**** BEGIN LOGGING AT Tue Mar 03 02:59:58 2015 Mar 03 03:06:55 Crofton: Yep, im working on updates for SPL and U-Boot at the moment, u-boot-dtb.img is the way forward it appears :) Mar 03 03:21:42 ah cool Mar 03 03:22:00 th exlnx u-boot branch works for me Mar 03 03:22:13 I'll not worry and wait for it to trickle down then Mar 03 03:29:58 hi, I want to use open-iscsi recipe, found it's inside meta-cloud-services and seems this need poky 1.7.x? Im using poky 1.6.x , any possible to use that open-iscsi recipe directly? or should I upgrde to 1.7.x? Mar 03 03:39:46 I built dizzy for beaglebone black today, and USB host doesn't work. (3.14 kernel) Mar 03 03:40:00 Is this broken for everyone or just me? Mar 03 05:18:20 is the meta-cloud-services ready for dizzy? I got a "no recipe for dhcp_4.3" error Mar 03 05:24:46 eddylaitw: looks like there is only support for master, not sure what the 'grizzly'/etc. branches refer too though :| Mar 03 05:28:13 thanks Mar 03 07:58:41 anyone did yocto on IMX6 Quad (element 14 PCB board) Mar 03 07:59:29 is there any step by step yocto document for new newbie like me? Mar 03 08:00:33 is it a sabrelite ? Mar 03 08:00:38 we use yocto to build an imx6 platform Mar 03 08:00:47 as abelloni said, using sabrelite is a good starr Mar 03 08:00:48 t Mar 03 08:00:55 *start; geez I can't typer Mar 03 08:02:19 bluepimple: https://github.com/Freescale/fsl-community-bsp-platform Mar 03 08:04:25 fsl stand for? Mar 03 08:04:28 set MACHINE to imx6qsabrelite Mar 03 08:04:39 bluepimple: freescale Mar 03 08:05:29 yes, i am using freescale imx6, thanks Mar 03 08:06:04 bluepimple: freescalle imx6 is too broad, it doesn't say which board you are using Mar 03 08:10:35 dual core Mar 03 08:13:17 Part number: MCIMX6Q-SL Mar 03 08:15:14 isn't it a quad core ? Mar 03 08:15:35 i had the BSP working once, a green GUI screen that can be swept by the mouse. Mar 03 08:18:14 bluepimple: http://www.element14.com/community/docs/DOC-49013?ICID=designcenter-sabre-space Mar 03 08:21:05 bluepimple: it is the one I use during my FSL training http://www.adeneo-embedded.com/Media/All-static-html/Invitation_Training_Linux_Milano_Marzo_2015 Mar 03 09:19:30 kernel-dev package has headers, but is there a package generated for the kernel source so I can include it as part of a generated SDK? Mar 03 09:31:29 Net147: add TOOLCHAIN_TARGET_TASK += " kernel-dev" Mar 03 09:31:31 iirc Mar 03 09:50:13 morning all Mar 03 10:00:48 bluelightning: hi Mar 03 10:01:57 * funman finally getting to build kernels for the odroid boards Mar 03 10:07:45 arm-poky-linux-gnueabi-ld.bfd: error: Source object drivers/amlogic/audiodsp/dsp_data.o has EABI version 5, but target drivers/amlogic/audiodsp/audiodsp.o has EABI version 0 Mar 03 10:08:18 How can I set up PATH properly to run make V=1 in kernel build dir? Mar 03 10:10:23 funman: dps_data is an object compiled by amlogic Mar 03 10:10:35 I guess you are out of louck :) Mar 03 10:10:38 luck Mar 03 10:52:04 Hi, I have managed to setup Yocto ADT, eclipse plugin and build/debug the hello world application. I would now like to experiment with pthread but when I build get an error "undefined reference to `pthread_create'", I assume I need to add the pthread library? How do I do this? Mar 03 10:52:15 rink_: I thought kernel-dev has only headers and not sources? Mar 03 10:53:33 euan: -lpthread ? Mar 03 10:54:20 where do I specify this? I have tried adding "AM_LDFLAGS = -lpthread" to Makefile.am but no success Mar 03 10:56:49 Net147: oh, right, sorry. Mar 03 10:57:15 i assumed you wanted to build kernel modules outside of yocto or similar - you just need the haders for that Mar 03 10:57:27 rink_: I would like to debug the kernel Mar 03 10:59:23 no experience with that apart from printk(), sorry Mar 03 11:00:01 printk is the best tool for the job :) Mar 03 11:00:44 assuming it works - which, incidentally, isn't the case during very early i.mx6 startup Mar 03 11:00:48 trying to find the cause of a panic that happens once a month or two... Mar 03 11:02:01 using gdb and kgdb to debug over serial Mar 03 11:02:16 <_4urele_> hi everyone! Mar 03 