**** BEGIN LOGGING AT Sun Nov 16 02:59:56 2008
Nov 16 07:49:37 <CIA-8> juhosg * r13226 /trunk/tools/firmware-utils/src/ (mkmylofw.c myloader.h): [tools] firmware_utils/mkmylofw: add definitions for the NP25G and WPE53G boards
Nov 16 08:59:54 <johnrw> __pci_data_abort
Nov 16 11:53:41 <CIA-8> florian * r13227 /packages/utils/sysstat/ (6 files in 2 dirs): Add sysstat (#2124)
Nov 16 12:05:01 <fish_> can i preconfigure the firewall when building an image from trunk? does putting a firewall.conf into trunk/files/etc still works?
Nov 16 12:05:21 <fish_> looks like the firewall is finally uci based
Nov 16 12:06:14 <fish_> uhm, the default config is in package/firewall/files/firewall.config.. is it possible to put a modified file in trunk/files?
Nov 16 12:26:21 <CIA-8> florian * r13228 /packages/net/iperf/Makefile: The original iperf website is offline, use the openwrt mirror instead
Nov 16 12:26:40 <[florian]> fish_: yes
Nov 16 12:26:50 <[florian]> fish_: you should put your custom files in trunk/files/
Nov 16 12:26:56 <[florian]> fish_: with the same directory structure
Nov 16 12:27:18 <fish_> [florian]: well yes, but does that work with uci config as well?
Nov 16 12:27:30 <[florian]> why would not it ?
Nov 16 12:27:50 <fish_> well because the uci default config is created on first boot afaik
Nov 16 12:28:07 <[florian]> files get copied in the end of the build process, so uci on first boot should read the file if it exists
Nov 16 12:28:16 <fish_> ah ok
Nov 16 12:28:26 <fish_> than i'll try that
Nov 16 12:28:44 <fish_> btw, could:
Nov 16 12:28:46 <fish_> 01:12 < fish_> can someone delete or fix toolchain/uClibc/patches/160-add_openat_and_friends.patch? is has broken  the toolchain...
Nov 16 12:29:02 <[florian]> let me have a look at it
Nov 16 12:29:44 <[florian]> which platform does it break with ?
Nov 16 12:32:05 <fish_> mom
Nov 16 12:33:02 <fish_> i tried to create a struct of __NR_openat which is undeclared
Nov 16 12:33:10 <fish_> i got libc/sysdeps/linux/common/openat.c:20: error: `__NR_openat' undeclared (first use in this function)
Nov 16 12:34:50 <fish_> i guess it break all platforms
Nov 16 12:35:22 <[florian]> it does not break here neither on mips or mipsel
Nov 16 12:38:01 <fish_> maybe its already fixed somewhere else..
Nov 16 12:38:09 <fish_> i had that problem right aber running make distclean
Nov 16 12:38:14 <fish_> s/aber/after/
Nov 16 12:38:22 <fish_> ..and building a fresh image
Nov 16 12:38:47 <fish_> someone else in #openwrt had the same problem
Nov 16 12:40:53 <fish_> [florian]: i'll make a fresh checkout and try that again
Nov 16 12:41:28 <[florian]> I cannot reproduce this problen with x86, so yes, seems like it got fixed
Nov 16 12:42:54 <fish_> maybe its using some header file from the host system which i'm missing
Nov 16 12:58:41 <fish_> [florian]: will the telnetd be bind to all interfaces after the first reboot or just to the lan-bridge? i have a router here and i'm connected to it via the WAN port but i want to flash it with mtd and access it afterwards without having to plug the cable in one of the lan ports
Nov 16 12:59:04 <fish_> if it binds to all interfaces i just had to open port 23 on WAN i guess...
