**** BEGIN LOGGING AT Thu Dec 05 02:59:58 2013 Dec 05 08:37:04 good morning Dec 05 09:38:32 hey Dec 05 09:40:24 i want to compile the kexecboot binary as a static linked program, there was former use the "inherit klibc" option Dec 05 09:40:35 this seems to be replaced with BBCLASSEXTEND = "klibc" Dec 05 09:40:54 but that does not create a static linked file. any ideas ? Dec 05 09:49:20 ping ant_work Dec 05 09:49:32 gm Dec 05 09:51:04 ant_work: i can not compile the kexecboot to a klibc linked file . ( after adding inherit klibc it seems to work ) Dec 05 09:53:04 have you a recent git pull? The -klibc variant is built with BBCLASSEXTEND Dec 05 09:54:24 anyway inherit klibc is thedefault way for single/problematic recipes (kexec-tools i.e.) Dec 05 09:56:23 i replaced BBCLASSEXTEND with inherit klibc and now it seems that its working Dec 05 09:56:47 that's what BBCLASSEXTEND ssupposedly does ;) Dec 05 09:57:10 it did not so long ago... Dec 05 09:58:14 i have my old patch to make kexec-hardboot and want to build an image to boot the ouya. Dec 05 09:59:34 iirc on eother dev has forked an d added that Dec 05 09:59:51 * ant_work switching to wired kb Dec 05 10:00:00 i just tested it. with inherit klibc -> kexeboot size:121539 static linked , with BBCLASSEXTEND size:242222 dun linked Dec 05 10:00:25 you have to check kexecboot-klibc not kexecboot Dec 05 10:00:44 the recipe produces two packages Dec 05 10:00:55 the kexecboot-klibc was removed Dec 05 10:01:10 'bitbake kexecboot' builds both Dec 05 10:01:18 hmmm Dec 05 10:01:46 is a class extension, like for -native Dec 05 10:03:05 whats about the idea to add an option to the boot.cfg to set the --hardboot options ? Dec 05 10:04:10 welcome Dec 05 10:04:21 check out github forks Dec 05 10:04:51 242 Dec 05 10:04:51 + const char *load_argv[] = { NULL, "--load-hardboot", item->kernelpath, NULL, "--mem-min=0x84000000", NULL, NULL }; Dec 05 10:04:51 243 Dec 05 10:04:51 + const char *exec_argv[] = { NULL, "-e", NULL}; Dec 05 10:05:03 seems similar stuff Dec 05 10:06:14 yea, but whats when you want to run it with --mem-min= 0x1000000 . the you have to recompile kexecboot ..... Dec 05 10:09:57 you might better ask Jay7, I think he as some ideas in that regard Dec 05 10:10:16 m? Dec 05 10:10:33 btw. how did you find a fork on kexecboot with the hardboot patch ? Dec 05 10:11:16 * Jay7 like any idea about improving kexecboot :) Dec 05 10:12:14 click 'network' on github repo Dec 05 10:12:21 it seems we need another option in boot.cfg to pass custom params to kexec binary :) Dec 05 10:12:47 anyway, subject to talk on another channel :) Dec 05 10:12:55 an a cinfigure option to set the maximim size for commandline parameters. Dec 05 10:14:12 btw there are two threads on kexec/kernel ML: Dec 05 10:14:14 [RFC PATCH] Signed kexec support Dec 05 10:14:32 and kexec: A new system call to allow in kernel loading Dec 05 10:14:46 so more changes coming Dec 05 10:16:49 morning all Dec 05 10:27:59 hey bluelightning Dec 05 10:28:05 hi koen Dec 05 10:28:12 ant_work: have you thought about supporting fastboot in kexecboot? Dec 05 10:31:47 koen: yes, we need to add some extensions Dec 05 10:36:03 koen: there are already some Android tablets running Dec 05 10:37:12 tada. a new layer was born: meta-ouya .. Dec 05 10:37:50 pwgen: could you please submit it to the layer index? Dec 05 10:46:56 ant_work: I'm not interested in android, fastboot is usefull in normal linux as well :) Dec 05 11:03:14 there, all 17 perl-module-foo.bb recipes renames to libfoo-perl and changed all the RDEPENDS to match Dec 05 12:09:09 pwgen: er... why does meta-ouya have copies of OE-Core conf/machine/include files for x86? Dec 05 12:54:57 bluelightning,: .. oops i used this recipes to build ubi based-initramfs images. and some other .. there is still work left Dec 05 13:07:09 Hi, I am compiling a package based of autoconf. I need to pass preprocessor for compiler like -DSAMPLE=1. How can I do this? My project is having cpp files. I tried using CPPFLAGS += "-DSAMPLE=1" and it is not working Dec 05 13:09:55 that *should* work Dec 05 13:21:47 hi all. Is there anyone who can help me with pulseaudio? Dec 05 13:44:33 no single person? Dec 05 13:46:39 dol ~ask Dec 05 13:46:44 ~ask Dec 05 13:46:44 Questions in the channel should be specific, informative, complete, concise, and on-topic. Don't ask if you can ask a question first. Don't ask if a person is there; just ask what you intended to ask them. Better questions more frequently yield better answers. We are