11:03:18 and there isn't a linux-kernel-dbg or something pacakge? Mar 03 11:03:21 <_4urele_> I'm using smart to update my target, i discovered that a package crash smart during install due to an unicode error... Mar 03 11:03:53 <_4urele_> does someone knows about this? does someone knows where I could get some help? Mar 03 11:03:57 rink_: that's what I am trying to figure out. at the moment I just do bitbake -c patch linux and then rsync the linux source directory over to the SDK machine. Mar 03 11:20:02 hi Mar 03 11:20:38 I am working with an imx28evk board and in the path : /home/pjo/fsl-community-bsp/build/tmp/work/imx28evk-poky-linux-gnueabi/core-image-sato-sdk/1.0-r0/rootfs/lib/modules/2.6.35.3-maintain+gb3912bb Mar 03 11:21:03 there are 2 files named modules.builtin and modules.order Mar 03 11:21:53 I suppose that the modules.builtin regroups all the modules available, whereas the second one is listing the ones available in the image Mar 03 11:22:34 this is the case for example for : kernel/drivers/power/mxs/mxs-battery.ko Mar 03 11:23:38 I tried to use bitbake -c menuconfig linux-imx Mar 03 11:24:50 and in the configuration menu : Power supply class support ---> Mxs SoC battery charger driver is already activated .... Mar 03 11:26:21 I do not understand what other otpion I should check to get it activated before I rebuild the kernel Mar 03 11:26:30 any help would be appreciated Mar 03 11:26:40 thank you in avance ;) Mar 03 11:31:57 rink_: looks like kernel-devsrc is what I need. it's in master but not in any official release yet... Mar 03 11:38:19 can anyone help me add the pthread library to the hello world application? I can now see there is a second Makefile.am in /src which has the line "AM_LDFLAGS = @euan_test_LIBS@" but I can't find when euan_test_LIBS comes from? Mar 03 11:42:39 morning Mar 03 11:43:47 If I'm generating a bbappend and want to create / manipulate /etc/default/dropbear, whats the "right" mechanism? It looks like image.bbclass generates / adds to it but scratching my head slightly Mar 03 11:45:05 Or should I just provide a complete replacement in the same way that interfaces gets over-ridden for ifupdown Mar 03 12:14:11 good morning Mar 03 12:23:27 khem, hello Mar 03 12:24:26 abelloni: hey, start-stop-daemon -S -b -n ntpd -a /sbin/ntpd -- -p pool.ntp.org -l -n will not start Mar 03 12:24:40 abelloni: is it because I am not connected to the internet and so pool.ntp.org cannot be accessed? Mar 03 12:29:02 abelloni: ntpd -p pool.ntp.org -l -n Mar 03 12:29:02 ntpd: bad address 'pool.ntp.org' Mar 03 12:29:35 perhaps it is because the hostname cannot be resolved? Mar 03 12:31:22 abelloni: there does not seem to be a force option either to at least get the server part running in such cases. :/ Mar 03 12:38:15 you are asking for conflicting things Mar 03 12:41:03 abelloni: what do you mean? Mar 03 12:43:42 its is either a stratum 1 server or not Mar 03 12:44:00 what you can do is start 2 servers Mar 03 12:44:15 /sbin/ntpd -I lo Mar 03 12:44:37 /sbin/ntpd -p pool.ntp.org -p 127.0.0.1 -I eth0 Mar 03 12:45:07 well, I was told that it can operate as a server as well as a client .... Mar 03 12:45:29 that is the case Mar 03 12:45:36 it is now more confusing than it has ever been. Mar 03 12:45:42 that documentation really could take some love. Mar 03 12:46:01 or maybe you could learn how ntp is working Mar 03 12:46:44 but bascially, what you are currently trying to do is bad Mar 03 12:47:20 either the time on your machine is correct and don't need to be synchronized Mar 03 12:47:32 or it needs to be synchronized Mar 03 12:47:39 that is just wrong Mar 03 12:47:43 you can't say both at the same time Mar 03 12:47:58 board one _is_ supposed to fetch its own time and also provide a time server for board two. Mar 03 12:48:39 then why do you want to use pool.ntp.org if board one has the correct time ? Mar 03 12:48:49 well, learning ntp can be done through a very simple documentation or the painful hard dway. Mar 03 12:48:52 hard way* Mar 03 12:49:07 and to be honest, even then, you may not know at all, how busybox decided to