**** BEGIN LOGGING AT Tue Jun 02 02:59:57 2009 Jun 02 03:31:50 nbd: would your over9000 madwifi patches work on 2.6.27? Jun 02 03:33:17 russell_: I apologise, but I cannot Jun 02 03:33:33 russel_: If it's images you want, I can get them Jun 02 03:37:57 mib_w02elm: no thanks, but i wouldn't start to try to port something to a device i don't have in my possession Jun 02 04:14:20 russell_: :(. Too bad Jun 02 05:17:02 hmm. what fixed mrd6? Jun 02 05:24:24 mysterious. either my memory is bad (yes) or something odd happened (or both) Jun 02 05:24:30 anyway, better now. Jun 02 05:27:55 current list of packages (as of r16290) not building for wgt634u (in packages and luci with a "Select All Packages"): iptables-snmp, xsupplicant, openswan, classpath, xinetd, weechat. Jun 02 06:42:25 kaloz * r16291 /trunk/target/linux/avr32/ (3 files in 2 dirs): upgrade avr32 to 2.6.30-rc7 Jun 02 07:37:30 juhosg: ping Jun 02 09:18:57 kaloz * r16292 /trunk/target/linux/ (cavium-octeon/ octeon/ octeon/Makefile): rename cavium-octeon to octeon Jun 02 09:50:44 kaloz * r16293 /trunk/include/target.mk: add pcie feature flag Jun 02 09:54:46 kaloz * r16294 /trunk/ (scripts/metadata.pl target/Config.in): missing changes for pcie handling Jun 02 10:02:25 kaloz * r16295 /trunk/package/kernel/modules/netdevices.mk: package the e1000e driver Jun 02 10:06:43 kaloz * r16296 /trunk/target/linux/ppc44x/config-default: enable PCIE on ppc44x as well Jun 02 10:19:56 mirko * r16297 /trunk/target/linux/s3c24xx/ (3 files in 2 dirs): add bootsplash to 2.6.30 as well and enable it per default Jun 02 13:43:16 mirko * r16298 /feeds/efl/enlightenment/files/uci-defaults/x11_standard: start e17 instead of xterm when profile 'standard' is chosen - this was accidently committed before Jun 02 14:43:38 i have a big-endian squashfs image, is there a way to mount it under little-endian box without resolving to qemu,etc Jun 02 14:45:05 <[florian]> I think you can mount it using some hacks Jun 02 14:46:59 trying to look at antonio's new robin code, he only releases .bin at this point. any hints on the 'hacks' Jun 02 14:47:29 <[florian]> hold on Jun 02 14:49:13 <[florian]> http://tldp.org/HOWTO/html_single/SquashFS-HOWTO/ Jun 02 14:49:41 <[florian]> "Both big and little endian architectures are supported; SquashFS can mount file systems created on different byte-order machines" Jun 02 14:50:09 <[florian]> so I think you should be able to mount the be filesystem anyway Jun 02 14:54:26 [florian]: could be true, as the error is "SQUASHFS error: Can't find a SQUASHFS superblock on loop1", though "file squashfs2.2065" showed "squashfs2.2065: Squashfs filesystem, big endian..." Jun 02 14:54:52 anyway Jun 02 14:55:19 i extracted squashfs from his .bin file, maybe something wrong there Jun 02 14:55:38 <[florian]> if file reports it to be a squashfs image it should not be a problem Jun 02 14:57:12 <[florian]> you have the squashfs module for your kernel right? Jun 02 14:57:36 <[florian]> which squashfs version are you trying to mount btw? Jun 02 14:57:53 hello everyone, I'm compiling openwrt brcm-2.6 include in it ipv6 and others things and I get some warning, that i have not idea if can be a future issue , there is it http://openwrt.pastebin.com/d6093f2ca Jun 02 14:58:01 yeah...maybe it's ubuntu9.04+custom2.6.29-4 