Nov 16 13:00:54 <[florian]> it will bind to all interfaces, the firewall dropping by default incoming connections on port 23 coming from the WAN interface
Nov 16 13:02:28 <fish_> yep, ok, than i'll open it copy putting a config to trunk/files/etc/config/firewall and it should be fine
Nov 16 13:02:36 <[florian]> genau ;)
Nov 16 13:02:44 <fish_> *hrhr*
Nov 16 13:03:29 <[florian]> Ich spreche nur ein bissen deutsch :p
Nov 16 13:04:02 <fish_> well, when i talk to nbd here i also stay in english
Nov 16 13:04:20 <fish_> but my english sucks ;)
Nov 16 13:04:36 <[florian]> general usage is if you cannot speak english, speak german, someone here probably understands you ;~
Nov 16 13:04:56 <fish_> i have especially problem with the tenses ;)
Nov 16 13:05:22 <fish_> sometime ago someone in a newsgroup told me i sound like yoda ;))
Nov 16 13:06:15 <fish_> well, my english is good enough the communicate but it would be nice if i were better at it ;)
Nov 16 13:06:19 <[florian]> lol !!
Nov 16 13:06:29 <fish_> trunk is still broken here
Nov 16 13:06:35 <fish_> libc/sysdeps/linux/common/openat.c: In function `__syscall_openat':
Nov 16 13:06:35 <fish_> libc/sysdeps/linux/common/openat.c:20: error: `__NR_openat' undeclared (first use in this function)
Nov 16 13:06:38 <fish_> libc/sysdeps/linux/common/openat.c:20: error: (Each undeclared identifier is reported only once
Nov 16 13:06:41 <fish_> libc/sysdeps/linux/common/openat.c:20: error: for each function it appears in.)
Nov 16 13:06:44 <fish_> make[5]: *** [libc/sysdeps/linux/common/openat.os] Error 1
Nov 16 13:07:01 <[florian]> ok will see what I can do to reproduce it
Nov 16 13:07:23 <fish_> [florian]: don't know where you have __NR_openat declared but i guess if you rebuild your whole toolchain it will happen too
Nov 16 13:07:48 <[florian]> I actually started building all the toolchains from scratch
Nov 16 13:08:23 <fish_> maybe its really a header file from your host system
Nov 16 13:08:51 <fish_> do you have __NR_openat declared there somewhere?
Nov 16 13:10:20 <fish_> [florian]: do you have it in <sys/unistd.h>?
Nov 16 13:10:58 <[florian]> nope
Nov 16 13:11:03 <[florian]> not on the host at least
Nov 16 13:12:09 <fish_> seems that its part of the linux headers at least in some versions: http://www.google.com/codesearch?hl=en&lr=&q=define\s*__NR_openat+lang%3Ac&sbtn=Search
Nov 16 13:12:35 <[florian]> yep
Nov 16 13:16:26 <russell_> can anyone tell my why kmod-hostap-pci isn't showing up in menuconfig?
Nov 16 13:19:12 <fish_> [florian]: that sure work for the firewall, right? http://perlpunks.de/paste/show/49201dc0.2a2f.3d0
Nov 16 13:20:02 <fish_> my router is on the cold balcony and i don't want to have to go out there if that went wrong *G*
Nov 16 13:24:33 <[florian]> this should :)
Nov 16 13:26:29 <[florian]> I am not that experienced with the new firewall
Nov 16 13:27:53 <fish_> yeah, the last time a was playing around with openwrt it has still the old firewall
Nov 16 13:32:57 <fish_> [florian]: do i still have to remove and rebuild the toolchain when switching the kernel version from 2.4 to 2.6?
Nov 16 13:33:29 <[florian]> that's better yeah
Nov 16 13:33:46 <[florian]> otherwise the build system somehow messes up compiler with kernel headers
Nov 16 13:33:51 <fish_> yep ;)
Nov 16 13:33:56 <[florian]> though it's been a while since I did not switch to test it
Nov 16 13:41:41 <CIA-8> florian * r13229 /trunk/target/linux/rdc/config-2.6.24: Remove some left-over debugging options making the kerkenl exceed 768KB, resulting in a non-bootable image
Nov 16 13:42:42 <CIA-8> florian * r13230 /trunk/target/linux/rdc/files/drivers/net/r6040.c: Fix small typo in the ethernet driver
Nov 16 16:04:08 <fish_> uhmm
Nov 16 16:04:33 <fish_> i can't find kmod-sit in menuconfig.. dunno why..
