[01:13] <CIA-17> nbd * r4617 /branches/buildroot-ng/openwrt/Makefile: fix up TMPDIR for initial prereq check to avoid creating a bogus build_ directory
[01:15] <CIA-17> nbd * r4618 /branches/buildroot-ng/openwrt/package/busybox/patches/911-ipkg.patch: revert [4599], the bug was fixed elsewhere
[01:57] <nbd> hmm... anyone ever thought about making a ports-like system for linux distros out of buildroot-ng?
[01:58] <nbd> i mean it would be really easy to make it build .deb, .tgz, .rpm or whatever the linux distro uses
[01:58] <Bartman007> you could call it portage-ng and piss off gentoo :-)
[01:58] <nbd> haha, yeah
[01:58] <nbd> and i could add some of the nice dependency checking from the debian build stuff
[01:59] <nbd> while it's nothing i'd start now (too much work when i have other things to do), i actually do think that making something like that would be a nice idea
[02:00] <Bartman007> I personally like portage better than apt, but I wouldn't want to maintain emerge.py it's ~3500 lines
[02:00] <nbd> i like to have binary packages that Just Work [tm]
[02:01] <Bartman007> on my own systems, I'm a fan of useflags, but on the servers at work, I agree.
[02:02] <BrainSlayer> we all would like to have that
[02:02] <nbd> and it would be cool to have something that works with all major linux distributions in their native packaging format
[02:04] <Bartman007> heh, autopackage is trying to achieve something similar, but they are an abomination.
[02:04] <nbd> yeah, autopackage is *nasty*
[02:05] <Bartman007> and it's followers are more rabid than gentoo, freebsd and openbsd combined.
[02:05] <nbd> haha
[07:16] <groz> kaloz ping
[07:16] <Kaloz> groz: pong
[07:16] <groz> is there a problem with the svn tonite ?
[07:16] <Kaloz> h3sp4wn: i was too busy yesterday, i make the image for you today
[07:17] <Kaloz> groz: afaik it's all okay
[07:18] <Bartman007> Kaloz: could I grab the image too? (I still have to hack this usb device driver, failed in my first attempt)
[07:19] <Kaloz> Bartman007: sure, i'll let you know when it's done
[07:21] <Bartman007> btw, does any of you have experience adding a new usb device id to a device driver?
[07:22] <Kaloz> should be quite simple
[07:22] <Kaloz> which driver you want to modify?
[07:22] <Bartman007> the pl2303 driver
[07:24] <Kaloz> Bartman007: https://dev.openwrt.org/changeset/2526
[07:24] <Bartman007> damnit, I just spotted my mistake
[07:24] <Bartman007> lol...
[07:25] <Kaloz> erm, https://dev.openwrt.org/browser/trunk/openwrt/target/linux/linux-2.4/patches/generic/202-pl2303_backport.patch?rev=2526
[07:25] <Kaloz> :)
[07:25] <Kaloz> that should give you an idea
[07:25] <Bartman007> Kaloz: it helps when I don't define the same string twice
[07:25] <Kaloz> :))
[07:25] <Kaloz> yeah, that can come handy
[07:26] <Bartman007> I should probably compile usb serial drivers as modules anyway since it's likely I'll be modifying them again as I get more devices
[10:31] <[florian]> any reasons why ipsec-tools is 0.6.4 and not 0.6. ?
[10:31] <[florian]> 6
[10:35] <nbd> because it hasn't been updated yet :)
[10:37] <murb> is iproute2 in -ng known to not build? or have i broken it lotcally?
[10:39] <malbon> murb: it builds for me.
[10:39] <malbon> mind you that's no guarentee :)
[11:07] <nbd> murb: show me the error, please
[11:13] <malbon> hmm, defconfig is causing me grief, it ignores my settings. :(
[11:20] <malbon> argh! I hate it when it turns out to be my fault.....
[11:48] <[florian]> by the way tc-ng is how
[11:48] <[florian]> out sorry
[11:50] <CIA-17> florian * r4619 /branches/buildroot-ng/openwrt/package/ipsec-tools/Makefile: Add --enable-natt and --enable-dpd configuration options, closes #700
[11:53] <[florian]> nbd: regarding ticket #571, did you merge the fix you proposed ?
