**** BEGIN LOGGING AT Thu Dec 16 02:59:57 2010 Dec 16 03:25:32 hello Dec 16 03:25:47 I have been hacking on Tor and using OpenWRT as a base - it's a really useful thing Dec 16 03:26:06 I wanted to submit some patches back to OpenWRT from the Tor Project proper - is the Tor OpenWRT person here? Dec 16 03:30:31 ioerror: I don't think their is a 'Tor OpenWRT person', just whoever submits patches Dec 16 03:31:01 ioerror: unless there is someone listed with a MAINTAINER field in the Makefile Dec 16 03:31:40 ioerror: but if you need I'll help I'll see if I can Dec 16 03:31:47 ok, I have three patches Dec 16 03:31:56 The first bumps the stable Tor package to the current stable version Dec 16 03:32:17 The second adds a tor-alpha package for the alpha series (better on embedded systems, better Tor bridge support, etc) Dec 16 03:32:29 The third adds upnp client support to tor-alpha Dec 16 03:33:11 Is it OK to send them as three different [PATCH] emails to the devel list as attachments? Or should I only send it inline? Dec 16 03:33:22 I guess inline, only? Dec 16 03:33:30 ok; I'm not sure about have tor-alpha as a package unless you're willing to maintain Dec 16 03:33:33 reading the docs implied inline Dec 16 03:33:36 yes, I am Dec 16 03:33:40 I am a paid Tor developer Dec 16 03:33:51 actually if you could do both that would be great Dec 16 03:33:53 I will maintain it and I will contribute work back to OpenWRT Dec 16 03:33:55 yes, I will do both Dec 16 03:33:59 cool! Dec 16 03:34:11 I'd actually like to ship an OpenWRT router and contribute money back to OpenWRT in my big picture Dec 16 03:34:18 could you let me know when you send the mail? Dec 16 03:34:30 I can send the first patch now Dec 16 03:34:46 ok Dec 16 03:35:22 should I gpg sign my emails? Dec 16 03:35:52 ioerror: you should also send mail to thepeople (see dev.openwrt.org) about being maintainter; I'm sure it would be appreciated Dec 16 03:36:09 ioerror: Well I do, but it's not required Dec 16 03:36:19 cshore: i will sign it - it includes my torproject.org email uid Dec 16 03:36:22 ioerror: we don't have an official submitter keyring Dec 16 03:36:25 ok Dec 16 03:37:15 build #44 of avr32 is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/avr32/builds/44 Dec 16 03:38:30 cshore: i just sent it Dec 16 03:38:43 cshore: i hope that's OK Dec 16 03:39:10 if it is OK, I have two other patches but they seem rather large for in-line emails Dec 16 03:40:03 I need to go now, feel free to either email or reply here, I will return after I move to a new location Dec 16 03:40:10 thanks for your help cshore! Dec 16 04:08:32 cshore * r24617 /packages/net/tor/Makefile: Dec 16 04:08:32 This is a version bump to build the newest stable Tor release. Dec 16 04:08:32 Signed-off-by: Jacob Appelbaum Dec 16 04:41:58 build #48 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/48 Dec 16 06:19:25 build #44 of sibyte is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/44 Dec 16 06:53:41 cshore: thank you :-) Dec 16 07:04:47 np Dec 16 09:00:41 does anyone here other than Hauke know about ssb? Dec 16 09:04:35 i have determined that without CONFIG_SSB_DEBUG=y, the wrt350n does not boot at all, and with it, it boots ok, but there is no eth or wifi, and lspci is erroneous Dec 16 09:29:47 sn9: that's strange, debug shouldn't change things Dec 16 09:32:59 KanjiMonster: i am now doing some more experiments Dec 16 09:35:38 sn9: I am currently trying to get a bcm4716 to boot, so I am kinda forced to read the ssb code ;) Dec 16 10:09:26 nbd: how does one usually use the wl driver under 2.6 in openwrt? Dec 