**** BEGIN LOGGING AT Mon Dec 11 03:00:02 2017
Dec 11 08:06:18 <nayfe> <lukma> maybe with something like require ${PN}_{MACHINE}.inc, not sure
Dec 11 08:56:25 <paulbarker> Hey, I've put up a wiki page for FOSDEM: https://www.openembedded.org/wiki/FOSDEM_2018. Please add yourself to this if you can help out with the OE stand or if you're attending and just want to say hello!
Dec 11 08:59:53 <RP> bluelightning: did you come to a conclusion on that patch?
Dec 11 09:00:31 <bluelightning> RP: I didn't look into it any more closely - we should just merge it I think
Dec 11 09:02:36 <RP> bluelightning: ok, I'm just hoping to run a few builds today...
Dec 11 09:05:24 <yocti> New news from stackoverflow: C++ Run a command detached from its parent so the parent can die w/o killing its child <https://stackoverflow.com/questions/47748990/c-run-a-command-detached-from-its-parent-so-the-parent-can-die-w-o-killing-its> || Yocto - Add Files to kernel-sources for every kernel version <https://stackoverflow.com/questions/47749022/yocto-add-files-to-kernel-sources-for-every-kernel-version>
Dec 11 09:48:36 <mckoan> khem: yes rpi3 graphics with x11 and mesa works
Dec 11 10:33:09 <sagner> nayfe: does imx-gst1.0-plugin compiles for you
Dec 11 10:33:33 <sagner> nayfe: I really feel everything rocko I touch is broken right now :-(
Dec 11 10:34:14 <nayfe> sagner: it does not fetch
Dec 11 10:34:40 <sagner> nayfe: yeah that is what I get too.. SRCREV seems to be missing.
Dec 11 10:35:13 <sagner> nayfe: when I set a valid SRCREV compile fails though...
Dec 11 10:35:43 <nayfe> that error on gstreamer1.0-plugins-imx is really strange
Dec 11 10:36:15 <nayfe> it should not compile anything in /usr/lib64 imo
Dec 11 10:37:50 <nayfe> sagner: i'm checking https://github.com/Freescale/gstreamer-imx/compare/4c73ba67174921164a0631f78910f0788544cb49...master for diffs between 0.12.3 and 0.13.0 nothing obvious
Dec 11 10:39:46 <cornel> hello
Dec 11 10:40:18 <cornel> is there a way to tell ALL binaries in a yocto project to compile using a predefined set of flags?
Dec 11 10:40:32 <nayfe> sagner: in <work>/../gstreamer1.0-plugins-imx/0.13.0-r0 when i do find . -name 'libgstimxcommon.so.0.13.0' i don't have it in any lib64 folder, do you ?
Dec 11 10:40:58 <LetoThe2nd> cornel: like a machine tune or such?
Dec 11 10:56:31 <cornel> LetoThe2nd,
Dec 11 10:56:49 <cornel> like in: Position Independent Executable
Dec 11 10:58:00 <LetoThe2nd> cornel: be inspired: https://github.com/openembedded/openembedded-core/blob/master/meta/conf/distro/include/security_flags.inc
Dec 11 10:59:34 <cornel> LetoThe2nd, thank you, Sir! But, will all the binaries use this?
Dec 11 11:00:42 <rheagar> hi , is there a way to add some conditonal judge for require/include ? thanks
Dec 11 11:01:22 <LetoThe2nd> cornel: unless some badly behavin recipe manually overrides it. and there is code that cannot be compiled with fPIE.
Dec 11 11:01:43 <cornel> LetoThe2nd, ok, thank you very much
Dec 11 11:01:50 <rheagar> by use distro_features , to choose whether require/include an inc file ?
Dec 11 11:05:32 <LetoThe2nd> rburton: can you please try to rephrase?
Dec 11 11:05:37 <LetoThe2nd> rheagar: ^^^^
Dec 11 11:13:38 <rheagar> LetoThe2nd: there is an data.inc used to create a data partiton image, I put require data.inc in a common bbfile,  for some platform, i want to include it , while not for others. Is there a way to handle this ?
Dec 11 11:16:04 <LetoThe2nd> mabye you could use packageconfigs in that common bbfile, and select them through machine and/or distro, but your description is a bit vague to me.
