**** BEGIN LOGGING AT Mon May 24 02:59:57 2010 May 24 07:54:05 Hi, I am getting this error when trying to compile backfire with lucid: May 24 07:54:06 cp: cannot stat `.../backfire/build_dir/target-mipsel_uClibc-0.9.30.1/luci-0.9.0/libs/lucid/dist/*': No such file or directory May 24 07:54:06 make[3]: *** [/home/nagydani/src/backfire/bin/brcm-2.4/packages/luci-lucid_0.9.0-1_brcm-2.4.ipk] Error 1 May 24 07:54:06 Any ideas why this might be? May 24 08:04:18 epointsystem my compile server ist xbuntu lucid x86_64 and here works fine, i followed this wiki http://wiki.openwrt.org/doc/howto/buildroot maybe you are missing some devs library/tools May 24 08:08:40 crow: it's not that lucid. :-) There's a package in OpenWRT called lucid (luci super-server daemon). May 24 08:08:49 But thanks for the link. May 24 08:09:19 epointsystem oh sorry didnt read your log long enough :/ May 24 08:09:22 I can build the firmware allright, except for this particular package. May 24 08:12:03 It is included in the binary distribution, so it should be certainly possible to compile (besided, it is a fairly central piece), but if you just select it from menuconfig, you get this error. May 24 08:15:38 crow: does your build have a web UI? May 24 08:20:47 epointsystem no i dont build luci May 24 10:40:34 nico * r21561 /packages/lang/luasql/ (. Makefile): packages: add luasql package - Lua SQL bindings for MySQL, PostgreSQL & SQLite (closes: #7353) May 24 11:37:24 epointsystem: this kind of error happens when luci features after it was already built, make package/luci/clean should fix it May 24 11:37:42 *features are changed May 24 14:15:47 xMff: thank you. Indeed. May 24 14:57:34 [florian]: you don't know by chance what's at 0xfff55000 on a bcm6348, do you? ;) May 24 15:05:54 nbd: regarding the mtd patch, there is a problem May 24 15:06:09 what problem? May 24 15:06:19 nbd: it doesn't work, because mmap seeme to be buggy on the wrt160nl May 24 15:06:37 details please May 24 15:07:49 nbd: basically, if I mmap the mtd device, the map seems to be empty from time to time May 24 15:08:29 nbd: it works, if I add the argument to keep the pages in memory all the time (not sure, how it's called), but not without out it May 24 15:08:39 nbd: empty meaning filled with zero May 24 15:08:47 odd May 24 15:09:09 nbd: it doesn't happen always, but most of the time May 24 15:09:28 nbd: and nearly always after I run sync on the command line May 24 15:09:28 is this happening with the code that's in svn right now? May 24 15:10:02 rtz2: so mmap basically only serves stuff out of the io cache? (if it does not work after sync) May 24 15:10:27 nbd: no it doesn't, but I don't think it mmaps anything on the ar71xx May 24 15:10:38 can you show me the code that's causing issues? May 24 15:10:50 xMff: somethink like that May 24 15:12:08 nbd: we had similar "weirdnesses" with nvram on ar71xx, there it was a file backed mmap and that only worked sometimes until MAP_LOCKED or so was passed to the mmap call May 24 15:12:36 nbd: do you still have my latest mtd patch around? May 24 15:12:41 schould be easy to create small testcase May 24 15:12:44 *should May 24 15:12:55 rtz2: subject of the email? May 24 15:13:23 nbd: MTD rework May 24 15:13:43 ok. i'll take another look May 24 15:14:13 nbd: one moment, there was another bug with the wrt160nl, let me take a look at it May 24 15:15:20 nbd: in trx.c, you have to remove the first #ifdef target_ar71xx block May 24 15:15:42 nbd: then run "mtd fixcrc" May 24 15:16:42 nbd: in trx_check, if you print out some random bytes from the map, it's empty May 24 15:18:51 nbd: I can also create a smaller testcase, if you want May 24 15:20:43 yes, that would be helpful May 24 15:20:44 rtz2: I bet its also the cause why mtd carshes when fixing up the fis table May 24 15:20:54 its also shared mmap May 24 15:22:13 xMff: never had to deal with it, but it seems likely May 24 15:23:24 the dangerous thing is that the mmap call always succeeds so sysupgrade happily moves on and then