16 10:14:30 similar to how it was used on 2.4 Dec 16 11:36:28 hi Hauke Dec 16 11:37:09 Hauke: ever touched an bcm4716 device? Dec 16 11:52:06 does bcm47xx support serial flash? Dec 16 11:54:59 hrm, doesn't look like it... Dec 16 11:55:20 it was added to brcm-2.4 at some point Dec 16 11:55:36 sflash Dec 16 11:56:19 yeah, just saw it Dec 16 11:58:08 looks like at least the bcm47162 hangs if a serial flash is attached and you try to read the flash_config register Dec 16 11:58:16 w00t, got console Dec 16 11:59:49 Is there a good writeup on how to test new packages? Dec 16 11:59:59 xMff: just got openwrt booted on this bcm47162 :) (2.6.37-rc5 kernel, of course no ethernet, wifi or writable flash ;) Dec 16 12:00:21 eg: I want to port libnatpnp to OpenWRT but I want to test it extensively before I publish the patch Dec 16 12:02:05 KanjiMonster: cool Dec 16 12:02:30 xMff: used berhards patches as a base Dec 16 12:04:29 xMff: the bcm47162 seems to be quite a "special" chip, the original broadcom code has several hacks for it in it that I needed to add ;) Dec 16 12:07:19 ioerror: what kind of tests do you want to run ? Dec 16 12:09:45 jow * r24618 /packages/net/nufw/Makefile: [packages] nufw: don't call autogen.sh Dec 16 12:15:49 blogic: i want to ensure that i run all the tests any reasonable person would run - eg - should I also build the packages on x86-64 or x86-32, etc? Dec 16 12:16:18 well test it on two arches, a be one and a le one Dec 16 12:16:33 I've read a bunch of the wiki and dev docs but there doesn't seem to be a good doc on how to add a new package and run through whatvever is expected Dec 16 12:16:48 xMff: i don't have any big endian hardware nearby Dec 16 12:16:56 I can probably find something Dec 16 12:18:04 mips-malta might be available in big endian Dec 16 12:18:22 brcm63xx is be Dec 16 12:18:44 i am currently working on this: Dec 16 12:18:46 mips-malta can be run in qemu, so it doesn't need big endian hardware ;) Dec 16 12:18:56 ah, I see Dec 16 12:19:10 https://trac.torproject.org/projects/tor/wiki/TheOnionRouter/Torouter Dec 16 12:19:38 The project sorta appears to be heading towards Buffalo WZR-HP-G300NH Dec 16 12:20:15 that would be ar71xx, so big endian Dec 16 12:21:22 ah ha, good to know Dec 16 12:21:30 well, that hardware is in my lab Dec 16 12:21:36 so it's not near me tonight Dec 16 12:21:58 But I'll hopefully have this tor-alpha package ready before i get there Dec 16 13:20:00 ping cshore Dec 16 13:20:20 Hauke: i have determined that without CONFIG_SSB_DEBUG=y, the wrt350n does not boot at all, and with it, it boots ok, but there is no ethernet, and i have not been able to figure out how to enable wl Dec 16 13:21:57 nbd: is wl supposed to coexist with the in-kernel ssb code? Dec 16 13:22:10 yes it is Dec 16 13:23:06 i know that on ubuntu on x86, wl will not load unless ssb is unloaded Dec 16 13:23:24 the ubuntu wl is not the same as the one we use Dec 16 13:23:42 i know, that one is sta-only Dec 16 13:23:48 pong xMff Dec 16 13:23:49 it has been modified under nda to work along the open ssb Dec 16 13:24:00 nice Dec 16 13:24:02 ours that is Dec 16 13:24:16 and the kernel integration part is open for hacking Dec 16 13:24:29 this includes the ssb glue, sprom processing etc. Dec 16 13:24:57 sprom has been the most common culprit until now Dec 16 13:24:59 i thought it was negotiated that they'd gpl the whole thing Dec 16 13:25:06 thats another matter Dec 16 13:25:29 cshore: I remembered what I wanted to ask, did you track down what broke collectd in backfire? Dec 16 13:26:22 xMff: sort of....collectd is a new