Dec 11 11:35:53 <yocti> New news from stackoverflow: Error: Nothing provides u-boot, Either UBOOT_MACHINE or UBOOT_CONFIG must be set Apalis imx6? <https://stackoverflow.com/questions/47751637/error-nothing-provides-u-boot-either-uboot-machine-or-uboot-config-must-be-set>
Dec 11 11:38:52 <sagner> nayfe: currently I reverted gstreamer1.0-plugins-imx to 0.12.2
Dec 11 11:41:54 <sagner> nayfe: yes, for me libgstimxcommon.so.0.13.0 is in lib64
Dec 11 11:42:48 <sagner> nayfe: I did not build a aarch64 machine with that build directory... and file ...libgstimxcommon.so.0.13.0 says its a 32-bit ARM EABI5 binary, so no idea why it got deployed into a lib64 folder
Dec 11 11:52:34 <otavio> rburton: there is one change pending, the mesa upgrade
Dec 11 11:53:02 <otavio> sagner: the lib dir is stange. Is it using any distro?
Dec 11 11:56:27 <sagner> otavio: yeah Ångstrom...
Dec 11 11:57:38 <sagner> otavio: waf seems to use the wrong lib dir...
Dec 11 12:02:12 <sagner> nayfe: waf got updated from 1.7.11 to 1.9.3, I guess that could be related. Still interesting that it works for some
Dec 11 12:09:23 <mcfrisk> hi, what is the status of multiple kernel support in poky master? can multiple kernels for a single machine be built and used nicely with also external modules?
Dec 11 12:12:42 <otavio> sagner: are you mixing branches?
Dec 11 12:32:49 <sagner> otavio: not really, all rocko or rocko-next
Dec 11 12:55:27 <nayfe> sagner:  do you have gstreamer installed on host ?
Dec 11 12:59:10 <sagner> nayfe: yes
Dec 11 12:59:53 <sagner> nayfe, otavio: Just could reporduce it using fslc-framebuffer distro and pure freescale community bsp, imx6qdlsabresd machine
Dec 11 13:00:47 <sagner> nayfe: I think it has to do with the waf update... waf --help says:
Dec 11 13:00:48 <nayfe> sagner: waf looks good candidate, what is your host distro ?
Dec 11 13:00:49 <sagner> nayfe: --libdir=LIBDIR     object code libraries [EXEC_PREFIX/lib64]
Dec 11 13:00:58 <sagner> nayfe: so it seems to default to lib64...
Dec 11 13:01:15 <otavio> sagner: likely you need to pass: --libdir=${libdir}
Dec 11 13:01:24 <otavio> sagner: make sure it is using the waf class
Dec 11 13:01:42 <sagner> nayfe: Fedora 26
Dec 11 13:01:51 <nayfe> sagner: you have waf on your host ?
Dec 11 13:02:28 <sagner> otavio: the gstreamer1.0-plugins-imx recipe is inheriting waf...
Dec 11 13:02:45 <sagner> nayfe: no, don't think so (at least just executing waf says command not found.
Dec 11 13:04:40 <sagner> otavio: according to temp/run.do_configure we do set --prefix=/usr... maybe we should also add --libdir=...
Dec 11 13:07:20 <nayfe> sagner: i dont have "object code libraries" stuff in my waf --help
Dec 11 13:07:44 <sagner> nayfe: bitbake gstreamer1.0-plugins-imx -c devshell, then ./waf --hel
Dec 11 13:08:39 <nayfe> --libdir=LIBDIR     object code libraries [EXEC_PREFIX/lib]
Dec 11 13:09:22 <sagner> nayfe: interesting... I guess that is some runtime detection which ends up to provide different results
Dec 11 13:09:28 <sagner> nayfe: what distro are you using?
Dec 11 13:09:44 <rburton> what package is this?
Dec 11 13:09:45 <nayfe> mint 18
Dec 11 13:09:55 <nayfe> rburton: gstreamer1.0-plugins-imx
Dec 11 13:10:12 <nayfe> rburton: it uses waf class
Dec 11 13:10:13 <rburton> remember some maintainers "are clever" and see the host is 64-bit fedora so set libdir
Dec 11 13:10:45 <sagner> rburton: yeah that sounds like such a case
Dec 11 13:11:49 <rburton> passing --libdir is the right fix, but for that recipe you need --plugin-install-path
Dec 11 13:11:54 <rburton> https://github.com/Freescale/gstreamer-imx/blob/master/wscript#L55
Dec 11 13:15:25 <sagner> rburton: you mean we should explicitly set --plugin-install-path? Default seems reasonable there...
