CVS lock ?

[Vladimir Dergachev]

On Thu, 16 Dec 2004, Michel [ISO-8859-1] Dänzer wrote:

> On Thu, 2004-12-16 at 07:43 -0500, Vladimir Dergachev wrote:
>>
>> +    /* TODO: Fix this more elegantly.
>
> Why do you commit code that contains so many question marks without any
> prior discussion?

The code contains so many question marks because we do not have 
documentation on this hardware.

This piece does not affect any users of earlier chips and it does not 
affect those that use MMIO.

It does not cause lockups (in fact prevents them) and is overly 
conservative which results in small decrease in performance.

I thought to preserve the original comment by Nicolai Haehnle so that 
later on when this piece finds its permanent place it is easier to 
remember why it was put in.

>> +     * However, I don't feel confident enough with the control flow
>> +     * inside the X server to implement either fix. -- nh
>
> RADEONEnterServer() gets called whenever the X server grabs the hardware
> lock after a 3D client held it.

Thank you, I was not aware of this :)

Btw, r300_demo is not a glx app so it can run simultaneously with X, 
especially on SMP machines - which is why I decided not to explore how to 
do this in X.

Once R300 Mesa 3d driver is closer to usable state this should no longer 
be necessary - instead the DRM driver would do cache flushes as 
appropriate.

Lastly, please feel free to correct this code - and clarify any register 
names if you feel like it :)

                      best

                         Vladimir Dergachev

>
>
> -- 
> Earthling Michel Dänzer      |     Debian (powerpc), X and DRI developer
> Libre software enthusiast    |   http://svcs.affero.net/rm.php?r=daenzer
>