[Intel-gfx] [PATCH 1/6] drm/i915: grab force_wake when restoring LCPLL

[Paulo Zanoni]

2013/8/19 Ben Widawsky <ben at bwidawsk.net>:
> On Mon, Aug 19, 2013 at 04:01:13PM -0300, Paulo Zanoni wrote:
>> From: Paulo Zanoni <paulo.r.zanoni at intel.com>
>>
>> If LCPLL is disabled, there's a chance we might be in package C8 state
>> or deeper, and we'll get a hard hang when restoring LCPLL (also, a red
>> led lights up on my motherboard). So grab the force_wake, which will
>> get us out of RC6 and, as a consequence, out of PC8+ (since we need
>> RC6 to get into PC8+).
>>
>> v2: Call the correct force_wake interface.
>>
>> Signed-off-by: Paulo Zanoni <paulo.r.zanoni at intel.com>
>
> Hi Paulo. As I recall, the steps in the bspec are laid out exactly so we
> don't have to call forcewake. I think the explanation sounds a little
> fishy.

Well, we seem to do exactly what BSpec says we need to do and we still
get the hard machine hang if we're in PC8 while restoring from LCPLL.


>
> Specifically, I thought D_COMP COMP_FORCE + waiting for PLL lock was
> sufficient.

Why? I don't see D_COMP documentation anywhere or any text that leads
me to believe that. The spec just says "enable these bits, disable
these others" without any explanations on what we're really doing.


> If it's not, I worry you're papering over another bug, or
> incorrect bspec sequence.
>
> Not opposed to the patch in the meanwhile, but it just feels a little
> unsafe to me.

I'll send some emails requesting for clarification. We certainly can't
merge PC8 without this patch or an equivalent fix.


>
> --
> Ben Widawsky, Intel Open Source Technology Center



-- 
Paulo Zanoni