**** BEGIN LOGGING AT Tue Oct 21 02:59:57 2008 Oct 21 03:36:55 hola Oct 21 06:55:44 G'morning Oct 21 07:37:39 How to build fennec when its always complaining about missing autoconf2.13? Oct 21 07:52:09 bonjour Oct 21 08:25:06 morning all Oct 21 08:25:58 hi RP Oct 21 08:26:06 and hi everyone :> Oct 21 08:28:23 our last movie : http://aldebaran-robotics.com/download/NaoAcademicsV3.mov Oct 21 08:28:31 running Openembedded , enjoy Oct 21 08:32:45 also 10 - 15uhr schweinfurt ... und halt immer hin und rueckfahrt Oct 21 08:32:46 also 10 - 15uhr schweinfurt ... und halt immer hin und rueckfahrt Oct 21 08:34:51 tag Oct 21 08:34:57 hi woglinde Oct 21 08:35:34 woglinde: I also tried glibc for sh4, I'll make a diff, maybe we could add those patches from STLinux to OE too Oct 21 08:36:11 jineld sure Oct 21 08:39:06 hi, in fennec there is used in mozconfig autoconf2.13 without '-' is it right? its not available in autoconf-native and my distro (gentoo) has only autoconf-2.13.. adding '-' to mozconfig seems working now Oct 21 08:39:52 jama mozilla foundation has so much money, but it is a shame that the do not fix theire abuse of autotools Oct 21 08:43:08 having much money is still no guarantee that people who have it know what they are doing :> Oct 21 08:49:05 Is the ap7000 supported in oe ? Oct 21 08:50:44 gerwini hm what machine is ap7000? Oct 21 08:51:30 I have this atngw100 Oct 21 08:51:46 woglinde: atngw100 Oct 21 08:52:00 hm ah there should be support Oct 21 08:52:05 woglinde: need to make some kind of mini pabx Oct 21 08:52:59 woglinde: seems to be an interesting device Oct 21 08:53:26 we one at work I think but I didnt have the time to play with it yet Oct 21 08:54:24 anyone working with TI DaVinci stuff, from what I can see on their homepage - they use gstreamer? is it used as the interface to the decoder? Oct 21 08:55:35 Yes I worked with the davinci stuff Oct 21 08:55:51 Had a lot of work to get the development kit working Oct 21 08:55:58 Had all kind of hardware problems Oct 21 08:56:28 I think I had about 4 new once before I had a good one Oct 21 08:56:43 heh :) Oct 21 08:57:09 but do they use gstreamer for decoding? because that would be kind of cool Oct 21 08:57:25 so far the most hw I have seen had a terrible interface to the proprietary decoder part Oct 21 08:57:31 Jin: As far as I know yes Oct 21 08:57:39 nice, very nice.. Oct 21 08:57:48 Jin: well nothing matches NXP if it concerns crappy hardware ;) Oct 21 08:58:16 jin: they have even a revision - (run out of characters :)) Oct 21 08:58:16 gerwinin: ...and the delivered bsp ;)))) Oct 21 08:58:47 hehe Oct 21 09:00:34 jin: and you get free indians with it :) who do not know how to get in the integrated mac working and when you do it they say you do it wrong :) Oct 21 09:01:12 hehe :))) Oct 21 09:02:36 well, I know, not that funny if you actually have to work on this and with them Oct 21 09:02:37 :> Oct 21 09:06:03 It is not funny at all , as long as they seem to have all an oscilloscoop which is way out of sync ;) Oct 21 09:06:34 :) Oct 21 09:19:42 gerwinin: btw just curious, which NXP hw did you use? Oct 21 09:20:03 hi florian Oct 21 09:20:42 yo woglinde & florian_ Oct 21 09:22:13 hi genesis Oct 21 09:23:57 good morning Oct 21 09:25:34 2378 Oct 21 09:28:55 not familiar with it, only had a little fun with the stb810 and stb225 Oct 21 10:03:49 03Stefan Schmidt  07org.openembedded.dev * rd05dca7b0d 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 21 10:03:49 03Stefan Schmidt  07org.openembedded.dev * r5d027e9e1f 10openembedded.git/conf/checksums.ini: checksums.ini: Add sums for hal 0.5.9 Oct 21 10:03:51 03Stefan Schmidt  07org.openembedded.dev * r14bf8220ea 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 21 10:03:54 03Stefan Schmidt  07org.openembedded.dev * r313753ac3e 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 21 10:03:56 03Stefan Schmidt  07org.openembedded.dev * r310ccc1071 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 21 10:16:03 morning all Oct 21 10:36:32 khem: can you commit yet? Oct 21 10:36:59 zecke: hi Oct 21 10:37:49 got your mail Oct 21 10:38:11 kalev: did you have push access before? Oct 21 10:38:16 zecke: have you added my key? Oct 21 10:38:35 zecke: no, I did not Oct 21 10:38:48 tmbinc: yes, I could have messed up (not that I have anything to believe that as I#m perfect) :) Oct 21 10:38:48 zecke: I thought Woglinde talked to you about that, at least he told me to send you the key Oct 21 10:39:06 kalev: wait a second :) Oct 21 10:39:36 tmbinc: can you git clone? git clone git@git.openembedded.net/openembedded? Oct 21 10:39:38 zecke: "git clone git@git.openembedded.net:openembedded" should work, yes? Oct 21 10:39:50 it asks me for a password Oct 21 10:40:03 tmbinc: hehe Oct 21 10:40:32 tmbinc: try again please (you could debug messages now) Oct 21 10:41:31 ssh debug? Oct 21 10:41:36 it just doesn't seem to like my key Oct 21 10:41:48 s/could/should get/ :) Oct 21 10:41:59 tmbinc: but feel free to pastebin your pub key again Oct 21 10:43:08 http://pastebin.ca/1232633 Oct 21 10:43:20 where should i get debug messages now? Oct 21 10:43:41 tmbinc: when you do git clone you should see gitosis messages Oct 21 10:44:18 no, all i see is ssh's "git@git.openembedded.net's password:" Oct 21 10:45:49 tmbinc: try again, your chat client wrapped the lines... and I messed up :) Oct 21 10:47:23 yay, works. thanks Oct 21 10:48:06 03Andreas Oberritter  07org.openembedded.dreambox * rf748502bf8 10openembedded.git/packages/projectx/projectx-0.90.4.00.bb: projectx: fix string comparison in .bb Oct 21 10:48:10 03Andreas Oberritter  07org.openembedded.dreambox * r63aee69781 10openembedded.git/packages/libcdio/libcdio_0.80.bb: libcdio: fix build dependencies Oct 21 10:48:11 03Andreas Oberritter  07org.openembedded.dreambox * rb8cbd5028b 10openembedded.git/packages/binutils/ (2 files in 2 dirs): binutils: fix possible buffer overflow in ar Oct 21 10:48:12 03Andreas Oberritter  07org.openembedded.dreambox * r8184cb7dd3 10openembedded.git/packages/gcc/ (3 files in 2 dirs): gcc: fix missing mode flags when opening response file Oct 21 10:48:12 03Andreas Oberritter  07org.openembedded.dreambox * r60ee2d806f 10openembedded.git/packages/gstreamer/gst-plugins.inc: gst-plugins: depend on flac to build gst-plugin-flac Oct 21 10:48:15 03Andreas Oberritter  07org.openembedded.dreambox * r34886ac0fc 10openembedded.git/packages/enigma2/enigma2.bb: enigma2: rdepend on gst-plugin-flac Oct 21 