**** BEGIN LOGGING AT Wed Nov 04 02:59:57 2009 Nov 04 05:05:50 xmff: around? Nov 04 15:31:27 hello guys, before install serial port to my router, I want to know if someone have a linksys WRT54GL and did build own image in last 2 days? Nov 04 15:33:08 fofware: many people did Nov 04 15:33:32 [Wed 04 Nov 2009 06:58:35 AM PST] < sn9 > revisions prior to r18289 will brick on 2.6, and revisions prior to r18287 will brick on 2.4 Nov 04 15:34:08 assuming trunk, that is Nov 04 15:34:32 because, I'm trying to get goodone but i can't, so I don't know if the problem is with my buildroot Nov 04 15:34:45 hello sn9 Nov 04 15:35:01 it does not build for you? Nov 04 15:35:09 yes I trying with trunk Nov 04 15:35:25 please show the error Nov 04 15:35:57 I have not error, the router not boot Nov 04 15:36:29 which svn revision? Nov 04 15:37:18 for this reason I thinking install a serial port in it to show error in channel so someone with knowledge can be fix if the issue is in openwrt Nov 04 15:37:33 which svn revision did you flash? Nov 04 15:38:53 I did try with 18282 Nov 04 15:39:01 [Wed 04 Nov 2009 06:58:35 AM PST] < sn9 > revisions prior to r18289 will brick on 2.6, and revisions prior to r18287 will brick on 2.4 Nov 04 15:39:23 ok I will try with a new one Nov 04 15:39:26 ok Nov 04 15:39:32 thanks sn9 Nov 04 15:39:58 yw Nov 04 15:42:59 so, early in the morning someone discober the issue :o) Nov 04 15:43:22 yesterday morning Nov 04 15:45:40 hmmm, I past the last 2 day trying... and no body told me about that, and I guess last night I did try with newer revisions but don't work too. Nov 04 15:46:17 After I did revert to an oldone, make distclean and try but don't work too Nov 04 16:35:55 Ok I have 18291 rev now, I did a distclean and started to build, I hope now work Nov 04 16:36:22 distclean will also remove .config Nov 04 16:38:00 yes, remove config and all prebuild staff Nov 04 16:51:46 jow * r18292 /trunk/feeds.conf.default: [feeds] LuCI main repo is down due to server crash, switch to repo mirror Nov 04 17:11:22 who here is in contact with Oleg? Nov 04 17:12:10 somebody said he might have broadcom 802.11n working under 2.6 Nov 04 17:12:27 that would mean 2.4 can finally be killed off for good Nov 04 17:12:47 binary blob on 2.6 or b43 ? Nov 04 17:13:02 blob Nov 04 17:13:05 ah Nov 04 17:13:40 yes, this would be awesome Nov 04 17:13:57 but it would be easier to substitute b43 into a working 2.6 than to maintain 2.4 Nov 04 17:14:33 probably, yes. But its not so great to kill 2.4 with b43 lacking support for multi-ssid or wds Nov 04 17:14:58 i thought that was done Nov 04 17:15:22 anyway, a blob would not be affected Nov 04 17:15:26 yes Nov 04 17:56:29 i think i may have registered for a nuwiki login before, but i don't remember my username Nov 04 17:57:14 never mind, got it Nov 04 18:36:43 xmff: whatever happened with the TOH? Nov 04 18:37:20 xmff: multi-ssid looks implemented to me Nov 04 18:37:57 sn9: must be very new then, didn't work three weeks ago Nov 04 18:38:32 you can define it and create multiple interfaces with "iw" but any attempt to actually bring them up just failed Nov 04 18:39:27 oh Nov 04 18:42:55 danage: not much Nov 04 18:43:06 why Nov 04 18:43:13 lack of time Nov 04 18:43:34 we should conspire and withhold the palinka from Bartman007 until it's finished Nov 04 19:53:26 since danage isn't here I can't talk to him, but xmff, I was hoping that others would take to filling in the ToH based on devices they own. I don't consider the old toh to be a trusted source and won't pull data from there. Nov 04 20:51:25 whoever is around: this is ready for commit http://openwrt.pastebin.com/f22567a7d Nov 04 20:51:32 [florian]: ^ Nov 04 20:53:44 I would like to build a recent trunk version for x86/alix, does anyone know which revision started causing the 'file not found' after deleting a file problem? Nov 04 20:54:58 I went back to r18266, thinking