**** BEGIN LOGGING AT Sat Sep 24 02:59:57 2011 Sep 24 14:34:05 Any pointers to stable kernel mirrors? Couldn't find any Sep 24 15:38:12 joelagnel ??? github Sep 24 15:39:17 woglinde, do you have the repository URL for stable kernel tree 3.0.y Sep 24 15:39:32 is searching on github this hard? Sep 24 15:39:56 https://github.com/mirrors/linux-2.6 Sep 24 15:39:58 ofcourse not Sep 24 15:40:01 first hit Sep 24 15:40:06 when searching linus Sep 24 15:40:06 I already did Sep 24 15:40:41 https://github.com/mirrors/linux Sep 24 15:40:47 one beer Sep 24 15:40:53 you own me Sep 24 15:41:15 woglinde, come on, I am looking for the stable kernel Sep 24 15:41:43 https://github.com/mirrors/linux/tree/v3.0 Sep 24 15:41:46 two beers Sep 24 15:42:02 for not finding the tag button Sep 24 15:42:16 * joelagnel gives up Sep 24 15:42:52 or you want 3.0.4? Sep 24 15:43:14 yeah, that was what I was trying to say :) Sep 24 15:44:36 http://vesta.informatik.rwth-aachen.de/ftp/pub/comp/Linux/kernel/v3.0/ Sep 24 15:45:03 google should find you a mirror near by you Sep 24 15:45:58 woglinde, thank you very much :-) Sep 24 15:46:19 otherwise this was discussed on oe-core too Sep 24 15:46:29 so its 3 beers now Sep 24 15:46:34 for not reading email Sep 24 15:51:34 * joelagnel has enough to read already :( Sep 24 20:52:00 what is the simplest/easiest distro for console-only tests in oe ? Sep 24 20:59:26 lautriv, maybe pick any but make a console-image Sep 24 20:59:40 angstrom is good for instance Sep 24 21:00:05 Hello, I thought about working on real-time MAC policy generation for android platform as part of my graduation project but I don't know how much time would this take (Not very much into Embedded but a very fast learner :) ) Sep 24 21:00:05 console-image in angstrom Sep 24 21:00:17 would 9 months be enough ? What do you think would be the best MAC technology to use ? SELinux or other ? I would really like to contribute to Linux and Open Source but I'm afraid the time I've won't be enough Sep 24 21:03:17 GNUtoo, i started with aurora and got some errors, switched to angstrom and that fails on kernel (3.0, source changed) machine is a new target, similar to hx4700. Sep 24 21:03:56 ok Sep 24 21:04:11 is your machine setup ok? Sep 24 21:04:42 what is the 3.0 kernel you are failing on? Sep 24 21:04:42 is it a beagle one? Sep 24 21:04:42 Xenin, hi Sep 24 21:05:05 GNUtoo: Hey :) Sep 24 21:05:07 Xenin, there are some android channels on freenode Sep 24 21:05:22 GNUtoo, basically i made a copy of hx4700 and changed screen (small 240x320) and disabled the special acx stuff, rest is unknown and i will dig it via some minimal functionality. Sep 24 21:05:42 if you want help on finding the correct channel just ask me Sep 24 21:05:53 lautriv, does it fails on beagleboard 3.0 kenrel? Sep 24 21:05:59 or hx4700 knerel? Sep 24 21:06:05 I'm searching around (about +20 channels now) Sep 24 21:06:19 ok Sep 24 21:06:21 At the moment I'ven't made my mind about my topic Sep 24 21:06:24 there is android-root Sep 24 21:06:25 GNUtoo, hx4700 kernel Sep 24 21:06:26 cyanogenmod Sep 24 21:06:30 and several others Sep 24 21:06:38 there is even one channel for developers Sep 24 21:06:49 Would MAC on Linux embedded (not Android specific) be a good graduation project ? Sep 24 21:06:50 maybe pick the developper channel Sep 24 21:07:04 yes Sep 24 21:07:24 let's call it GNU/Linux embedded Sep 24 21:07:31 because there is GNU/Linux and Android Sep 24 21:07:48 