[meta-freescale] Help - new to Yocto

Nikolay Dimitrov picmaster at mail.bg
Fri Aug 28 03:00:06 PDT 2015

Hi Terry,

On 08/28/2015 08:29 AM, Terry Farnham wrote:
> I attempted a clean build today to see if I could get past this
> issue.  I am still encountering the same compile problem.
> I’m not exactly sure where I must be going wrong with the steps for
> setting up to build OE and the Yocto packages, but it always seems to
> fail to build the ptpd daemon.  Looking at the Makefiles that are
> generated, it seems to me that the mechanics which sets up the
> compiler include directives are generating an include reference into
> the host machines /usr/include folder rather than into the proper arm
> version of the includes.  I attached the Makefile from my
> "<build_folder>/tmp/work/cortexa9hf-vfp-neon-poky-linux-gnueabi/ptpd/2.3.1-rc2-r0/build"
> folder.  You can see that the includedir seems to improperly
> reference /usr/include.  The error that gets generated is a result of
> including from nested includes that eventually land on
> /usr/include/bits/byteswap.h which happens to have x86 specific
> assembler in it.  I’m surprised that I am not seeing more compile
> errors than this, as I suspect that a lot of the code being compiled
> from the cortexa9hf-vfp-neon-poky-linux-gnueabi is referencing into
> the host machines /usr/include instead of the hardware specific
> version.
> I'm hoping someone has a clue that will help me get past this
> problem.
> Thanks,
> Terry

Just tried to build ptpd on a Debian 7 host and it works like a charm.
Here's what I'm doing:

// Get Yocto
mkdir ~/yocto-master
cd ~/yocto-master
repo init -u https://github.com/Freescale/fsl-community-bsp-platform -b fido
repo sync
. setup-environment build

// Change PACKAGE_CLASSES to package_ipk to avoid waiting for RPMs
nano conf/local.conf

// Add meta-networking and meta-python to BBLAYERS
nano conf/bblayers.conf

// Default MACHINE is imx6qsabresd, so no need to set it explicitly
bitbake ptpd

This worked on both master and fido branches.


More information about the meta-freescale mailing list