[12:14] <nbd> [florian]: not yet. i lost the patch and i don't even know if it worked
[12:15] <nbd> ah, found the patch again
[12:15] <nbd> if somebody tells me it works, i will merge it :)
[12:38] <murb> nbd: ip/tc depend on iproute2 which doens't show up.
[12:39] <nbd> ah, right
[12:39] <nbd> i think i fixed that already
[12:39] <nbd> no, wait
[12:40] <nbd> i didn't
[12:40] <CIA-17> nbd * r4621 /branches/buildroot-ng/openwrt/package/iproute2/Makefile: remove bogus dependency line
[12:40] <nbd> fixed now
[13:38] <[florian]> nbd: actually, I hoped the guy would have a try at your patch
[13:38] <nbd> if you see any forum threads or know any other guy with the same problem, please tell him to install the package referenced in the ticket
[13:42] <[florian]> sure
[13:46] <CIA-17> florian * r4622 /branches/buildroot-ng/openwrt/package/ipset/Makefile: Set KERNEL_DIR to avoid inclusion of host-side linux sources, closes #702
[13:49] <[florian]> nbd: do you think we can close #694 as invalid ?
[13:50] <nbd> i don't know. it might be a valid bug
[13:50] <nbd> and i think i've seen it before
[13:51] <[florian]> I wonder if the guy is not confusing non authoritative DNS answer with a classic answer
[13:52] <[florian]> I think I have always been seeing this message under winxp
[13:52] <nbd> oh, wait
[13:52] <nbd> misread the ticket
[13:52] <nbd> i thought in one case it didn't give a valid answer at all
[13:52] <nbd> well, in this case we can mark it as invalid
[13:53] <nbd> or wontfix
[13:59] <CIA-17> nbd * r4623 /branches/whiterussian/openwrt/package/wificonf/wificonf.c: disable watchdog functionality (except for distance) in ad-hoc mode (fixes #373)
[14:03] <[florian]> I just discovered an amazingly well working webmail : roundcube
[14:06] <CIA-17> florian * r4624 /branches/ (2 files in 2 dirs): Add maxfail option to avoid pptp giving up, closes #663
[14:07] <[florian]> motorola devices represent a lot of tickets actually :(
[14:14] <CIA-17> nbd * r4625 /branches/whiterussian/openwrt/package/webif/files/usr/lib/webif/webif.sh: allow numbers in webif scripts (#619)
[14:17] <CIA-17> florian * r4626 /branches/whiterussian/openwrt/package/webif/files/usr/lib/webif/lang/hu/ (. common.txt): Add hungarian translation, closes #642
[14:19] <CIA-17> nbd * r4627 /branches/ (2 files in 2 dirs): enable mknod by default (#490) - some drivers don't register with devfs properly
[14:24] <CIA-17> nbd * r4628 /branches/whiterussian/openwrt/package/base-files/default/etc/functions.sh: fix typo (#693)
[14:26] <CIA-17> florian * r4629 /packages/libs/libjpeg/ (Makefile patches/200-crop.patch): Add crop features, and build jpeg tools, closes #665
[14:40] <CIA-17> florian * r4630 /branches/buildroot-ng/openwrt/ (include/modules-2.6.mk package/kernel/modules.mk): Port kmod-gre to the new module layout
[14:40] <CIA-17> florian * r4631 /packages/net/pptpd/Makefile: Add missing dependencies for pptpd, closes #656
[16:10] <[florian]> I never figured out openswan was such as painful software :(
[16:21] <CIA-17> florian * r4632 /branches/ (7 files in 6 dirs): Upgrade to openswan-2.4.6, fix showhostkey script (closes #444), *should* fix #647, #671
[16:24] <[florian]> do we have a special target that could allow us recompiling every package that depends on foo ?
[18:44] <[florian]> re
[18:49] <[florian]> humm users are boring with stupid questions and no faq reading
[18:52] <murb> [florian]: well if you have lots of users there will always be a percentage who can write but can't read.