Nov 16 16:16:12 <CIA-8> florian * r13231 /trunk/target/linux/rdc/config-2.6.24: Since profiles kernel configuration are not concatenated to the one, resulting images will panic due to no platform flash driver, prevent this from happening
Nov 16 16:16:51 <fish_> [florian]: does kmod-sit shows up on your menuconfig?
Nov 16 16:17:09 <[florian]> yes
Nov 16 16:18:34 <fish_> weird..
Nov 16 16:30:31 <fish_> ah
Nov 16 16:30:34 <fish_> <- stupid
Nov 16 16:30:41 <fish_> # sit is compiled in the 2.4 ipv6 stack
Nov 16 16:31:27 <CIA-8> florian * r13232 /packages/net/wifidog/Makefile: Add missing dependency to wifidog, thanks florida (#4132).
Nov 16 16:36:07 <fish_> but it seems that there is another issue. ipv6 can't resolve tcp_destroy_sock
Nov 16 16:36:19 <fish_> nm shows that there is no such symbol in any of my modules
Nov 16 16:40:05 <CIA-8> florian * r13233 /packages/net/wifidog/Makefile: Bump wifidog release since last commit
Nov 16 16:46:33 <[florian]> fish_: toolchain problem confirmed for brcm-2.4
Nov 16 16:46:53 <[florian]> fish_: this system call does only exist with 2.6 kernels, thus the error
Nov 16 16:47:00 <fish_> ah ok
Nov 16 16:47:32 <fish_> [florian]: and it seems that ipv6 is broken for 2.4..
Nov 16 16:47:45 <[florian]> will check that later
Nov 16 16:47:45 <fish_> [florian]: i get insmod: unresolved symbol tcp_destroy_sock
Nov 16 16:48:07 <[florian]> ok
Nov 16 16:48:43 <fish_> [florian]: and i can't find that symbol in any modules
Nov 16 16:49:06 <[florian]> you are mising some kernel configuration option then
Nov 16 16:49:11 <[florian]> grep the source for tcp_destroy_sock
Nov 16 16:49:16 <fish_> ok
Nov 16 16:49:25 <[florian]> which will certainly show it being defined for a particular kernel config symbol
Nov 16 16:51:51 <fish_> its definied in net/ipv4/tcp.c and exported by net/netsyms.c (well, at least there is a line EXPORT_SYMBOL(tcp_destroy_sock); but i must admit that i don't know really what that means ;))
Nov 16 16:52:19 <fish_> [florian]: but both looks like it should be included in the kernel image..
Nov 16 16:54:00 <[florian]> ok
Nov 16 17:03:20 <CIA-8> florian * r13234 /trunk/toolchain/uClibc/patches/160-add_openat_and_friends.patch: openat is not an implemented syscall on linux-2.4, so make it only available for linux-2.6
Nov 16 17:03:51 <fish_> [florian]++
Nov 16 17:03:52 <fish_> ;)
Nov 16 17:03:59 <[florian]> this will fix the compilation error
Nov 16 17:04:13 <[florian]> I fix the other problems, when they come up :p
Nov 16 17:05:14 <fish_> if i can help to debug that ipv6 problem futher than tell me what i should try
Nov 16 17:05:36 <[florian]> I got a router right there with ipv6 connectivity to test, but thanks
Nov 16 17:36:16 <pensador> Where are 9uclib 0.9.28 patches?
Nov 16 17:36:30 <pensador> ls: cannot access ./patches/0.9.28: No such file or directory
Nov 16 17:40:52 <CIA-8> florian * r13235 /trunk/package/iptables/ (Makefile patches/1.3.8/006-ipt_conntrack.patch): Finally fix iptables -m conntrack (#988), bump release numnber
Nov 16 17:50:04 <[florian]> fish_: I am having no issue with ipv6
Nov 16 17:50:12 <[florian]> IPv6 v0.8 for NET4.0
Nov 16 17:50:12 <[florian]> IPv6 over IPv4 tunneling driver
Nov 16 17:58:00 <fish_> [florian]: do you have a module which provides that symbol?