Dec 11 13:16:01 <sagner> rburton: add --libdir to waf.bbclass
Dec 11 13:32:32 <sagner> rburton: FWIW, it works with setting only --libdir. gstreamer plugins seem to end up in the right location with that (/usr/lib/gstreamer-1.0)
Dec 11 13:38:05 <nayfe> otavio rburton sagner: Thanks Stefan, do you submit patch?
Dec 11 13:39:26 <kosta> Guys, I'm building safplus and at "configure" stage I get error "checking for glib-2.0 2.2.0... .../SAFplus/configure: line 7505: pkg-config: command not found". But doing "configure" from devshell is ok. Why?
Dec 11 13:40:23 <nayfe> kosta: do your recipe inherit pkg-config ?
Dec 11 13:40:30 <sagner> nayfe: will do
Dec 11 13:40:50 <kosta> nayfe: No. Only autotools.
Dec 11 13:41:17 <nayfe> kosta: try to add it :)
Dec 11 13:41:38 <kosta> nayfe: Tried: ERROR: ParseError at /home/mrv/My/MRV/yocto_all/rocko_all/poky/meta-openclovis/recipes-extended/safplus/safplus_6.1.bb:20: Could not inherit file classes/pkg-config.bbclass
Dec 11 13:42:14 <kosta> nayfe: Worked now!
Dec 11 13:42:37 <kosta> nayfe: I missplelled. The right way is "inherit pkgconfig" (without dash).
Dec 11 13:42:43 <kosta> nayfe: Thanks man!!
Dec 11 13:43:02 <nayfe> kosta: np :)
Dec 11 13:43:39 <rburton> the inherit is just a glorified way of doing depends = pkgconfig-native
Dec 11 13:43:46 <rburton> which is why it couldn't find pkgconfig :)
Dec 11 13:44:14 <kosta> rbuton: Thanks. Yeah...
Dec 11 13:44:40 <nayfe> rburton: can you explain why you think we need --plugin-install-path stuff in gstreamer recipe ?
Dec 11 13:44:55 <rburton> you don't if the defaut works after --libdir is passed
Dec 11 13:45:29 <nayfe> rburton: ok ty
Dec 11 14:36:31 <yocti> New news from stackoverflow: Bitbake AttributeError: 'module' object has no attribute 'getVar'? <https://stackoverflow.com/questions/47754767/bitbake-attributeerror-module-object-has-no-attribute-getvar>
Dec 11 15:07:19 <rburton> moto-timo: around?
Dec 11 15:44:25 <lukma> Maybe somebody would know........
Dec 11 15:44:30 <lukma> :)
Dec 11 15:45:03 <lukma> Is there a way to have base-files_%.bbappend to be executed only for COMPATIBLE_MACHINE_foo = "foo" ?
Dec 11 15:46:12 <RP> lukma: not really at this point in time
Dec 11 15:46:53 <RP> lukma: include somefilename-${MACHINE}.inc ?
Dec 11 15:47:49 <lukma> RP: How this include somefilename-${MACHINE}.inc works?
Dec 11 15:48:15 <lukma> do you mean by putting all the "Machine specific stuff" to thins *.inc file
Dec 11 15:48:36 <lukma> and the it would be resolved for the specific machine?
Dec 11 15:49:11 <lukma> RP: In my case I do have base-files_ (which are rather large recipe)
Dec 11 15:49:27 <lukma> I would like to add/modify fstab on one machine
Dec 11 15:50:07 <lukma> do_install_append_foo () {
Dec 11 15:50:07 <lukma>         install -m 0644 ${WORKDIR}/fstab_foo ${D}${sysconfdir}/fstab
Dec 11 15:50:07 <lukma> }
Dec 11 15:50:09 <RP> lukma: I think that might allow you to include that file only for the matching machine. But I've not tried it and it may not work
Dec 11 15:50:39 <nayfe> <lukma>: maybe try to just create some temp vars and use override ?
