[Intel-gfx] [PATCH 2/6] drm/i915/guc: Make guc_enable/disable_communication functions public

[Michal Wajdeczko]

On Thu, 14 Sep 2017 17:56:17 +0200, Kamble, Sagar A  
<sagar.a.kamble at intel.com> wrote:

>
>
> On 9/14/2017 9:01 PM, Michal Wajdeczko wrote:
>> On Thu, 14 Sep 2017 11:55:04 +0200, Sagar Arun Kamble  
>> <sagar.a.kamble at intel.com> wrote:
>>
>>> From: "Kamble, Sagar A" <sagar.a.kamble at intel.com>
>>>
>>> This patch is moving guc_enable_communication and
>>> guc_disable_communication to intel_guc.c and making it available for
>>> use through intel_guc.h.
>>> Intent is to reuse this function for calling from intel_uc_init_hw
>>> and also as part of intel_uc_fini_hw where it will be coupled with
>>> other teardown related to GuC in the upcoming patch.
>>
>> Hmm, but both intel_uc_init_hw() and intel_uc_fini_hw() shall stay in
>> intel_uc.c and the best place for guc_[enable|disable]_communication()
>> functions is also in intel_uc.c as they provide generic switch between
>> MMIO/CT send mechanism, so I'm not so sure that we need this patch.
>>
>> Michal
> Hi Michal,
>
> Plan was to disable communication while entering rpm/system suspend,  
> reset, unload
> and enable back when resuming. If we don't want to update these  
> mechanisms across rpm/system suspend or reset will
> remove this patch then.

We should disable any communication with GuC, as Guc will be unavailable,
but since suspend/resume/unload code should stay (or be defined) in
intel_uc.c file, I see no point of moving related functions out of that  
file.

Michal