[Mesa-dev] [PATCH 15/16] nine: Add drirc options

Axel Davy axel.davy at ens.fr
Mon Oct 20 00:31:44 PDT 2014


On 20/10/2014 09:12, Michel Dänzer wrote :
> On 20.10.2014 14:47, Axel Davy wrote:
>> On 20/10/2014 04:32, Michel Dänzer wrote :
>>> On 18.10.2014 20:55, David Heidelberger wrote:
>>>> From: Axel Davy <axel.davy at ens.fr>
>>>
>>> [...]
>>>
>>>> +#define DRI_CONF_NINE_VSYNC(def) \
>>>> +DRI_CONF_OPT_BEGIN_V(vsync_force,enum,def,"0:2") \
>>>> +        DRI_CONF_DESC_BEGIN(en,gettext("Force the usage of vsync")) \
>>>> +                DRI_CONF_ENUM(0,gettext("Obey application's 
>>>> choice")) \
>>>> +                DRI_CONF_ENUM(1,gettext("Never synchronize with
>>>> vertical refresh")) \
>>>> +                DRI_CONF_ENUM(2,gettext("Always synchronize with
>>>> vertical refresh")) \
>>>> +        DRI_CONF_DESC_END \
>>>> +DRI_CONF_OPT_END
>>>
>>> Can't the existing vblank_mode be used for this?
>>>
>> The problem is that d3d9 apps always specifies the mode, which makes
>> vblank_mode not appropriate, since vblank_mode has options to set
>> default when when application doesn't specify anything.
>
> That just means that vblank_mode values 1 and 2 would have the same 
> effect, namely honouring the application choice?
Ok, we can do that
>
>
>> So instead of reusing vblank_mode with different definition of the
>> arguments, introducing a new vsync_force seemed a better option.
>
> If the above is the only reason, I respectfully disagree. How are 
> users supposed to know which of these almost-but-not-quite identical 
> options applies to which apps?
>
As long as default values for this option for gl and d3d9 are the same, 
it should be fine.


More information about the mesa-dev mailing list