behave. Mar 03 12:49:36 (which is the case here, pretty much, I could implement an ntp daemon the way I told you) Mar 03 12:49:53 and that would be plain wrong Mar 03 12:50:01 I can say and _want_ to say both at the same time. Mar 03 12:50:20 I really do not want both board to go the internet for several reasons, some of them are in no particular order: Mar 03 12:50:24 1) Limit internet access Mar 03 12:50:40 2) Keep the boards in sync even when the device is not connected to the internet. Mar 03 12:50:57 to a certain definition of "plain wrong". It is different. Mar 03 12:51:25 I do not like the complexity of busybox's ntpd, but it is not plain wrong, it is just not how I would do it. Mar 03 12:51:57 you can also use another ntp server like I suggest Mar 03 12:52:10 but I don't believe you'll find a simpler one Mar 03 12:52:25 and you keep changing your spec Mar 03 12:52:33 anyway, this was exactly the reason why I asked the problem of sysvinit and ntpd the other day. Mar 03 12:52:50 when I started asking about it. I wanted to run two processes initially until I was told one process could do both. Mar 03 12:52:53 if board one is connected to internet, then let it contact a stratum1 server Mar 03 12:53:02 if it is notl et it serve its own time Mar 03 12:53:04 so I guess am back to root square after 1-2 days investigation. Mar 03 12:53:17 to the same problem: start-stop-daemon is not nice for multiple instances. Mar 03 12:53:23 to track the instance's lifecycle on their own. Mar 03 12:53:27 if it is only connected sporadically, then let it serve its own time and synchronize the time another way Mar 03 12:53:41 I have never ever changed the spec. Mar 03 12:53:47 it is the same from day one. Mar 03 12:53:49 yes you do Mar 03 12:54:00 your understanding may change, but that is a different thing. Mar 03 12:54:06 sometimes you say board one has internet acces, sometimes no Mar 03 12:54:09 this is exactly what I have been trying to achieve from day first; just read back. Mar 03 12:54:16 Let me quote my first questions: Mar 03 12:54:51 12:38 < lpapp> but now I have another board and I would like to sync that up from the one which goes to the internet Mar 03 12:55:06 12:38 < lpapp> therefore, the "main" board would have a client daemon as well as an ntpd server running Mar 03 12:55:19 so ntpd -p pool.ntp.org -l is working fine Mar 03 12:57:27 no it is not. Mar 03 12:57:45 because pool.ntp.org is not accessible if the internet is dropped Mar 03 12:57:59 but the whole point of internal syncing is not relying on the internet. Mar 03 12:58:09 whereas, when connected, the server should still sync up from the remote ntp server. Mar 03 12:58:19 Priority 1: Internal sync Mar 03 12:58:32 Priority 2 (nice to have): sync them to the external. Mar 03 13:00:37 ntpd -l on board one with ntpd -p pool.ntp.org -q from time to time Mar 03 13:00:46 when connected to the internet Mar 03 13:01:45 this is a shame that busybox does not make this simple. Mar 03 13:02:00 patches are welcome Mar 03 13:02:01 there ought to be an option or something for this. Mar 03 13:03:26 I really believe you should stop using ntpd from busybox Mar 03 13:06:37 hi guys Mar 03 13:07:13 abelloni: why? We do not have much flash. Mar 03 13:07:26 does anyone have a clue about my post 12:25 ? Mar 03 13:07:42 abelloni: a few megs storage and we already use busybox for other things. Mar 03 13:07:50 although I am thinking about rewriting the minimal tools that I need. Mar 03 13:07:56 but that is not a discussion for today.... Mar 03 13:08:27 is that so difficult to patch it so that it doesn't abort when not finding peers ? Mar 03 13:10:00 abelloni: ah, it is not that easy. Mar 03 13:10:11 abelloni: plus the busybox codebase is horrible IMHO. Mar 03 13:10:22 it is also disappointing not getting replies many times. Mar 03 13:10:39 but what makes it more difficult is the fact that you can online anytime and it should detect that somehow, you see.... Mar 03 13:10:43 come on, it is a two liner Mar 03 13:10:44 go* Mar 03 13:11:00 most definitely not two lines. Mar 03 13:11:26 I can do it in two Mar 03 