**** BEGIN LOGGING AT Fri Apr 30 02:59:57 2010 Apr 30 07:07:23 acoul * r21264 /trunk/package/grub/Makefile: package/grub: fix x86_64 issue (closes #7248) Apr 30 07:53:03 gmorning ppl Apr 30 08:26:45 good morning Apr 30 09:02:37 ping nbd Apr 30 09:05:30 pong Apr 30 09:06:38 nbd: by the way, did you have time to look at patch for the cache access on brcm47xx? Apr 30 09:07:21 no, where can i find it again? Apr 30 09:07:36 mail subject on openwrt-devel? Apr 30 09:08:09 <[florian]> rtz2: about that, we are not providing an "optimized" kernel if we stick to having a single kernel for the various flavors of bcm47xx Apr 30 09:08:23 <[florian]> rtz2: the best we can do right now is have a mips74k aware user-space Apr 30 09:08:34 nbd: BRCM4716 support Apr 30 09:08:42 nbd: it's attached there Apr 30 09:09:25 what register is at SSB_ENUM_BASE? Apr 30 09:09:25 [florian]: well, it doesn work at all, because the cache managemt code accesses a backplane register, which isn't there on the 4716 Apr 30 09:09:52 <[florian]> rtz2: cannot you use the cpuid to runtime detect which CPU version you are running on? Apr 30 09:10:34 nbd: ssb_enum_base, is the start of the ssb config space, so this is register 0 of core 0 Apr 30 09:10:34 i'm injecting traffic on a monitor interface using madwifi. The point is: if I use atheros headers it works fine on both x86 and ixp4xx targets, instead if I use radiotap headers it works only on x86, while on the ixp4xx target the interface beahves like it was set to use wifi headers Apr 30 09:10:46 nbd: which should be the core id of the chipcommon core Apr 30 09:10:56 ok Apr 30 09:11:17 then that patch can go in Apr 30 09:11:33 [florian]: in don't think it really matters, which register is used, it's only a dummy read Apr 30 09:11:41 i'll take a quick look at the other patches Apr 30 09:12:08 nbd: acould already commit them, only the one from yesterday is still missing Apr 30 09:12:57 acoul Apr 30 09:13:39 ok Apr 30 09:14:02 btw. since we're gathering more and more ssb patches, could you try to push some of them upstream? Apr 30 09:14:24 not the cache related stuff Apr 30 09:14:34 but at least the parts that touch drivers/ssb and its includes Apr 30 09:14:48 nbd: I plan to do this, but I would like to keep them in openwrt for a few weeks, to give them some testing Apr 30 09:15:01 nbd: i only have this one brcm4716 device Apr 30 09:15:16 ok Apr 30 09:17:28 nbd: by the way, what's the story with those cache access patches? Apr 30 09:17:39 nbd: they aren't in the original broadcom code Apr 30 09:18:32 they were in earlier versions Apr 30 09:18:42 i think broadcom dropped support for bcm4710 Apr 30 09:19:45 nbd: they aren't needed for newer socs? Apr 30 09:24:15 rtz2: right Apr 30 09:25:40 hmmm, ok Apr 30 09:29:42 hmm, I still am not able to get usermode linux target to work on a x86_64 host, compilation fails Apr 30 09:30:05 on 32-bit works with no problems whatsoever Apr 30 09:36:28 <[florian]> Kimvais: yes, there is a problem here as well {Nico} is working on it Apr 30 09:48:57 [florian]: cool. I think I just got it working Apr 30 09:50:00 what I did was ln -s lib lib64 in ~/openwrt/trunk/staging_dir/toolchain-x86_64_gcc-4.4.3_uClibc-0.9.30.1 and this makefile change: https://dev.openwrt.org/attachment/ticket/5832/uml-target-cross.patch Apr 30 09:50:18 false alarm, still failed :) Apr 30 09:50:45 @Kimvais: can you please also try with latest trunk, there is a relative patch that should be dealing with this issue Apr 30 09:51:12 save your .config file Apr 30 09:51:17 make distclean Apr 30 09:51:23 and make V=99 Apr 30 09:53:13 does make distclean keep saved environments? Apr 30 09:53:25 scripts/env thingis Apr 30 09:53:49 I think it kills everything Apr 30 09:53:51 At revision 21264. Apr 30 09:53:58 I was already using that Apr 30 09:55:18 can you pastebin the error you get? Apr 30 09:55:32 you still need a clean setup to test latest trunk though Apr 30 09:55:53 that's why the distclean Apr 30 09:56:21 one sec, let me make a copy Apr 30 09:59:30 