[Intel-gfx] [PATCH 01/10] drm/i915: Add i915 device based MISSING_CASE macro

[Jani Nikula]

On Thu, 27 Feb 2020, "Laxminarayan Bharadiya, Pankaj"	<pankaj.laxminarayan.bharadiya at intel.com> wrote:
> Hi Chris,
>
>> -----Original Message-----
>> From: Chris Wilson <chris at chris-wilson.co.uk>
>> Sent: 25 February 2020 19:32
>> To: David Airlie <airlied at linux.ie>; Joonas Lahtinen
>> <joonas.lahtinen at linux.intel.com>; Laxminarayan Bharadiya, Pankaj
>> <pankaj.laxminarayan.bharadiya at intel.com>; Vivi, Rodrigo
>> <rodrigo.vivi at intel.com>; daniel at ffwll.ch; dri-devel at lists.freedesktop.org;
>> intel-gfx at lists.freedesktop.org; jani.nikula at linux.intel.com
>> Cc: Laxminarayan Bharadiya, Pankaj
>> <pankaj.laxminarayan.bharadiya at intel.com>
>> Subject: Re: [Intel-gfx][PATCH 01/10] drm/i915: Add i915 device based
>> MISSING_CASE macro
>> 
>> Quoting Pankaj Bharadiya (2020-02-25 13:47:00)
>> > Now that we have struct drm_device based drm_WARN, introduce struct
>> > drm_i915_private based i915_MISSING_CASE macro which uses
>> drm_WARN so
>> > that device specific information will also get printed in backtrace.
>> >
>> > i915_MISSING_CASE macro should be preferred over MISSING_CASE,
>> > wherever possible.
>> 
>> Whatever for? MISSING_CASE() itself should be a complete picture for the
>> forgotten code.
>
> Are you saying, no need to have a new device specific macro?
>
> We want convert all the calls of WARN* with device specific drm_WARN* 
> in i915, hence I introduced new i915_MISSING_CASE macro.
>
> Jani, Will you please share your opinion on this?

In general, many or most WARNs are device specific, and the device
information is useful. However MISSING_CASE is about the *code*. That
was the intent anyway. Perhaps there are cases where the device
information might be useful, but for most cases probably not.

BR,
Jani.


-- 
Jani Nikula, Intel Open Source Graphics Center