issue, as squashfs-source is broken on ubuntu9.04/2.6.28(could not compile m-a a-i), at home i have 8.04 with squashfs loaded, let me reverse ssh home Jun 02 14:58:34 <[florian]> fofware: that should be reported to the upstream kernel I do not think we modify those files Jun 02 14:58:47 <[florian]> fofware: well in fact we add those files with a patch Jun 02 14:58:54 <[florian]> fofware: but it should not be a problem Jun 02 14:59:46 ok, thanks <[florian]> Jun 02 14:59:59 description: squashfs 4.0, a compressed read-only filesystem Jun 02 15:00:34 <[florian]> squashfs4 is not backward compatible with squashfs3 Jun 02 15:01:51 sigh. thanks. try 8.04 at home now Jun 02 15:09:11 lars * r16299 /feeds/efl/ (2 files in 2 dirs): Jun 02 15:09:11 Don't call evas_object_image_scale_hint_set on Evas_Rectangle objects. This Jun 02 15:09:11 caused random segfaults previously, which could be workedaround by enabling Jun 02 15:09:11 evas-magic-debug. Not that the problem fixed we can disable it again. Jun 02 15:12:13 [florian]: tried on 8.04 with squashfs3.3, "[ 434.080597] SQUASHFS: Mounting a different endian SQUASHFS filesystem on loop0" Jun 02 15:12:25 [ 434.080984] SQUASHFS error: zlib_inflate returned unexpected result 0xfffffffd, srclength 8192, avail_in 231, avail_out 8192 Jun 02 15:12:36 i think it did not work. Jun 02 15:13:53 going to revive rb532a as a mips testbed to avoid this hassle Jun 02 15:37:44 <[florian]> xxiao: rb532a is mipsel Jun 02 16:09:59 [florian]: that's probably right, recall i was using debian on it. thanks! Jun 02 16:13:13 hmm Jun 02 16:13:20 /aux/src/openwrt/staging_dir/toolchain-i386_gcc-4.1.2_uClibc-0.9.30.1/usr/lib/gcc/i486-openwrt-linux-uclibc/4.1.2/../../../../i486-openwrt-linux-uclibc/lib/libc.a(libc_pthread_init.os): In function `__libc_pthread_init': Jun 02 16:13:20 libc_pthread_init.c:(.text+0xe): undefined reference to `__libc_multiple_threads' Jun 02 16:13:39 on x86-alix (r16290) Jun 02 18:20:49 [florian]: did you just renamed target/linux/cavium-octeon to octeon? saw a 4MB diff between r16290 and r16299 Jun 02 18:21:38 yes Jun 02 18:21:49 no wonder..thanks Jun 02 18:22:03 actually Kaloz did it, r16292 Jun 02 18:24:07 the old name was ghey Jun 02 18:44:21 can someone update wireless to 0602 release, 80211s is not working with ath9k(lock-up) with compat-wireless0521 Jun 02 18:45:34 plus ath5k crashes when trying 80211s Jun 02 18:48:55 hauke * r16300 /trunk/package/mac80211/ (7 files in 2 dirs): [mac80211] Update to compat-wireless-2009-06-02 Jun 02 18:52:09 Hauke: thanks! waiting for just that, your 06-02 commit Jun 02 18:54:46 why does x.ipkg exist under bin/packages after I selected "x" as compiled-in(i.e. not a module-style package)? Jun 02 18:56:03 in the openwrt menuconfig * means create package and install it into the image, M means create package but do not install into image. Jun 02 18:56:28 Bartman007: ok, thanks. Jun 02 18:57:41 I'm not sure if anyone saw this :make -r world: build failed. Please re-run make with V=99 to see what's going on, when I added "V=99", no errors and it compiles to the end Jun 02 18:57:52 i saw this a couple of times already Jun 02 18:58:24 most of the time V=99 does fail, though Jun 02 19:12:13 xxiao: are you using -j for parallelization? Jun 02 19:13:41 if so then there is a missing