radeon/r200 color tiling ddx / drm questions

[Roland Scheidegger]

Michel Dänzer wrote:
>>>> By making the initialization fail, e.g.
>> 
>> Do you have some pointers how you'd do that? All code I see for 
>> initialization is on the dri side, except the generic stuff in
>> dri.c. When it comes to initialization, there seems to be some
>> black magic involved :-).
> 
> 
> E.g., the sequence of XF86DRI...() calls could be changed along the 
> lines of: If the DDX DRI extension version determined by 
> XF86DRIQueryVersion() indicates that it's supported, the client calls
>  something like XF86DRISetClientVersion() to indicate the
> capabilities it has. XF86DRIGetDeviceInfo() could then fail if the
> client's advertised capabilities don't allow direct rendering with
> the DDX configuration. Of course, there's a lot of details to be
> worked out around this idea, but I don't see anything that would
> prevent it altogether, except maybe that it may not be possible while
> preserving the interface between libGL and the 3D driver. :\
ah, a dri api change (would be the first since 2.5 years, btw), so we'd
bump the XF86DRI_MINOR_VERSION to 2 and just introduce that 
XF86DRISetClientVersion function (sounds like a good name to me).
Sounds good to me in theory, if not slightly complicated (the part about 
basing the decision on current ddx configuration sounds iffy). I have no 
clue though how you'd do that, how it travels through libGL and so on. I 
guess someone familiar with that interface stuff would need to implement 
it, it doesn't sound like something I could do...

> 
>>>>> If there is some two-way handshake somewhere, I must have
>>>>> missed it.
>>>> 
>>>> There's none yet in the DDX, it would have to be added (if the
>>>> one in the DRM doesn't suffice). It would probably be useful
>>>> for other stuff like dynamic buffer offset updates (for
>>>> Composite, pbuffers, ...) as well.
>> 
>> Actually, I can't see a two-way handshake in drm neither, any
>> pointers?
> 
> 
> See drmSetInterfaceVersion() in xf86drm.c.
so, this is actually a handshake (I always thought that's just to figure
out what version the drm is)! If I see that correctly, when calling this
you can specify both a drm minor/major and actual driver major/minor
version. drm will then return an error if the major versions don't
match, or if the requested minor versions are larger than what the
drm/driver themself have. Additionally, if there is no error it will
also hopefully restrict drm features to those supported by the user
supplied drm minor/major values (the same is not done for the actual
driver, maybe that's not necessary).
Doesn't look like it could be used for what we need though.

Roland