the squashfs changes were the problem, but bug still there Nov 04 20:56:13 there was some work done on union mount support which also changed the vfs layer, maybe it is related to that Nov 04 21:00:36 sn9: I'll handle that commit Nov 04 21:00:39 xmff: thanks, looks like around r17724-17778, will try before that Nov 04 21:01:14 netprince: apart from the annoyance factor, did you see actual problems? Nov 04 21:07:38 I was having problem with installing packages via opkg, I just assumed it was related Nov 04 21:07:54 what kinds of problems? Nov 04 21:08:21 sn9: to confirm, this patch adds wl153 flashmap and squashfs image support? Nov 04 21:08:37 actually, for the wl176 Nov 04 21:08:50 xmff: leme boot up my alix just to double check it... Nov 04 21:09:20 netprince: ok, because I added some patches to opkg recently, maybe they're the actual culprit Nov 04 21:12:24 <[Fate]> doh... in target/linux/brcm-2.4/modules.mk KernelPackage/lp/brcm-2.4 is defined, but does not end up in the final kmod-lp package (related to my parport not working on wl-500g) Nov 04 21:13:47 <[Fate]> this mechanism does not seem to be used by any other platform Nov 04 21:20:32 xmff: opkg seems to be working fine right now, I'm not sure what my problem with it was yesterday, sorry about that Nov 04 21:21:37 netprince: ok, if you see segfaults or weird stuff with opkg then tell me :) Nov 04 21:21:57 xmff: ok Nov 04 21:42:15 <[Fate]> nbd: ping Nov 04 21:42:20 [Fate]: pong Nov 04 21:44:32 <[Fate]> nbd: in package/kernel/Makefile we include $(TOPDIR)/target/linux/*/modules.mk. This affects eg. target/linux/brcm-2.4/modules.mk Nov 04 21:44:53 <[Fate]> there we define KernelPackage/lp/brcm-2.4 but I think this gets never evaled Nov 04 21:45:17 <[Fate]> it _should_ be called from include/kernel.mk's line $(eval $(call KernelPackage/$(1)/$(BOARD)-$(KERNEL))) Nov 04 21:45:29 <[Fate]> thing is wl-500g compile doesn't pick up parport_splink Nov 04 21:45:44 <[Fate]> 8.09.1 svn rev 18291 that is Nov 04 21:46:27 <[Fate]> if I put bullshit in target/linux/brcm-2.4/modules.mk I get an error as expected, so the include works Nov 04 21:46:27 yeah, makes sense Nov 04 21:46:39 <[Fate]> but if I put bullshit inside the define block I get no error Nov 04 21:46:40 $(BOARD)-$(KERNEL) means brcm-2.4-2.4 Nov 04 21:46:50 so this is something old that wasn't forward ported Nov 04 21:46:51 trivial to fix Nov 04 21:47:34 <[Fate]> should be just $(BOARD) then, I guess Nov 04 21:47:54 <[Fate]> ack Nov 04 21:47:59 <[Fate]> pls fix :) Nov 04 21:48:36 will do Nov 04 21:49:16 xmff: luci fails to download now; makefile points at defunct server Nov 04 21:50:25 sn9: regarding the wl176 patch, is the 32KB padding at the beginning of the image neccessary? (if so, what a waste of space) Nov 04 21:50:48 Bartman007: bootloader looks for kernel 0x8014 in Nov 04 21:51:05 same on the wl153, fwiw Nov 04 21:51:56 and the wl153 has only 2M flash Nov 04 21:52:16 that's part of why I was asking, 32KB is a big deal on the wl153. Nov 04 21:52:35 xmff: can you please fix? Nov 04 21:53:51 nbd * r18293 /trunk/package/ (hostapd/files/hostapd.sh mac80211/files/lib/wifi/mac80211.sh): mac80211: fill some important capabilities into ht_capab in the hostapd config Nov 04 21:54:27 sn9: the server has hardware issues, use my git repo in the mean time Nov 04 21:54:39 it'll probably be fixed tomorrow or so Nov 04 21:54:49 nbd * r18294 /trunk/package/mac80211/ (10 files in 2 dirs): mac80211: update to 2009-11-03 Nov 04 21:54:55 nbd * r18295 /trunk/include/kernel.mk: fix platform specific overrides for kernel modules, fixes lp on brcm-2.4 Nov 04 21:55:53 nbd: irrelevant Nov 04 21:56:06 it's not downloadable Nov 04 21:56:14 the makefile needs fixing Nov 04 21:56:29 it is fixed now Nov 04 21:57:25 nbd * r18296 /branches/8.09/include/kernel.mk: merge kernel module fix from r18295 to 8.09 Nov 04 21:57:42 sn9: why would the makefile need fixing? if