**** BEGIN LOGGING AT Sun Jan 25 02:59:58 2009 Jan 25 11:52:59 so Jan 25 11:53:05 is the build still broken? Jan 25 11:53:40 juhosg: hey, mac80211 has distfile checksum mismatches Jan 25 11:54:15 also, good morning! Jan 25 12:29:00 does continue in a do { } while (0) construct continue or does it evaluate the condition and stop? Jan 25 12:29:17 * jb42 wonders Jan 25 12:36:38 I think it equals break in that case Jan 25 12:36:46 yeah Jan 25 13:31:42 <[florian]> jb42: seems like mac80211 build is broken, let me commit a fix Jan 25 13:33:49 juhosg * r14172 /trunk/package/mac80211/ (23 files in 2 dirs): [package] mac80211: update compat-wireless to 2009-01-24 (patch by Hauke Mehrtens ) Jan 25 13:35:17 <[florian]> hei NLeiten Jan 25 13:36:31 florian * r14173 /trunk/target/linux/rb532/ (21 files in 2 dirs): [rb532] update 2.6.27 patches with upstream fixes from git repository and mailing-lists. via-rhine is still broken, thus not switching to 2.6.27 as default. Jan 25 13:37:27 hi, i got some problems with cross-compiled kernel for mips (adm5120 - compex wp54). I know that issue is simple, but need help. I build selfmade FW from openwrt kamikaze, board loads it well. But it hangs up with "turn off boot console early0" Jan 25 13:37:58 <[florian]> did you make any change to the kernel configuration ? Jan 25 13:38:11 As I saw source-codes, it hangs somewhere in do_initcall function from init/main.c Jan 25 13:39:22 I got vanilla kernel 2.6.27, put patches from openwrt and thats it. I've compiled with openwrt-config and with my own one Jan 25 13:39:35 <[florian]> ok Jan 25 13:39:38 result is the same... Jan 25 13:41:18 <[florian]> last I tried it worked fine for me on a compex-wp54g-wrt Jan 25 13:41:25 <[florian]> w/ 2.6.27 as wlel Jan 25 13:42:08 ok, I'll try this one config. Jan 25 13:42:41 <[florian]> ermm, no you should use the profile that is appropriate for your board Jan 25 13:44:54 ok Jan 25 13:46:34 as I can understand the problem is generous, not special for this board/arch... It seems to be some misconfiguration when building it. Jan 25 13:47:50 it initilizes the mtd partirions but not going forward to mount them (especially root_fs) Jan 25 13:49:00 so I checked out for init sequence and got problem in do_initcalls... A put there some printk in loop to see where it hangs, but it just tells nothing to me Jan 25 13:50:20 <[florian]> printk is only available if either an early console is registered or a full console driver Jan 25 13:51:54 well i got it Jan 25 13:52:14 the last messages is Jan 25 13:52:16 NET: Registered protocol family 15 Jan 25 13:52:16 do_initcalls: result 0 Jan 25 13:52:16 do_initcalls: Calling do_one_initcall(8027577c) Jan 25 13:52:16 do_initcalls: result 0 Jan 25 13:52:16 do_initcalls: Calling do_one_initcall(80275780) Jan 25 13:52:18 turn off boot console early0 Jan 25 13:52:32 <[florian]> ok, then the switch from console early0 to /dev/ttyS0 did not work Jan 25 13:54:19 it didn't work, yes.. because no root_fs mounted. And that's because no 'prepare_namespace' function called... Jan 25 13:54:47 <[florian]> no, that's not related Jan 25 13:55:00 <[florian]> console should work no matter if there is a filesystem available or not Jan 25 13:55:11 <[florian]> could you please post the full bootlog of your device ? Jan 25 13:55:57 it's very big because of my "traceinitcall" injection Jan 25 13:56:13 <[florian]> put it on pastebin.com Jan 25 13:56:14 <[florian]> not here Jan 25 13:56:27 ok Jan 25 13:57:30 http://pastebin.com/d10773296 Jan 25 13:58:18 <[florian]> NLeiten: I do not see what is wrong with your kernel,everything seems to work properly, which type of image did you build ? Jan 