[18:53] <[florian]> sure :)
[18:57] <[florian]> this raises the questions of how they use a computer by the way ;)
[18:58] <Bartman007> not really. I've seen many people who can use a computer, but don't know how to use it properly
[18:59] <thepeople_work> bartman007: it should read *most people*
[18:59] <Bartman007> double-clicking weblinks, confusing logging onto the computer and logging into their webmail, etc....
[18:59] <thepeople_work> ohh no my paypal account has been suspended
[19:00] <[florian]> ah Bartman007 this is not the same thing ? ;)
[19:00] <thepeople_work> i better follow the link and update it
[19:00] <thepeople_work> wait do I even have a paypal account?
[19:02] <[florian]> lol
[19:03] <[florian]> this is probably spam ?
[19:45] <h3sp4wn> Kaloz: If the images are ready than I can try them
[21:25] <nbd> [florian]: ping
[21:28] <[florian]> nbd: pong
[21:29] <nbd> have you looked into the new kernel packaging system?
[21:29] <nbd> because i could use some help with porting over all the old kmod-* packaging stuff
[21:29] <[florian]> nbd: a bit yes, did I make a mistake with the recent kmod-gre ?
[21:30] <nbd> ah, didn't check the log and didn't notice you already added some stuff
[21:30] <nbd> forgot to svn update :)
[21:30] <nbd> no, gre looks fine
[21:31] <[florian]> ok :)
[21:32] <[florian]> by the way, I have just added rssi to wlcompat, I should actually to the list of private ioctl's right ?
[21:34] <nbd> yeah
[21:34] <[florian]> maybe you have already done it ?
[21:34] <nbd> done what?
[21:34] <[florian]> adding rssi and assoclist to wlcompat N
[21:34] <[florian]> ?
[21:35] <nbd> no
[21:35] <[florian]> ok
[21:36] <[florian]> regarding kernel modules, there is not much work to do on it actually
[21:37] <nbd> yeah, just translating all the stuff from modules-2.*.mk and removing it from target/linux/Config.in afterwards
[21:37] <[florian]> I will port kmod-tun to the new layout right now, because it is a dependency for several user-space packages
[21:38] <Bartman007> [florian]: thanks for doing that, that's one of my issues rightnow.
[21:44] <CIA-17> florian * r4633 /branches/buildroot-ng/openwrt/ (3 files in 3 dirs): Port kmod-tun to the new module layout
[21:55] <[florian]> nbd: actually the whole wl binary could be easily rewritten with the help of wlioctl.h right ?
[21:55] <nbd> not all of it is documented, but yes
[21:55] <nbd> but there's no need to rewrite wl
[21:56] <nbd> wlc is better
[21:56] <[florian]> sure
[21:56] <[florian]> what is the goal of having both wlcompat and wlc ?
[21:56] <[florian]> it is just in the case we want to use the iwlib right ?
[21:56] <nbd> yeah, and wireless-tools is a convenient interface for some parts
[21:57] <[florian]> that's right
[21:58] <[florian]> wl could probably be spied to get the missing ioctls ?
[21:59] <[florian]> but it's really no use now with bcm43xx
[22:01] <Bartman007> so is bcm43xx to a point that it can be used in mastermode?
[22:01] <[florian]> not yet I think
[22:01] <[florian]> all the more as they are working on v4 specs
[22:02] <h3sp4wn> can the dscape version run in master mode (on x86 even)
[22:03] <[florian]> I think so
[22:04] <[florian]> from what I have read on the ml, looks like some people are running it
[22:10] <[florian]> nbd: I added rssi to wlcompat, but it does not seem to work : http://pastebin.ca/145024
[22:10] <[florian]> any hints ?
[22:11] <nbd> what happens if you try it?
[22:11] <[florian]> Interface doesn't accept private ioctl...
[22:11] <[florian]> get_rssi (8BEA): Invalid argument
[22:13] <nbd> does 'wl rssi' work?
[22:14] <[florian]> ah, I forgot that wl rssi expects the mac address as argument !