you put my git url into your feeds.conf it should work without it trying to download from the luci server Nov 04 21:57:53 nbd: no the makefile references the rpoi Nov 04 21:57:55 repo Nov 04 21:58:09 xmff: i know that, but once he downloads from my git, it finds the sources locally Nov 04 21:58:21 and then no longer tries to pull from the repo Nov 04 21:58:23 ah indeed, forgot that Nov 04 21:58:27 at least that's how it *should* work ;) Nov 04 21:58:57 xmff: thx Nov 04 22:09:31 <[Fate]> is there some host-side "opkg" that can be used to peek into the generated binary packages? Nov 04 22:10:02 [Fate]: they are just .tar.gz files. Nov 04 22:10:59 <[Fate]> oh :) Nov 04 22:19:08 <[Fate]> now parport_splink is included, but autoload is still going wrong Nov 04 22:20:13 how? Nov 04 22:21:15 florian * r18297 /trunk/target/linux/generic-2.6/ (2 files in 2 dirs): [kernel] add missing ledtrig-netdev dependency on the networking stack Nov 04 22:21:57 <[Fate]> the generated etc/modules.d/50-lp does not contain parport_splink Nov 04 22:22:04 <[Fate]> let me check if the AUTOLOAD is not eval'd Nov 04 22:30:29 <[Fate]> yep Nov 04 22:41:00 <[florian]> sn9: got an ar525w? Nov 04 22:41:34 <[Fate]> nbd: you oversaw another place in target/kernel.mk, line 103 (near "define KernelPackage") Nov 04 22:42:24 oh, right Nov 04 22:43:07 florian * r18298 /trunk/ (2 files in 2 dirs): [rdc] make rdc321-wdt become a package, saves some space in the kernel Nov 04 22:43:08 [Sat 31 Oct 2009 09:03:02 PM PDT] < Bartman007 > am I correct in recalling that you don't have an ar525w to test? Nov 04 22:43:08 [Sat 31 Oct 2009 09:03:39 PM PDT] < sn9 > it needs some smd rework Nov 04 22:43:08 [Sat 31 Oct 2009 09:07:20 PM PDT] < Bartman007 > damn. Nov 04 22:43:08 [Sat 31 Oct 2009 09:09:24 PM PDT] < sn9 > it was the first board i ever tried to connect serial to; live and learn. that's how i know that serial needs a header, not hookup wires directly to the board Nov 04 22:43:36 florian * r18299 /trunk/target/linux/rdc/Makefile: [rdc] include kmod-rdc321x-wdt by default Nov 04 22:43:54 <[florian]> sn9: Nov 04 22:43:55 <[florian]> sn9: ok Nov 04 22:44:04 agb * r18300 /trunk/target/linux/rdc/ (3 files in 3 dirs): [rdc] 2.6.28: add flashmap and image support for the sitecom wl-176. Thanks Daniel Gimpelevich Nov 04 22:44:19 <[florian]> Bartman007: please tell me when you do something like this Nov 04 22:44:48 i did warn about that Nov 04 22:44:55 [florian]: will do in the future. Nov 04 22:45:23 <[florian]> Bartman007: no problems Nov 04 22:48:23 <[Fate]> nbd: could you commit this second oneliner pls? Nov 04 22:48:27 sure Nov 04 22:51:10 nbd * r18301 /trunk/include/kernel.mk: fix one more instance of r18295 Nov 04 22:52:00 nbd * r18302 /branches/8.09/include/kernel.mk: merge r18301 to 8.09 Nov 04 23:07:58 florian * r18303 /trunk/target/linux/rdc/config-2.6.28: [rdc] remove useless configuration options Nov 04 23:15:33 [florian]: won't the kernel config hang without CONFIG_IBM_ASM? Nov 04 23:15:56 <[florian]> sn9: no it's in generic now Nov 04 23:16:07 <[florian]> sn9: or it was last I checked Nov 04 23:19:59 <[Fate]> nbd: thx and good n8 Nov 04 23:30:41 Nbd: thanks for the madwifi refcount fix: no more crashes for another 12 hours! Nov 04 23:33:07 agb * r18304 /branches/8.09/package/madwifi/Makefile: [8.09] madwifi: bump package release number Nov 04 23:33:10 agb * r18305 /branches/8.09/package/opkg/files/opkg.conf: [8.09] opkg: bump repository url to 8.09.2 Nov 05 01:24:34 nbd * r18306 /trunk/scripts/metadata.pl: metadata: process dependencies for targets that have subtargets as well - hides the ramips target unless the user has requested to play with broken packages/targets (related to #6051) Nov 05 01:49:22 nbd * r18307 /trunk/toolchain/binutils/patches/2.20/200-mips_non_pic.patch: binutils: fix build errors in 2.20 (patch from #6094) **** ENDING LOGGING AT Thu Nov 05 02:59:56 2009