Dec 11 15:50:55 <lukma> The problem is that the base-files_%.bbappend is always appended no matter if I do use compatible machine or not
Dec 11 15:51:40 <lukma> And when from machine specific package I try to call  install -m 0644 ${WORKDIR}/fstab_foo ${D}${sysconfdir}/fstab, OE compains that I try to provide fstab from two sources (which is true)
Dec 11 15:53:02 <lukma> I could hack it a bit and remove the original fstab -> but this poses dependency on which recipe (base_file or my_adjust_pkg) is executed first
Dec 11 15:55:03 <nayfe> lukma: what you should do is put the file in <recipe>/files/<machine>/fstab so it is only resolved for the correct machine
Dec 11 15:58:25 <lukma> nayfe: But for the other machine I do want to use default fstab
Dec 11 15:59:58 <nayfe> lukma: when using FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:" it choose file in multiple folders, so it will resolve fstab in default folder
Dec 11 16:00:05 <nayfe> or FILESEXTRAPATHS_prepend := "${THISDIR}/files:"
Dec 11 16:00:49 <lukma> So the ${THISDIR}/${PN}: must be different in appended file than in the base one?
Dec 11 16:01:45 <nayfe> in your base-files_%.bbappend, you should already have set FILESEXTRAPATHS_prepend right ?
Dec 11 16:01:53 <lukma> yes
Dec 11 16:02:13 <lukma> it is set to FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:
Dec 11 16:02:30 <lukma> and then I do have SRC_URI += "file://fstab_foo"
Dec 11 16:03:06 <nayfe> you should keep "fstab" file name
Dec 11 16:03:22 <nayfe> and just let path override choose the right file for you
Dec 11 16:03:40 <lukma> nayfe: ...... I will check it
Dec 11 16:04:10 <lukma> and one more problem SRC_URI_foo += "file..." only gives me file on the overall SRC_URI ....
Dec 11 16:04:11 <nayfe> so if you put some files in meta-foo/<recipe>/base-files/<machine>/fstab it will choose it, then it will choose poky/....../fstab if none is found
Dec 11 16:05:22 <nayfe> with that solution, you just need to put FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:" in base-file_%.bbappend, as fstab is already set in SRC_URI and FILES_${PN}
Dec 11 16:09:10 <lukma> nayfe: Is the <machine> part searched by default?
Dec 11 16:09:26 <nayfe> yes
Dec 11 16:09:36 <lukma> does it has higher precedence ?
Dec 11 16:10:17 <nayfe> yes
Dec 11 16:11:11 <lukma> nayfe: Indeed, the <machine>/fstab was choosen
Dec 11 16:11:40 <lukma> is it written somewhere ? or it is a rule that if present stuff is took from <machine> subdir?
Dec 11 16:12:48 <nayfe> http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#best-practices-to-follow-when-creating-layers
Dec 11 16:14:08 <lukma> nayfe: thanks, this simplifies things a lot
Dec 11 16:14:27 <nayfe> yocto rocks :p
Dec 11 16:16:47 <nayfe> a lot of things are in manual, but it's big ;)
Dec 11 16:17:10 <nayfe> -a
Dec 11 16:17:21 <lukma> nayfe: Yocto/OE is very powerful, if you know how to use it
Dec 11 16:17:53 <lukma> nayfe: And I do wonder, how it will behave when one would need to fix some thing in e.g. 6 years time
Dec 11 16:18:28 <lukma> would I need a VM with nowaday distro or I could run OE/Yocto on a "then" contemporary machine
Dec 11 16:21:03 <nayfe> <lukma> in 6 years we will have a yocto chip in our  brain :)
Dec 11 16:24:16 <Phlogistique> My understanding of RDEPENDS is that it's purely a "packages" thing; and my understanding of the nativesdk feature is that it's purely a "sysroot" thing, i.e. the SDK is generated from the sysroot without using the packages; however the file-rdeps check uses RDEPENDS even for nativesdk packages, why is that?
Dec 11 16:44:53 <ythl> Is there a way to tell bitbake not to include in the kernel in the RootFS?
Dec 11 16:45:36 <ythl> By default it seems to include the Image in `/boot` which is an extra ~13M uncompressed
Dec 11 16:46:08 <ythl> Right now I'm manually unzipping the rootfs, removing the kernel, and rezipping it
Dec 11 16:52:25 <majuk> Hey all, have an issue I don't understand. Trying to change a kernel module to built-in. When I run 'bitbake -c menuconfig linux-fslc-imx', the module is shown as built-in. However, when I change to linux-fslc-imx's working directory and 'make menuconfig', it's shows it's a module.
