[00:31] <CIA-17> nbd * r4087 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/config: enable CRC32 in brcm-2.4 kernel config - no idea why it even worked without
[00:35] <[mbm]> crc32 was probably forced on by something else
[00:37] <nbd> but even after oldconfig it was disabled in the .config file
[00:37] <nbd> and the autogenerated header files said #undef CONFIG_CRC32
[00:37] <[mbm]> weird
[00:44] <CIA-17> nbd * r4088 /branches/buildroot-ng/openwrt/package/base-files/default/etc/functions.sh: don't rely on temporary variables in config() as they may get overwritten by config_cb()
[01:52] <CIA-17> nbd * r4089 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/001-bcm47xx.patch: add missing exports
[02:36] <CIA-17> nbd * r4090 /branches/buildroot-ng/openwrt/package/base-files/ (40 files in 28 dirs): add copyright headers to base-files scripts and config files
[02:36] <CIA-17> mbm * r4091 /branches/buildroot-ng/openwrt/ (90 files in 72 dirs): credit where credit is due
[02:39] <CIA-17> nbd * r4092 /branches/buildroot-ng/openwrt/package/broadcom-wl/src/kmod/patchtable.pl: add missing copyright
[02:40] <CIA-17> nbd * r4093 /branches/buildroot-ng/openwrt/package/switch/src/gpio.h: add missing copyright
[02:44] <CIA-17> nbd * r4094 /branches/buildroot-ng/openwrt/toolchain/ (16 files in 13 dirs): add missing copyright
[02:52] <CIA-17> nbd * r4095 /branches/buildroot-ng/openwrt/package/busybox/patches/ (7 files): add copyright headers to busybox patches
[03:09] <common> nbd?
[03:10] <common> habt ihr irgendnen feedback von "fon" erhalten, offensichtlich nutzen die openwrt als firmware
[03:11] <nbd> sure. some of us are working for them occasionally
[03:21] <common> is there any way to debug the linksys firmware itself?
[03:21] <common> i think i found a remote flaw in the upnp service
[03:22] <common> i can shot the daemon
[03:23] <nbd> you could patch in a telnet daemon or something like that
[03:25] <nbd> wouldn't surprise me if the upnp daemon contains lots of security holes
[03:26] <common> i'd need gdb & gdbserver, but i did not even manage to get gdbserver working with openwrt
[03:26] <nbd> i have no clue about gdb
[03:29] <nbd> anyway... the source to upnpd is open, right?
[03:29] <nbd> that would mean you could do 'printf() debugging' :)
[03:32] <common> no
[03:32] <common> afaik the upnpd is closed source?
[03:32] <nbd> did you check?
[03:32] <common> i was told
[03:33] <nbd> i was told it was open source
[03:34] <common> hrm
[03:34] <common> i'll download right away
[03:37] <common> wget http://download.fon.com/heartbeat.php?mac=$MAC
[03:37] <common> wow
[03:38] <common> really advanced checking
[03:47] <common> hr
[03:47] <common> the fon docs state you can insert a coax cable into a wrt54
[17:37] <h3sp4wn> Does the current buildroot-ng sdk not have the ability to support Build/InstallDev ?
[17:41] <nbd> yes
[17:41] <nbd> it should have the ability
[17:42] <nbd> but i just noticed, it might not have been updated for the latest makefile moves
[17:42] <nbd> right... need to fix that
[17:42] <nbd> oh, wait
[17:42] <nbd> i already did
[17:44] <h3sp4wn> If a package compiles with the sdk it should just drop in and make config or make menuconfig should find it or is there something else I need to do ?
[17:46] <[mbm]> I think we completely dropped support for the old package makefiles
[17:46] <[mbm]> but if you have a buildroot-ng package then yes, you just place it in the packages directry
[17:47] <h3sp4wn> I am using the buildroot-ng generated sdk - It worked this time (but I have to use make config instead of make menuconfig or I get make: *** [scripts/config/mconf] Error 2
[17:49] <[mbm]> so use the 'make V=99' and figure out why mconf didn't build
[17:50] <h3sp4wn> mconf.o: In function `conf':
[17:50] <h3sp4wn> mconf.c:(.text+0x153d): undefined reference to `conf_reset'
[17:50] <h3sp4wn> (Thats just from make menuconfig not my package though)
[17:54] <nbd> run make -C scripts/config clean
[17:54] <nbd> and then try menuconfig again
[17:54] <nbd> it has something to do with a change that i made to menuconfig
[17:56] <h3sp4wn> #include "lkc_proto.h" (instead of #include "lkc.h ?)