http://pastebin.com/NhYVmXhw Apr 30 09:59:53 that was prior to distclean though, running it again but will take a few minuts Apr 30 10:00:03 <[florian]> Kimvais: there is a patch upstream to fix this, hold on Apr 30 10:00:32 [florian]: yay Apr 30 10:00:35 <[florian]> Kimvais: try this patch: __rela_iplt_end Apr 30 10:00:40 <[florian]> Kimvais: https://patchwork.kernel.org/patch/95408/ Apr 30 10:05:12 [florian]: worked, thanks! :) Apr 30 10:05:30 <[florian]> Kimvais: you are welcome, now if you manage to run it, that will be even better ;) Apr 30 10:05:42 <[florian]> Kimvais: last I tried, uml segfaulted once the rootfs was mounted Apr 30 10:05:57 [florian]: it runs allright Apr 30 10:06:05 <[florian]> cool Apr 30 10:06:23 root@OpenWrt:/# uname -a Apr 30 10:06:24 Linux OpenWrt 2.6.30.10 #4 Fri Apr 30 13:04:05 EEST 2010 x86_64 GNU/Linux Apr 30 10:30:46 ping rtz2 Apr 30 11:05:44 acoul * r21265 /trunk/target/linux/brcm47xx/patches-2.6.33/ (6 files): brcm47xx: refresh 2.6.33 kernel patches Apr 30 11:10:06 acoul * r21266 /trunk/target/linux/brcm47xx/config-2.6.33: brcm47xx: update config-2.6.33 options Apr 30 11:48:12 <[florian]> Kimvais: is your x86_64 machine an amd64 one? Apr 30 12:01:12 [florian]: model name : Intel(R) Core(TM)2 Duo CPU E6750 @ 2.66GHz Apr 30 12:01:25 <[florian]> Kimvais: ok, thanks Apr 30 12:20:26 florian * r21267 /trunk/target/linux/uml/ (3 files in 3 dirs): [uml] add upstream patch to fix linking failures on '__rela_iplt_end' Apr 30 13:58:53 nbd: there is a patch on the mac80211 ML to support the RADIOTAP_RATE field. Is this patch suitable to be merged in openwrt? I've tested it and it works fine. Apr 30 14:31:21 jow * r21268 /packages/net/arpd/ (3 files in 2 dirs): Apr 30 14:31:22 [packages] arpd: Apr 30 14:31:22 - fix compilation against newer libevent by dropping obsolete external references Apr 30 14:31:22 - fix uninitialized pointer to pcap pattern leading to invalid filter patterns even if -a is not used Apr 30 14:31:22 - remove dependency on @BROKEN Apr 30 15:53:11 nbd * r21269 /trunk/target/linux/ (18 files in 7 dirs): kernel: bring ssb in sync with upstream (as of current wireless-testing) Apr 30 15:55:10 jow * r21270 /trunk/package/swconfig/Makefile: [package] swconfig: mask on brcm47xx Apr 30 15:55:26 nbd * r21271 /trunk/package/mac80211/ (13 files in 2 dirs): mac80211: update to wireless-testing 2010-04-28 Apr 30 15:55:42 jow * r21272 /branches/backfire/package/swconfig/Makefile: [backfire] merge r21270 Apr 30 16:01:51 craigc * r21273 /packages/net/multiwan/ (Makefile files/etc/config/multiwan files/usr/bin/multiwan): [packages] multiwan: Modified failover, implemented a check to see if resolv.conf.auto is current, added a cache file Apr 30 16:11:33 nbd * r21274 /trunk/package/vsc73x5-ucode/Makefile: vsc73x5-ucode: use the mirrored ucode files from my server to replace an 127 MB download with a 14k one Apr 30 16:11:57 hehe Apr 30 16:54:55 nbd * r21275 /branches/backfire/package/vsc73x5-ucode/Makefile: [backfire] merge vsc73x5-ucode download patch from r21274 Apr 30 16:54:59 nbd * r21276 /branches/backfire/target/linux/ (9 files in 3 dirs): [backfire] backport ssb updates from r21269, this is required for future mac80211 updates Apr 30 16:55:23 hey has anyone been looking for me lately...I've been forgetting to check my IRC Apr 30 17:00:50 cshore: how can you forget that :P Apr 30 17:01:57 heh Apr 30 17:02:14 no rest for the wicked Apr 30 18:43:41 nico * r21277 /trunk/target/linux/x86/image/Makefile: target/x86: add to changes missed from [21246] preventing .vdi & .vmdk image generation Apr 30 21:13:48 cshore * r21278 /packages/net/git/Makefile: packages: net/git: Fixed build error on some systems due to looking for mkstemps. Closes #6023. Apr 30 22:18:13 hey the backfire images for brcm63xx look like they missed a change, at least for one image Apr 30 22:18:38 ping thepeople Apr 30 22:27:08 ping florian Apr 30 23:14:16 <{Nico}> cshore: known problem Apr 30 23:28:51 {Nico}: ok, thanks **** ENDING LOGGING AT Sat May 01 02:59:56 2010