version compared to when it worked....I didn't track down why the new version fails though Dec 16 13:26:23 but then why is lspci wrong on the wifi unless b43 is loaded? Dec 16 13:27:01 b43 has some ssb bridging code that might flip some bits which wl does not do by itself Dec 16 13:30:32 so, wl now requires b43??? Dec 16 13:30:39 also depending on the format of your nvram, wl might fail to build the sprom Dec 16 13:31:03 no it doesn't Dec 16 13:33:13 i think the problem of the missing bit-flipping, the non-creation of eth0, and the requirement of CONFIG_SSB_DEBUG=y are all somehow related Dec 16 13:33:40 build #51 of at91 is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/51 Dec 16 13:36:11 xMff: so, who currently maintains all the ssb stuff? Dec 16 13:36:20 kernel.org ? Dec 16 13:36:34 xMff: I'm mostly working with the working build because I'm trying to finish up the freeswitch GUI stuff....once it's ready I'll just be waiting for confirmation from the top guy at Vertical to confirm what I did when I was contracted there can be open sourced (because I'm being cautious on that regard as I currently don't have backup proof (i.e. something written or whatever rather than just IM logs) that it's allowed and I don't want Dec 16 13:36:39 with the openwrt patches, i mean Dec 16 13:36:49 nbd synced it from time to time Dec 16 13:37:15 cshore: sure, no worries Dec 16 13:37:25 well, something in it broke from 2.6.33 to 2.6.34 Dec 16 13:37:59 sn9: hi Dec 16 13:38:13 hello Dec 16 13:38:45 sn9: when building an image with CONFIG_SSB_DEBUG=y it does not boot and when building with CONFIG_SSB_DEBUG=n it boots just ethernet is not working? Dec 16 13:38:58 no, the reverse Dec 16 13:39:35 with CONFIG_SSB_DEBUG=y it boots just ethernet is not working Dec 16 13:39:43 sn9: so you have to activate CONFIG_SSB_DEBUG to make it boot? Dec 16 13:39:48 yes Dec 16 13:39:52 strange Dec 16 13:41:32 sn9: what svn revision did you test? Dec 16 13:41:46 i looked at the code, and there is nothing obvious that CONFIG_SSB_DEBUG=y adds that would make it work Dec 16 13:42:00 different timing? Dec 16 13:43:31 i tested r22119 with 2.6.32 and 2.6.33, r22294 with 2.6.33 and 2.6.34, as well as current trunk with 2.6.36 Dec 16 13:44:37 CONFIG_SSB_DEBUG is what makes or breaks it every time Dec 16 13:50:12 i think if we first find why eth0 is not created, we'll be closer to seeing why this happens, since it all appears to be related to the pci core Dec 16 13:50:47 missing mac address? Dec 16 13:51:00 that would be realted to nvram / sprom again Dec 16 13:52:14 with CONFIG_SSB_DEBUG=y, the first line that appears in the boot after the point where the boot fails without it is: ssb: PCIcore in host mode found Dec 16 13:53:07 sn9: culd you giv me the content of your nvram "nvram show" Dec 16 13:53:27 at the cfe prompt? Dec 16 13:53:41 no from openwrt Dec 16 13:53:47 ok Dec 16 13:55:58 interesting, on the second boot, wl works, but not on the first Dec 16 13:59:37 bad core reset? Dec 16 14:01:19 jow * r24619 /trunk/package/libtool/patches/110-dont-use-target-dir-for-relinking.patch: Dec 16 14:01:19 [packages] libtool: I don't know what crackhead thought that relinking with /usr/lib during cross compilation is a good idea, so just remove this crap. Dec 16 14:01:19 It should solve a few instances of "/usr/lib/libc.a: could not read symbols: File format not recognized" when building stuff. Dec 16 14:07:41 somehow i cant compile anymore toolchain http://openwrt.pastebin.com/jEeFVBPv Dec 16 14:08:55 about gmp.h not found Dec 16 14:11:07 should be in /home/user/openwrt/trunk/staging_dir/host/include Dec 16 14:12:59 yea but it isnt Dec 16 14:13:03 odd, now even on first boot, wl works Dec 16 14:13:22 crow: then rebuild gmp Dec 16 14:13:24 xMff thats reason i did clean everything to try compile toolchain new, as it fails Dec 16 14:13:58 should tthis not compile it " make tools/compile" Dec 16 14:14:15 make tools/{clean,compile,install} Dec 16 14:17:21 Hauke: http://www.gimpelevich.san-francisco.ca.us/danielg4/nvram.txt Dec 16 14:18:56 this looks normal to me Dec 16 14:19:22 do you get any messages when the ethernet driver wants to start, but fails? Dec 16 14:19:42 btw, "reboot" from openwrt does not work Dec 16 14:20:03 no, no messages about failed eth Dec 16 14:22:12 crow: you may end having to distclean...some .configs are disliked by current trunk even with make oldconfig Dec 16 14:22:36 xMff fail to Dec 16 14:22:42 *end up Dec 16 14:23:08 cshore i did "make clean" then "make distclean" then "svn up && ./scripts/feeds update" and tryed to compile tools Dec 16 14:24:15 crow: parallel compile or single tasking? Dec 16 14:24:29 just "make tools/{clean,compile,install}" Dec 16 14:25:42 crow what order is it trying to compile tools in? Dec 16 14:27:16 i think it still fails on gmp.h , and it is compiled before mpfr Dec 16 14:29:24 cshore just a sec to rerun proces. lost history Dec 16 14:33:59 http://openwrt.pastebin.com/mxKwxaSa Dec 16 14:34:30 also gmp is compiled but not installed Dec 16 14:36:36 crow: try make tools/install (I think the problems is tools/{clean,compile,install} means clean everything then compile everything, then install everything Dec 16 14:36:57 and dependencies aren't used Dec 16 14:37:14 cshore well if log is correct and its doing that in correct order, gmp is but skiped to install Dec 16 14:37:48 no, that's my point....it compiling everything....no installs until everything is compiled Dec 16 14:38:40 well line 29,30 shows mp4 compiling and installing Dec 16 14:38:41 but that ends up not working because compiling mpfr depends on gmp, but we said compile everything before installing anything Dec 16 14:39:16 doing make tools/install right now will report Dec 16 14:39:47 crow: hmmm...I think you're right....it ought to work, but it's not Dec 16 14:42:20 https://dev.openwrt.org/browser/trunk/tools/Makefile line 45: mpfr should call gmp install Dec 16 14:43:14 or maybe i am looking to wrong file.. Dec 16 14:44:41 yeah, I found make tools worked but not {clean,compile,install} . I assume this causes some type of dependency problem due to specification of what to do instead of figure it out command Dec 16 14:45:56 cshore tools/install worked as to seen here line 45: http://openwrt.pastebin.com/PvtehJr1 Dec 16 14:47:44 crow: I think the {clean,compile,install} is somehow overriding the following of dependencies that involve a different order than clean,compile,install in batches Dec 16 14:48:39 Hauke: i have confirmed that if flashed with an image without CONFIG_SSB_DEBUG and then an image with it, wl does not work, and lspci is erroneous, until the device is power-cycled, after which it works Dec 16 14:48:41 make tools/compile didnt worked also, here it just works as "make tools/install" which will compile and install. Dec 16 14:50:17 build #50 of s3c24xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/50 Dec 16 15:26:35 Hauke: ideas? Dec 16 15:38:57 build #52 of atheros is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/52 Dec 16 15:55:27 Hauke: are you still there? Dec 16 15:56:57 build #45 of ps3 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/45 Dec 16 16:08:57 