crashes in the final step rendering the device unbootable May 24 15:42:32 nbd: it doesn't happen anymore :/ May 24 15:43:05 nbd: I didn't change anything, I have no clue, what the problem is May 24 15:43:20 nbd: maybe the nvram problem will be easier to reproduce May 24 15:44:24 wait, it did happen again May 24 15:44:42 but it's pretty rare now May 24 16:12:24 nbd: ok, got it May 24 16:12:30 http://openwrt.pastebin.com/KswAJ6C8 May 24 16:12:37 nbd: works 90% of the time May 24 16:12:49 nbd: if not, run sync and try again May 24 16:13:17 nico * r21562 /packages/lang/ruby/Makefile: [packages] ruby: update to 1.9.1-p376, cleanup May 24 16:13:38 nbd: the mtd number and sice are for the wrt160nl, so you may have to change it May 24 16:25:41 13:27:50 < Weedy> nbd: apply this mkay? http://pastebin.com/raw.php?i=rwtcN673 May 24 16:26:02 Weedy: stuff like this should go through upstream review, imho May 24 16:26:40 after more testing you can delete the first and last blocks, it's the change from IFs to CASE that broke it May 24 16:26:47 broke minitor mode May 24 16:26:54 i'm reluctant to add non openwrt-specific patches to mac80211 that didn't go through upstream review May 24 16:27:13 it's a revert of ... uhh May 24 16:27:15 one sec May 24 16:27:34 especially if it is a revert, it needs to be discussed on linux-wireless@vger.kernel.org May 24 16:27:42 please start there May 24 16:27:50 and i'll merge the patch as soon as it looks like it's going to get accepted May 24 16:27:54 56d1de0a21db28e41741cfa0a66e18bc8d920554 May 24 16:28:00 i have a thread there May 24 16:28:15 i just saying, if you want it working NOW then you can use that May 24 16:28:48 jow * r21563 /trunk/package/pptp/ (Makefile files/pptp.sh): [package] pptp: add "buffering" uci option to enable --nobuffer if needed (#7080) May 24 16:29:25 ok May 24 16:29:29 i'll look into it May 24 16:29:39 some of this stuff looks wrong to me too May 24 16:29:47 nbd: http://article.gmane.org/gmane.linux.kernel.wireless.general/51265 May 24 16:30:03 yeah, just found the thread in my email folder May 24 17:07:05 jow * r21564 /branches/backfire/package/ (15 files in 12 dirs): May 24 17:07:05 [backfire] merge ppp, pppoe, pptp and 3g fixes May 24 17:07:05 r21283, r21284, r21285, r21287, r21302, r21303, r21341, May 24 17:07:05 r21361, r21379, r21383, r21386, r21387, r21388, r21389, May 24 17:07:05 r21390, r21391, r21392, r21393, r21426, r21556, r21563 May 24 18:13:01 joy, finally figured the issue with my console not working inside kvm :> May 24 18:46:07 Weedy there was some Hunks with your patch http://openwrt.pastebin.com/H6jEf50k May 24 20:35:50 nico * r21566 /trunk/target/linux/ (42 files in 9 dirs): target/linux: refresh 2.6.32 patches May 24 20:38:38 nico * r21567 /trunk/toolchain/kernel-headers/Makefile: toolchain/kernel-headers: simplify linux headers installation (remove legacy hacks, cleanup) May 24 21:44:45 build #57 of octeon is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/57 May 24 23:47:29 build #53 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/53 May 25 00:29:34 hi! is UCI still being actively developed? May 25 00:30:03 yes, i just don't have much time for it lately May 25 00:30:50 last checkin was in december? Just asking if the project maybe took a different route or so May 25 00:32:41 mh how do I crosscompile uci... it seems setting CC does not work? May 25 00:32:58 take a look at the openwrt makefile May 25 00:33:16 ok I will eat a banana first :) May 25 00:33:50 by the way I have some patches around for UCI that makes it easier to write queries in C May 25 00:34:02 ok, feel free to send them to me May 25 00:34:26 basically is makes an interface very similar to using printf so one does not have to build up a string upfront May 25 00:44:28 nbd, the trick to crosscompile is $(TARGET_CONFIGURE_OPTS) ?? May 25 01:04:43 alesan: that one sets CC, LD, ... May 25 01:10:12 Is the architecture string stored somewhere in the installation? **** ENDING LOGGING AT Tue May 25 02:59:57 2010