[meta-freescale] [meta-fsl-demos][PATCH 01/12] packagegroup-fsl-benchmark.bb: New packagegroup for benchmarks

[Nikolay Dimitrov]

Hi guys,

On 08/25/2015 08:06 PM, Otavio Salvador wrote:
> On Tue, Aug 25, 2015 at 11:49 AM, Dominic Sacré <dominic.sacre at gmx.de> wrote:
>> On 2015-08-25 16:20, Otavio Salvador wrote:
>>> On Tue, Aug 25, 2015 at 5:22 AM, Nikolay Dimitrov <picmaster at mail.bg> wrote:
>>>> Can you please exclude "fio" from RDEPENDS_${PN}-extended, at least for
>>>> ARM boards? Here's more info about the issue:
>>>>
>>>> https://github.com/Freescale/meta-fsl-demos/commit/836bdf5a9d500ed258f1ddc07d89eca74a704098
>>>
>>> This is the very same packagegroup as
>>> packagegroup-fsl-tools-benchmark; there are two differences though:
>>>
>>>   a. it is split in core and extended
>>>   b. it includes fio
>>>
>>> The core and extended for a benchmark packagegroup seems pointless as
>>> both offer same functionality. So I think what we have is more
>>> suitable for general use.
>>>
>>> I thus ask you to remove this patch for next patchset. The 'fio'
>>> package has been removed on purpose as it was causing builds to fail
>>> so if it needs to be added back a more careful test needs to be done
>>> before reverting the patch which has removed it.
>>
>> The issue with fio was just temporary (a QA fix for x86 broke ARM
>> builds) and is fixed in meta-openembedded master:
>>
>> https://github.com/openembedded/meta-openembedded/commit/eb01b36cb3f2f686bc05c690b06d1e0680dda1f5
>>
>> There's no need to exclude fio from packagegroup-fsl-tools-benchmark.
>
> This is indeed what I thought and it must to be proper tested so we
> can revert the patch which removes it with a proper commit log. As you
> been testing it, can you propose the revert?
>

Just did a fresh build and can confirm that now fio compiles OK on
master and fido for all supported ARM machines.

Btw I won't be able to test the package. If someone can test it, please
share.

Also, if noone else wants to revert my patch, I can do it :D.

Regards,
Nikolay