25 13:59:45 sorry, don't understand image of what? FW binary? Jan 25 14:00:20 <[florian]> in menuconfig "Target images" what did you choose ? Jan 25 14:01:20 <[florian]> seems to me like you do not have a filesystem available Jan 25 14:01:44 well, I made up own sdk. Not building from openwrt kamikaze one... I'm using patches, tolchains from openwrt, but wrote own scripts to compile. Jan 25 14:01:57 <[florian]> ok, then your problem is on your side :) Jan 25 14:02:10 <[florian]> clearly, the kernel works fine, but your init is broken Jan 25 14:02:12 :) Jan 25 14:03:22 hey [florian] thanks for the mac80211 fix Jan 25 14:03:32 it seems to work now Jan 25 14:03:40 i'll try to make openwrt firmware Jan 25 14:04:05 <[florian]> Acinonyx_: thank juhosg instead ;) Jan 25 14:32:58 florian * r14174 /packages/net/ntpd/Makefile: [package] prevent ntpdate from using librt, bump release number to 2, should have been 1 when updated to 4.2.4p6 Jan 25 14:59:28 jow * r14175 /packages/net/miniupnpd/ (Makefile files/miniupnpd.init files/upnpd.config): packages/net/miniupnpd: implement option for "secure" mode and enable it by default Jan 25 15:51:05 how to force make with user root? i'm making export 'FORCE=1' before start 'make' or type 'make FORCE=1' Jan 25 15:51:27 but it says 'Prerequisite check failed. Use FORCE=1 to override.' Jan 25 16:00:42 nbd: /win 21 Jan 25 16:00:50 ... Jan 25 16:13:30 got problem with autocon Jan 25 16:13:33 *autoconf Jan 25 16:13:53 make[4]: Entering directory `/home/firmware/atheros/ThorFW/openwrt/kamikaze/build_dir/host/autoconf-2.62' Jan 25 16:13:53 There seems to be no Makefile in this directory. Jan 25 16:13:53 You must run ./configure before running `make'. Jan 25 16:13:53 make[4]: *** [abort-due-to-no-makefile] Error 1 Jan 25 16:14:22 if I pass ./configure in autoconf - it nees some special parameters to pass? Jan 25 16:16:51 solved Jan 25 16:17:02 old configuration passed to autoconf Jan 25 16:21:28 mirko * r14176 /packages/net/ (shorewall-common/Makefile shorewall-shell/Makefile): upgrade shorewall - thanks to puchu Jan 25 17:01:58 florian * r14177 /trunk/target/linux/rb532/files/: [rb532] delete files directory, could overlap with newer files and is redundant with per kernel-version directory Jan 25 17:07:02 ? Jan 25 17:24:30 <[florian]> mirko: they actually overlaped in the case of a 2.6.27 kernel, where there are no files-2.6.27 directory Jan 25 17:24:43 <[florian]> so files/* get copied as well, resulting in incoherencies Jan 25 17:28:15 blogic * r14178 /trunk/include/quilt.mk: if quilt failed whilst patching, no output was shown Jan 25 18:17:55 ah - ok Jan 25 18:33:25 juhosg * r14179 /trunk/ (10 files in 10 dirs): [kernel] update to 2.6.27.13 Jan 25 18:53:26 * jb42 svn updates Jan 25 19:00:47 nbd * r14180 /trunk/ (include/prereq-build.mk scripts/md5sum tools/Makefile): improve support for building on mac os x by improving detection of missing components Jan 25 19:03:06 nbd * r14181 /trunk/tools/quilt/Makefile: detect quilt configure errors properly (their broken configure script does not return proper error codes) Jan 25 19:05:10 nbd * r14182 /branches/8.09/ (include/prereq-build.mk tools/Makefile tools/quilt/Makefile): merge r14180,r14181 to 8.09 Jan 25 19:10:45 marek * r14183 /packages/net/batman/Makefile: update batman package to 0.3.1 release Jan 25 19:47:40 juhosg * r14184 /trunk/target/linux/generic-2.6/config-2.6.28: [kernel] generic-2.6/2.6.28: add missing config symbol Jan 25 20:31:29 juhosg * r14185 /trunk/include/kernel-version.mk: [kernel] remove trailing whitespace from the md5sum of the 2.6.27.13 tarball (closes #4502) Jan 25 20:42:13 matein4 * r14186 /trunk/scripts/flashing/jungo-image.py: Added support WRV54G. Jan 25 20:50:30 lars * r14187 /feeds/desktop/libs/libburn/ (. Makefile): Add libburn. Jan 25 20:52:05 lars * r14188 /feeds/desktop/libs/libisofs/ (. Makefile): Add libisofs package. Jan 25 21:04:21 juhosg * r14189 /trunk/target/linux/adm5120/ (28 files in 10 dirs): [adm5120] preliminary support for 2.6.28 (based on the patches by Alexandros C. Couloumbis) Jan 25 21:07:20 juhosg * r14190 /trunk/target/linux/adm5120/Makefile: [adm5120] oops, revert to 2.6.27.13 Jan 25 21:45:44 how to select adm5120 target for kernel in "make kernel_menuconfig"? Jan 25 21:47:12 on what key it depends? Jan 25 21:50:23 [florian]: are you here? Jan 25 23:27:11 FYI: compat-wireless is faulting on me in most recent svn releases. trunk/build_dir/linux-ar71xx/compat-wireless-2009-01-24/net/mac80211/main.c:161 Jan 25 23:27:55 I'm working on a package for msmtp -- I find ssmtp to be buggy / bloated / unmaintained Jan 25 23:28:01 http://msmtp.sourceforge.net/ Jan 25 23:28:04 Anyone interested? Jan 25 23:29:41 My build yesterday was SVN Revision: 14158 and used compat-wireless-2009-01-23 and had no such crash Jan 26 00:27:29 nbd * r14191 /trunk/package/madwifi/patches/393-mbss_vap_auth.patch: madwifi: fix a noderef problem in the mbss vap cleanup Jan 26 00:27:41 nbd * r14192 /trunk/package/madwifi/patches/395-ath_ff_unmap.patch: madwifi: fix fast-frame related dma unmap bug Jan 26 00:27:50 nbd * r14193 /trunk/package/madwifi/patches/ (5 files): madwifi: fix a race condition in the wds sta separation, which is triggered by using wpa authentication on the ap Jan 26 00:33:20 nbd * r14194 /branches/8.09/package/madwifi/patches/395-ath_ff_unmap.patch: merge r14192 to 8.09 Jan 26 01:31:59 is any developer available? Jan 26 01:32:28 maybe Jan 26 01:32:31 ;) Jan 26 01:32:34 ^) Jan 26 01:32:36 :) Jan 26 01:33:10 i have made some improvements in my patched kernel from openwrt.... Jan 26 01:33:21 well, just look into rootfs_split patch Jan 26 01:33:40 and add Endianess check in squashfs_detect part Jan 26 01:34:21 because of using little endian it just cannot check with SQUASHFS_MAGIC, but need to check SQUASHFS_MAGIC_SWAP also Jan 26 01:34:37 as it is done in native squashfs driver in kernel Jan 26 01:35:38 ok, feel free to send patches against trunk Jan 26 01:37:15 ouch, sorry but i don't know how actually this is done Jan 26 01:37:41 i'm not quietly get understanding with svn/git/cvs work Jan 26 01:37:52 how did you generate your patch? Jan 26 01:37:58 did you use quilt in the kernel tree? Jan 26 01:38:00 so I can take patch Jan 26 01:38:59 I just wrote some code to kernel part... and then diff Jan 26 01:40:32 ok Jan 26 01:42:28 i can send patch somewhere else and you can get it to trunk :) Jan 26 01:43:48 ok Jan 26 01:50:30 Can someone package my package? I submitted to -devel. Thanks, :) Jan 26 01:53:08 nbd: well, it some new question - do i need to send patch with only my changes or in one file of rootfs_split? Jan 26 01:53:17 only the changes Jan 26 01:53:59 nbd: do! :P Jan 26 01:55:06 ok Jan 26 02:06:32 where to send? Jan 26 02:08:34 nbd are you here? Jan 26 02:13:04 NLeiten: openwrt-devel@lists.openwrt.org along with a short description what was changed and why Jan 26 02:13:54 xMff ok Jan 26 02:28:11 i sent patch with description Jan 26 02:57:46 nbd * r14195 /trunk/package/madwifi/patches/ (14 files): madwifi: preserve the interface mode correctly when using wds with sta separation Jan 26 02:59:15 well, the reply is that my patch is rejected... and i'm not allowed to submit patches... Jan 26 02:59:26 subscribe to the list Jan 26 02:59:37 :) **** ENDING LOGGING AT Mon Jan 26 02:59:57 2009