sn9: No I do not know what happened there Dec 16 16:09:42 making eth work is still my priority atm Dec 16 16:26:38 jow * r24620 /packages/libs/libsigc++/patches/002-missing-macros.patch: [packages] libsigc++: supply missing macros Dec 16 16:34:34 build #47 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/47 Dec 16 16:41:00 I have trouble with make menuconfig in Trunk http://dpaste.com/hold/287366/ Dec 16 16:41:49 ./scripts/feeds uninstall freepops Dec 16 16:41:52 retry Dec 16 16:42:41 jow * r24621 /packages/libs/glibmm/patches/001-missing-macros.patch: [packages] glibmm: supply missing macros Dec 16 16:57:30 the same http://dpaste.com/287377/ Dec 16 16:59:17 I don't see an error Dec 16 16:59:29 the select thing is gone Dec 16 17:01:36 this is only after make menuconfig Dec 16 17:01:51 so what is the issue again? Dec 16 17:02:45 only is this cosmetic I think Dec 16 17:02:52 yes, it is Dec 16 17:04:35 jow: is order maintained for uci list items, or can it be different than in the config file? Dec 16 17:04:52 maintained Dec 16 17:12:40 Hauke: still there? i found why eth doesn't work: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=patch;h=67b284d476bcb3d100e946da23d6cf9acfd0465c Dec 16 17:13:28 ... Dec 16 17:13:29 sn9: your device is supported by tg3? Dec 16 17:13:43 it's 14e4:1676 Dec 16 17:13:58 PCI_DEVICE_ID_TIGON3_5750 Dec 16 17:14:58 kernel maintainers blindly committed that stupid patch supplied by broadcom Dec 16 17:14:59 have you tried to reverse apply the patch? Dec 16 17:15:04 i will now Dec 16 17:15:43 normally broadcom should know at best what chips they released and what not Dec 16 17:16:30 they never released them to the general public Dec 16 17:17:00 they DID release them to the odm market Dec 16 17:17:43 they see tg3 as being for the general public only Dec 16 17:17:49 yes as boradcom does not support these devices in the mainline kernel Dec 16 17:24:04 acoul * r24622 /packages/net/quagga/patches/999-PW-8083-Intern_communities_in_route_maps.patch: net/quagga: "Intern" communities in route maps Dec 16 17:25:10 acoul * r24623 /packages/net/quagga/patches/999-PW-8261-memleak_del_community.patch: net/quagga: fix memory leak when deleting a community Dec 16 17:27:41 build #54 of brcm63xx is complete: Failure [failed shell compile_10] Build details are at http://tksite.gotdns.org:8010/builders/brcm63xx/builds/54 Dec 16 17:49:49 KanjiMonster: no I do not have a bcm4716 device Dec 16 17:51:12 http://openwrt.pastebin.com/FNH1fh37 why setting two time crda code? Dec 16 17:55:26 Hauke: ok, everything works, you can commit this: http://www.gimpelevich.san-francisco.ca.us/danielg4/r24623.patch Dec 16 17:56:28 sn9: ok thank you Dec 16 17:57:08 Hauke: got it to boot, it seems it doesn't like its flash_control read if a serial flash is attached (if it even supports parallel flashes) Dec 16 17:58:23 after skimming the broadcom code for the nth time, I finally noticed a "return" in the serial flash section ;) Dec 16 18:05:53 udevadm control --reload-rules give me this: Dec 16 19:04:22 OpenWrt user.err udevadm[1450]: error sending message: Connection refused Dec 16 18:11:22 seems udev inst starting at startup?? after doing "udevd" i see in dmesg this Dec 16 18:11:23 udevd (1458): /proc/1458/oom_adj is deprecated, please use /proc/1458/oom_score_adj instead. Dec 16 18:11:23 udev: starting version 142 Dec 16 18:12:19 actualy udevd gives Dec 16 18:12:20 http://openwrt.pastebin.com/G7d4BMyT Dec 16 18:16:15 udevd --daemon give me same error Dec 16 18:19:51 Hauke: is there something else that needs to happen before it's