Re: [PATCH] mm/gup_benchmark: GUP_BENCHMARK depends on DEBUG_FS

[Randy Dunlap <rdunlap@infradead.org>]

On 11/7/20 11:16 AM, John Hubbard wrote:
> On 11/7/20 11:05 AM, Song Bao Hua (Barry Song) wrote:
>>> -----Original Message-----
>>> From: John Hubbard [mailto:jhubbard@nvidia.com]
> ...
>>>>    config GUP_BENCHMARK
>>>>        bool "Enable infrastructure for get_user_pages() and related calls
>>> benchmarking"
>>>> +    depends on DEBUG_FS
>>>
>>>
>>> I think "select DEBUG_FS" is better here. "depends on" has the obnoxious
>>> behavior of hiding the choice from you, if the dependencies aren't already met.
>>> Whereas what the developer *really* wants is a no-nonsense activation of the
>>> choice: "enable GUP_BENCHMARK and the debug fs that it requires".
>>>
>>
>> To some extent, I agree with you. But I still think here it is better to use "depends on".
>> According to
>> https://www.kernel.org/doc/Documentation/kbuild/kconfig-language.txt
>>
>>     select should be used with care. select will force
>>     a symbol to a value without visiting the dependencies.
>>     By abusing select you are able to select a symbol FOO even
>>     if FOO depends on BAR that is not set.
>>     In general use select only for non-visible symbols
>>     (no prompts anywhere) and for symbols with no dependencies.
>>     That will limit the usefulness but on the other hand avoid
>>     the illegal configurations all over.
>>
>> On the other hand, in kernel there are 78 "depends on DEBUG_FS" and
>> only 14 "select DEBUG_FS".
>>
> 
> You're not looking at the best statistics. Go look at what *already* selects
> DEBUG_FS, and you'll find about 50 items.

Sorry, I'm not following you. I see the same 14 "select DEBUG_FS" as Barry.

In general we don't want any one large "feature" (or subsystem) to be enabled
by one driver. If someone has gone to the trouble to disable DEBUG_FS (or whatever),
then a different Kconfig symbol shouldn't undo that.

-- 
~Randy