**** BEGIN LOGGING AT Mon Jan 31 02:59:57 2011
Jan 31 03:13:22 <cshore> ping KanjiMonster
Jan 31 03:13:29 <cshore> ping xMFf
Jan 31 03:49:01 <builder> build #77 of brcm47xx is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/77
Jan 31 05:12:20 <CIA-84> cshore * r25259 /trunk/package/block-extroot/files/extmount.sh: [package]: block-extroot: Fixed hang when no modules in on either squashfs or jffs2 and suppressed extraneous not found error message when no modules on jffs2.
Jan 31 08:58:56 <KanjiMonster> cshore: pong
Jan 31 08:59:06 <cshore> hey, what's up?
Jan 31 08:59:44 <KanjiMonster> currently working with the telekom spw303v, and its tag doesn't have information fields
Jan 31 09:00:21 <cshore> you mean it's already used for something?
Jan 31 09:00:46 <KanjiMonster> yepp, for length/size fields for an embedded certificate
Jan 31 09:01:54 <cshore> bah, I had hoped we'd be able to have one-size-fits-all
Jan 31 09:01:59 <KanjiMonster> cshore: http://pastebin.com/WbXGSNEk
Jan 31 09:03:03 <cshore> hmm....we might actually be okay
Jan 31 09:03:05 <KanjiMonster> cshore: at least its using a custom board name, and the device needs special handling anyway - the image crc is calculated by appending 64 bytes for the calculation, which don't get actually added to the image
Jan 31 09:03:37 <KanjiMonster> and the image is expected to be xor'd with "12345678"
Jan 31 09:03:45 <cshore> bizarre
Jan 31 09:03:53 <KanjiMonster> yeah
Jan 31 09:05:15 <KanjiMonster> I already got it to accept an image, but cfe wants to flash the image to 0xbe400000 instead of 0xbe010000 (where my first flash is). Probably has something to do with the image sequence
Jan 31 09:05:38 <cshore> ah, sounds like it's dual image thing
Jan 31 09:05:56 <cshore> if you've got one good image and you're flashing on top, then it flashes to the second half
Jan 31 09:06:13 <cshore> if you want it to flash to the first half you have to get rid of the 'good' image
Jan 31 09:06:16 <KanjiMonster> well, the original images are 5MB+
Jan 31 09:06:18 <cshore> (erase
Jan 31 09:06:30 <KanjiMonster> so it shouldn't work in that cas
Jan 31 09:06:33 <KanjiMonster> e
Jan 31 09:06:41 <KanjiMonster> that's why I am a bit confused
Jan 31 09:06:41 <cshore> I have  Telsey that had a 5MB and 3MB image
Jan 31 09:06:58 <cshore> but when you flash on top of a good image it's like I described
Jan 31 09:07:11 <KanjiMonster> ah, ok
Jan 31 09:07:37 <cshore> It's only the cfe that does that IIRC
Jan 31 09:07:58 <KanjiMonster> I guess so, too
Jan 31 09:08:46 <KanjiMonster> after flashing, it also boots from 0xbe400000 (which somewhat succeeds - it uses the squashfs from the original flash ;), but cold boots still boot from 0xbe010000
Jan 31 09:08:56 <cshore> so, yeah, you'll have to hack an alternate struct
Jan 31 09:09:12 <cshore> yeah, it gets confused
Jan 31 09:09:26 <KanjiMonster> I wrote a small c util that "fixes" the header
Jan 31 09:09:37 <KanjiMonster> (and xors the image)
Jan 31 09:11:32 <KanjiMonster> the cfe is a PITA on this board, you have to keep reset pressed on boot (which first erases the nvram); canceling the boot through serial doesn't work at all
Jan 31 09:12:13 <cshore> sounds like they're trying to lock down the board excessively
Jan 31 09:12:42 <KanjiMonster> in the original firmware the serial is also "read only"
Jan 31 09:14:05 <KanjiMonster> also this device warrants a patch to linux-mips; the voip-dsp uses active low and active high gpios; but the dev-dsp treats all gpios as one of them (I don't remember which one ;)
Jan 31 09:14:55 <cshore> sounds like fun
Jan 31 09:14:56 <CIA-84> swalker * r25260 /packages/net/tor-alpha/Makefile: [packages] tor-alpha: update to 0.2.2.22-alpha, add archive mirror
Jan 31 09:16:05 <EqUaTe> fnord
Jan 31 09:16:32 <KanjiMonster> cshore: the cfe is even somewhat evil, it still prints the "press any key to stop auto run" despite not accepting input ;)
Jan 31 09:17:36 <cshore> I see....so you could think you serial is broken
Jan 31 09:17:45 <KanjiMonster> yeah
Jan 31 09:18:30 <cshore> was telnet able let you get a shell, or did you have to flash first?