[22:15] <[florian]> so I should probably have a try at the assoclist rather than rssi for the moment
[22:17] <nbd> you can integrate rssi in the assoclist
[22:18] <[florian]> so that it reports the associated clients and their rssi
[22:19] <CIA-17> nbd * r4634 /branches/buildroot-ng/openwrt/target/linux/ (2 files in 2 dirs): don't try to create mtd partitions on bcm947xx when the trx header is not found - should fix #318
[22:20] <nbd> [florian]: yes
[22:36] <[florian]> nbd: when calling wl_ioctl, the value is then stored in the void * buf right ?
[22:39] <nbd> it's stored in a buffer with the size that you specify
[22:39] <[florian]> ok
[22:39] <[florian]> and where can I access it after ?
[22:39] <nbd> if you want to know what kind of buffer is used for the rssi stuff, use wlcompat-debug
[22:39] <[florian]> ok
[22:39] <nbd> since you supply the buffer, you get the result
[22:39] <nbd> it doesn't allocate something new
[22:40] <[florian]> ok
[23:03] <[florian]> nbd: why do I send and receive exactly the same things ? http://pastebin.ca/145134
[23:05] <nbd> which command is that from?
[23:06] <[florian]> iwpriv eth1 assoclist
[23:06] <nbd> are you using a static buffer?
[23:07] <nbd> or a kmalloc'd one?
[23:07] <[florian]> I the variable extra, like for other ioctls, should not I ?
[23:08] <nbd> returning binary data in an iwpriv call is not a good idea
[23:08] <nbd> and when you actually make the call, you should fill the buffer with 0
[23:10] <[florian]> ok, so I probably should use another buffer for testing purposes ?
[23:13] <nbd> if you use a static buffer, it will keep the data from the last request
[23:13] <nbd> which is probably what you're seeing thereb
[23:13] <[florian]> right
[23:13] <nbd> because at a first glance, it does look like an assoclist buffer
[23:13] <[florian]> I ran wl assoclist in order to figure out which size I should use as buffer size
[23:14] <nbd> doesn't wl usually use 8k for something like that?
[23:15] <[florian]> precisely
[23:41] <[florian]> nbd: sorry to bother you again, but was maybe wrong with this : http://pastebin.ca/145225
[23:46] <[florian]> what I don't understand is that using mutliple times symbol extra seems to work for other ioctls though used as int
[23:53] <nbd> you should make assoclist return a string
[00:00] --- Wed Aug 23 2006
[01:15] <CIA-17> nbd * r4618 /branches/buildroot-ng/openwrt/package/busybox/patches/911-ipkg.patch: revert [4599], the bug was fixed elsewhere
[01:57] <nbd> hmm... anyone ever thought about making a ports-like system for linux distros out of buildroot-ng?
[01:58] <nbd> i mean it would be really easy to make it build .deb, .tgz, .rpm or whatever the linux distro uses
[01:58] <Bartman007> you could call it portage-ng and piss off gentoo :-)
[01:58] <nbd> haha, yeah
[01:58] <nbd> and i could add some of the nice dependency checking from the debian build stuff
[01:59] <nbd> while it's nothing i'd start now (too much work when i have other things to do), i actually do think that making something like that would be a nice idea
[02:00] <Bartman007> I personally like portage better than apt, but I wouldn't want to maintain emerge.py it's ~3500 lines
[02:00] <nbd> i like to have binary packages that Just Work [tm]
[02:01] <Bartman007> on my own systems, I'm a fan of useflags, but on the servers at work, I agree.
[02:02] <BrainSlayer> we all would like to have that
[02:02] <nbd> and it would be cool to have something that works with all major linux distributions in their native packaging format
[02:04] <Bartman007> heh, autopackage is trying to achieve something similar, but they are an abomination.
[02:04] <nbd> yeah, autopackage is *nasty*
[02:05] <Bartman007> and it's followers are more rabid than gentoo, freebsd and openbsd combined.
[02:05] <nbd> haha
[07:16] <groz> kaloz ping
[07:16] <Kaloz> groz: pong
[07:16] <groz> is there a problem with the svn tonite ?
[07:16] <Kaloz> h3sp4wn: i was too busy yesterday, i make the image for you today
[07:17] <Kaloz> groz: afaik it's all okay
[07:18] <Bartman007> Kaloz: could I grab the image too? (I still have to hack this usb device driver, failed in my first attempt)
[07:19] <Kaloz> Bartman007: sure, i'll let you know when it's done
[07:21] <Bartman007> btw, does any of you have experience adding a new usb device id to a device driver?