10:48:18 03Andreas Oberritter  07org.openembedded.dreambox * r9b7a0e6333 10openembedded.git/packages/images/dreambox-image.bb: dreambox-image: don't use bash extensions Oct 21 10:48:23 03Andreas Oberritter  07org.openembedded.dreambox * rbbcaf4aee6 10openembedded.git/classes/base.bbclass: remove oesources.org from list of mirrors Oct 21 10:48:26 03Andreas Oberritter  07org.openembedded.dreambox * r0f3677f28c 10openembedded.git/packages/dreambox/dreambox-secondstage.bb: dreambox-secondstage: DM8000 #66 Oct 21 10:48:29 03Andreas Oberritter  07org.openembedded.dreambox * ra0bfd41f85 10openembedded.git/packages/initscripts/ (2 files in 2 dirs): Oct 21 10:48:32 initscripts: update bootup script for DM8000 Oct 21 10:48:34 - don't modprobe lcd.ko (already loaded by rcS) Oct 21 10:48:36 - don't mount proc (already mounted by rcS) Oct 21 10:48:38 - use /dev/discs/disc0/disc instead of for-loop Oct 21 10:48:38 hehe Oct 21 10:48:40 - don't modprobe obsolete info.ko Oct 21 10:48:42 03Andreas Oberritter  07org.openembedded.dreambox * r2fb41ab6c7 10openembedded.git/packages/linux/linux-dm8000/dm8000_defconfig: linux-dm8000: enable CONFIG_VT for keyboard input Oct 21 10:48:51 03Andreas Oberritter  07org.openembedded.dreambox * r8e827dc111 10openembedded.git/packages/linux/linux-dm8000.bb: linux-dm8000: bump revision Oct 21 10:48:54 03Andreas Oberritter  07org.openembedded.dreambox * r6b4beddc76 10openembedded.git/packages/linux/linux-dm800/dm800_defconfig: linux-dm800: enable CONFIG_VT for keyboard input Oct 21 10:48:57 03Andreas Oberritter  07org.openembedded.dreambox * rbc63965c09 10openembedded.git/packages/linux/linux-dm800.bb: linux-dm800: bump revision Oct 21 10:49:02 03Andreas Oberritter  07org.openembedded.dreambox * r30e19b4177 10openembedded.git/conf/bitbake.conf: set sourceforge mirror to downloads.sourceforge.net Oct 21 10:49:05 03Andreas Oberritter  07org.openembedded.dreambox * recfd19c2c8 10openembedded.git/packages/dreambox/dreambox-dvb-modules.bb: dreambox-dvb-modules: update DM8000 drivers to 20081015 Oct 21 10:49:39 tmbinc: I'm happy to add obi's key too Oct 21 10:51:20 yeah, but he wanted me as a firewall against bogus commits. (not that I'm a good firewall...) Oct 21 10:51:57 zecke: You're the one to talk to regarding keys? Never got around to sending my key. Oct 21 10:52:00 yeah, i heard humans and fire don't mix well Oct 21 10:52:18 NAiL: send it to mickeyl, koen, rp, me, the first of us will add it Oct 21 10:52:36 zecke: seems my key works, at least I could check out using git@ Oct 21 10:52:38 thanks Oct 21 10:53:07 kalev: welcome, keep in mind it is a privilege :) Oct 21 10:53:14 yes Oct 21 10:55:15 tmbinc: we can create obi/org.openembedded.dreambox.firewall for him :) Oct 21 10:55:45 there's the MAINTAINERS file. Should I add an entry for myself and put those new recipes I've added so far into the list? Oct 21 10:55:56 kalev: if you want to Oct 21 10:59:40 zecke: How can I convince you to give commit-access to another developer? :) Oct 21 10:59:59 tmbinc: is there any plan to merge the dreambox branch into dev? Oct 21 11:00:20 NAiL: we have a process for that. propose the developer get a second for that Oct 21 11:00:28 johncylee: hey! Oct 21 11:00:34 johncylee: we need to talk Oct 21 11:01:03 zecke: sure. what?> Oct 21 11:02:34 thesing: we will "soon" get rid of what's currently in the dreambox branch, and start with a version derived from a current .dev. It's not yet working. Then we intend to minimize the differences between those branches by porting stuff from .dreambox to .dev, but a real merge will probably not happen soon, unless somebody fixes all the evil hacks we did. (actually, there are not too many anymore) Oct 21 11:04:09 great. Oct 21 11:04:10 thesing: so far we have http://git.opendreambox.org/?p=openembedded.git;a=summary, but at least for now it's more a sandbox for us, that's why we want to have that repository separate. we are all git-noobs. Oct 21 11:04:53 johncylee: how you feel, how you sleep :) Oct 21 11:24:33 zecke: oh I feel good~ Oct 21 11:24:40 zecke: and sleep well Oct 21 11:24:57 zecke: Should we just get on and merge that FILE_PR branch? Oct 21 11:25:22 zecke: (thanks for the aquavit you give me) Oct 21 11:25:42 johncylee: giving? I just forgot it :) Oct 21 11:26:27 RP: yes, maybe do another rebase to catch all FILE_PR's again? I can do this after class Oct 21 11:26:45 johncylee: I hope you guys enjoy the aquavit Oct 21 11:48:03 morning Oct 21 11:50:24 hrw: morning Oct 21 11:53:48 someone tried 'bitbake world' recently? its mess Oct 21 11:55:30 http://www.webitpr.com/release_detail.asp?ReleaseID=10258 Oct 21 11:59:04 zecke: ok. I'm a little caught up in other things atm :/ Oct 21 12:00:20 hrw: seems like a zaurus 3100 Oct 21 12:01:26 thesing: bit better, bit worse Oct 21 12:01:39 thesing: wireless in but qvga screen Oct 21 12:02:11 and no internal hd and no cf slot. Oct 21 12:02:20 I think zaurus wins. Oct 21 12:03:01 and a 2.4.19 kernel... Oct 21 12:04:10 with old qtopia Oct 21 13:40:44 is there a resaon why the openembedded X11 recipie does not cover GLX? Oct 21 14:04:31 cpo_: Traditionally there hasn't been much embedded glx hardware. See poky for glx capable recipes Oct 21 14:04:38 hi rp :) Oct 21 14:06:21 ph.ph.cox.net Oct 21 14:06:22 hm Oct 21 14:06:27 this looks like an error Oct 21 14:07:05 ~curse BB_ENV_WHITELIST stuff Oct 21 14:07:05 May the fleas of a thousand camels infest your most sensitive regions, BB_ENV_WHITELIST stuff ! Oct 21 14:07:46 my buildbot is capable of making Poky builds but OE fails on environment Oct 21 14:09:35 time to rewrite auto.conf on each machine change :( Oct 21 14:14:15 hi steliosk Oct 21 14:17:24 uf. finally buildbot for angstrom started Oct 21 14:19:25 now my machine covers Poky and Angstrom builds Oct 21 14:20:52 hi robert Oct 21 14:22:05 bye guys Oct 21 14:22:10 bye hrw Oct 21 14:23:02 hi woglinde Oct 21 14:23:09 he kalev Oct 21 14:23:32 I've got my own push access now, so I won't ask you to commit my changes any more Oct 21 14:23:37 thanks for all the help Oct 21 14:24:45 kalev *g* I was next to zecke as he added your key Oct 21 14:27:38 kalev: btw. why did you get access so fast? I thought it should involve mails to ML Oct 21 14:30:14 maybe because I've been contributing the project for a while Oct 21 14:36:05 hehe Oct 21 14:37:44 my x server was compiled without --disable-glx but still i get the __glxActive_Screens missing symbol error when starting up x Oct 21 