13:11:34 and one of those is '}' Mar 03 13:11:53 I can do many things in one-line, too, that is not good enough :) Mar 03 13:12:10 do not forget about the usage either. Mar 03 13:13:17 actually, I don't see the issue with using 'ntpd -p pool.ntp.org -q' in an ifup script Mar 03 13:14:25 I do. Mar 03 13:14:52 remember that -q will only query once and then stop once it is sync'd. Mar 03 13:15:07 I modified src/Makefile.am replacing "AM_LDFLAGS = @euan_test_LIBS@" with "AM_LDFLAGS = -lpthread" and my application now builds and I can run with qemu. This doesn't feel like the correct way as I have bypassed the euan_test_LIBS variable, does anyone know where this is set or how to properly include a library with autotools? Mar 03 13:15:09 but these machines may have very long runs. Mar 03 13:15:18 therefore -q is insufficient. Mar 03 13:15:35 so it is a bit more subtle than that. Mar 03 13:15:38 then start ntpd -p pool.ntp.org -l from the ifup script Mar 03 13:15:50 that is not good either ;-) Mar 03 13:16:12 that will likely stop when we do ifdown. Mar 03 13:16:29 therefore, we lose the sync up internally. Mar 03 13:16:35 unless you put some magic into ifdown, too. Mar 03 13:16:50 and more magic into ifup (stop, start) and the same but vice versa into ifdown. Mar 03 13:17:03 abelloni: I didn't see a .o in the git log Mar 03 13:17:20 it has to come from somewhere though Mar 03 13:17:22 funman: it is a .z that get copied as a .o Mar 03 13:17:27 ok Mar 03 13:17:54 lpapp: it wouldn't stop unless you explicitely do it in ifdown Mar 03 13:20:10 abelloni: but we need to stop it Mar 03 13:20:41 when ? Mar 03 13:26:27 something to note - you shouldn't in general be using pool.ntp.org in any default configuration: http://www.pool.ntp.org/en/vendors.html Mar 03 13:26:56 (if you aren't doing that, no worries - just thought I would mention it) Mar 03 13:27:23 bluelightning: yeah, I am using it for local testing now, but it will be up to the customer to set the remote server. Mar 03 13:27:33 if it is wrong, I am happy to use something else for testing. Mar 03 13:27:40 abelloni: I think I will need a complex state machine :/ Mar 03 13:27:59 there was some discussion about registering a yocto project or OE pool but I don't think we ever got around to it Mar 03 13:28:29 abelloni: if the customer plugs the "internet" cable from one board to another in the device, I will also need to swap the local sync logic between the master and slave.... it is all complex a bit with two baords supporting both internet and non-internet case... Mar 03 13:28:39 I'm not even sure that is a good idea Mar 03 13:29:06 abelloni: see, it is up to our customers how they use our product. We have two boards in it. Mar 03 13:29:07 (registering for an oe pool) Mar 03 13:29:25 they have to be kept in sync irrespective of the remote internet connection. Mar 03 13:29:42 and we would like to minimize the outbound internet traffic, too. Mar 03 13:30:10 abelloni: right, I'm not 100% convinced by that either - our current approach is no default, forcing those using ntp recipes to set one themselves Mar 03 13:32:07 hello Mar 03 13:32:25 I'd like to use this : http://bec-systems.com/site/967/setting-the-root-password-in-an-openembedded-image Mar 03 13:32:51 but ${IMAGE_ROOTFS} seems to give the wrong path Mar 03 13:33:10 where do I have to put the code ? Mar 03 13:33:27 Retza: in the image recipe you are attempting to build Mar 03 13:33:33 (currently it's in core-image-base.bb, it's dirty) Mar 03 13:33:47 bluelightning: in core-image-base.bb ? Mar 03 13:33:57 if that's the image you are building then yes Mar 03 13:34:03 ok Mar 03 13:34:13 but it doesn't work though Mar 03 13:35:26 I have this : Mar 03 13:35:31 WARNING: Function mv /home/pi20/poky/freescale/tmp/work/imx6qsabresd-poky-linux-gnueabi/core-image-base/1.0-r0/rootfs/etc/shadow.new /home/pi20/poky/freescale/tmp/work/imx6qsabresd-poky-linux-gnueabi/core-image-base/1.0-r0/rootfs/etc/shadow doesn't exist Mar 03 13:36:25 and /home/pi20/poky/freescale/tmp/work/imx6qsabresd-poky-linux-gnueabi/core-image-base/1.0-r0/rootfs doesn't exists, even if