[07:22] <Kaloz> should be quite simple
[07:22] <Kaloz> which driver you want to modify?
[07:22] <Bartman007> the pl2303 driver
[07:24] <Kaloz> Bartman007: https://dev.openwrt.org/changeset/2526
[07:24] <Bartman007> damnit, I just spotted my mistake
[07:24] <Bartman007> lol...
[07:25] <Kaloz> erm, https://dev.openwrt.org/browser/trunk/openwrt/target/linux/linux-2.4/patches/generic/202-pl2303_backport.patch?rev=2526
[07:25] <Kaloz> :)
[07:25] <Kaloz> that should give you an idea
[07:25] <Bartman007> Kaloz: it helps when I don't define the same string twice
[07:25] <Kaloz> :))
[07:25] <Kaloz> yeah, that can come handy
[07:26] <Bartman007> I should probably compile usb serial drivers as modules anyway since it's likely I'll be modifying them again as I get more devices
[10:31] <[florian]> any reasons why ipsec-tools is 0.6.4 and not 0.6. ?
[10:31] <[florian]> 6
[10:35] <nbd> because it hasn't been updated yet :)
[10:37] <murb> is iproute2 in -ng known to not build? or have i broken it lotcally?
[10:39] <malbon> murb: it builds for me.
[10:39] <malbon> mind you that's no guarentee :)
[11:07] <nbd> murb: show me the error, please
[11:13] <malbon> hmm, defconfig is causing me grief, it ignores my settings. :(
[11:20] <malbon> argh! I hate it when it turns out to be my fault.....
[11:48] <[florian]> by the way tc-ng is how
[11:48] <[florian]> out sorry
[11:50] <CIA-17> florian * r4619 /branches/buildroot-ng/openwrt/package/ipsec-tools/Makefile: Add --enable-natt and --enable-dpd configuration options, closes #700
[11:53] <[florian]> nbd: regarding ticket #571, did you merge the fix you proposed ?
[12:14] <nbd> [florian]: not yet. i lost the patch and i don't even know if it worked
[12:15] <nbd> ah, found the patch again
[12:15] <nbd> if somebody tells me it works, i will merge it :)
[12:38] <murb> nbd: ip/tc depend on iproute2 which doens't show up.
[12:39] <nbd> ah, right
[12:39] <nbd> i think i fixed that already
[12:39] <nbd> no, wait
[12:40] <nbd> i didn't
[12:40] <CIA-17> nbd * r4621 /branches/buildroot-ng/openwrt/package/iproute2/Makefile: remove bogus dependency line
[12:40] <nbd> fixed now
[13:38] <[florian]> nbd: actually, I hoped the guy would have a try at your patch
[13:38] <nbd> if you see any forum threads or know any other guy with the same problem, please tell him to install the package referenced in the ticket
[13:42] <[florian]> sure
[13:46] <CIA-17> florian * r4622 /branches/buildroot-ng/openwrt/package/ipset/Makefile: Set KERNEL_DIR to avoid inclusion of host-side linux sources, closes #702
[13:49] <[florian]> nbd: do you think we can close #694 as invalid ?
