[Intel-gfx] [PATCH 2/3] drm/i915: Generate all IS_<platform> macros

Jani Nikula jani.nikula at intel.com
Thu Dec 8 13:40:11 UTC 2016


On Thu, 08 Dec 2016, Jani Nikula <jani.nikula at intel.com> wrote:
> On Thu, 08 Dec 2016, Tvrtko Ursulin <tvrtko.ursulin at linux.intel.com> wrote:
>> On 08/12/2016 10:46, Jani Nikula wrote:
>>> On Thu, 08 Dec 2016, Tvrtko Ursulin <tursulin at ursulin.net> wrote:
>>>> From: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>>>
>>>> Instead of listing them individually we can generate them
>>>> using the new i915_platforms.h header.
>>>>
>>>> Also convert them to a static inline function which
>>>> interestingly makes the code smaller as well.
>>>>
>>>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>>> Cc: Chris Wilson <chris at chris-wilson.co.uk>
>>>> Cc: Jani Nikula <jani.nikula at intel.com>
>>>> Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
>>>
>>> NAK. Absolutely opposed to this.
>>
>> Gee, sounds a bit to harsh to me. :) Didn't we say we are not doing NAKs 
>> any longer?
>
> Only when dropped without rationale. I needed to make it clear in no
> uncertain terms how important this is to me.
>
>>> A large part of my work involves digging through the source tree, and a
>>> crucial part of that is looking up definitions and references, both for
>>> macros and functions. Not having the macro/function definitions breaks
>>> that workflow. Losing that, source code archeology gets *much*
>>> harder. The losses are much greater than the gains.
>>
>> Hm, I struggle to see that point on the same magnitude of a disaster 
>> scale as you. I would have thought we all know what IS_SKYLAKE & co are 
>> so it would be no big deal.
>
> Sure we know what they are; I want to be able to see all the
> *references* to them as well, using GNU global. That fails if they're
> not defined in the first place. And no, git grep is not the same.
>
>> Imagine if we changed it to IS_PLATFORM(SKYLAKE) for instance.
>
> Then all the things passed as parameter would have to be defined.

Oh, btw, patch 1/3 also fails this because the INTEL_FOO enumerations
aren't defined as-is but via macros. This is not as important as the
IS_FOO ones, because the former aren't really used throughout the
source, while the latter are.

BR,
Jani.

>
>
> BR,
> Jani.

-- 
Jani Nikula, Intel Open Source Technology Center


More information about the Intel-gfx mailing list