it's the path given by ${IMAGE_ROOTFS} Mar 03 13:37:31 any idea bluelightning ? Mar 03 13:39:16 retza1: you need to create a function and call that from ROOTFS_POSTPROCESS_COMMAND, rather than adding the command to that variable value Mar 03 13:40:13 bluelightning: have you an example of syntaxe ? Mar 03 13:44:10 retza1: yes, see meta/classes/insserv.bbclass (ignore the first line) Mar 03 13:45:34 parsing ok, thanks Mar 03 13:45:47 now I wait to see if I have the same WARNING Mar 03 13:50:02 bluelightning: yay, thx a lot ! Mar 03 13:51:07 Retza: you're welcome :) Mar 03 13:51:54 by the way, do you know how to add an user (with password and /home/user) ? Mar 03 13:52:28 I guess I can't do it with ROOTFS_POSTPROCESS_COMMAND Mar 03 13:53:29 abelloni: http://git.busybox.net/busybox/commit/?id=3aef814c0b08d9703280b4772060ce5016c683c4 Mar 03 14:01:13 Retza: you can instruct yocto to add users to your system Mar 03 14:01:20 in the image's passwd/group/etc Mar 03 14:14:37 Retza: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#ref-classes-extrausers Mar 03 14:14:54 assuming you are using a new enough version of the system that provides that class Mar 03 14:27:13 thanks bluelightning. Mar 03 14:29:55 abelloni: CONFIG_AEABI seems to do the trick Mar 03 14:30:04 not sure why it's not in their defconfig Mar 03 15:04:23 i think it's rather a buggy recipe Mar 03 15:04:33 kernel_conf_variable Mar 03 15:22:56 hi Mar 03 15:36:45 what tool there is to "benchmark" a linux system? I realize that "benchmarking" is vague, but I'm open to hear about anything - still I think I'd like to start with something that can bench the CPU with a tool that tries a lot of different operation (like floating point math etc) Mar 03 15:37:37 google Mar 03 15:48:40 s/google/I don't know Mar 03 15:50:40 there are a lot of stress testing and benchmarking tools available Mar 03 15:50:58 but most seem to have some kind of focus (integer, float performance, etc) Mar 03 15:51:19 I found this one, seems pretty cool https://code.google.com/p/byte-unixbench/ Mar 03 15:51:23 * rink_ thinks the tool to use depends quite a lot on what you want to know Mar 03 15:51:55 acidfu: no, if you're too lazy to bother looking, you really don't need to be wasting our time anymore Mar 03 15:51:57 go do some research Mar 03 15:52:14 kergoth, if you don't know just not bother to answer me Mar 03 15:52:20 thats it Mar 03 15:52:29 I know how to use google... Mar 03 15:52:39 apparently not, if you waste people's time on irc before looking yourself Mar 03 15:52:50 kergoth, just don't speak to me, thats it Mar 03 15:52:56 so you won't feel you are loosing anything Mar 03 15:53:13 acidfu: there are a bunch of recipes in meta-openembedded/meta-oe/recipes-benchmark/ that may be of interest Mar 03 15:53:15 agreed, thank god for /ignore Mar 03 15:53:29 bluelightning, cool I'll have a look thank you Mar 03 15:54:50 * paulg munches on some popcorn Mar 03 15:56:08 I hate it when people value others time less than their own. Pet peeve of mine. Mar 03 15:56:23 I hear ya. Mar 03 15:56:33 * vmeson points to: http://www.catb.org/esr/faqs/smart-questions.html Mar 03 15:56:38 ERROR: The recipe linux-hardkernel is trying to install files into a shared area when those files already exist. Those files and their manifest location are: /media/fun/1T/poky/build/tmp/sysroots/odroid-c1/sysroot-providers/virtual_kernel Mar 03 15:56:56 * paulg fines vmeson $2 for using an ESR reference. Mar 03 15:57:03 hah Mar 03 15:57:19 I think I understand, it means I already built a dummy kernel, and then modified my sources to build the real kernel Mar 03 15:57:20 what happened to the beer fines, paulg? Mar 03 15:57:55 I fine you in cash and then use the $ to buy beer. Same as it has always been. Mar 03 15:58:49 * funman unbuilds some stuff Mar 03 16:02:13 vmeson: hadn't seen that link before. I like it! Mar 03 16:02:31 i didn't realize esr updated it, i read it years ago. i see it mentions stackexchange now Mar 03 16:02:48 darknighte: don't tell paulg, he'll be drunk before we know it. :) Mar 03 16:03:21 hah! Mar 03 16:03:47 * darknighte