14:37:54 anyone got an idea? Oct 21 14:49:41 kalev: Don't get me wrong. We always need more contributers. I just thought that was some more complicated voting involved or so. Oct 21 14:50:00 kalev: what areas in oe are you working on? Oct 21 14:50:41 thesing : http://aldebaran-robotics.com/download/NaoAcademicsV3.mov ( one year i work there on OE ) Oct 21 14:53:05 thesing: we have our own OE based distro, which we maintain separately. It runs on embedded x86 hardware. So far we're been quietly doing it on our own, but now it's time to give something back to the community :) Oct 21 14:54:00 thats good. I think x86 could use some work in oe ;) Oct 21 14:54:15 * Genesis works on x86 target too Oct 21 14:54:35 I think I'll try to get xorg drivers into a better shape next Oct 21 14:54:45 a lot of them don't even compile now Oct 21 14:54:54 anybody working with a mips target? Oct 21 14:56:46 Genesis: this robot is powered by oe? Oct 21 14:57:02 x86 works great in OE Oct 21 14:57:36 ot maybe anyone know how to build the libglx.so file Oct 21 14:57:42 s/ot/or/ Oct 21 15:04:35 Genesis: really cool video. Did the blue robot lie to the other about the duck? Scary :) Oct 21 15:05:15 ^^ Oct 21 15:05:25 thesing : yes of course=) Oct 21 15:06:17 Genesis: but those robots are not for general sale yet, are they? Oct 21 15:06:23 no no Oct 21 15:06:28 academic sales Oct 21 15:36:48 re Oct 21 15:53:41 hi sweetlilmre Oct 21 15:53:43 woglinde: you around? Oct 21 15:53:46 :) Oct 21 15:53:48 lol Oct 21 15:53:48 thesing: yes, we just got some mips stuff for playing around Oct 21 15:54:28 woglinde: I got an open tyrian (vertical SDL shooter) recepie working under OE! :) Just some weird dependency issues Oct 21 15:55:31 woglinde: it depends on libsdl-net which is in DEPENDS, but does not get added to the image unless I specify it explicitly. I thought libs got added automatically? Oct 21 15:56:26 hm Oct 21 15:56:30 woglinde: tried RDEPENDS, but that didn't work either Oct 21 15:56:39 florian: do you build for them regularly? Oct 21 15:57:06 sweetlimre hm could be that the image or task does seomthing strange Oct 21 15:58:02 sweetlimre but I do not know exactly, had to have to test myself Oct 21 15:58:37 woglinde: I will ask in #beagle as it is koen's demo image and I have had other similar issues Oct 21 15:59:06 thesing: not yet, but I know that a colleague had quite good success with a mips (big endian) target lately. Oct 21 15:59:09 if it stuff gets automatic into an image Oct 21 15:59:09 args Oct 21 15:59:09 my english is this bad Oct 21 15:59:52 dreamboxes are mips Oct 21 15:59:57 mostly Oct 21 16:02:53 so Oct 21 16:02:58 I have to shop some bread Oct 21 16:03:01 till later Oct 21 16:09:12 woglinde : hi ! Oct 21 16:35:16 thesing: I once had added qemu mips in .dev never committed that stuff. I do not have any mips hardware Oct 21 16:35:45 I was thinking of adding all possible qemu targets Oct 21 16:36:09 thats a great idea. Oct 21 16:37:01 we could also have some scripts that run those images automatically. Oct 21 16:37:26 so one could run "bitbake console-image -c run-qemu" Oct 21 16:38:24 zecke: hail zecke Oct 21 16:56:25 if I want to include a module in a kernel, can I do a 'bitbake virtual/kernel -c devshell' and then a 'make menuconfig' to change kernel options? Oct 21 16:56:44 yes, that should work fine Oct 21 16:57:46 sweetlilmre: what is wrong with bitbake -c menuconfig virtual/kernel (or config) Oct 21 16:58:10 sweetlilmre: bitbake -c listtasks (or list_tasks should list you the available tasks) Oct 21 16:58:29 zecke: this is why I ask in this channel, better answers :) Oct 21 16:59:00 :) Oct 21 16:59:08 thanks. looking at it now :) Oct 21 16:59:49 heh, there's a -c menuconfig now eh, that's nifty Oct 21 17:00:56 0926149ccd7fb1fbf2810a04a9929df3e1266ed4 RP added it :) Oct 21 17:01:37 this oe stuff just gets better and better. in fact its almost converted me to Linux as a desktop I have learnt so much! Oct 21 17:03:09 hehe Oct 21 17:05:17 cute Oct 21 17:05:40 man, I have a hard enough time wrapping my head around OE as a fairly experienced desktop linux user. I can't imagine jumping into it from the windows world Oct 21 17:06:52 hi kergoth Oct 21 17:07:29 * kergoth wonders how things should have been done to make oe easier to grasp, conceptually Oct 21 17:07:31 hey pb__ Oct 21 17:07:41 slothlove: I love a challenge :) Oct 21 17:08:51 kergoth: I think a gui for configuration would have made the trick. Oct 21 17:09:26 we did have one of those at one point, but it doesn't really solve the intellectual difficulty of understanding how the thing works. Oct 21 17:09:51 in fact, it probably obfusticates it Oct 21 17:09:54 oecommander or something, iirc. Oct 21 17:09:55 true, but it simplifies the setup. Oct 21 17:10:16 kergoth: It hasn't really been THAT hard, I've been at it about a week in my spare time and I've been able to get the cross-compiling environment set up and I'm working creating some custom packages. Mostly just figuring out how bitbake does things, how the directory tree is structured, going from svn to git, that sort of thing Oct 21 17:10:24 thesing: right. that's a different problem though. Oct 21 17:10:25 if you have something running its easier to learn how it works. Oct 21 17:10:47 I disagree Oct 21 17:10:49 being able to get it out of the box and have it running quickly is one thing. understanding how it works in order to modify/maintain it is a different thing. Oct 21 17:10:52 if you learn how it works, its easier to make it run Oct 21 17:11:10 slothlove: I find OE relatively simple compared to the insane complexities of git... Oct 21 17:11:23 yes but you can't learn how it works unless you make it run Oct 21 17:11:33 untrue Oct 21 17:11:38 git's very simple actually, at least compared to a lot of the other scms around Oct 21 17:11:42 go play with clearcase for a while Oct 21 17:11:43 then we'll talk Oct 21 17:12:37 kergoth: I only have svn experience, and it doesn't get simpler than than (I have heard of clearcase nightmares tho...) Oct 21 17:12:50 *than that Oct 21 17:13:22 the only thing that i'd see as being even somewhat difficult about git is the whole concept of the index. once you've wrapped your head around that, the rest is cake Oct 21 17:13:24 mm, cake Oct 21 17:14:04 kergoth: I need a git for dummies tut, or a very patient individual to walk me through it Oct 21 17:14:33 kergoth: e.g. why after a clone do you do a checkout? Oct 21 17:14:38 