I'm running GNU/Linux :) Sep 24 21:07:51 they are different so it's easier to differentiate that way(calling GNU/Linux) Sep 24 21:07:56 because android uses a modified linux kenrel Sep 24 21:08:23 ok Sep 24 21:08:24 That's my point, would this modification make it hard to implement MAC for android plateform ? Sep 24 21:08:54 no ideas, I guess that would be a long project Sep 24 21:08:59 android != GNU/Linux Sep 24 21:09:03 everything is different Sep 24 21:09:11 even the kernel userland interfaces are different Sep 24 21:09:21 let's take selinux Sep 24 21:09:36 in android selinux won't run out of the box Sep 24 21:09:42 I've about 9 months Sep 24 21:09:43 you would need support for it in the lowest layers in userspace Sep 24 21:09:54 like the apps themselves Sep 24 21:09:59 for instance su Sep 24 21:10:03 or the package manager Sep 24 21:10:05 beside.... Sep 24 21:10:13 apps usually run in some sort of sandbox Sep 24 21:10:27 it would be more usefull and easier for GNU/Linux Sep 24 21:10:46 however some distros don't even run as user.... Sep 24 21:10:52 like SHR for instance Sep 24 21:10:54 they run as root Sep 24 21:11:06 other distros like angstrom run ok as users I guess Sep 24 21:11:24 if you have the chance to have a user creation GUI Sep 24 21:11:32 like for instance for the opie images Sep 24 21:12:20 I've created a .bbappend for u-boot with SRC_URI_append_beagleboard. When I try to compile it says "ERROR: Function 'The URL: '\' is invalid and cannot be interpreted' failed. Sep 24 21:12:52 hmm ok, so after I finish research work on MAC embedded etc... what do you think would be a good target for applying the research on ? Sep 24 21:12:54 tasslehoff, tell more about what you did Sep 24 21:13:07 Meego, Android or other idea ? Sep 24 21:13:17 Xenin, I think the best target would be to harden SHR for instance Sep 24 21:13:44 *)it's a phone distro Sep 24 21:13:51 *)it lack every security feature Sep 24 21:13:57 for instance it runs as root Sep 24 21:14:22 the only security is that ssh doesn't accept empty passwords Sep 24 21:14:25 but that's all Sep 24 21:15:03 GNUtoo: SRC_URI_append_beagleboard = "file://0001-patchone.patch \ file://0002-patchtwo.patch" Sep 24 21:15:06 if you want to do security stuff with embedded devices it's ones of the best targets because there is a lot to do Sep 24 21:15:16 we lack policy framework for security too Sep 24 21:15:25 that, and I added a FILESEXTRAPATHS_prepend Sep 24 21:15:42 tasslehoff, you added that in one line only? Sep 24 21:15:48 I mean not like that: Sep 24 21:15:48 "file://0001-patchone.patch \ Sep 24 21:15:48 file://0002-patchtwo.patch" Sep 24 21:15:52 GNUtoo: no, two. should have said that Sep 24 21:15:54 two lines Sep 24 21:16:27 then I suspect that you have a space after \ Sep 24 21:16:27 ah Sep 24 21:16:47 try to see if you have a space Sep 24 21:16:49 if so remove it Sep 24 21:16:54 GNUtoo, i had a read at the specs of the beagleboard, nice thingie but far away from mine. i know this -> PXA270@624, ARMv5te, 128M flash,120M RAM, 240x320 touchscreen, samsung wlan, unknown bluetooth, GSM/GPRS capable, cam/barcodereader, some multicolor-led................maybe you know a tool running on win-mob4 to get more details like memmap/devices ? Sep 24 21:17:04 GNUtoo: no, but I had a trailing ' \' after the second patch Sep 24 21:17:20 the stuff I wrote to you was all wrong :) Sep 24 21:17:31 tasslehoff, ok Sep 24 21:17:35 * tasslehoff is typing this on