Jan 31 09:19:32 <KanjiMonster> no telnet access
Jan 31 09:19:51 <KanjiMonster> gpl sources are available though
Jan 31 09:19:57 <cshore> ah, ok
Jan 31 09:22:00 <KanjiMonster> hm, maybe braking into cfe doesn't require resetting nvram - I just notices that normal keys do stop the bootprocess (and print) until you press return, so it seems it expects some magic password
Jan 31 09:24:56 <KanjiMonster> cshore: oh, the sig1 of the tag also isnt broadcom, but "AAAAAAAA Corporation" ;)
Jan 31 09:29:22 <cshore> that part's easy enough....the tag program is pretty flexible as far as that goes, but the certificate things really puts a cramp on the information fields
Jan 31 09:33:12 <KanjiMonster> filling the fields with zeroes makes the cfe complain about a missing "encrypted key", but it flashes it anyway
Jan 31 09:37:55 <KanjiMonster> cshore: didn't you use qemu once to decompress a cfe?
Jan 31 09:38:45 <cshore> Right now I'm working on an alternative to extroot .... instead of running entirely on external storage I'm working on automating the setup where the flash is root, but extra packages can run off external storage
Jan 31 09:38:51 <cshore> no, but I have a tag analyzer
Jan 31 09:39:01 <cshore> I wrote
Jan 31 09:39:17 <cshore> oh cfe
Jan 31 09:39:24 <cshore> no, never touched the cfe
Jan 31 09:39:30 <cshore> maybe florian?
Jan 31 09:46:01 <KanjiMonster> cshore: I patches the analyzer locally to test some things on generated images/official stock images (like did I correctly modify the crc generation)
Jan 31 09:46:53 <cshore> the analyzetag thing I wrote or something else?
Jan 31 09:47:14 <KanjiMonster> that one
Jan 31 09:47:22 <cshore> cool....it's being used!
Jan 31 09:47:39 <KanjiMonster> it was quite helpful
Jan 31 09:48:33 <cshore> you just made my day (well night...it's 4:45am here)
Jan 31 09:49:22 <KanjiMonster> it was ~4 am here when you pinged me ;)
Jan 31 09:50:05 <cshore> did I send it to you or do you find it on the wiki?
Jan 31 09:52:23 <KanjiMonster> wiki
Jan 31 09:52:41 <cshore> cool...I knew it was good for something :)
Jan 31 09:52:55 <EqUaTe> heh
Jan 31 09:53:23 <EqUaTe> god damn i'm tired..
Jan 31 09:53:54 <EqUaTe> KanjiMonster: i'll sort out sending you that wnr2000v1 in the next few days..
Jan 31 09:54:15 <KanjiMonster> EqUaTe: ah, cool
Jan 31 09:54:30 <EqUaTe> and hopefully get mine hooked back up again to start poking at some more.
Jan 31 09:56:12 <KanjiMonster> btw, I saw the wnr2200 is basically a wnr2000v1 but without the stupid rootfs crc-check and with "real" image tag parsing
Jan 31 09:56:47 <EqUaTe> heh
Jan 31 09:56:59 <EqUaTe> sounds like someone told them how retarded they were being and they listened
Jan 31 11:09:14 <Acinonyx> gmorning
Jan 31 11:23:56 <CIA-84> hcg * r25261 /packages/devel/cppunit/Makefile: [packages] cppunit: Fix install
Jan 31 11:57:30 <builder> build #64 of sibyte is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/64
Jan 31 12:11:46 <builder> build #76 of ubicom32 is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/76
Jan 31 12:20:06 <builder> build #63 of avr32 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/avr32/builds/63
Jan 31 12:48:24 <WTGPhoben> nbd:  I think this is your domain...  I have an ubiquiti nanostation locoM5, and in sta mode on any of the 30dBm channels (option 'country' 'US')  The maximum power I am allowed is 21dBm.  In AP or adhoc I can get 30.  Is this a known bug? (didn't find in defects)
Jan 31 12:49:04 <WTGPhoben> running r25061
Jan 31 14:42:44 <builder> build #60 of ixp4xx is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/60
Jan 31 14:59:05 <builder> build #69 of pxcab is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/69
Jan 31 18:22:34 <CIA-84> obsy * r25262 /packages/net/transmission/Makefile: [packages] transmission: update to 2.20b3
Jan 31 19:58:12 <CIA-84> nbd * r25263 /trunk/package/libnl-tiny/src/ (Makefile include/unl.h unl.c): libnl-tiny: add unl a convenience wrapper around the libnl api
Jan 31 19:58:22 <Acinonyx> g'evening
Jan 31 20:16:43 <CIA-84> nbd * r25264 /trunk/tools/ (Makefile cmake/ cmake/Makefile): add tools/cmake
Jan 31 20:16:46 <CIA-84> nbd * r25265 /trunk/include/cmake.mk: add a template for cmake packages
Jan 31 20:34:49 <philipp64|laptop> nbd: can you commit my libnl patches while you're in there? thanks.