i think oe would have done better to be slightly less flexible. -everything- being in the metadata, coupled with the lack of enforcement of metadata policy and the whole layered nature of everything (overrides, collections, etc).. there's too many ways things could be done, and too little enforcement and documentation of the way things are done Oct 21 17:14:40 sweetlilmre: what? Oct 21 17:15:02 a clone == git init, git remote add, git fetch, then creation of a local branch tracking the remote, and a checkout of that branch Oct 21 17:15:12 checkout is after the remote content is fetched, not before Oct 21 17:15:38 sweetlilmre: http://book.git-scm.com/ seems like a good start on introductory material Oct 21 17:15:42 sweetlilmre: maybe try that out Oct 21 17:16:00 thanks will check it out Oct 21 17:16:11 kergoth: I really don't think thats the problem. There are really many examples to understand these. Oct 21 17:16:32 you're familiar with svn, right sweetlilmre ? Oct 21 17:16:40 clone is like downloading the remote svn repo, the one in /var/lib/svn Oct 21 17:16:46 and then checking that out into a local working copy Oct 21 17:17:00 i think you need to grasp the whole distributed scm thing Oct 21 17:17:40 thesing: the examples don't let you grasp the system in its entirety though, they just show you a particular piece of syntax Oct 21 17:17:51 * kergoth shrugs Oct 21 17:20:14 yes but they are enough to get started. For the remaining stuff on can ask on irc ;) Oct 21 17:21:23 kergoth: yep, its the distributed / disconnected model that is so foreign. I have a mate at work who is also interested. I think we will go and blow up some test repos to figure this out. Learn by doing is the only real way Oct 21 17:21:35 there have always been problems with people grasping the system, conceptually. the number and consistency of these problems implies a problem with the design, imo Oct 21 17:21:42 re Oct 21 17:21:42 sweetlilmre: heh, good plan :) Oct 21 17:21:50 &away Oct 21 17:21:51 i always end up doing that with a new scm tool, just try out various test workflows Oct 21 17:21:53 ups Oct 21 17:22:06 woglinde: got the bread? :) Oct 21 17:22:22 sweetlimre sure Oct 21 17:22:51 now opened me a beer Oct 21 17:24:13 ~seen zecke Oct 21 17:24:14 zecke is currently on #oe (46m 13s). Has said a total of 5 messages. Is idling for 21m 5s, last said: 'hehe'. Oct 21 17:26:47 woglinde: ja? Oct 21 17:27:05 zecke how was your talk? Oct 21 17:27:20 woglinde: my computer didn't work, but I think it was okay Oct 21 17:27:30 aus dem naehkaestchen geplaudert Oct 21 17:27:30 hu? Oct 21 17:27:35 didnt work? Oct 21 17:27:43 hm Oct 21 17:27:44 okay Oct 21 17:27:56 woglinde: X is busted... didn't initialize the screen properly Oct 21 17:28:02 woglinde: and then didn't find the projector Oct 21 17:28:28 haha Oct 21 17:28:32 bloody mac Oct 21 17:29:07 woglinde: I like the intel gfx driver's Oct 21 17:33:27 hi gregoiregentil Oct 21 17:37:31 Hello woglinde! Oct 21 17:38:20 Woglinde: would you have any recommendation for http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-October/006599.html? Oct 21 17:40:11 gregoiregentil hm you do not need pkgconfig anymore Oct 21 17:40:18 autotools alreadty takes care of it Oct 21 17:41:33 ok. actually, I tested with autotools only and it was the same result. so I agree. Oct 21 17:41:59 I am just looking what are the depends really for it Oct 21 17:42:03 nevertheless, I still need to modify the firefox recipe to produce those missing files, especially the .pc. How to do that? Oct 21 17:42:04 thanks Oct 21 17:42:46 okay gettext Oct 21 17:43:02 args Oct 21 17:43:09 AM_GLIB_GNU_GETTEXT Oct 21 17:43:14 I hate this sucking macro Oct 21 17:43:43 hm I think xulrunner should be enough Oct 21 17:44:34 I was afraid of xulrunner but it means to completely modify Firefox recipe to use xulrunner compiled engine Oct 21 17:44:41 s/but/because Oct 21 17:44:56 hi all! Oct 21 17:45:12 How do I fix the error message "| checking LD_LIBRARY_PATH variable... contains current directory Oct 21 17:45:13 | configure: error: Oct 21 17:45:15 | *** LD_LIBRARY_PATH shouldn't contain the current directory when Oct 21 17:45:16 | *** building glibc. Please change the environment variable Oct 21 17:45:18 | *** and run configure again. Oct 21 17:45:19 | FATAL: oe_runconf failed"? Oct 21 17:45:53 unset LD_LIBRARY_PATH? Oct 21 17:47:08 hmm, just export LD_PATH to an empty string? Oct 21 17:47:38 why not using 'unset'? Oct 21 17:50:36 hmm Oct 21 17:51:18 Is there anyway to use the packages I already downloaded inside my source directory instead of redownloading them? Oct 21 17:51:44 It took quite a while to download them again.l Oct 21 17:51:55 zecke: I did 'bitbake -c menuconfig virtual/kernel' added the module I needed, but bitbake virtual/kernel after that does nothing? i.e. it doesn't seem to pick up the change Oct 21 17:53:22 gregoiregentil hm seems we have no xulrunner in oe yet Oct 21 17:54:01 exactly, it's why I would like to create firefox-xpcom.pc instead of xulrunner-xpcom.pc. If you take a look at the gecko-mediaplayer configure, Oct 21 17:54:23 it can use both. it would be much easier just to create the firefox-xpcom.pc file Oct 21 17:54:45 most of the headers are already exported. what i don't know to do, is how to create the pc file? Oct 21 17:54:47 gregoiregentil yes I already did Oct 21 17:55:00 should I do it "manually"? Oct 21 17:55:09 you already did? Meaning? Oct 21 17:55:10 but firefox abuses autotools and relays on autoconf-2.13 Oct 21 17:55:32 gregoiregentil read the configure.in of gecko-media Oct 21 17:56:39 sweetlilmre: force a compile? Oct 21 17:57:07 kergoth: I think the main problem with understanding the oe repository is the weird mixture of apparently static, declarative things and dynamic, mutable variables, coupled with the fact that the processing order for various things is a little bit unpredictable without knowing all the details of how bitbake operates. Oct 21 17:57:42 kergoth: for example, the long-standing thing where "FILES_${PN}" and "FILES_foo" are not exactly synonyms for package foo, although it appears at first glance like they ought to be. Oct 21 17:57:54 zecke: I tried a -f -c build, still nothing Oct 21 17:58:16 unless you meant -f -c compile? Oct 21 17:58:27 sweetlilmre: -f -c compile :) Oct 21 17:58:36 :) Oct 21 17:58:39 kergoth: in fact, I guess the way that overrides operate are at the root of a lot of the misunderstandings. Oct 