another machine than the one that's building Sep 24 21:17:48 lautriv, yes I do, the tool is called haret Sep 24 21:18:01 ok Sep 24 21:18:42 GNUtoo, ran haret but found just some register/acpi functions ? Sep 24 21:19:24 GNUtoo, you know a lot the subject. Do you think that 9 months would be enough to fast pick up embedded dev, one MAC technology, study the internals of a distro (Android, SHR or other GNU/Linux) and apply the work. Sep 24 21:19:33 and write the thesis. Sep 24 21:19:46 a lot about* Sep 24 21:20:09 Xenin, no idea on the time schedule, it depend on how good you are and on how much blockers you encounter Sep 24 21:20:46 Xenin, I also know about selinux btw Sep 24 21:20:53 I used to run it and to do some policies Sep 24 21:21:26 that's good. Someone told me that a good project would be: real-time selinux policy generation for android platform Sep 24 21:21:48 yes but as I said before android is not very well suited for selinux: Sep 24 21:21:49 could be not android, but another distro Sep 24 21:21:53 *)lots and lots of part to patch Sep 24 21:22:12 *)Android already have an app policy framework Sep 24 21:22:20 for instance you install an app Sep 24 21:22:34 and it tells you the app wants to access the network or get a gps fix Sep 24 21:22:41 *wants permission to Sep 24 21:23:46 hmm, that's true Sep 24 21:26:04 else you may find other ideas not involving security + embedded devices Sep 24 21:26:13 but making a secure phone from SHR is a good idea Sep 24 21:26:19 else maybe: Sep 24 21:26:29 *GSM voice encryption system Sep 24 21:26:41 along with sms encryption system Sep 24 21:27:02 that would require a phone where you can control the audio Sep 24 21:27:06 for instance like the n900 Sep 24 21:27:08 from nokia Sep 24 21:27:24 so you would: Sep 24 21:27:36 capture the audio from the microphone->encrypt it->feed it to the modem Sep 24 21:27:44 and do the countrary on the remote side Sep 24 21:28:08 and beware that's not possible on all phones as I said before Sep 24 21:28:48 for instance in some android phones like the ones with a qualcomm SOC, you have no access to the audio Sep 24 21:30:13 that is to say you have: Sep 24 21:30:13 CPU<->[modem<->audio] Sep 24 21:30:14 on other phones like the freeurnner the cpu controls the sound card but still the modem pcm is routed directly to the sound card Sep 24 21:30:54 Xenin, what do you think about voice encryption for phones? Sep 24 21:32:15 ah but let me think Sep 24 21:32:15 on the gta02 it might be possible to route the PCM from the sound card to the phone Sep 24 21:32:16 and even to record the sound from the modem Sep 24 21:32:16 I was mistaken sorry Sep 24 21:32:29 Xenin, are you still there? Sep 24 21:35:20 he's takin' his dose of dopamin :-) Sep 24 21:35:39 good nite Sep 24 21:35:43 good night Sep 24 21:35:51 n8 Sep 24 21:41:37 back Sep 24 21:41:40 GNUtoo, do you know about some good sites where i could get more details about hw-specific details for X|PDA resp. smartphones ? Sep 24 21:42:14 lautriv, what pda do you have? Sep 24 21:42:28 GNUtoo wow, you've lots of ideas. Sep 24 21:42:52 I've also ideas on how to implement the things I talk about.... Sep 24 21:43:07 GNUtoo, the one i'm talking about is a motorola HCe700g Sep 24 21:43:30 I've access to an Android Nexus S Sep 24 21:43:46 Xenin, ok I know a bit this phone Sep 24 21:43:53 2 key people have it: Sep 24 21:44:08 morphis in #openmoko-cdevel and #replicant Sep 24 21:44:46 and paulk