feb 10 23:06:44 xMff: according to the rc4 changelog, there was an arm machtype update :) Feb 10 23:06:45 what kind of client mode? Feb 10 23:07:09 I require IPv6 for the dockstar in a local network Feb 10 23:07:47 xMff: if you mean me the stateless one Feb 10 23:07:59 I mean how is the client mode configured Feb 10 23:08:55 only eth0 has dhcp Feb 10 23:09:25 but since the local ipv6 is assigned it should assign the global too Feb 10 23:09:31 (00:04:56) das-Iro: ipv6 in client mode seems to not wor propperly Feb 10 23:09:47 so you're not talking about wifi? Feb 10 23:09:52 i somtimes get an global one Feb 10 23:09:57 no ethernet Feb 10 23:10:04 i have a dockstar Feb 10 23:10:05 has the device ipv6 forwarding enabled? Feb 10 23:10:36 why should the dockstar forward ipv6? Feb 10 23:11:10 fw needs to be disabled on non-routers Feb 10 23:11:12 xMff: ah, wait.. maybe I am too tired.. let me check again after a git pull Feb 10 23:11:16 I didn't ask whether it makes sense, I wanted to know whether the sysctl is on or not Feb 10 23:11:26 if it is on you need to override accept_ra Feb 10 23:11:28 sysctl -w net.ipv6.conf.eth0.accept_ra=2 Feb 10 23:11:33 i removed the entire firewall stuff Feb 10 23:11:41 this is about sysctl, not firewall Feb 10 23:12:05 I will try that when i'm home Feb 10 23:12:51 forwarding disables listening to router advertisements (unless you explicitly enable it again) Feb 10 23:17:58 xMff: nope, still breaks Feb 10 23:18:11 KanjiMonster: ok Feb 10 23:33:53 everything else seems to apply with a bit of fuzzyness Feb 11 00:08:27 thanks xMff an mb__ now it works :) Feb 11 00:10:46 is there a way to disable it at compiletime? Feb 11 00:11:06 no Feb 11 00:11:47 but you can embed a sysctl.conf Feb 11 00:12:03 with net.ipv6.conf.eth0.accept_ra=2 added Feb 11 00:12:16 i did forward=0 Feb 11 00:12:21 or that Feb 11 00:12:26 since i do not require it Feb 11 00:40:15 jow * r25454 /trunk/package/base-files/ (3 files in 3 dirs): Feb 11 00:40:15 [package] base-files: Feb 11 00:40:15 - properly disable IPv6 on bridge ports Feb 11 00:40:15 - introduce accept_ra option, defaults to true for dhcp interfaces, false for others Feb 11 00:40:15 - introduce send_rs option, defaults to true for static interfaces, false for others Feb 11 00:40:45 build #73 of gemini is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/73 Feb 11 00:43:01 jow * r25455 /branches/backfire/package/base-files/ (3 files in 3 dirs): [backfire] backport r25454 Feb 11 01:59:49 nbd * r25456 /trunk/package/libubox/Makefile: libubox: update to 2011-02-11, includes a few uloop fixes Feb 11 02:00:00 nbd * r25457 /trunk/package/ubus/ (. Makefile): add ubus - the new openwrt message bus / rpc daemon Feb 11 02:09:04 nbd * r25458 /trunk/package/libubox/Makefile: libubox does not depend on libjson Feb 11 02:10:31 nbd * r25459 /trunk/package/libubox/Makefile: libubox: add a package for libblobmsg-json Feb 11 02:12:14 nbd * r25460 /trunk/package/libubox/Makefile: add missing chunk from the last commit **** ENDING LOGGING AT Fri Feb 11 02:59:57 2011