[17:58] <h3sp4wn> That seems to be working thanks
[18:09] <florian> hallo
[18:15] <Kaloz> http://www.marvell.com/press/pressNewsDisplay.do?releaseID=581
[18:19] <florian> Kaloz: so we had a chance to have an opensource support; now it will probably be full proprietary :(
[18:23] <nbd> why?
[18:29] <florian> because I don't think marvell is opensource friendly regarding from what they have done in the past with wrt54gp2 and sipura boxes for instance
[18:31] <florian> but maybe I am wrong
[18:50] <florian> nbd: do we use the recvudp utility ?
[18:57] <nbd> it's used on the pc in case you need failsafe access
[19:14] <florian> ok
[19:14] <florian> I meant is there a procedure, are the users aware they can use it ?
[19:15] <nbd> i don't know if its's documneted
[19:24] <florian> ok
[19:25] <florian> in fact, as I mentionned, I will make an openwrt presentation on july the 6th
[19:25] <nbd> great
[19:25] <nbd> florian: https://snr.freifunk.net/trac/wcc/ -- can you come?
[19:26] <florian> nbd: unfortunately no, I am working this summer, and they accepted to give me the day for the presentation
[19:26] <nbd> ok
[19:27] <florian> such a pity
[19:27] <nbd> maybe we can all meet on 23c3
[19:27] <florian> for the presentation I would like to talk about several things hence it lasts between 1h and 1h3à
[19:27] <florian> 1h30 sorry
[19:27] <nbd> sounds good
[19:27] <florian> nbd: sure I will try to come this time ;)
[19:27] <florian> so, basically here is what I would like to talk about :
[19:28] <florian> - short OpenWrt history : origin, goals
[19:28] <florian> - porting new devices : legal issues, were to find code, how to test and integrate
[19:28] <florian> - making packages
[19:28] <florian> - a demo probably
[19:29] <nbd> i'm sure you can fill the time with all these topics
[19:30] <florian> and maybe if I have time, I would like to talk about general legal concerns and considerations about manufacturers and their position with opensource
[19:30] <nbd> yeah
[19:30] <florian> what I really don't know is if I should have a strong technical presentation (can be boring) or just introduce the buildroot system
[19:31] <florian> I also don't know if I should talk about whiterussian as well
[19:31] <nbd> i think you should have a mix of both
[19:32] <nbd> too many details will make it boring (i fell for that trap with my last presentation) and too much noob-only information annoys technical people
[19:33] <florian> right, that's the difficulty with such project where there is a mixture of technical knowledges and general concerns
[19:34] <florian_2> your presentation was great to me, very well explained, but I agree, putting a makefile template was not probably the best thing to do
[19:35] <florian_2> was happens here :(
[19:37] <nbd> no idea
[19:37] <nbd> weird
[19:37] <florian___> I think the school server I run irssi on was lagging or getting disconnected briefly
[19:39] <florian___> I will try to make an english version of the presentation so that you can read it before
[22:44] <florian___> are there any tools able to measure the kernel side load average ?
[22:48] <florian___> or maybe it is irrelevant ?
[23:11] <dragorn> should show up under system load
[23:13] Action: [mbm] plans on writing some buildroot-ng docs in the next week
[23:13] <nbd> great
[23:14] Action: h3sp4wn wishes he had waited a week to start trying to understand it (but its easier than the old system)
[23:15] <malbon> [mbm]: yay. docs. :)
[23:15] <[mbm]> h3sp4wn: looking at a package makefile things are pretty self explainitory; I don't expect people to be able to understand some of the nasty stuff that makes it all work
[23:15] <[mbm]> ...
[23:16] <[mbm]> oh, and we also need content for the openwrt.org news page; articles about using openwrt or possibly thoughts on the various hardware platforms
[23:16] <[mbm]> any volunteers?
[23:17] <florian___> [mbm], nbd I will try to add generic init.d script for packages which do not have one, and in the case it's relevant to have one, does it make sense ?