[13:50] <nbd> i don't know. it might be a valid bug
[13:50] <nbd> and i think i've seen it before
[13:51] <[florian]> I wonder if the guy is not confusing non authoritative DNS answer with a classic answer
[13:52] <[florian]> I think I have always been seeing this message under winxp
[13:52] <nbd> oh, wait
[13:52] <nbd> misread the ticket
[13:52] <nbd> i thought in one case it didn't give a valid answer at all
[13:52] <nbd> well, in this case we can mark it as invalid
[13:53] <nbd> or wontfix
[13:59] <CIA-17> nbd * r4623 /branches/whiterussian/openwrt/package/wificonf/wificonf.c: disable watchdog functionality (except for distance) in ad-hoc mode (fixes #373)
[14:03] <[florian]> I just discovered an amazingly well working webmail : roundcube
[14:06] <CIA-17> florian * r4624 /branches/ (2 files in 2 dirs): Add maxfail option to avoid pptp giving up, closes #663
[14:07] <[florian]> motorola devices represent a lot of tickets actually :(
[14:14] <CIA-17> nbd * r4625 /branches/whiterussian/openwrt/package/webif/files/usr/lib/webif/webif.sh: allow numbers in webif scripts (#619)
[14:17] <CIA-17> florian * r4626 /branches/whiterussian/openwrt/package/webif/files/usr/lib/webif/lang/hu/ (. common.txt): Add hungarian translation, closes #642
[14:19] <CIA-17> nbd * r4627 /branches/ (2 files in 2 dirs): enable mknod by default (#490) - some drivers don't register with devfs properly
[14:24] <CIA-17> nbd * r4628 /branches/whiterussian/openwrt/package/base-files/default/etc/functions.sh: fix typo (#693)
[14:26] <CIA-17> florian * r4629 /packages/libs/libjpeg/ (Makefile patches/200-crop.patch): Add crop features, and build jpeg tools, closes #665
[14:40] <CIA-17> florian * r4630 /branches/buildroot-ng/openwrt/ (include/modules-2.6.mk package/kernel/modules.mk): Port kmod-gre to the new module layout
[14:40] <CIA-17> florian * r4631 /packages/net/pptpd/Makefile: Add missing dependencies for pptpd, closes #656
[16:10] <[florian]> I never figured out openswan was such as painful software :(
[16:21] <CIA-17> florian * r4632 /branches/ (7 files in 6 dirs): Upgrade to openswan-2.4.6, fix showhostkey script (closes #444), *should* fix #647, #671
[16:24] <[florian]> do we have a special target that could allow us recompiling every package that depends on foo ?
[18:44] <[florian]> re
[18:49] <[florian]> humm users are boring with stupid questions and no faq reading
[18:52] <murb> [florian]: well if you have lots of users there will always be a percentage who can write but can't read.
[18:53] <[florian]> sure :)
[18:57] <[florian]> this raises the questions of how they use a computer by the way ;)
[18:58] <Bartman007> not really. I've seen many people who can use a computer, but don't know how to use it properly
[18:59] <thepeople_work> bartman007: it should read *most people*
[18:59] <Bartman007> double-clicking weblinks, confusing logging onto the computer and logging into their webmail, etc....
[18:59] <thepeople_work> ohh no my paypal account has been suspended
[19:00] <[florian]> ah Bartman007 this is not the same thing ? ;)
[19:00] <thepeople_work> i better follow the link and update it
[19:00] <thepeople_work> wait do I even have a paypal account?
[19:02] <[florian]> lol
[19:03] <[florian]> this is probably spam ?
[19:45] <h3sp4wn> Kaloz: If the images are ready than I can try them
[21:25] <nbd> [florian]: ping
[21:28] <[florian]> nbd: pong
[21:29] <nbd> have you looked into the new kernel packaging system?
[21:29] <nbd> because i could use some help with porting over all the old kmod-* packaging stuff
[21:29] <[florian]> nbd: a bit yes, did I make a mistake with the recent kmod-gre ?
[21:30] <nbd> ah, didn't check the log and didn't notice you already added some stuff
[21:30] <nbd> forgot to svn update :)
[21:30] <nbd> no, gre looks fine
[21:31] <[florian]> ok :)
[21:32] <[florian]> by the way, I have just added rssi to wlcompat, I should actually to the list of private ioctl's right ?
[21:34] <nbd> yeah
[21:34] <[florian]> maybe you have already done it ?
[21:34] <nbd> done what?
[21:34] <[florian]> adding rssi and assoclist to wlcompat N
[21:34] <[florian]> ?
[21:35] <nbd> no
[21:35] <[florian]> ok
[21:36] <[florian]> regarding kernel modules, there is not much work to do on it actually
[21:37] <nbd> yeah, just translating all the stuff from modules-2.*.mk and removing it from target/linux/Config.in afterwards
[21:37] <[florian]> I will port kmod-tun to the new layout right now, because it is a dependency for several user-space packages
[21:38] <Bartman007> [florian]: thanks for doing that, that's one of my issues rightnow.
