Re: [PATCHv1] firmware: intel_stratix10_service: add hardware dependency

[Dinh Nguyen]

On 1/23/19 11:01 AM, Alan Tull wrote:
> On Wed, Jan 23, 2019 at 10:42 AM Dinh Nguyen <dinguyen@xxxxxxxxxx> wrote:
>>
>>
>>
>> On 1/23/19 10:37 AM, Alan Tull wrote:
>>> On Wed, Jan 23, 2019 at 10:00 AM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>>>
>>> Hi Greg,
>>>
>>>>
>>>> On Wed, Jan 23, 2019 at 09:47:56AM -0600, richard.gong@xxxxxxxxxxxxxxx wrote:
>>>>> From: Richard Gong <richard.gong@xxxxxxxxx>
>>>>>
>>>>> Add a Kconfig dependency to ensure Intel Stratix10 service layer driver
>>>>> can be built only on the platform that supports it.
>>>>>
>>>>> Signed-off-by: Richard Gong <richard.gong@xxxxxxxxx>
>>>>> ---
>>>>>  drivers/firmware/Kconfig | 2 +-
>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig
>>>>> index f754578..cac16c4 100644
>>>>> --- a/drivers/firmware/Kconfig
>>>>> +++ b/drivers/firmware/Kconfig
>>>>> @@ -218,7 +218,7 @@ config FW_CFG_SYSFS_CMDLINE
>>>>>
>>>>>  config INTEL_STRATIX10_SERVICE
>>>>>       tristate "Intel Stratix10 Service Layer"
>>>>> -     depends on HAVE_ARM_SMCCC
>>>>> +     depends on ARCH_STRATIX10 && HAVE_ARM_SMCCC
>>>>
>>>> That's lame, what about building for testing?
>>>
>>> Do you mean this instead?
>>>
>>> depends on (ARCH_STRATIX10 && HAVE_ARM_SMCCC) || COMPILE_TEST
>>>
>>>>
>>>> And is this needed now, for 5.0-final, or can it wait for 5.1?
>>>
>>> This change will reduce kernel size for most arm64.  It can go into
>>> whichever kernel.  We can resubmit allowing for COMPILE_TEST.
>>>
>>
>> I don't see how this is true? ARM64 has a single defconfig and
>> ARCH_STRATIX10 is included in that defconfig. I don't see how adding
>> this dependency will reduce the kernel size for most arm64?
> 
> Sorry, I wasn't clear.  By default it will be built in since there's
> one arm64 defconfig.  But adding ARCH_STRATIX10 dependency here
> supports users who turn off all the ARCH_* they don't care about in
> their own defconfig that won't get upstreamed.
> 

Ah yes, forgot to consider that.

Dinh