in #replicant Sep 24 21:44:46 they can tell you how it works Sep 24 21:44:46 because I know a bit the modem Sep 24 21:44:47 I know there is a sound CODEC Sep 24 21:44:47 that is controlled by the CPU Sep 24 21:44:48 but I've no ideas on how to feed the modem Sep 24 21:44:49 or maybe I don't remember Sep 24 21:45:55 hmm, and for the encryption feature, the app should be installed on both devices communicating right ? Sep 24 21:46:13 or could this be introduced via a ROM ? Sep 24 21:46:21 yes on both devices Sep 24 21:46:26 Xenin, also with a common key. Sep 24 21:46:31 via ROM? what do you mean? Sep 24 21:47:21 ie. Cyanogenmod Sep 24 21:48:16 could they for example ship it as a feature for example: Two Cyanogenmod phones use the encryption by default Sep 24 21:48:18 I think you need low level access Sep 24 21:48:18 so trough a ROM Sep 24 21:48:39 but as I said it will depend a lot on if it's possible at the hardware level Sep 24 21:48:54 you should check with paulk, ius or morphis in #replicant Sep 24 21:49:29 I can look at the sources too Sep 24 21:49:29 but I've no nexus S yet Sep 24 21:49:29 I looked at it to help them Sep 24 21:49:29 *at the sources Sep 24 21:49:33 fu**, it's harder to get the sources because kernel.org's maintenance than to compile the whole.....................anyone a clue when they'll be back ? Sep 24 21:50:41 hmm, another point is could the device with such a feature be able to detect if the other phone does/doesn't support the encryption mode Sep 24 21:52:44 I will check with them by the way Sep 24 21:53:02 ok Sep 24 21:53:22 to detect if the other phone support: you can add a basic mode at first with a button Sep 24 21:53:29 for instance you ear scrambled audio Sep 24 21:53:34 =>you press the button Sep 24 21:55:42 Thanks GNUtoo. You're very helpful. :) Sep 24 21:56:22 np Sep 24 21:59:54 hi, does someone knows how to cleansstate everything that was built Sep 24 22:00:13 it's beause I am trying to add support for armv6 without vfp Sep 24 23:45:16 anyone know what could cause -I/usr/include to be given to g++ and g++ to fail finding ? Sep 24 23:45:38 ? Sep 24 23:45:52 usually -I/usr/include is bad Sep 24 23:46:17 it's not arm Sep 24 23:46:20 it's x86 Sep 24 23:48:21 --oldincludedir=/usr/include, is that normal for autotools? Sep 24 23:49:13 no idea Sep 24 23:57:05 Hi, I've got a tarball which decompress all of it's files into the current directory. Any idea how I can write a bitbake recipe to unpack it in a separate source dir ? Sep 25 00:00:05 I was thinking about S = "${WORKDIR}", but that's not very nice Sep 25 00:07:03 wouldn't that be S = "${WORKDIR}/seperate"? Sep 25 00:11:11 nope S doesn't change where the sources are unpacked Sep 25 00:11:26 it just tells bitbake where to find the sources once they're unpacked Sep 25 00:35:56 the wrong sdl-config is being called too :( Sep 25 00:58:10 Zoxc, sdl-config...are you in games porting? Sep 25 00:58:19 if so where to put games? Sep 25 00:58:24 is there a recipes-games? Sep 25 00:58:28 or a meta-games? Sep 25 01:03:24 julian_: perhaps specify destdir=subdirname Sep 25 01:03:40 in SRC_URI Sep 25 01:03:42 in the SRC_URI line ? Sep 25 01:03:45 julian_: yep Sep 25 01:03:49 mmmh, cool Sep 25 01:04:13 will try that Sep 25 01:40:18 ugh! i modify this chumby bitbake recipee by adding gcc to the main bb recipee in the end or the IMAGE_INSTALL and i cant find a gcc bin in the image **** ENDING LOGGING AT Sun Sep 25 02:59:57 2011