committed? i need to disconnect soon Dec 16 18:20:26 sn9: I think it is really strange that CONFIG_SSB_DEBUG=y is needed Dec 16 18:20:50 does it print out anything? Dec 16 18:21:09 agreed, but prior to 2.6.34 it was always there, and with it, it always works Dec 16 18:21:31 does what print out anything? Dec 16 18:32:15 what kernel log messages do you get when booting? Dec 16 18:33:30 KanjiMonster: do you know any cheep device with a recent broadcom 47xx or 53xx SoC ? Dec 16 18:34:28 Hauke: the asus rt-12 has a 5356; I think that one should be quite cheap, but I think it has only 16 MB ram Dec 16 18:34:40 Hauke: i showed you the log without debug Dec 16 18:35:32 Hauke: besides that I only know of devices where only some versions have a 47xx or 53xx Dec 16 18:36:03 sn9: It would just be interesting if the the ssb debug option results in any printks after the position where it failed for you Dec 16 18:36:32 Hauke: ah, wait, its the rt-n10, not n12 Dec 16 18:36:35 the first line after that position is: Dec 16 18:36:39 the n12 is ralink Dec 16 18:36:58 ssb: PCIcore in host mode found Dec 16 18:37:01 Hauke: ^^^^^^ Dec 16 18:39:55 Hauke: if the fcc images are from a final device it should be fairly easy to add jtag and serial to it (assuming these aren't populated) Dec 16 18:40:04 Hauke: see http://infodepot.wikia.com/wiki/Asus_RT-N10 for more info Dec 16 18:41:41 Hauke: oh, I really have a bad memory, the rt-n12 would be fine, too: http://infodepot.wikia.com/wiki/Asus_RT-N12 Dec 16 18:42:28 KanjiMonster: there are two versions of the rt-n12 available, one is with ralink and one with boradcom http://en.wikipedia.org/wiki/Asus_routers Dec 16 18:42:45 no the ralink is the rt-n13 Dec 16 18:43:28 but there seems to be a second rt-n10 with ralink Dec 16 18:45:04 I will look into the source code released by asus Dec 16 18:46:02 Hauke: the rt-n10 tarball is quite complete Dec 16 18:48:21 there is just one firmware available for the rt-n10 so there is probably only one architecture used Dec 16 18:53:12 Hauke: yeah, the ralink one might be called rt-n10+ or rt-n10b. The rt-n10 seems to run 2.4, while the n12 runs 2.6. besides that the tarballs are almost identical, with the code in the n10 being slightly newer (e.g. n10's siutils.c are from oct 09, while the n12's are from jan 09) Dec 16 18:55:11 hrm, well, asking asus might have been fastest, the ralink one is the rt-n10+, so easily avoidable ;) Dec 16 18:55:57 build #39 of octeon is complete: Failure [failed shell_13] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/39 Dec 16 19:28:08 hauke * r24624 /trunk/target/linux/brcm47xx/patches-2.6.37/030-b44-fix-workarround-for-wap54g10.patch: brcm47xx: this patch went upstream Dec 16 19:32:18 hauke * r24625 /trunk/target/linux/brcm47xx/ (2 files in 2 dirs): Dec 16 19:32:18 brcm47xx: add some missing pci ID for tg3 driver Dec 16 19:32:18 Broadcom removed these pci id, but at least the wrt350n has a Ethernet Dec 16 19:32:18 controller with a pci id of 14e4:1676 Dec 16 19:32:18 Thank you sn9 for fixing this problem. Dec 16 19:37:12 hauke * r24626 /trunk/target/linux/brcm47xx/ (config-2.6.34 config-2.6.36 config-2.6.37): Dec 16 19:37:12 brcm47xx: add ssb debug option. Dec 16 19:37:12 This fixes a problem with wrt350n. Dec 16 19:37:12 It boots only if this config option is set, otherwise it reboots after "Switching to clocksource MIPS" Dec 16 19:37:12 Thank you sn9 for fixing this problem. Dec 16 20:08:40 why openwrt dont have udev in boot proces? or is something wrong on my side Dec 16 20:34:17 anyone around to handle some trac spam? Dec 16 20:55:03 