Nov 16 17:58:46 <fish_> [florian]: maybe selecting ipv6 as module has enabled some code in the kernel image
Nov 16 17:58:51 <[florian]> yes
Nov 16 17:58:53 <[florian]> it does
Nov 16 17:59:44 <fish_> oh ok, than thats the reson. i build i image without ipv6 support and after flashing that image to my router i decided to build ipv6 support as a ipk module..
Nov 16 18:00:29 <fish_> [florian]: but imo thats not very nice.. shouldn't the additional code be in another kernel module? like it is on a vanilla kernel?
Nov 16 18:03:22 <[florian]> I agree, but the culprit is not probably openwrt but rather the config symbol that got used
Nov 16 18:03:41 <CIA-8> jow * r13236 /trunk/package/broadcom-wl/files/lib/wifi/broadcom.sh:
Nov 16 18:03:41 <CIA-8> package/broadcom-wl:
Nov 16 18:03:41 <CIA-8>  - use [tr]xantenna instead of [tr]xant to unify config
Nov 16 18:03:41 <CIA-8>  - implement txpower option
Nov 16 18:24:27 <loswillios> Bartman007: have you seen this before: http://pastebin.ca/1257839 ? 2.6.27.4 on WGT634U without swap (normal build, with only 2.6.27-patch applied)
Nov 16 18:25:45 <CIA-8> cyrus * r13237 /packages/utils/collectd/Makefile: Updated collectd to 4.4.4
Nov 16 19:31:50 <CIA-8> mirko * r13238 /packages/ (337 files in 152 dirs):
Nov 16 19:31:50 <CIA-8> - update to Xorg X11R7.4 and reorganization of the xorg-section
Nov 16 19:31:50 <CIA-8> - added tslib
Nov 16 19:31:50 <CIA-8> - added e17 and related efl-libs
Nov 16 19:35:00 <crazy_imp> mhm, e17 for openwrt?
Nov 16 19:39:03 <CIA-8> jow * r13239 /trunk/package/broadcom-wl/files/lib/wifi/broadcom.sh: move the txpower option to wifi-iface to mimic madwifi configuration
Nov 16 19:57:56 <CIA-8> jow * r13240 /trunk/docs/wireless.tex: update wireless documentation to reflect broadcom config changes
Nov 16 20:09:46 <larsc> datenmaulwurf: http://rafb.net/p/EKuWBE49.html
Nov 16 20:13:40 <glp> crazy_imp: e17 can be interesting when it comes to small screens: http://www.embeddedlinuxconference.com/elc_europe08/sessions.html#Barbieri
Nov 16 20:32:25 <crazy_imp> true
Nov 16 20:44:38 <pensador> i'm having problems
Nov 16 20:45:07 <pensador> in a linked library with gpsd. At start uclib does segfault
Nov 16 20:59:34 <CIA-8> jow * r13241 /branches/8.09/ (2 files in 2 dirs): merge r13236, r13239 and r13240
Nov 16 21:06:52 <pensador> hcidump.c:749: error: 'AI_ADDRCONFIG' undeclared (first use in this function)
Nov 16 21:07:20 <pensador> Anyone have this problem using uclibc 28.2
Nov 16 21:08:51 <datenmaulwurf> larsc: thx
Nov 16 22:09:41 <CIA-8> florian * r13242 /branches/8.09/package/iptables/ (Makefile patches/1.3.8/006-ipt_conntrack.patch): Merge r13235 to 8.09
Nov 16 22:10:11 <CIA-8> florian * r13243 /branches/8.09/target/linux/rdc/files/drivers/net/r6040.c: Merge r13230 to 8.09
Nov 16 22:15:19 <datenmaulwurf> crazy_imp: it's a project ;)
Nov 16 22:15:33 * datenmaulwurf is mirko
Nov 16 22:17:57 <crazy_imp> hehe
Nov 16 22:32:08 <CIA-8> mirko * r13244 /packages/Xorg/lib/libXft/Makefile: upgrade to 2.1.13
Nov 16 23:01:18 <mskzo>  segfault at bf91affc ip b7eeb836 sp bf91b000 error ffff0006 in libuClibc-0.9.29 in a hello world program
Nov 16 23:01:23 <mskzo> only have included lgps
Nov 16 23:01:42 <mskzo> is a linker problem,
Nov 16 23:01:46 <mskzo> how can i resolve?