[23:17] <[mbm]> sounds good
[23:18] <h3sp4wn> mbm: Are you doing to explain the nasty stuff ? (I think I basically understand it) - also from a clean checkout the 2.4.32 header fail unless --wildcards is added to the tar command (bzcat $(DL_DIR)/$(PKG_SOURCE) | tar -C $(TOOL_BUILD_DIR) $(TAR_OPTIONS) - \ (I don't know whether it needs to be in TAR_OPTIONS or just prefixed to that one command because everything else works
[23:18] <h3sp4wn> mbm: (basically understand it other than the nasty stuff I mean)
[23:22] <florian___> buildroot-ng is really clean compared to uclibc's buildroot for instance
[23:22] <[mbm]> h3sp4wn: I might explain some of the nasty stuff but I don't want to start with that and have people think they're required to know it
[23:23] <[mbm]> it really should be a black box style -- write according to these guidelines and everything will work magically
[23:24] <h3sp4wn> Apart from if the Makefile has tests which Makes it fail if cross-compiling (as alot of things do in my experience)
[23:25] <[mbm]> for some of those there's nothing you can do other than write a patch that fixes the bug or creates a makefile, skipping the bug
[23:37] <florian___> I also think about adding a qemu target for ppl who want to test their filesystem ?
[23:40] <florian___> not sure it's a good idea, but the idea is basically to be able to test for instance how daemons are starting ... if the changes you made in base-files are working
[23:40] <florian___> the ideal qemu target would require not recompiling of the toolchain
[23:40] <[mbm]> the qemu thing looked interesting
[23:40] <florian___> but it needs another kernel config most of the time
[23:41] <[mbm]> haven't had a chance to try it yet
[23:42] <florian___> well at work I am working on xscale pxa255 and I added a qemu target because it's far convenient for testing my filesystem changes before burning the flash ;)
[23:42] <florian___> fortunately they use 2 different kernel sources
[23:44] <florian___> and there is a qemu-arm emulation
[23:44] <florian___> I think it was with malbon we discussed about adding build system dependencies, the most famous one is sablevm
[23:45] <florian___> it requires jikes on your build system
[23:45] <h3sp4wn> florian: I never knew that you need procmail though to compile fetchmail
[23:46] <florian___> neither did I actually
[23:46] <florian___> that's why we could probably need a kind of configure script before running make in top of the openwrt buildroot
[23:47] <florian___> I am not fond of autoconf, it's mostly E.T language
[00:00] --- Wed Jun 28 2006
[00:35] <[mbm]> crc32 was probably forced on by something else
[00:37] <nbd> but even after oldconfig it was disabled in the .config file
[00:37] <nbd> and the autogenerated header files said #undef CONFIG_CRC32
[00:37] <[mbm]> weird
[00:44] <CIA-17> nbd * r4088 /branches/buildroot-ng/openwrt/package/base-files/default/etc/functions.sh: don't rely on temporary variables in config() as they may get overwritten by config_cb()
[01:52] <CIA-17> nbd * r4089 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/001-bcm47xx.patch: add missing exports
[02:36] <CIA-17> nbd * r4090 /branches/buildroot-ng/openwrt/package/base-files/ (40 files in 28 dirs): add copyright headers to base-files scripts and config files
[02:36] <CIA-17> mbm * r4091 /branches/buildroot-ng/openwrt/ (90 files in 72 dirs): credit where credit is due
[02:39] <CIA-17> nbd * r4092 /branches/buildroot-ng/openwrt/package/broadcom-wl/src/kmod/patchtable.pl: add missing copyright
[02:40] <CIA-17> nbd * r4093 /branches/buildroot-ng/openwrt/package/switch/src/gpio.h: add missing copyright
[02:44] <CIA-17> nbd * r4094 /branches/buildroot-ng/openwrt/toolchain/ (16 files in 13 dirs): add missing copyright
[02:52] <CIA-17> nbd * r4095 /branches/buildroot-ng/openwrt/package/busybox/patches/ (7 files): add copyright headers to busybox patches
[03:09] <common> nbd?
[03:10] <common> habt ihr irgendnen feedback von "fon" erhalten, offensichtlich nutzen die openwrt als firmware
[03:11] <nbd> sure. some of us are working for them occasionally
[03:21] <common> is there any way to debug the linksys firmware itself?
[03:21] <common> i think i found a remote flaw in the upnp service
[03:22] <common> i can shot the daemon
[03:23] <nbd> you could patch in a telnet daemon or something like that
[03:25] <nbd> wouldn't surprise me if the upnp daemon contains lots of security holes
[03:26] <common> i'd need gdb & gdbserver, but i did not even manage to get gdbserver working with openwrt
[03:26] <nbd> i have no clue about gdb
[03:29] <nbd> anyway... the source to upnpd is open, right?