21 17:58:55 Is there anyway to use the packages I already downloaded inside my source directory instead of redownloading them? Oct 21 17:59:11 hubar???? Oct 21 17:59:16 hubar: DL_DIR? Oct 21 18:00:01 hmm? Oct 21 18:00:15 let me try DL_DIR :) Oct 21 18:00:26 you have one in your local.conf Oct 21 18:00:40 and thats where all downloaded files are put Oct 21 18:00:54 hubar: the only reason I can think of that it is redownloading them is that OE conf is looking for a different dir, which is specified by DL_DIR Oct 21 18:00:55 bbiab Oct 21 18:01:15 pb_: that's a good point. overrides was always a bit odd, both syntactically and when it affects things. _prepend/_append's evaluation time was a part of all that mess too. it's pretty obvious i started going from a gmake like syntax, not exactly one thing or another Oct 21 18:01:22 hubar: default to TMPDIR/download afaik Oct 21 18:01:26 ah well Oct 21 18:01:39 so you mean OE won't try to download new files if they are already downloaded? Oct 21 18:01:46 * kergoth jots down pb's comments for future reference in case he comes across the necessary motivation to do something about it Oct 21 18:01:56 hubar: yep, if hashes match etc. Oct 21 18:02:12 hubar: *warning* I am a n00b here.. :) Oct 21 18:02:17 alright! :) Oct 21 18:02:27 I am a bigger n00b then you :) Oct 21 18:02:36 kergoth: right, the _prepend/_append thing is also a fertile source of confusion, though luckily oe doesn't use it much anymore. Oct 21 18:02:43 * kergoth nods Oct 21 18:02:50 hubar: which install guide did you follow? Oct 21 18:03:11 pb_ hm what to use instead? Oct 21 18:03:31 if I was designing a putative "oe 2.0" kind of format then I think I would have conditional declarations be directly supported in the metadata Oct 21 18:03:36 sweetlilmre: wiki. Oct 21 18:03:55 pb_: I thing in some cases one has to use _append because += doesn't work. Oct 21 18:03:58 hubar: err, yeah, but which page, and what is your target system? Oct 21 18:04:07 pb_: i think it was a mistake to let 'inherit' be order dependent from the start. it acted too much like include Oct 21 18:04:26 thesing: right, but that is just another symptom of the underlying problem. Oct 21 18:04:37 kergoth: yah, exactly Oct 21 18:04:38 to point out a specific example of that sort of format confusion.. are we declarative or not Oct 21 18:05:41 quite. Oct 21 18:06:11 sweetlilmre: http://wiki.openembedded.net/index.php/Getting_Started Oct 21 18:06:22 sweetlilmre: targetting system is x86. Oct 21 18:06:41 right. This gave me some headache. Oct 21 18:07:03 I first thought that order of declaration doesn't matter in bbfiles. Oct 21 18:07:51 thesing nope Oct 21 18:07:58 inherit it matters Oct 21 18:09:03 yes. if you want to declare sth. that is used by the class you inherit you have to declare it before the inherit stmt. Oct 21 18:09:57 thesing not only that Oct 21 18:10:06 if you inhert foo1 foo2 Oct 21 18:10:17 and you declare the same variable in both Oct 21 18:10:21 the last one is taken Oct 21 18:10:29 depends on when the class's functionality is evaluated, which is what pb was pointing out, processing order ..it might not need the variables until task executuion, or it might be used in an immediately exapnded variable or anonymous function Oct 21 18:10:29 what LICENSE is this: http://pastebin.ca/1232999 Oct 21 18:10:38 wondering what to put in the OE recipe Oct 21 18:11:08 smells like MIT or X11, but i'm rusty Oct 21 18:11:31 check the osi page Oct 21 18:11:38 kergoth: ok, thanks Oct 21 18:11:43 hm hplip package Oct 21 18:11:51 woglinde: yup Oct 21 18:11:56 woglinde: you ever work with it? Oct 21 18:12:50 cbrake http://hplipopensource.com/node/296 Oct 21 18:13:06 cbrake if you have a printer from hp Oct 21 18:13:12 you will not live without it Oct 21 18:14:26 woglinde: I am working on a scaled down printing solution for an embedded system Oct 21 18:14:59 woglinde: hoping to print more or less directly from a cario surface to the HP IJS driver and skip most of hplip Oct 21 18:15:05 woglinde: and postscript Oct 21 18:15:43 zecke: that fixed it, and I now have my module :) thanks! Oct 21 18:15:51 cbrake hm yes Oct 21 18:16:05 maybee you only need the cups lpr command line Oct 21 18:16:05 gn all, thanks for the help Oct 21 18:16:10 instead of the full server Oct 21 18:18:33 woglinde: yeah, I've been debating. You still have to have a filter of some type that transforms a raster image to a printer compatible format Oct 21 18:18:55 woglinde: some I'm not sure how to eliminate the IJS part Oct 21 18:19:04 cbrake hm sorry Oct 21 18:19:18 I only do apt-get install hplip sofar Oct 21 18:20:06 woglinde: yeah, I wish I could do that here. It gets pretty hairy under the hood :-) Oct 21 19:05:46 re Oct 21 19:07:24 re hrw Oct 21 19:13:29 he mickeyl Oct 21 19:14:40 yo Oct 21 19:20:04 yo mickey|munich Oct 21 19:22:50 http://tinderbox.openembedded.net/builders/hrw-buildbot-angstrom/ Oct 21 19:23:39 this is my buildbot which will build Ångström images (base/console/x11) for qemuarm/qemux86/om-gta01/c7x0 Oct 21 19:24:09 I plan to have full build (2-3 times per week) and incremental (daily) Oct 21 19:28:32 Does angstrom target x86? Oct 21 19:31:44 re Oct 21 19:31:48 re florian Oct 21 19:34:50 03Koen Kooi  07org.openembedded.dev * rf685a83827 10openembedded.git/conf/checksums.ini: checksums.ini: add more checksums Oct 21 19:35:18 hubar: it does Oct 21 19:35:43 hubar: angstrom targets arm, powerpc, x86 and can target other archs as long as OE supports it Oct 21 19:40:18 hrw: thanks, I didn't look in http://www.linuxtogo.org/gowiki/OeDeviceList only saw the much shorter list on angstrom's wiki. :) Oct 21 19:40:22 my bad Oct 21 19:41:22 is there an easy way to get bitbake to only see the important core packages instead of all 6067? I'm usually only compiling super-minimal images. Oct 21 19:42:14 slothlove: there is BBMASK were you can mask out packages. Oct 21 19:46:28 thesing: can I mask out whole sets of packages, or somehow mask everything except dependencies for a certain package like minimal-image? Oct 21 19:47:06 slothlove: I never used it. I think you can enter a regular expression of packages you don't want. Oct 21 19:47:46 slothlove: another solution is to delete the packages you don't want ;) Oct 21 19:49:19 if I'm just building a single package can I skip the dependency checking? Oct 21 19:55:52 n/m figured it out. "bitbake -b" Oct 21 19:57:03 hrm... why does my build error out complaining about a