Nov 16 23:07:30 <xMff> perhaps run strace on it to see where it segfaults?
Nov 16 23:07:33 <larsc> use a debugger and find out what causes the segfault. you probaly need to rebuild everything with debugging symbols enabled
Nov 16 23:13:19 <mskzo> Ok
Nov 16 23:13:20 <mskzo> 1 moment
Nov 16 23:27:04 <mskzo> another question, when i downgrade uclib i have hcidump.c:749: error: 'AI_ADDRCONFIG' undeclared
Nov 16 23:27:05 <mskzo> error
Nov 16 23:43:32 <mskzo> xMff, can i send to you the strace output?
Nov 16 23:46:34 <CIA-8> lars * r13245 /packages/Xorg/lib/pango/Makefile: Update pango to 1.22.2
Nov 16 23:48:27 <CIA-8> lars * r13246 /packages/Xorg/lib/cairo/Makefile: Update cairo to 1.8.2
Nov 16 23:49:45 <CIA-8> lars * r13247 /packages/Xorg/lib/librsvg/ (Makefile patches/): Update librsvg to 2.22.3
Nov 16 23:50:48 <CIA-8> lars * r13248 /packages/Xorg/lib/gtk2/Makefile: Update gtk2 to 2.14.4
Nov 16 23:56:58 <xMff> mskzo: put it in pastebin and post here, I'm not such a great C coder
Nov 16 23:57:06 <mskzo> http://rafb.net/p/IZRTZ957.html
Nov 16 23:58:23 <xMff> looks pretty innocent
Nov 16 23:59:14 <mskzo> ??
Nov 16 23:59:19 <xMff> I think your only option to debug that is a debugger, as larsc suggested
Nov 16 23:59:30 <xMff> there is nothing obvious to see in strace
Nov 16 23:59:53 <mskzo> how can i enable debug?
Nov 17 00:00:16 <xMff> good question, I guess that depends on the lib / program
Nov 17 00:00:27 <xMff> -DDEBUG perhaps?
Nov 17 00:02:13 <mskzo> ah ok
Nov 17 00:02:53 <mskzo> i readed a lot of tickets saying that downgrading uClibc to version 0.9.28 fixes the problem
Nov 17 00:03:10 <xMff> hmmm
Nov 17 00:04:17 <mskzo> https://dev.openwrt.org/ticket/3139 for example
Nov 17 00:05:11 <larsc> so its a good idea to debug the problem and fix it in uclibc 0.9.29 too ;)
Nov 17 00:06:06 <mskzo> larsc, can you help?
Nov 17 00:06:46 <mskzo> =)
Nov 17 00:06:46 <larsc> yes
Nov 17 00:07:08 <mskzo> so how can i enable debug mode?
Nov 17 00:08:08 <mskzo> i set enable debugging
Nov 17 00:08:11 <mskzo> in buildroot
Nov 17 00:08:50 <larsc> mom
Nov 17 00:08:59 <mskzo> and gdbserver
Nov 17 00:09:21 <mskzo> something more?
Nov 17 00:10:19 <larsc> make menuconfig -> advanced configuration options -> toolchain options -> build gdb
Nov 17 00:10:58 <larsc> and advanced configuration options -> build options -> enable debugging
Nov 17 00:11:28 <mskzo> Ok
Nov 17 00:11:30 <mskzo> building!
Nov 17 00:16:48 <mskzo> built
Nov 17 00:19:21 <xMff> hmm is this related? http://bugs.busybox.net/view.php?id=5694
Nov 17 00:20:54 <mskzo> i don know
Nov 17 00:21:13 <CIA-8> lars * r13249 /packages/Xorg/lib/vte/Makefile: Update vte to 0.17.4
Nov 17 00:21:42 <mskzo> larsc, running
Nov 17 00:22:04 <mskzo> but i dont have gdb binary
Nov 17 00:22:12 <mskzo> only gdbserver
Nov 17 00:22:24 <mskzo> how can i do the debug?
