[PATCH 1/2] drm/radeon/kms/r600+: do not set the same voltage as current one

Alex Deucher alexdeucher at gmail.com
Mon Jun 7 09:24:10 PDT 2010


2010/6/7 Rafał Miłecki <zajec5 at gmail.com>:
> W dniu 7 czerwca 2010 18:09 użytkownik Alex Deucher
> <alexdeucher at gmail.com> napisał:
>> 2010/6/7 Rafał Miłecki <zajec5 at gmail.com>:
>>> W dniu 7 czerwca 2010 07:00 użytkownik Alex Deucher
>>> <alexdeucher at gmail.com> napisał:
>>>> 2010/6/6 Rafał Miłecki <zajec5 at gmail.com>:
>>>>> We have similar tracking for engine and memory. We could think about some
>>>>> solution for older GPUs as well, however there are not strict voltage values
>>>>> set.
>>>>>
>>>>> Signed-off-by: Rafał Miłecki <zajec5 at gmail.com>
>>>>
>>>> I actually had almost the exact same patch in my tree (broken out from
>>>> my voltage step patch), I just hadn't gotten around to sending it.
>>>> Mine is against drm-linus and has the default voltage level set.  See
>>>> attached.
>>>
>>> I know, actually I've stolen your name (curr_vddc).
>>>
>>> Do you plan to include this in .35? If yes, it's fine. If not, I'd
>>> prefer to split your patch into two. The one I've sent and second that
>>> adds you step setting.
>>
>> I'm not planning to include the whole step setting patch, but I did
>> break out the current_vddc part and planned to push that (attached to
>> the last email) for 2.6.35.
>
> Ahh, indeed, sorry. I just wonder about setting current_vddc even when
> we do not call radeon_atom_set_voltage. This way we will have false
> value in current_vddc, if voltage change is not available (type !=
> VOLTAGE_SW).

The default voltage level is set when the asic is initialized at post.
 I've changed the default behaviour to no set the default mode
unconditionally at start up since it's already set by asicinit and it
avoids the possibility that someone might get a hang on module load
due to pm.  See this patch:
http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=commitdiff;h=aed70738a5e5563276cd75eb1ae2801cae8128c8
Only VOLTAGE_SW uses current_vddc so it's irrelevant for other voltage types.

Alex


More information about the dri-devel mailing list