Jan 31 21:04:32 <builder> build #61 of kirkwood is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/61
Jan 31 21:22:05 <[florian]> philipp64|laptop: have you compared the size difference between libnl-1.1 and libnl-2.0?
Jan 31 21:26:04 <philipp64|laptop> let's see... for libnl.so.1.1 for x86, 219524 bytes.
Jan 31 21:27:26 <philipp64|laptop> for libnl.so.2.0 ... 60242 bytes for x86.
Jan 31 21:28:12 <nbd> because libnl 2.0 is split up some more compared to 1.0
Jan 31 21:28:24 <philipp64|laptop> yeah, I'm just noticing that...
Jan 31 21:28:37 <philipp64|laptop> I'm wondering why ldd didn't complain.
Jan 31 21:28:43 <philipp64|laptop> or ldcache
Jan 31 21:28:59 <[florian]> cool, then that's a net improvement for size
Jan 31 21:29:04 <[florian]> is that a reduced build?
Jan 31 21:30:12 <philipp64|laptop> no, it's because the makefile needs to be modified to copy all of the other libraries. sigh.
Jan 31 21:30:20 <philipp64|laptop> what does PKG_INSTALL_DIR expand out to be?
Jan 31 21:31:22 <xMff> the staging dir for packaging the .ipk
Jan 31 21:43:16 <philipp64|laptop> inside the build_dir for that package I take it.
Jan 31 21:45:44 <philipp64|laptop> ok, rebuilt: libnl-genl: 10286, libnl-nf: 45014, libnl-route: 134343, libnl: 60242
Jan 31 21:47:00 <philipp64|laptop> so about 260K.
Jan 31 21:47:12 <philipp64|laptop> 30KB increase from before.
Jan 31 21:47:28 <nbd> i guess that's ok, since only one package uses libnl
Jan 31 21:47:35 <nbd> the other ones use libnl-tiny
Jan 31 21:47:39 <philipp64|laptop> the good news is that most things just link against libnl+libnl-genl, so the memory footprint will be lower.
Jan 31 21:47:54 <philipp64|laptop> actually, 2 do.  swconfig, and...
Jan 31 21:48:19 <philipp64|laptop> nope.  just swconfig. I take it back.
Jan 31 21:48:35 <nbd> swconfig uses libnl-tiny as well
Jan 31 21:49:53 <philipp64|laptop> you can link hostapd against it if you wanted to... oh, and there's l2tpv3tun ...
Jan 31 21:50:22 <nbd> well, libnl-tiny is less than 45k on mips (which is usually bigger than x86), and it has core + genl functionality
Jan 31 21:53:52 <philipp64|laptop> right, I'm just pointing out that the full libnl-1.1 is not that much smaller on disk that the full libnl-2.0...
Jan 31 21:54:12 <philipp64|laptop> and that the running libnl-2.0 is actually smaller (since most people don't need libnl-route.so)
Jan 31 21:54:43 <philipp64|laptop> ok, I'll send patched diffs.
Jan 31 21:55:37 <philipp64|laptop> not sure why the makefile uses $(CP) instead of $(INSTALL_BIN) or $(INSTALL_DATA)...
Jan 31 21:56:07 <xMff> symlinks
Jan 31 21:58:09 <russell--> "Building GCC requires GMP 4.2+, MPFR 2.3.1+ and MPC 0.8.0+"
Jan 31 22:06:34 <CIA-84> jow * r25266 /packages/libs/xmlrpc-c/Makefile: [packages] xmlrpc-c: don't use underscores in package names, it will confuse opkg
Jan 31 22:07:01 * russell-- found it
Jan 31 22:07:17 <CIA-84> jow * r25267 /packages/net/rtorrent/Makefile: [packages] rtorrent: don't use fix dep on xmlrpc-c after last commit
Jan 31 22:07:37 <xMff> erm... message fail
Jan 31 22:07:44 <russell--> libmpc-dev
Jan 31 22:11:53 <philipp64|laptop> posted...
Jan 31 22:30:16 <CIA-84> acinonyx * r25268 /trunk/package/base-files/files/etc/hotplug2-common.rules: [package] base-files: Remove deprecated hotplug2 commands, update old commands to new
Jan 31 22:39:38 <loswillios> nbd: ah, cmake support. nice!
Jan 31 23:42:21 <builder> build #78 of brcm47xx is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/78
Jan 31 23:55:16 <builder> build #63 of uml is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/uml/builds/63
Jan 31 23:57:14 <builder> build #64 of x86 is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/x86/builds/64
Feb 01 02:52:40 <builder> build #74 of ppc40x is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/ppc40x/builds/74
**** ENDING LOGGING AT Tue Feb 01 02:59:57 2011