Nov 17 00:24:12 <larsc> you need a network connection between your target and the host
Nov 17 00:25:09 <mskzo> Ok
Nov 17 00:25:16 <mskzo> connection is ready =)
Nov 17 00:25:43 <larsc> and then on the target run `gdbserver :12345 /path/to/your/executable`
Nov 17 00:25:59 <larsc> where 12345 is the port on which gdbserver will listen
Nov 17 00:26:12 <larsc> on the host run gdb
Nov 17 00:26:59 <mskzo> done
Nov 17 00:27:04 <mskzo> and now?
Nov 17 00:27:07 <larsc> and enter 'set solib-absolute-prefix /path/to/your/openwrt/checkout/staging_dir/arch'
Nov 17 00:27:21 <larsc> this will tell gdb were to find the libs for the target
Nov 17 00:27:56 <larsc> then do a 'file /path/to/your/openwrt/build/dir/your_exectuable'
Nov 17 00:28:21 <larsc> and then 'target remote target_ip:12345'
Nov 17 00:28:58 <larsc> now you should have a connection and should be able to debug the app
Nov 17 00:29:12 <larsc> enter continue to run it
Nov 17 00:29:43 <mskzo> sorry, a 'file /path/to/your/openwrt/build/dir/your_exectuable' i dont understand
Nov 17 00:29:50 <mskzo> my executable in buildroot dir?
Nov 17 00:30:39 <larsc> yes
Nov 17 00:31:32 <CIA-8> lars * r13250 /packages/Xorg/lib/pixman/Makefile: Cleanup pixman package.
Nov 17 00:31:55 <mskzo> oh i did a make clean :|
Nov 17 00:32:44 <mskzo> building
Nov 17 00:53:20 <mskzo> larsc, is possible run debuger without executable in desktop?
Nov 17 00:58:43 <larsc> hm... not sure, but i guess not
Nov 17 00:59:05 <mskzo> ok
Nov 17 01:04:13 <larsc> what is your target plattform?
Nov 17 01:04:21 <mskzo> i386
Nov 17 01:04:34 <mskzo> is a WRAP.2E
Nov 17 01:06:48 <larsc> ah... i386 makes debugging easier
Nov 17 01:07:39 <larsc> you've build the image and not cleaned the buildir?
Nov 17 01:08:31 <CIA-8> lars * r13251 /packages/Xorg/lib/libXi/ (. Makefile): Add libXi.
Nov 17 01:09:39 <mskzo> larsc,  yer
Nov 17 01:09:41 <mskzo> larsc,  yes
Nov 17 01:10:22 <mskzo> but now is doing a clean build
Nov 17 01:11:34 <larsc> when the build is finished you should have a folder called build_dir/i386/root-*
Nov 17 01:12:49 <larsc> you can chroot there and then run gdbserver from there
Nov 17 01:13:02 <mskzo> Ok
Nov 17 01:13:25 <mskzo> nice =)
Nov 17 01:30:37 <mskzo> larsc, done
Nov 17 01:32:39 <larsc> what exactly?
Nov 17 01:32:50 <larsc> building?
Nov 17 01:33:24 <mskzo> Remote debugging using localhost:12345
Nov 17 01:33:24 <mskzo> warning: Unable to find dynamic linker breakpoint function.
Nov 17 01:33:24 <mskzo> GDB will be unable to debug shared library initializers
Nov 17 01:33:24 <mskzo> and track explicitly loaded dynamic code.
Nov 17 01:33:24 <mskzo> 0xb8049910 in ??
Nov 17 01:33:28 <mskzo> built
Nov 17 01:33:38 <larsc> hm
Nov 17 01:34:06 <larsc> bt show what?
Nov 17 01:34:14 <mskzo> how can i debug now?
Nov 17 01:34:17 <mskzo> trying
Nov 17 01:35:16 <mskzo> no stack
Nov 17 01:35:50 <mskzo> (gdb) bt
Nov 17 01:35:50 <mskzo> #0  0xb8061910 in ?? ()
Nov 17 01:40:24 <mskzo> larsc, and now?