lacking checksums.ini entry?! Oct 21 19:57:20 I just got the same error Oct 21 19:57:33 flo_lap: you need to set OE_ALLOW_INSECURE_DOWNLOADS = "1" Oct 21 19:57:55 pb_:thanks Oct 21 19:58:06 florian / vs. // ? Oct 21 19:58:10 pb_: thanks... was this on the list lately? Oct 21 19:58:17 yeah, about a week ago Oct 21 19:58:23 * flo_lap feels old Oct 21 19:58:49 the "secure downloads" feature used to be off by default, but zecke changed that to on. Oct 21 19:58:59 hehe Oct 21 19:59:07 he forces us to fill checksum.ini Oct 21 19:59:33 well, or just to turn it back off again :-} Oct 21 19:59:55 I guess I already have enough OE_UNBREAK_MY_BUILD = "1" stuff in local.conf that one more variable doesn't matter. Oct 21 19:59:56 heh Oct 21 20:00:23 * flo_lap is not sure if this is a good idea at all Oct 21 20:00:41 checksums.ini, you mean? Oct 21 20:01:09 dammn. Oct 21 20:01:12 NOTE: fetch http://tinylogin.tinylogin.net/downloads/tinylogin-1.4.tar.bz2 Oct 21 20:01:13 --15:53:37-- http://tinylogin.tinylogin.net/downloads/tinylogin-1.4.tar.bz2 Oct 21 20:01:15 pb_: no, enforcing the global checksums.ini, I do not see any use in it Oct 21 20:01:15 => `/WorkSpace/oe/sources/tinylogin-1.4.tar.bz2' Oct 21 20:01:17 Resolving tinylogin.tinylogin.net... failed: Name or service not known. Oct 21 20:01:54 * pb_ dinner time now Oct 21 20:01:55 bbia Oct 21 20:01:56 b Oct 21 20:02:27 till later pb Oct 21 20:02:29 oe error messages are way too bad anyway. we need to get rid of this mess and get a single and clear error message. Oct 21 20:02:33 pb_: enjoy Oct 21 20:02:55 hmm how do I fix it? the correct place to download tinylogin should be at http://tinylogin.busybox.net/ Oct 21 20:03:26 hubar edit the .bb Oct 21 20:03:41 and look if a bugreport is in the bugtracker for it Oct 21 20:03:44 alright, let me try :) Oct 21 20:03:44 and if not Oct 21 20:03:46 file a bug Oct 21 20:04:50 Is this considered a bitbake bug or OE bug? Oct 21 20:05:27 oe bug Oct 21 20:08:17 I posted the patch, good positive replies, waited :) Oct 21 20:08:52 zecke which patch? Oct 21 20:09:20 enforcing to have a checksum Oct 21 20:10:00 I am fine with it Oct 21 20:15:26 03Rodrigo Vivi  07org.openembedded.dev * r132645ba13 10openembedded.git/MAINTAINERS: MAINTAINERS: add my entry Oct 21 20:25:11 bye Oct 21 20:30:38 hi mickey|munich ! Oct 21 20:30:51 hey gremlin[it] Oct 21 20:30:55 how are things? Oct 21 20:31:12 about good ... Oct 21 20:31:35 i'm preparing for LinuxDay in italy Oct 21 20:32:12 mhh question, what can i tell to people (i have a booth and will do a talk) about OE foundation ? Oct 21 20:32:12 cool! Oct 21 20:32:30 coming real soon now. Oct 21 20:32:45 we've been requested to change some statutes Oct 21 20:32:52 it exist and i'm a member ? :) Oct 21 20:32:52 so we did Oct 21 20:33:06 it's still pending legal approval Oct 21 20:33:11 but we're close now Oct 21 20:33:47 mhh ok ... Oct 21 20:34:01 what about u ? Oct 21 20:34:54 everything fine here Oct 21 20:35:02 just came back from 3 weeks taipei Oct 21 20:35:14 pushed some things forward wrt. Openmoko Oct 21 20:35:14 work or vacation ? Oct 21 20:35:20 work of course :) Oct 21 20:35:21 vacation? :) Oct 21 20:35:23 hahah :) Oct 21 20:47:26 re Oct 21 20:51:57 hmm where do I find the image after OE is done building? Oct 21 20:52:26 And how do I start it? Oct 21 20:52:46 tmp/deploy/... Oct 21 20:53:11 hm google made small c library Oct 21 20:54:04 for android? Oct 21 20:54:14 crazy google, you would think there were enough of those already Oct 21 20:54:14 pb yes Oct 21 20:54:20 http://android.kernel.org/?p=platform/bionic.git;a=tree;f=libc;h=142272f2219af2e7bdea5457b2869265b99fa598;hb=HEAD Oct 21 20:54:40 only for x86 and arm Oct 21 20:55:43 mm, seems like madness. Oct 21 20:55:57 pb_: own libstdc++, supposingly "faster" pthreads Oct 21 20:56:21 pb_: some API additions... yeah one can not add another lib... getproperty, LOGGING (what is wrong with syslog one might ask) Oct 21 20:56:49 zecke: ah, a small libstdc++ sounds like a good idea. Oct 21 20:56:51 pb_: their syscall glue is autogenerated but I couldn't find their python script... Oct 21 20:57:18 hm the sanitized kernel headers Oct 21 20:57:18 faster pthreads is probably feasible but you would have thought it would be easier to do that on top of one of the existing libraries. Oct 21 20:57:18 pb_: yes, but I didn#t look their... maybe this is only for webkit, which means std::min and std::max :) Oct 21 20:57:19 so they can directcly included by user space Oct 21 20:57:25 http://android.kernel.org/?p=platform/bionic.git;a=blob;f=libc/kernel/README.TXT;h=da8c5d8b6a1ab673f82eeebfef6a0d3d15be39c2;hb=HEAD Oct 21 20:57:29 pb_: their goal is to have no GPL in userspace... :( Oct 21 20:57:30 as for autogenerated syscall glue, heh, we have had that in glibc since about 1994. Oct 21 20:58:17 ah, I see Oct 21 20:58:19 pb_: instead they have a "hal" that allows to implement wifi, audio, foo drivers in userspace Oct 21 20:58:22 do they object to lgpl as well Oct 21 20:58:23 ? Oct 21 20:58:31 *GPL* Oct 21 20:58:43 right Oct 21 20:58:44 what the fuck Oct 21 20:58:51 libstdc++ is only 4 files Oct 21 20:59:01 heh Oct 21 20:59:10 I always thought the standard libstdc++ was bloated Oct 21 20:59:35 hm new is return malloc(size); Oct 21 21:00:18 in last months (6 about) i spend more time talking about OE than using it ... damn ... Oct 21 21:01:17 pb_: their is some reason behind androids libc Oct 21 21:01:39 pb_: basicly every app in the system is linking only to that piece of library, and all their userspace is supposed to be java Oct 21 21:01:57 pb_: so keeping their "legacy" small (no private data or such) is a reasonable goal Oct 21 21:02:06 pb_: obviously this should be possible within glibc/eglibc :) Oct 21 21:02:14 oh libm they stole from freebsd Oct 21 21:02:24 zecke: yah, given their antipathy towards gpl I guess that what they have done is reasonable. Oct 21 21:02:37 woglinde: hehe, their dns code is from netbsd Oct 21 21:03:08 zecke: if you were just starting from the perspective of wanting a small libc to provide minimal system glue, and didn't have a chip on your shoulder about gpl, then something like dietlibc or klibc would probably be a better bet than inventing your own. Oct 21 21:03:09 hm no log2 function Oct 21 21:03:23 that means qte will not be buildable Oct 21 21:03:32 but, if no gpl is the overriding concern, porting the bsd libc is not an unreasonable way to proceed Oct 21 21:04:12 hi? Oct 21 