Re: [PATCH v2] watchdog: Compile possible drivers with COMPILE_TEST

[Krzysztof Kozlowski]

2015-10-15 7:40 GMT+09:00 Guenter Roeck <linux@xxxxxxxxxxxx>:
> On 10/14/2015 03:11 PM, Luis de Bethencourt wrote:
>>
>> On 14/10/15 22:35, Guenter Roeck wrote:
>>>
>>> On 10/14/2015 11:55 AM, Luis de Bethencourt wrote:
>>>>
>>>> These drivers have depends that aren't build dependencies, so it's
>>>> a good idea to allow these drivers to always be built when the
>>>> COMPILE_TEST option is enabled.
>>>>
>>>> That way, the drivers can be built with a config generated by make
>>>> allyesconfig and check if a patch would break the build.
>>>>
>>>> Signed-off-by: Luis de Bethencourt <luisbg@xxxxxxxxxxxxxxx>
>>>
>>>
>>> What architectures have you actually tested this for ?
>>>
>>> Guenter
>>>
>>
>> I've compiled it in x86 and then I tried ARM
>> (with CROSS_COMPILE="ccache arm-linux-gnueabihf-" ARCH=arm)
>>
>> But ARM kept failing due to other drivers, first Lustre then an Ethernet
>> driver (I40E), and more.
>>
> arm:allmodconfig should build fine. Either case, you can always run "make
> -i"
> to ignore errors, to ensure that you at least don't introduce new problems.
>
>> Is there a better way of doing this?
>>
> Sure, at least run allmodconfig for _all_ architectures, or at least
> for the major architectures.

At least 32-bit and 64-bit of mixtures. x86 and ARMv7 is not enough.
For example something like:
$ apt-get install build-essential gcc-4.8-multilib
gcc-arm-linux-gnueabi gcc-aarch64-linux-gnu libc6-dev-arm64-cross
gcc-powerpc-linux-gnu
(PPC can be built for 32- and 64-bit)
And you can get MIPS from CodeSourcery.

Toggling COMPILE_TEST with this may break someone's else allyesconfig.

Best regards,
Krzysztof
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/