Re: [PATCH 1/1] kconfig: warn if an unknown symbol is selected

[Michal Marek <mmarek@suse.cz>]

On 2015-01-28 22:32, Paul Bolle wrote:
> On Wed, 2015-01-28 at 16:26 +0100, Michal Marek wrote:
>> On 2015-01-28 10:54, Paul Bolle wrote:
>>> Now that we've been told Yann has disappeared, would you consider taking
>>> this patch into one of your trees? It would be nice to have people
>>> actually use it for a while.
>>>
>>> Or should we first clean up (most of) the warnings it generates? There
>>
>> It seems your fixes have been accepted in the meantime, but there is one
>> new:
>> *** Default configuration is based on 'x86_64_defconfig'
>> drivers/misc/cxl/Kconfig:8:warning: 'CXL_BASE' selects unknown symbol
>> 'PPC_COPRO_BASE'
>>
>> PPC_COPRO_BASE is not known on x86, but at the same time, there is no
>> way for CXL_BASE to be selected on x86:
>>
>> config CXL_BASE
>> 	bool
>> 	default n
>> 	select PPC_COPRO_BASE
>>
>> config CXL
>> 	tristate "Support for IBM Coherent Accelerators (CXL)"
>> 	depends on PPC_POWERNV && PCI_MSI
>> 	select CXL_BASE
>> 	default m
>>
>> Shouldn't we only warn about a select when it is triggered? An
>> allyesconfig would still report all bogus selects for given architecture.
> 
> Good catch! I think checking only for selects done by Kconfig symbols
> that are not "n" might do the trick. Expect a v2 one of these days.
> 
> The current load of issues (for allyesconfig on all arches except um) is
> pasted at the bottom of this message (with some comments added). Looking
> at that load I think it might be preferable to first most of fix those
> issues before adding this to kconfig.

Maybe. OTOH, you have no warnings on x86 and powerpc and you already
have patches for the arm(64) warnings, so I think this can go in.

Michal