21:04:34 obviously you're right that everything is possible with glibc though :-) Oct 21 21:04:41 I'm trying to follow the how-to on http://www.oesf.org/index.php?title=Q18:_I_am_using_a_C3x00%2C_and_I_want_to_use_kernel_2.6.23.1_yonggun_with_uboot._How_can_I_install_Debian_EABI_on_the_internal_drive%3F Oct 21 21:05:00 oh no Oct 21 21:05:01 but the download link for 'zaurus-debian-eabi-afterinstall_1.1.tar.gz' is unavailable Oct 21 21:05:05 oe special forces again Oct 21 21:05:15 fosco: sorry, we can not help Oct 21 21:05:55 pb_: have you ever toyed with interpreters? know computed goto? Oct 21 21:05:58 http://discuz-android.blogspot.com/2008/10/google-android-native-libc-bionic.html Oct 21 21:06:03 it's not your site, that's it? :) Oct 21 21:06:22 fosco: it is not our site, it has nothing to do with us besides using our name Oct 21 21:06:59 ok, sorry for beeing annoying you, as a lot of other are doing Oct 21 21:07:20 don't worry Oct 21 21:07:34 pb_: "It doesn’t support certain POSIX features, like C++ exception" I knew C++ is part of posix... :} Oct 21 21:07:39 zecke hm why is webkit working with bionic? Oct 21 21:08:09 zecke: I don't know a huge amount about interpreters, though I do know what computed goto is. why do you ask? Oct 21 21:08:26 zecke: heh, all sorts of things are part of posix Oct 21 21:08:44 (even, maybe someone who have a validated account on their website could be kind enough to download the file for me?) Oct 21 21:08:56 you could probably have bitbake made into a posix standard if you tried hard enough Oct 21 21:09:31 pb_: hehe, only if POSIX.23 is still available Oct 21 21:09:56 pb_: dalvik has a nice arm variant for the dispatch Oct 21 21:10:07 pb_: mostly one memory read Oct 21 21:10:19 hm Oct 21 21:10:34 why they didnt use l4 kernel for somewhat Oct 21 21:10:44 woglinde: android.kernel.org... :) Oct 21 21:11:00 woglinde: they couldn't buy the whole domain to make it host l4 :) Oct 21 21:11:25 heh Oct 21 21:11:44 it is a bit sad that all our kernels are belong to google. Oct 21 21:11:47 * zecke seriously wonders how kernel.org can have such a subdomain Oct 21 21:12:11 hm they paid Oct 21 21:12:13 I guess Oct 21 21:13:35 re florian Oct 21 21:15:52 zecke: mm, I wonder why they don't just use jazelle Oct 21 21:16:00 I guess they have their reasons Oct 21 21:16:35 ah good, the cat has gotten bored of sitting on my lap Oct 21 21:17:52 hehe Oct 21 21:19:02 I have no idea how jazelle work and I think linux drivers will behard (license wise)? and I'm glad this way dalvik will work on armv4 Oct 21 21:20:46 pb_: I'm disappointed they did not write a new c/c++ compiler Oct 21 21:21:58 yeah, I guess armv4 compatibility is a good thing, though armv5tej is fairly widespread nowadays. Oct 21 21:22:24 jazelle support in linux is an interesting question, I have no idea what the status is with that. Oct 21 21:22:28 surely someone must have done it. Oct 21 21:22:59 zecke: is your gta02 still using arm920t, or does that have a newer cpu? Oct 21 21:23:45 pb_: I think the s3c2442 is an arm920t Oct 21 21:24:17 pb_: we look into the s3c6410 for future devices Oct 21 21:24:29 ah, right Oct 21 21:24:46 s3c2412 is arm926ej-s, and I had kind of assumed s3c2442 was as well. Oct 21 21:25:05 but I don't have any reason for thinking that other than the similarity in numbers :-} Oct 21 21:25:56 s3c6410 seems like a nice cpu. that's arm1176jzf-s, right? Oct 21 21:27:32 you can get jazelle in linux... if you pay ARM! :) Oct 21 21:27:40 $$$$ Oct 21 21:28:31 pb_: I still need to learn a lot about the actual arm core's and their features... Oct 21 21:28:54 pb_: I think it is armv6 or even armv7 (???), has vfp, and let me google Oct 21 21:29:03 zecke armv6 Oct 21 21:29:15 with arm trustzone Oct 21 21:29:21 arm1176*-s is armv6 plus trustzone Oct 21 21:30:03 the f is for the optional vfp Oct 21 21:30:07 wow Oct 21 21:30:12 this looks a nice site Oct 21 21:30:14 http://www.pdadb.net/index.php?m=cpu&id=a6410 Oct 21 21:31:43 iirc, arm1176 was the first cpu that we ever ran nptl on natively. Oct 21 21:31:43 heh Oct 21 21:32:08 pb_ hm??? Oct 21 21:35:14 woglinde: when we first started porting nptl to arm, it needed a tls register to operate because the kernel emulation didn't exist yet. Oct 21 21:35:40 pb sorry that I ask what is the tls register for? Oct 21 21:36:27 it holds the pointer to the private data for each thread. Oct 21 21:36:50 woglinde: tls == thread locale storage Oct 21 21:36:53 on arm, it lives in cp15 somewhere. I forget exactly where. Oct 21 21:36:57 local even Oct 21 21:36:58 there is privacy in threads? ;) Oct 21 21:37:02 * florian *hides* Oct 21 21:37:18 zecke: right, thread locales would be something altogether different Oct 21 21:37:33 I'm not quite sure what that would be. Oct 21 21:37:40 pb_: I have good excuses :) Oct 21 21:37:42 heh Oct 21 21:37:44 pb_: multi language support Oct 21 21:37:51 zecke: beer? Oct 21 21:38:11 zecke: indeed, you are well internationalised Oct 21 21:38:15 pb hm ah and there are cpu which support them native and some not and on the machine they do not have one the kernel emulate it similar like fpu Oct 21 21:40:45 woglinde: right. obviously the fpu-type emulation would be a massive performance loss, so nobody does this in reality. Oct 21 21:41:55 oh the S3C6410 has crypto enginge support hm this was trustpath? Oct 21 21:42:23 iirc, gcc always does its tls accesses via function calls to libc (and/or libgcc) so that you can swap them out with an appropriate implementation. Oct 21 21:43:00 I think there might have been some kind of -muse-eleet-tls option to do the register bashing directly but I don't think it was ever used for much. Oct 21 21:46:37 hm seems samsung measured trustzone with crypto Oct 21 21:46:44 but it is not Oct 21 21:47:34 no, they are different Oct 21 21:47:39 jepp Oct 21 21:47:46 s3c6410 has both Oct 21 21:47:52 hm Oct 21 21:48:03 trustzone is part of the core, the crypto engine is external Oct 21 21:48:33 hmm.. i will go to bed now... Oct 21 21:48:37 nite zecke Oct 21 21:48:43 zecke: sweet dreams Oct 21 21:49:51 rp would be nice to see the blueprints of the cpu Oct 21 21:50:23 ups pb Oct 21 21:50:25 args Oct 21 21:50:29 its to late Oct 21 21:55:30 pb_: a bit late but: trustzone is about separation of states of different "security domains" but has nothing to do with crypto per se, right? Oct 21 21:55:59 kensan right Oct 21 21:56:09 I just look at the specs brochure Oct 21 21:56:22 it says the hw crypot