[meta-freescale] [meta-freescale PATCH 3/3] qoriq-base.inc: Use 'fsl-dynamic-packagearch' class

[Zhenhua Luo]

Hello Otavio, 

> -----Original Message-----
> From: Otavio Salvador [mailto:otavio.salvador at ossystems.com.br]
> Sent: Monday, September 19, 2016 6:48 PM
> 
> On Sun, Sep 18, 2016 at 6:13 AM, Zhenhua Luo <zhenhua.luo at nxp.com> wrote:
> ...
> >> +MACHINE_SOCARCH_FILTER_append_qoriq = " cryptodev-qoriq-linux
> >> cryptodev-qoriq-module cryptodev-qoriq-test openssl-qoriq"
> > [Zhenhua Luo] I don't understand why the MACHINE_SOCARCH definition is
> needed for those packages, can you please elaborate?
> 
> Because it needs to use a different provider for QorIQ machine and as
> consequence the packages depending on those also need to be split out.
>
> Without this, following error happens:
> 
> ERROR: openssl-qoriq-1.0.2h-r0 do_package_write_rpm: If the above message
> is too much, the simpler version is you're advised to wipe out tmp and rebuild
> (reusing sstate is fine). That will likely fix things in most (but not all) cases.
> ERROR: openssl-qoriq-1.0.2h-r0 do_package_write_rpm: Function failed:
> sstate_task_postfunc
> ERROR: Logfile of failure stored in:
> .../build-wayland/tmp/work/cortexa7hf-neon-poky-linux-gnueabi/openssl-
> qoriq/1.0.2h-r0/temp/log.do_package_write_rpm.5308
> recipe openssl-qoriq-1.0.2h-r0: task do_package_write_rpm: Failed
> ERROR: Task (.../sources/meta-freescale/recipes-connectivity/openssl/openssl-
> qoriq_1.0.2h.bb:do_package_write_rpm)
> failed with exit code '1'
[Zhenhua Luo] I still don't know why MACHINE_SOCARCH is needed for openssl and cryptodev, what's the negative impaction without this definition? I didn't meet the build error when openssl-qoriq is built for ls1021atwr, may I know how the issue can be reproduced?

> >> +INHERIT += "fsl-dynamic-packagearch"
> > [Zhenhua Luo] According to the following code in fsl-dynamic-
> packagearch.bbclass, the MACHINE_SOCARCH of qoriq-ppc targets will be set to
> SOCARCH of ARM? This seems to be problematic.
> >
> >         ARM_EXTRA_SOCARCH =
> "${ARMPKGARCH}${ARMPKGSFX_DSP}${ARMPKGSFX_EABI}${ARMPKGSFX_ENDI
> AN}${ARMPKGSFX_FPU}${MACHINE_SOCARCH_SUFFIX}"
> >         THUMB_EXTRA_SOCARCH =
> "${ARMPKGARCH}${ARM_THUMB_SUFFIX}${ARMPKGSFX_DSP}${ARMPKGSFX_E
> ABI}${ARMPKGSFX_ENDIAN}${ARMPKGSFX_FPU}${MACHINE_SOCARCH_SUFFIX
> }"
> >
> >         MACHINE_SOCARCH = "${@bb.utils.contains('ARM_INSTRUCTION_SET',
> 'thumb', '${THUMB_EXTRA_SOCARCH}', '${ARM_EXTRA_SOCARCH}', d)}"
> 
> Good catch, I did not catch this  as all test cases I had were ARM-based. I will
> work on a fix for it.
[Zhenhua Luo] Thanks. 


Best Regards, 

Zhenhua