[21:44] <CIA-17> florian * r4633 /branches/buildroot-ng/openwrt/ (3 files in 3 dirs): Port kmod-tun to the new module layout
[21:55] <[florian]> nbd: actually the whole wl binary could be easily rewritten with the help of wlioctl.h right ?
[21:55] <nbd> not all of it is documented, but yes
[21:55] <nbd> but there's no need to rewrite wl
[21:56] <nbd> wlc is better
[21:56] <[florian]> sure
[21:56] <[florian]> what is the goal of having both wlcompat and wlc ?
[21:56] <[florian]> it is just in the case we want to use the iwlib right ?
[21:56] <nbd> yeah, and wireless-tools is a convenient interface for some parts
[21:57] <[florian]> that's right
[21:58] <[florian]> wl could probably be spied to get the missing ioctls ?
[21:59] <[florian]> but it's really no use now with bcm43xx
[22:01] <Bartman007> so is bcm43xx to a point that it can be used in mastermode?
[22:01] <[florian]> not yet I think
[22:01] <[florian]> all the more as they are working on v4 specs
[22:02] <h3sp4wn> can the dscape version run in master mode (on x86 even)
[22:03] <[florian]> I think so
[22:04] <[florian]> from what I have read on the ml, looks like some people are running it
[22:10] <[florian]> nbd: I added rssi to wlcompat, but it does not seem to work : http://pastebin.ca/145024
[22:10] <[florian]> any hints ?
[22:11] <nbd> what happens if you try it?
[22:11] <[florian]> Interface doesn't accept private ioctl...
[22:11] <[florian]> get_rssi (8BEA): Invalid argument
[22:13] <nbd> does 'wl rssi' work?
[22:14] <[florian]> ah, I forgot that wl rssi expects the mac address as argument !
[22:15] <[florian]> so I should probably have a try at the assoclist rather than rssi for the moment
[22:17] <nbd> you can integrate rssi in the assoclist
[22:18] <[florian]> so that it reports the associated clients and their rssi
[22:19] <CIA-17> nbd * r4634 /branches/buildroot-ng/openwrt/target/linux/ (2 files in 2 dirs): don't try to create mtd partitions on bcm947xx when the trx header is not found - should fix #318
[22:20] <nbd> [florian]: yes
[22:36] <[florian]> nbd: when calling wl_ioctl, the value is then stored in the void * buf right ?
[22:39] <nbd> it's stored in a buffer with the size that you specify
[22:39] <[florian]> ok
[22:39] <[florian]> and where can I access it after ?
[22:39] <nbd> if you want to know what kind of buffer is used for the rssi stuff, use wlcompat-debug
[22:39] <[florian]> ok
[22:39] <nbd> since you supply the buffer, you get the result
[22:39] <nbd> it doesn't allocate something new
[22:40] <[florian]> ok
[23:03] <[florian]> nbd: why do I send and receive exactly the same things ? http://pastebin.ca/145134
[23:05] <nbd> which command is that from?
[23:06] <[florian]> iwpriv eth1 assoclist
[23:06] <nbd> are you using a static buffer?
[23:07] <nbd> or a kmalloc'd one?
[23:07] <[florian]> I the variable extra, like for other ioctls, should not I ?
[23:08] <nbd> returning binary data in an iwpriv call is not a good idea
[23:08] <nbd> and when you actually make the call, you should fill the buffer with 0
[23:10] <[florian]> ok, so I probably should use another buffer for testing purposes ?
[23:13] <nbd> if you use a static buffer, it will keep the data from the last request
[23:13] <nbd> which is probably what you're seeing thereb
[23:13] <[florian]> right
[23:13] <nbd> because at a first glance, it does look like an assoclist buffer
[23:13] <[florian]> I ran wl assoclist in order to figure out which size I should use as buffer size
[23:14] <nbd> doesn't wl usually use 8k for something like that?
[23:15] <[florian]> precisely
[23:41] <[florian]> nbd: sorry to bother you again, but was maybe wrong with this : http://pastebin.ca/145225
[23:46] <[florian]> what I don't understand is that using mutliple times symbol extra seems to work for other ioctls though used as int
[23:53] <nbd> you should make assoclist return a string
[00:00] --- Wed Aug 23 2006