is spererate Oct 21 21:56:29 doing 3des/des aes and so on Oct 21 21:56:30 Kensan: right Oct 21 21:56:34 khem: ping Oct 21 21:56:58 but any the way this cpu package relly rockz Oct 21 21:57:22 woglinde: I was just skimming through the backlog. Will take a closer look at the hw crypto specs. Oct 21 21:57:35 woglinde: we will have to see if we are using it in future products Oct 21 21:57:53 woglinde: omap3 is also there, or maybe even marvel(l) Oct 21 21:58:02 hm I only wonder about the board options Oct 21 21:58:15 · 2Gb OneNAND + 1Gb mDDR Oct 21 21:58:16 and Oct 21 21:58:35 2Gb NAND + 512Mb mDDR + 512Mb OneDRAM Oct 21 21:58:41 whats Onedram? Oct 21 21:59:15 some crazy samsung proprietary memory thing, I think Oct 21 21:59:17 http://www.samsung.com/global/business/semiconductor/products/fusionmemory/Products_OneDRAM.html Oct 21 21:59:21 found it Oct 21 21:59:25 hm this seems nice Oct 21 21:59:44 faster for communication btw cpu and external co procs Oct 21 22:00:53 hm Oct 21 22:01:09 onenand seems the same but for stored data Oct 21 22:01:23 so you can choose btw. faster file or faster mem access Oct 21 22:01:50 Hm I can't seem to find any real information on the hw crypto except for the brochures. I guess detailed info is not freely accessible or does anybody have a link / google hints? Oct 21 22:02:02 Kensan guess wjy Oct 21 22:02:50 samsung is not ibm which pushed the some proz pictures of her z10 cpu with 2 cryptos Oct 21 22:02:54 security by obscurity? ;) Or just closed? Oct 21 22:03:00 woglinde: afaik, onenand is basically just regular nand wrapped up with a buffer ram and a nor-type interface. Oct 21 22:03:13 nite Oct 21 22:03:14 pb_ hehe yes marketing Oct 21 22:03:18 nite zecke Oct 21 22:03:21 I'm not sure it has much in common with onedram except the name Oct 21 22:03:33 pb faster access Oct 21 22:04:41 * pb_ bedtime as well now Oct 21 22:05:44 kensan z10 -> http://www.flickr.com/photos/21317126@N04/2291626868/ Oct 21 22:06:01 hmm, should I use qemu or qemu-i386 to start my minimal-image? Oct 21 22:06:19 hubar hehe what you like Oct 21 22:06:48 woglinde: I don't really care. I like it to be simple :) Oct 21 22:06:48 woglinde: nice :) Oct 21 22:07:22 hubar try minimal-image Oct 21 22:07:29 well I hope if OM decides to use it, that they take a closer look at the crypto documentation... Oct 21 22:07:52 oh I already created the image already. Oct 21 22:07:59 hubar *g* Oct 21 22:08:17 I just wonder if I wanna test it, should I use qemu or qemu-i386... Oct 21 22:08:28 hubar whats your arch? Oct 21 22:08:37 hm test in vmware Oct 21 22:08:39 or so Oct 21 22:08:43 virtualbox Oct 21 22:09:48 woglinde: How do I turn the tar file into a vmware/virtualbox image? Oct 21 22:10:03 arch is 386. Oct 21 22:10:16 hm right Oct 21 22:10:18 sorry Oct 21 22:10:20 okay Oct 21 22:10:25 use qemu Oct 21 22:10:41 [ 18:05:40 ] wweng «/distro/tmp/deploy/images $ ls Oct 21 22:10:42 bzImage-2.6.21-r12-x86.bin minimalist-image-x86-20081021203137.rootfs.jffs2 minimalist-image-x86.jffs2@ modules-2.6.21-r12-x86.tgz Oct 21 22:10:44 vut you have to get out a qemu fsimage to Oct 21 22:10:44 bzImage-x86.bin@ minimalist-image-x86-20081021203137.rootfs.tar minimalist-image-x86.tar@ Oct 21 22:11:12 let me see where a documentation is there Oct 21 22:11:14 hm wait Oct 21 22:11:20 you can build a qemu image Oct 21 22:11:22 I think Oct 21 22:11:23 mom Oct 21 22:11:32 what should I use to produce a qemu image? It seems it asks for a dot image file, which I don't have. Oct 21 22:12:02 ah tight Oct 21 22:12:07 try machine qemux86 Oct 21 22:12:30 this should build you a qemu fs Oct 21 22:14:36 woglinde: sorry i got disconnected. Oct 21 22:15:18 woglinde: What did you say after I asked how to produce a dot image file? Oct 21 22:15:39 try machine qemux86 Oct 21 22:15:46 in your local.conf Oct 21 22:16:17 hmm. what about minimal-image? Oct 21 22:16:51 How do you test a non-x86 image to see if it is alright? Oct 21 22:19:28 on the device Oct 21 22:19:42 in my case Oct 21 22:19:48 others using qemu Oct 21 22:20:53 hmm. Oct 21 22:21:26 So is there a way to test x86 image w/o building with qemux86 machine? Oct 21 22:22:11 untar the tar gz Oct 21 22:22:24 boot it on a real machine, or boot it on a vmware instance (both of which you have to do manually by untaring the image to a disk or virtual disk) Oct 21 22:22:47 or copy the content of oetmp/rootfs into a virtual app diskfromat Oct 21 22:22:55 and try to boot in the virtual app Oct 21 22:23:05 And for x86 I think you also have to do something with the bootloader, but I have not done that myself so I cannot say for certain. Oct 21 22:23:10 may for virtualbox it is easier then for vmware Oct 21 22:23:35 ... which is why qemu rocks Oct 21 22:23:51 hmm, so can not use qemu to test that image? Oct 21 22:23:52 qemu-system-i386 -kernel bzImage -hda minimal-image.ext2 -append "root=/dev/hda" Oct 21 22:24:02 qemu is easiest Oct 21 22:24:13 ahh Oct 21 22:24:16 Everything else is just as complex as real hardware always is Oct 21 22:24:41 I need to try current OE w/ qemux86 Oct 21 22:24:54 * mwester has never tried to use qemu to emulate the host system; always used vmware instead -- that does look easier, though. Oct 21 22:24:55 the snapshot i'm on now has a very goofy kernel config Oct 21 22:24:57 How do you get the ext2 file? Oct 21 22:25:12 hubar, IMAGE_FSTYPES = "ext2" in your local.conf Oct 21 22:25:26 then bitbake minimal-image Oct 21 22:25:57 ah right Oct 21 22:25:59 sorry Oct 21 22:26:04 forgot about ext2 Oct 21 22:27:59 mwester? Oct 21 22:30:05 okay I am going to bed too Oct 21 22:30:06 nite Oct 21 22:31:19 good night Oct 21 22:33:55 nite wog! Oct 21 22:54:29 re Oct 21 22:55:15 does someone here ever used uboot? :\ Oct 21 22:55:55 i'm sure quite a few people have. you'll probably want to be more specific Oct 21 22:56:54 I'm trying to install it on my C3000 Oct 21 22:58:39 but, after writing to the NAND, it never reboot (no charge led ever, and I need to re-flash vie Fn+D+M) Oct 21 22:59:13 I tried several build, from CF en SD, via the update menu or even manually Oct 21 22:59:15 i doubt this is the best place to be asking about device specific stuff Oct 21 22:59:24 you probably want the channel for your distribution? Oct 21 22:59:24 I can't get it to work Oct 21 22:59:46 wow Oct 21 22:59:48 specifically, I'm searching support fork uboot :\ Oct 21 22:59:51 -k Oct 21 22:59:51 someone else still haz a zeee **** ENDING LOGGING AT Wed Oct 22 02:59:57 2008