[03:29] <nbd> that would mean you could do 'printf() debugging' :)
[03:32] <common> no
[03:32] <common> afaik the upnpd is closed source?
[03:32] <nbd> did you check?
[03:32] <common> i was told
[03:33] <nbd> i was told it was open source
[03:34] <common> hrm
[03:34] <common> i'll download right away
[03:37] <common> wget http://download.fon.com/heartbeat.php?mac=$MAC
[03:37] <common> wow
[03:38] <common> really advanced checking
[03:47] <common> hr
[03:47] <common> the fon docs state you can insert a coax cable into a wrt54
[17:37] <h3sp4wn> Does the current buildroot-ng sdk not have the ability to support Build/InstallDev ?
[17:41] <nbd> yes
[17:41] <nbd> it should have the ability
[17:42] <nbd> but i just noticed, it might not have been updated for the latest makefile moves
[17:42] <nbd> right... need to fix that
[17:42] <nbd> oh, wait
[17:42] <nbd> i already did
[17:44] <h3sp4wn> If a package compiles with the sdk it should just drop in and make config or make menuconfig should find it or is there something else I need to do ?
[17:46] <[mbm]> I think we completely dropped support for the old package makefiles
[17:46] <[mbm]> but if you have a buildroot-ng package then yes, you just place it in the packages directry
[17:47] <h3sp4wn> I am using the buildroot-ng generated sdk - It worked this time (but I have to use make config instead of make menuconfig or I get make: *** [scripts/config/mconf] Error 2
[17:49] <[mbm]> so use the 'make V=99' and figure out why mconf didn't build
[17:50] <h3sp4wn> mconf.o: In function `conf':
[17:50] <h3sp4wn> mconf.c:(.text+0x153d): undefined reference to `conf_reset'
[17:50] <h3sp4wn> (Thats just from make menuconfig not my package though)
[17:54] <nbd> run make -C scripts/config clean
[17:54] <nbd> and then try menuconfig again
[17:54] <nbd> it has something to do with a change that i made to menuconfig
[17:56] <h3sp4wn> #include "lkc_proto.h" (instead of #include "lkc.h ?)
[17:58] <h3sp4wn> That seems to be working thanks
[18:09] <florian> hallo
[18:15] <Kaloz> http://www.marvell.com/press/pressNewsDisplay.do?releaseID=581
[18:19] <florian> Kaloz: so we had a chance to have an opensource support; now it will probably be full proprietary :(
[18:23] <nbd> why?
[18:29] <florian> because I don't think marvell is opensource friendly regarding from what they have done in the past with wrt54gp2 and sipura boxes for instance
[18:31] <florian> but maybe I am wrong
[18:50] <florian> nbd: do we use the recvudp utility ?
[18:57] <nbd> it's used on the pc in case you need failsafe access
[19:14] <florian> ok
[19:14] <florian> I meant is there a procedure, are the users aware they can use it ?
[19:15] <nbd> i don't know if its's documneted
[19:24] <florian> ok
[19:25] <florian> in fact, as I mentionned, I will make an openwrt presentation on july the 6th
[19:25] <nbd> great
[19:25] <nbd> florian: https://snr.freifunk.net/trac/wcc/ -- can you come?
[19:26] <florian> nbd: unfortunately no, I am working this summer, and they accepted to give me the day for the presentation
[19:26] <nbd> ok
[19:27] <florian> such a pity
[19:27] <nbd> maybe we can all meet on 23c3
[19:27] <florian> for the presentation I would like to talk about several things hence it lasts between 1h and 1h3à
[19:27] <florian> 1h30 sorry
[19:27] <nbd> sounds good
[19:27] <florian> nbd: sure I will try to come this time ;)
[19:27] <florian> so, basically here is what I would like to talk about :
[19:28] <florian> - short OpenWrt history : origin, goals
[19:28] <florian> - porting new devices : legal issues, were to find code, how to test and integrate
[19:28] <florian> - making packages
[19:28] <florian> - a demo probably
[19:29] <nbd> i'm sure you can fill the time with all these topics
[19:30] <florian> and maybe if I have time, I would like to talk about general legal concerns and considerations about manufacturers and their position with opensource
[19:30] <nbd> yeah
[19:30] <florian> what I really don't know is if I should have a strong technical presentation (can be boring) or just introduce the buildroot system
[19:31] <florian> I also don't know if I should talk about whiterussian as well
[19:31] <nbd> i think you should have a mix of both
[19:32] <nbd> too many details will make it boring (i fell for that trap with my last presentation) and too much noob-only information annoys technical people
[19:33] <florian> right, that's the difficulty with such project where there is a mixture of technical knowledges and general concerns
[19:34] <florian_2> your presentation was great to me, very well explained, but I agree, putting a makefile template was not probably the best thing to do
[19:35] <florian_2> was happens here :(
[19:37] <nbd> no idea
[19:37] <nbd> weird
[19:37] <florian___> I think the school server I run irssi on was lagging or getting disconnected briefly
[19:39] <florian___> I will try to make an english version of the presentation so that you can read it before
[22:44] <florian___> are there any tools able to measure the kernel side load average ?