Dec 11 17:01:51 <nayfe> majuk: when doing bitbake -c menuconfig linux-fslc-imx it loads defconfig from path, and when you do make menuconfig it probably loads default defconfig
Dec 11 17:09:30 <nayfe> <ythl> good question :p
Dec 11 17:13:42 <nayfe> ythl: one way is to remove /boot with ROOTFS_POSTPROCESS_COMMAND
Dec 11 17:33:33 <BubuIIC> I there a way to easily specify a local directory as source for a recipe? i.e. skip all remote fetching and checkout and just build the source that I I want to test?
Dec 11 17:33:35 <otavio> nayfe: I've been using Docker for builds, so it is easy to replicate the setup in future if need
Dec 11 17:33:59 <otavio> majuk: which module?
Dec 11 17:34:12 <rburton> BubuIIC: for local development? use externalsrc.  if its just a simple bit of source, you can drop it alongsie the recipe.
Dec 11 17:34:27 <majuk> otavio: I got it worked out, thanks. Getting USB Gadget support
Dec 11 17:34:36 <otavio> majuk: ok
Dec 11 17:35:21 <BubuIIC> rburton: externalsrc looks like what I've been looking for, thx
Dec 11 17:36:26 <otavio> BubuIIC: I usually prefer using a toolchain
Dec 11 17:36:38 <otavio> BubuIIC: but it is personal preference
Dec 11 17:37:44 <BubuIIC> otavio: ah, right I guess that would be an option as well
Dec 11 17:39:15 <nayfe> otavio: nice, i'll need to test Docker one day ...
Dec 11 17:45:06 <oni> Hello there. Does anyone know how to increase the u-boot "press any key to enter setup" delay?
Dec 11 17:45:23 <oni> It seems to be stuck on 0 seconds which gives me no time to do anything
Dec 11 17:45:48 <oni> I can see the poky recipe for uboot-config but nothing in there regarding the time delay
Dec 11 17:46:08 <nayfe> oni: you need to check for CONFIG_BOOTDELAY
Dec 11 17:46:28 <oni> Yeah, that isn't set anywhere
Dec 11 17:46:52 <nayfe> oni: you can add CONFIG_BOOTDELAY=5 if you want 5sec delay
Dec 11 17:47:01 <oni> ah yes, but where does one add that?
Dec 11 17:47:26 <oni> poky/meta/classes/uboot-config.bbclass maybe?
Dec 11 17:47:46 <oni> Im working with the imx28evk so its all in the meta-freescale layer I suspect
Dec 11 17:48:01 <nayfe> oni: you need to customize your u-boot defconfig
Dec 11 17:48:18 <oni> defconfig implies a make menuconfig?
Dec 11 17:50:13 <oni> imx28evk-fslc-linux-gnueabi/u-boot-fslc/v2016.11+gitAUTOINC+c44711d91e-r0/git/configs
Dec 11 17:50:26 <oni> Looks likek there are a few defconfigs in there.
Dec 11 17:51:47 <oni> thanks. I'll give that a try. Presumably bitbake is smart enough to rebuild u-boot and repackage if I change things?
Dec 11 17:53:24 <nayfe> oni: in ./meta-freescale/conf/machine/imx28evk.conf you have UBOOT_CONFIG variable
Dec 11 17:53:37 <nayfe> oni: it points to u-boot defconfig
Dec 11 17:53:52 <oni> Yeah, I saw that. I wasn't sure what it did
Dec 11 17:54:22 <nayfe> oni: so, you probably need to copy that file and make your own machine configuration
Dec 11 17:54:40 <nayfe> and provide your defconfig for it
Dec 11 17:54:57 <oni> hmmmm, I think I see what you mean
Dec 11 17:55:05 <oni> I've altered the existing one
Dec 11 17:55:36 <nayfe> maybe you can just make a patch for the defconfig and put it in SRC_URI
Dec 11 17:55:57 <oni> SRC_URI? Yes, I heard that mentioned. Ideally that will be the thing
Dec 11 17:56:09 <oni> If I can see a change I've made reflected that will be a great start
Dec 11 17:56:53 <oni> Damn! That didn't work
Dec 11 17:57:07 <oni> Its as if the changes I make just dont make it onto the board
Dec 11 17:57:26 <oni> uboot appears to be rebuilt as does the image
Dec 11 17:57:51 <nayfe> you can create a u-boot-fslc_%.bbappend file in a custom layer and create a patch for correct defconfig file, can be mx28evk_config, mx28evk_nand_config or mx28evk_auart_console_config
Dec 11 18:00:05 <oni> thats a neat idea. I've no idea how best to go about that. Im just bewildered as to why a change to a defconfig makes absolutely no effect
Dec 11 18:01:44 <oni> Im trying to get that 'helloworld' of embedded working. If I can see a change I know where I'm at
Dec 11 18:04:04 <nayfe> to make a patch, you can use quilt tool, go in mx28evk-fslc-linux-gnueabi/u-boot-fslc/v2016.11+gitAUTOINC+c44711d91e-r0/git then quilt new 0001-add-bootdelay.patch quilt add configs/mx28evk* modify defconfig files, quilt refresh copy patch/0001-add-bootdelay.patch into meta-foo/recipes-bsp/u-boot-fslc/files/ then create a  u-boot-fslc_%.bbappend with  FILESEXTRAPATHS_prepend := "${THISDIR}/files:" and SRC_URI +=
Dec 11 18:04:05 <nayfe> "file://0001-add-..."