dependency in a package and by the second runthrough it has been compiled Jun 02 19:20:15 Bartman007: yes -j=4 Jun 02 19:21:05 i guess that's the problem, as I don't recall this problem when I forgot to add -j Jun 02 19:22:11 wierd. gpsd build fails for x86 but built for wgt634u (circa r16290) Jun 02 19:22:50 e.g. ./.libs/libgps.so: undefined reference to `operator delete(void*)' Jun 02 19:23:23 try a different gcc version Jun 02 19:25:03 that was gcc-4.1.2 i guess Jun 02 19:26:58 svn up from r16290 to r16299 rerun make ok, pulled in r16300(compat-wireless) rerun make failed at bridge package(early), weird as new compat-wireless should not impact that Jun 02 19:27:13 is there a preferred gcc version? Jun 02 19:29:25 russell_: 4.1.2 on most stuff, but maybe x86 is stupid.. try with 4.2.4 or 4.3.3 Jun 02 19:29:40 i'm building 4.4.0 atm Jun 02 20:10:31 hauke * r16301 /trunk/target/linux/ifxmips/files/drivers/mtd/maps/ifxmips.c: [ifxmips] Fix accessing array after it end. Jun 02 20:52:07 hmm. not liking 4.4.0: Jun 02 20:52:07 configure: error: Building GCC requires GMP 4.1+ and MPFR 2.3.2+. Jun 02 20:52:07 Try the --with-gmp and/or --with-mpfr options to specify their locations. Jun 02 20:52:16 trying with 4.3.3 next Jun 02 20:55:10 <[florian]> wtf ? we build gmp and mpfr when you choose gcc-4.3 or gcc-4.4 and it even works fine when switching between compilers Jun 02 20:55:42 couldn't find maybe? or just some build system problem? dunno. Jun 02 20:56:03 <[florian]> rm -rf tmp .config; make menuconfig ? Jun 02 20:56:31 well, in the middle of a 4.3.3 build now Jun 02 20:56:40 when/if that fails, will do Jun 02 21:33:42 hmm. gcc-4.3.3 gives me the same error on gpsd Jun 02 21:34:07 could be a gcc vs cpp issue Jun 02 21:35:01 ./.libs/libgps.so: undefined reference to `__gxx_personality_v0'' Jun 02 21:35:01 ./.libs/libgps.so: undefined reference to `vtable for __cxxabiv1::__class_type_info' Jun 02 21:35:01 ./.libs/libgps.so: undefined reference to `operator new(unsigned int)' Jun 02 21:35:05 also Jun 02 21:35:31 when I search for "operator delete(void*)" in google, I get some references to gpsd Jun 02 21:35:46 I mean undefined reference to ... Jun 02 21:38:12 looks vaguely like a gcc vs g++ thing Jun 02 21:39:53 erm yes g++, not cpp Jun 02 21:41:45 try adding -lstdc++ to the ldflags Jun 02 22:42:26 xMff: you know i think the ulibc upgrade to .30 is causing my segfaults Jun 02 22:43:38 except that when I did .30 before trunk no problems, but now I do Jun 02 22:43:46 could it be kernel related? Jun 02 22:44:02 there's recent changes I understand Jun 02 22:47:51 have you imq loaded? Jun 02 22:47:59 yes Jun 02 22:48:13 Weedy, you? Jun 02 22:50:35 maybe it's realted Jun 02 22:50:47 you could check whether it makes a difference if imq is not loaded Jun 02 22:50:57 will do Jun 02 22:52:25 you're testing brcm63xx ? Jun 02 22:52:30 yes Jun 02 22:52:36 ok Jun 02 22:53:47 I'm working on a build with changes up to last night Jun 02 23:00:44 libtool: ignoring unknown tag CXX Jun 03 00:16:28 xMff: why would imq segfault dropbear and nsclient Jun 03 00:52:57 ok, I'm not sure my last message made it; I just said that rmmod imq stopped the segfaults and insmod imq brought them back, so at least for me it is definitely imq **** ENDING LOGGING AT Wed Jun 03 02:59:57 2009