[22:48] <florian___> or maybe it is irrelevant ?
[23:11] <dragorn> should show up under system load
[23:13] Action: [mbm] plans on writing some buildroot-ng docs in the next week
[23:13] <nbd> great
[23:14] Action: h3sp4wn wishes he had waited a week to start trying to understand it (but its easier than the old system)
[23:15] <malbon> [mbm]: yay. docs. :)
[23:15] <[mbm]> h3sp4wn: looking at a package makefile things are pretty self explainitory; I don't expect people to be able to understand some of the nasty stuff that makes it all work
[23:15] <[mbm]> ...
[23:16] <[mbm]> oh, and we also need content for the openwrt.org news page; articles about using openwrt or possibly thoughts on the various hardware platforms
[23:16] <[mbm]> any volunteers?
[23:17] <florian___> [mbm], nbd I will try to add generic init.d script for packages which do not have one, and in the case it's relevant to have one, does it make sense ?
[23:17] <[mbm]> sounds good
[23:18] <h3sp4wn> mbm: Are you doing to explain the nasty stuff ? (I think I basically understand it) - also from a clean checkout the 2.4.32 header fail unless --wildcards is added to the tar command (bzcat $(DL_DIR)/$(PKG_SOURCE) | tar -C $(TOOL_BUILD_DIR) $(TAR_OPTIONS) - \ (I don't know whether it needs to be in TAR_OPTIONS or just prefixed to that one command because everything else works
[23:18] <h3sp4wn> mbm: (basically understand it other than the nasty stuff I mean)
[23:22] <florian___> buildroot-ng is really clean compared to uclibc's buildroot for instance
[23:22] <[mbm]> h3sp4wn: I might explain some of the nasty stuff but I don't want to start with that and have people think they're required to know it
[23:23] <[mbm]> it really should be a black box style -- write according to these guidelines and everything will work magically
[23:24] <h3sp4wn> Apart from if the Makefile has tests which Makes it fail if cross-compiling (as alot of things do in my experience)
[23:25] <[mbm]> for some of those there's nothing you can do other than write a patch that fixes the bug or creates a makefile, skipping the bug
[23:37] <florian___> I also think about adding a qemu target for ppl who want to test their filesystem ?
[23:40] <florian___> not sure it's a good idea, but the idea is basically to be able to test for instance how daemons are starting ... if the changes you made in base-files are working
[23:40] <florian___> the ideal qemu target would require not recompiling of the toolchain
[23:40] <[mbm]> the qemu thing looked interesting
[23:40] <florian___> but it needs another kernel config most of the time
[23:41] <[mbm]> haven't had a chance to try it yet
[23:42] <florian___> well at work I am working on xscale pxa255 and I added a qemu target because it's far convenient for testing my filesystem changes before burning the flash ;)
[23:42] <florian___> fortunately they use 2 different kernel sources
[23:44] <florian___> and there is a qemu-arm emulation
[23:44] <florian___> I think it was with malbon we discussed about adding build system dependencies, the most famous one is sablevm
[23:45] <florian___> it requires jikes on your build system
[23:45] <h3sp4wn> florian: I never knew that you need procmail though to compile fetchmail
[23:46] <florian___> neither did I actually
[23:46] <florian___> that's why we could probably need a kind of configure script before running make in top of the openwrt buildroot
[23:47] <florian___> I am not fond of autoconf, it's mostly E.T language
[00:00] --- Wed Jun 28 2006