all here voluntarily or against our will. Dec 05 13:46:45 mckoan: Error: "ask" is not a valid command. Dec 05 13:47:49 mckoan, actually I already asked my question to http://stackoverflow.com/questions/20370418/pulseaudio-connection-failure-connection-refused Dec 05 13:48:11 and I tried pulseaudio-discuss and also alsa-devel groups but no help really Dec 05 13:48:29 could you please have a look at it shortly? Dec 05 13:49:35 dol: are you using OE ? Dec 05 13:49:39 yes Dec 05 13:50:20 dol: I'd rather ask to pulseaudio mailing list Dec 05 13:50:43 mckoan, I have already done it. It is not even posted there by the moderator Dec 05 13:50:52 perhaps I am doing something very stupid Dec 05 13:50:56 but yet not realized Dec 05 13:50:59 How long did you wait? Dec 05 13:51:06 2 days Dec 05 13:51:13 The moderation is just for non-subscribers so if you subscribe it'll get posted. Dec 05 13:51:23 That's not that long, the moderation is a bit more intermittent than that. Dec 05 13:51:52 but, let me ask you guys something about the audio Dec 05 13:52:09 I have two sound cards. I have an application which drives one of them perfectly using alsa Dec 05 13:52:26 I wanted to have the same audio from the other sound card as well Dec 05 13:52:31 just a copy of it Dec 05 13:52:52 I did it with alsa but got some cut outs after some time Dec 05 13:53:06 that was because of the two difffrerent clocks in two different sound cards. Dec 05 13:53:13 then I jumped into pulseaudio to do this Dec 05 13:53:30 perhaps you guys have another idea with which I can do this task easily Dec 05 13:53:45 Probably the two sound cards don't have synchronous clocks so one's idea of the clock rate is slightly different from the other and they drift over time. Dec 05 13:53:54 exactly Dec 05 13:54:04 do tyou have any solution in your mind for this? Dec 05 13:54:12 Pulse makes sense. Dec 05 13:54:30 broonie, is there any way to do this in alsa? Dec 05 13:54:41 I want to keep it in this asound.conf Dec 05 13:54:48 if possible of course Dec 05 13:55:21 It's possible some plugin may be able to handle it but the sample rate matching and conversion is a relatively complex job. Dec 05 13:55:43 I see. so you also suggest pulseaudio Dec 05 13:57:01 It'd be my first idea. Dec 05 13:57:10 There's also #pulsaudio on this server BTW. Dec 05 13:57:19 I really need your second idea :) Dec 05 13:57:33 I will check that Dec 05 13:58:40 broonie, can I do it in sofware? Dec 05 13:58:53 using the same asound.conf without playing with pulseaudio Dec 05 13:59:15 It might be possible but I don't know. Dec 05 13:59:45 ok. please let me know if you have any other thoughts about this Dec 05 14:04:34 dol: there is a reason its not supported, its a really complex issue Dec 05 14:10:15 XorA, what is your suggestion? pulseaudio? Dec 05 14:11:25 dol: you will need to write code even for pulse I think, I have never succeeded it getting it to output to two cards at once Dec 05 14:12:53 XorA, I had an impression that I could do it with a configuration file but apprrently not possible Dec 05 14:13:07 that problem is bigger that I thought Dec 05 14:13:17 dol: you can do it in config file for ALSA but one of your cards will skip Dec 05 14:13:54 you combine two cards into one virtual 4 channel one, then use the plugin to echo stereo to front and back channels Dec 05 14:13:59 XorA, that's the problem. I hava already done it in asound.conf. I did dmixing and multiplexing but the audio cuts out after some time in on sound card Dec 05 14:14:22 on = one Dec 05 14:14:41 you will at least need something constantly sample stuffing one of the cards (the el-cheapo method) Dec 05 14:16:00 then why can't I do the following Dec 05 14:16:23 what if I set the rate specifically for that sound card in asound.conf, Dec 05 14:16:39 I mean, I will let alsa to convert the rate for that specific sound card? Dec 05 14:16:53 dol: the issue is sound cards are only ever approx X rate, so they will still get out of sync Dec 05 14:16:55 but I guess alsaconf, there must be only one rate for all sound cards Dec 05 14:17:00 The problem is that the rates that you're seeing won't exactly reflect the named rate. Dec 05 14:17:37 then, how come pulseaudio will solve my issue? Dec 05 14:17:38 Getting sound cards with really accurate clocks would also help. Dec 05 14:17:40 el-cheapo method you monitor the dma queues and which card is going to underrun you stuff in a couple of repeat samples Dec 05 14:17:53 but that messes with your frequency response Dec 