considers how many times he's asked "non-smart" questions Mar 03 17:05:17 I'm trying to add net-snmp to a core-image-lsb build. I'm getting the following error: Mar 03 17:05:28 ERROR: Nothing RPROVIDES 'net-snmp-server' (but /home/dan/sabre/sources/poky/meta/recipes-extended/images/core-image-lsb.bb RDEPENDS on or otherwise requires it) Mar 03 17:05:53 I added the following to my local.conf: Mar 03 17:06:10 VIRTUAL-RUNTIME_snmp-manager = "net-snmp-server" Mar 03 17:06:10 IMAGE_INSTALL_append = " net-snmp-server net-snmp-libs net-snmp-mibs" Mar 03 17:06:59 The meta-openembedded layer is enabled Mar 03 17:13:16 cyclist: meta-openembedded is a repository, meta-oe is a layer within that - but I think you'll find the net-snmp recipe is part of meta-networking Mar 03 17:16:09 it's in there, but I'm geting the error when I add it Mar 03 17:16:44 Am I missing something else in my local.conf? Mar 03 17:17:10 meta-oe is in my bblayers.conf Mar 03 17:18:21 cyclist: you have both meta-oe and meta-networking? Mar 03 17:19:51 i.e. both are listed in bblayers.conf? Mar 03 17:20:10 no they aren't Mar 03 17:20:21 ok, you will also need to add meta-networking Mar 03 17:20:22 I get it now (I think) Mar 03 17:22:28 I thought that meta-networking was part of meta-oe, but now I see that it's not Mar 03 17:25:58 It's building now. I had to add meta-python, too Mar 03 17:26:18 thanks for your help Mar 03 17:34:39 np Mar 03 17:35:23 moin Mar 03 18:37:54 RP: there Mar 03 19:20:37 is there a specific libpthread recipe, or does poky build it by default? thanks ! Mar 03 19:22:30 acidfu: libc has it Mar 03 19:22:44 ok thanks Mar 03 20:32:42 Can I force a package to be rebuilt if a package it depends on has it's version change? Mar 03 20:32:49 Or does this always happen? Mar 03 20:44:44 fishey1: usually bitbake will notice direct deps Mar 03 20:56:33 khem: I just logged a bug for it 7388, but the journal appears to be empty after boot. Mar 03 21:01:48 khem: kind of Mar 03 21:47:18 hum, I'm building a yocto recipe that needs that call -lpthread while compiling, but still I'm getting stuff like this: undefined reference to `pthread_mutexattr_init' Mar 03 21:47:29 anyone have an idea where I could investigate? thanks! Mar 03 21:49:05 (using poky and arm-poky-linux-gnueabi) Mar 03 21:49:34 -pthread is the appropriate way to use libpthread when using gcc Mar 03 21:52:06 ah ok I'll try this ! Mar 03 23:19:43 I'm trying to add a package from github where the name of the package and the directory created are different. How do I handle thaat in my recipe? Mar 03 23:21:14 example: the package is "opendnp3", and the dircctory fetched by git is "dnp3" Mar 03 23:23:21 cyclist: typically the source is fetched from git to a subdirectory called "git" Mar 03 23:23:54 thanks Mar 03 23:23:55 well, unpacked to that subdirectory, technically Mar 03 23:23:58 which means you would set S ="${WORKDIR}/git" in the recipe to point to that Mar 03 23:25:44 so it's smart enough to determine the path downloaded by git Mar 03 23:31:38 Hey all, I'm in the process of standing up a new image on an x86-64 target and am having trouble with networking Mar 03 23:33:18 I can see my CPU's networking components in lspci, but when I try to add any interfaces (ip link set eth0 up) I get SIOCGIFFLAGS: No such device Mar 03 23:35:15 I have a custom distro that is closely derived from poky, and am building for the genericx86-64 machine Mar 03 23:35:21 cyclist: not really, it specifies it when running git Mar 03 23:36:28 The only interface available is lo/loopback. Not really sure where to get started on this one. Any advice, nudges in the right direction, or general methodology for approaching this kind of problem would be super helpful Mar 03 23:40:36 frankie99: if I had to guess I would say the appropriate driver isn't enabled in the kernel (either built in or as a module) Mar 03 23:41:29 bluelightning: In general, do networking drivers tend to have both built-in and modular components, or is it either-or? Mar 03 23:43:45 either-or, like most drivers in the kernel **** ENDING LOGGING AT Wed Mar 04 02:59:58 2015