Dec 11 18:06:32 <oni> wowzers! thanks
Dec 11 18:07:39 <nayfe> oni: it's a bit dense
Dec 11 18:08:14 <oni> Whne you say modify the defconfig files, what do you mean?
Dec 11 18:08:34 <oni> do you mean modify the patch file?
Dec 11 18:10:08 <moto-timo> rburton: pong
Dec 11 18:10:12 <oni> Ah, alter the config files themselves. Gotcha
Dec 11 18:16:52 <oni> nayfe so I think I follow you. Cheers! I shall try a new rebuild
Dec 11 18:19:38 <rburton> moto-timo: not 100% sure about the manifest changes yet, can the pypi stuff be applied without the manifest changes?
Dec 11 18:20:27 <moto-timo> rburton: I would have to go back to my old branch and make it work on top of existing master
Dec 11 18:20:45 <moto-timo> rburton: which I can do if the json manifest is going to wait awhile
Dec 11 18:24:21 <moto-timo> rburton: by old branch I mean oe-core-contrib/ttorling/pypi
Dec 11 18:24:39 <rburton> if that rebases easily then that would be awesome :)
Dec 11 18:24:55 <moto-timo> rburton: sure thing
Dec 11 18:38:57 <oni> Damn! Still no change
Dec 11 18:39:30 <oni> its still 0 seconds. I must admit, this is really infuritating! This yocto stuff - it's very hard to make a simple change :(
Dec 11 18:44:47 <oni> patch file looks good, u-boot-fslc appears to be rebuilt
Dec 11 18:49:20 <oni> I get an sdcard image which I manage to copy to the sdcard. The board appears to boot from it but what I'm seeing gives me pause for thought
Dec 11 18:49:23 <oni> U-Boot 2015.10-00407-g83bf005-dirty (Oct 26 2016 - 09:39:34 +0100)
Dec 11 18:50:13 <oni> I get the impression this is not the correct u-boot
Dec 11 19:04:31 <moto-timo> rburton: I will also add my change to add newlines to the existing manifest generator so it won't barf on git send-mail after this round
Dec 11 19:16:45 <nayfe> oni: back
Dec 11 19:17:14 <nayfe> oni: well, it seems your board does not boot from sdcard
Dec 11 19:18:33 <nayfe> oni: you can do something if you have fw_setenv tool
Dec 11 19:19:15 <nayfe> oni: fw_setenv bootdelay=5 or fw_setenv bootdelay 5 (i don't remember correct syntax everytime)
Dec 11 19:21:32 <nayfe> oni: maybe you have u-boot in special partition mmcblk0boot0 or in another memory type, dont know mv28evk board
Dec 11 19:28:06 <oni> The worst thing about this is that the supplier of the board is not very forthcoming at all
Dec 11 19:28:16 <oni> its based on the imx28evk.
Dec 11 19:28:50 <oni> I know it loads some things from the sdcard thats for sure.