05 14:18:18 its used in voice systems Dec 05 14:18:33 XorA, I see your point Dec 05 14:18:36 Pulse is the main place where fancy DSP stuff ends up these days and it does have at least some stuff for actual rate estimation. Dec 05 14:18:45 Since it uses that for interruptless playback. Dec 05 14:18:56 expensive method you need to constantly resample and thats where broonie ex employer makes their $$$ Dec 05 14:19:30 :) Dec 05 14:20:14 broonie: you said "Getting sound cards with really accurate clocks would also help." what is the way to go for this? Dec 05 14:20:21 personally I would solve the issue in the analogue domain Dec 05 14:20:33 XorA, how? Dec 05 14:20:44 dol: we used to call them Y splitter cables :-D Dec 05 14:21:05 dol: I don't know how realistic that is going to be but basically just try to find things with consistent clocks, which will probably be high end devices. Dec 05 14:21:16 XorA, :D I know that cable but it is out of question in my case Dec 05 14:21:41 broonie, no way. I have already sold 2000 units with these cards in :D Dec 05 14:22:27 dol: Right, that's not going to work then - this would be a hardware change. Dec 05 14:23:07 if this was embedded Id say use one codec to master the bitclock for all the rest Dec 05 14:23:38 XorA, one of them is embedded Dec 05 14:24:21 actually the problem is, I use the sound card which has an analog jack. But the second one that I say actually is the one with the displayport connector Dec 05 14:24:41 so my primary aim was to enable the audio on displayport Dec 05 14:25:48 anyway, thanks a lot for you guys. I appreciated it. you guys were the only ones who responded me Dec 05 14:27:07 dol: you just caught the ex audio guys awake for once :-D Dec 05 14:27:36 XorA: :)) Dec 05 14:27:49 refreshed your knowledge :) Dec 05 14:28:35 have a great day Dec 05 14:28:36 nbye Dec 05 17:42:16 moin Dec 05 21:43:05 anyone familiar with meta-fsl-arm-*? Dec 05 21:43:15 ERROR: Logfile of failure stored in: /home/balister/oe-repo/build/tmp-eglibc/work/nitrogen6x-oe-linux-gnueabi/u-boot-boundary/v2013.07+gitAUTOINC+f89bc066db-r0/temp/log.do_fetch.19109 Dec 05 21:46:28 Crofton|work: what is the error? Dec 05 21:46:46 ERROR: Fetcher failure: Unable to find revision f89bc066db0569008418fac1a0408a9a57f2232f in branch master even from upstream Dec 05 21:47:04 It was discussed on ML today Dec 05 21:47:17 I'm not in tune with the bsp,s just adding them to a manifest by request and running a test build Dec 05 21:47:19 probably points to different branch and bitbake is now more strict about it Dec 05 21:47:26 ah Dec 05 21:48:42 JaMa: hmm. that makes me wonder why bitbake would care about the branch in the first place, if the commit exists. do you kow? Dec 05 21:48:44 know? Dec 05 21:52:22 branch is important only for AUTOREV imho Dec 05 21:52:35 but as it's there then it should be consistent with the commit specified in SRCREV Dec 05 21:52:42 some crzy user just extended his sdk with Ha! Just ran "sudo make install DESTDIR=" Dec 05 21:54:18 JaMa: but branches are 'temporary' as opposed to commits Dec 05 21:55:53 ndec: but still if your recipe says branch=2.x; you want to AUTOREV to stay in 2.x branch and not to drift to master branch Dec 05 21:56:28 JaMa: yes, for AUTOREV, i agree. Dec 05 21:56:51 but failing a 'fetch' because the branch name was wrong, while the commit is found looks weird Dec 05 21:57:10 and I agree that it's a bit unnecassary for locked SRCREV Dec 05 21:57:22 ok. Dec 05 21:57:24 but even if you remove it from SRC_URI it will get default value "master" Dec 05 21:57:46 so as long as it's there it's better to keep it consistent with SRCREV commit Dec 05 21:58:41 and to just remove it together with default value is bad for AUTOREVs which are often selected outside recipe files in some foo-autorev.inc in distro layer Dec 05 21:58:57 and they would need to add branch information for every AUTOREVed recipe Dec 05 21:59:36 as RP said only very few recipes don't use "master" and he already fixed 2 cases in oe-core Dec 05 21:59:44 so maybe few more to go, but that's all Dec 05 21:59:52 well, ok. maybe you're right.. it's better if it's consistent... so better fail now that in a more crazy way later... Dec 05 22:01:24 ndec: the issue is the fetcher was hitting the network in weird ways behind the scenes as the branch it thought it was supposed to use didn't contain the revision and hence might need updating Dec 05 22:01:51 yeah. i read that earlier today. **** ENDING LOGGING AT Fri Dec 06 02:59:58 2013