xMff: is there any reason to not close #8408 as fixed? Dec 16 21:20:26 done Dec 16 21:36:18 swalker: anything else? Dec 16 21:37:20 crow: openwrt uses hotplug2 not udev by default Dec 16 21:39:06 cshore aha I see, well i have some usb rules which are handled by udev rule.. Dec 16 21:54:24 crow: you'll have to add udev and remove hotplug and I don't know what else Dec 16 21:54:37 cshore: not really other than the frequent spam issue(s) Dec 16 22:13:34 Hauke: still awake? Dec 16 22:15:02 i have brcm-wl putting up a signal, but i can't associate with it, even unencrypted Dec 16 22:56:56 build #49 of orion is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/49 Dec 16 23:56:29 nbd * r24627 /trunk/package/mac80211/patches/540-ath9k_aggr_depth_fix.patch: ath9k: use the number of pending aggregates instead of the queue depth to determine when it is time to form the next A-MPDU Dec 16 23:56:37 nbd * r24628 /trunk/package/mac80211/patches/550-ath9k_disable_hw_rfkill.patch: ath9k: remove code for setting up gpio based baseband rfkill - it was never used, nor configured properly Dec 16 23:56:46 nbd * r24629 /trunk/package/mac80211/patches/ (2 files): ath9k: update initvals for ar9003 Dec 16 23:57:26 nbd: is it a common issue for brcm-wl to simply not associate, even unencrypted? Dec 16 23:58:46 no idea Dec 17 00:00:51 what could cause that? Dec 17 00:01:26 i don't know Dec 17 00:04:16 build #48 of gemini is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/48 Dec 17 01:44:08 jow * r24630 /packages/utils/collectd/ (Makefile patches/002-fix-ltdl-linking.patch): [packages] collectd: solve "cannot relink convenience library", it was a wrongly applied fixup Dec 17 01:55:44 nbd: is the brcm-wl currently in openwrt not the old brcm-wl-mimo? Dec 17 02:00:47 nbd: you still working on the broadcom proprietary drivers or have you decided to move on the drivers for devices where the manufacturer is somewhat supportive of Linux? Dec 17 02:01:06 *on to Dec 17 02:02:23 i'm not actively working on it. i put enough work into it to make the binaries independent of the kernel versions so that people can work on the part that interacts with the rest of the kernel Dec 17 02:02:32 but i'm not really interested in putting much work into that driver myself Dec 17 02:03:18 sn9: i don't know what you mean Dec 17 02:03:42 nbd: its the same problem I have on the whr-hp-g128 Dec 17 02:03:54 ssid is broadcasted, nas running no assoc possible Dec 17 02:04:06 seems to affect all newer bcm chips Dec 17 02:04:30 nbd: under 2.4, there was brcm-wl and brcm-wl-mimo Dec 17 02:04:45 oh nvm, I was one question behind Dec 17 02:05:12 the current blob has mimo capabilities builtin Dec 17 02:05:39 current? is that the "mini" ? Dec 17 02:06:09 probably not, its been a while since I hacked on it, don't recall which variant has what Dec 17 02:31:37 build #51 of brcm47xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/51 Dec 17 02:38:12 build #46 of ramips is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ramips/builds/46 Dec 17 02:42:32 jow * r24631 /trunk/tools/ (Makefile flex/ flex/Makefile): [tools] build flex on the host, some pakcages need it in the newest version Dec 17 02:51:01 build #40 of etrax is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/40 Dec 17 02:53:43 xMff: ping Dec 17 02:56:24 pong Dec 17 02:57:52 sent an email a few days ago about changing how we invoke the compiler from 4.4 and onwards... did you see it? Dec 17 02:58:08 yes, but we have < 4.4 in production Dec 17 02:58:18 for what platforms? **** ENDING LOGGING AT Fri Dec 17 02:59:59 2010