Dec 11 19:30:00 <oni> Annoyingly, this is the least of my problems. The kernel itself doesnt load at all. It halts at "Starting kernel ..." or at least that is what appears to happen
Dec 11 19:30:36 <oni> I thought if I could start with the bootloader that would be best
Dec 11 19:32:49 <nayfe> oni: on a lot of imx boards, there are dip switch to change how board boots
Dec 11 19:33:57 <oni> Good tip. I cant see any dips but there are plenty of headers
Dec 11 19:35:30 <oni> This would tally with other things I've seen, e.g uEnv.txt not working
Dec 11 19:36:18 <oni> tbh, I wouldnt mind but I can't appear to load the mainline fslc kernel. all this messing with the bootloader is to try and get more messages out of the kernel
Dec 11 19:37:25 <nayfe> https://www.nxp.com/docs/en/user-guide/IMX28EVKHUG.pdf
Dec 11 19:37:38 <nayfe> page 10
Dec 11 19:39:14 <oni> Ah yes. this board is much worse :p
Dec 11 19:41:36 <nayfe> oni: maybe you can try to use mfgtool/imx_loader to load u-boot
Dec 11 19:41:44 <oni> http://www.noritake-itron.com/NewWeb/TFT/Overview/Overview.asp
Dec 11 19:42:22 <oni> I'm not familiar with mfgtool/imx_loader - thanks. I'll have a look
Dec 11 19:49:05 <nayfe> oni: thing is, if you have a custom board, you should ask vendor for support, as you will probably need to customize ddr timings, device tree etc ...
Dec 11 19:49:19 <oni> absolutely, and they've been awful
Dec 11 19:49:39 <oni> debating that a different supplier might be needed
Dec 11 20:05:41 <oni> nayfe apparently there is a jumper
Dec 11 20:06:05 <oni> according to the PDF
Dec 11 20:06:07 <oni> 10, J11, J12, J14
Dec 11 20:06:07 <oni> Boot Mode
Dec 11 20:06:08 <oni> Internal boot selection, do not change
Dec 11 20:11:39 <oni> I suspect that means I'm locked out of the bootloader. These jumpers are hidden behind another board
Dec 11 20:13:20 <oni> nayfe thanks a lot for your help. I may have to admit defeat
Dec 11 20:15:12 <nayfe> oni: np and sorry for your defeat, but without complete documentation and sometimes its not enough, it can be impossible :)
Dec 11 20:15:33 <oni> Yeah :(
Dec 11 20:15:50 <oni> Is there no other way to get information out of the kernel or to find out why it halts?
Dec 11 20:16:17 <oni> The instructions given rely on the dizzy branch of yocto. This is quite old sadly
Dec 11 20:16:55 <oni> Although I've managed to adapt this older version to work - getting the screen to fire up remains a problem
Dec 11 20:17:47 <oni> My thinking was to try the latest software instead, but sadly, it appears to just hang with the serial console not doing very much
Dec 11 20:19:57 <nayfe> oni: if you don't have access to u-boot due to bootdelay and you can't load another u-boot by other way, then it will be hard to know what's wrong
Dec 11 20:20:23 <oni> This is my suspicion yes
Dec 11 20:40:27 <seebs> nooo rsync
Dec 11 20:40:32 <seebs> don't give generic error messages
Dec 11 20:40:34 <seebs> bad rsync
Dec 11 20:40:46 <seebs> (not really yocto-related, i just felt like sharing)
Dec 11 21:38:35 <paulbarker> Anyone else getting an error from https://www.yoctoproject.org/ ?
Dec 11 21:39:14 <nayfe> Same
Dec 11 21:40:27 <neverpanic> worksforme
Dec 11 21:44:52 <paulbarker> its back now :)
Dec 11 21:49:48 <RP> armpit: are the pyro-next patches ready to merge?
Dec 11 21:51:22 <armpit> RP, yes
Dec 11 21:51:31 <armpit> stable/pyro-next
Dec 11 21:52:19 <armpit> I would have sent a pull request bug I needed to double check the binutils changes to make sure Rocky was not affected
Dec 11 21:56:53 <RP> armpit: that was mainly what I was wondering :/
Dec 11 21:57:57 <armpit> RP. all good. I put version info in the commit message Rocko is @ 2.29 and all the fixes for for <= 2.28
Dec 11 21:58:37 * armpit can't believe i had my head screwed on correct for that one ; )
Dec 11 22:00:16 <RP> armpit: :D
Dec 11 22:00:37 <RP> armpit: cool, thanks. I'll merge those and then we can perhaps get the pyro and morty point releases done :)
Dec 11 22:01:06 <armpit> that would be nice... been a long road
Dec 11 22:03:00 <RP> armpit: it has indeed...
Dec 12 00:01:19 <rburton> note to self: when firing builds to run overnight, actually press the Run button
**** ENDING LOGGING AT Tue Dec 12 03:00:01 2017