Nov 17 01:41:06 <larsc> hm, i think last time i used gdb on uclibc i added -g3 in the uclibcs config in toolchain/uClibc/config/
Nov 17 01:42:26 <mskzo> :\
Nov 17 01:42:34 <mskzo> so i need to add ?
Nov 17 01:42:37 <CIA-8> jow * r13252 /packages/utils/lcd4linux/Makefile: bumpt lcd4linux to r899
Nov 17 01:47:50 <mskzo> larsc, i need to add -g3 in my config to use gdb in uclibc?
Nov 17 01:48:39 <larsc> probaly yes
Nov 17 01:48:51 <mskzo> Ok
Nov 17 01:48:54 <mskzo> i cant do debug
Nov 17 01:49:27 <larsc> i'm building gpsd now, lets see if i can reproduce the crash
Nov 17 01:50:06 <mskzo> larsc, can i send my simple code?
Nov 17 01:50:10 <mskzo> simple package
Nov 17 01:50:11 <mskzo> ?
Nov 17 01:50:57 <larsc> sure
Nov 17 01:51:07 <larsc> will make it easier for me
Nov 17 01:54:57 <mskzo> www.mskzo.org/spcri_client.tar.gz
Nov 17 01:59:57 <larsc> building
Nov 17 02:04:37 <larsc> nope you dont have to add -g3 to uclibc
Nov 17 02:05:12 <larsc> it should be fine if 'building with debug symbols' was enabled in menuconfig when building the toolchain
Nov 17 02:17:12 <mskzo> Ok
Nov 17 02:18:35 <mskzo> where are debug symbols option?
Nov 17 02:20:26 <mskzo> is busybox with extra Debugging symbols?
Nov 17 02:20:54 <larsc> "Build options" -> "Enable Debugging"
Nov 17 02:21:00 <mskzo> ah ok
Nov 17 02:21:01 <mskzo> is enable
Nov 17 02:21:22 <larsc> the crash is related to a endless recursion in uClibc
Nov 17 02:22:06 <mskzo> Ohh finally founded the problem!
Nov 17 02:22:44 <mskzo> so, how can i resolve that problem?
Nov 17 02:23:01 <larsc> i'm not sure how to fix it though. but http://rafb.net/p/HiT2YA18.html should at least avoid the segfault
Nov 17 02:24:15 <larsc> pthread_cond_broadcast calls .ptr_pthread_cond_broadcast and with it being pthread_cond_broadcast it runs in a endless recursion.
Nov 17 02:24:38 <mskzo> ok
Nov 17 02:25:08 <mskzo> so if i apply this patch can fix the problem?
Nov 17 02:26:03 <larsc> yes
Nov 17 02:29:21 <mskzo> usually what is the directory where i can find pthread patches?
Nov 17 02:30:39 <larsc> toolchain/uClibc/patches
Nov 17 02:31:20 <larsc> ha!
Nov 17 02:32:23 <larsc> after reading some more of the uclibc pthreads source if come to the conclusion that the real fix would be to link your application against pthreads
Nov 17 02:32:27 <larsc> and it does works
Nov 17 02:32:46 <mskzo> very nice
Nov 17 02:33:27 <mskzo> -lpthread fixes my big problem?
Nov 17 02:34:06 <larsc> yes
Nov 17 02:34:49 <mskzo> but this is a bug no?
Nov 17 02:35:00 <mskzo> of uclibc
Nov 17 02:35:32 <larsc> maybe
Nov 17 02:45:26 <mskzo> you are the best!
Nov 17 02:45:28 <mskzo> works!
Nov 17 02:45:33 <mskzo> it works!!
Nov 17 02:45:48 <mskzo> fantastic!
Nov 17 02:46:58 <mskzo> Now i can sleep quiet!
Nov 17 02:47:47 <mskzo> thanks!!
Nov 17 02:48:22 <larsc> np
**** ENDING LOGGING AT Mon Nov 17 02:59:57 2008