[Openchrome-devel] ttm_branch.

[Xavier Bachelot]

Thomas Hellstr?m wrote:
> Xavier Bachelot wrote:
>> Thomas Hellstr?m wrote:
>>> Hi!
>>>
>> Hi Thomas,
>>
>>> I'll slowly be pulling our openchrome work into the ttm_branch.
>>> As you will  notice, some of the cursor work will duplicate the RandR 
>>> branch work, but
>>> I'll make sure in the end the latest version will merge in the latest 
>>> trunk stuff.
>>>
>> randr_branch stuff and thus cursor work was already merged to the 
>> trunk revision you copied to create the ttm_branch.
> It looks to me like the randr_branch was merged in r593, whereas the 
> ttm_branch is started from trunk at r528 and synced up to r592. Hence 
> still missing the RandR work.

Indeed. I saw the branch was copied from trunk today, and I wrongly
assumed it was today's trunk.

>>
>>> At some point in branch revision history, of course things will stop 
>>> working as it needs the new drm and a user-space library.
>>>
>> What are the non-openchrome pre-requisites (kernel, drm, xserver, 
>> etc...) needed to test the ttm_branch ? Is libwsbm already available ?
> Kernel will be at least 2.6.22 I think. 
Ok

> Completely new drm (not released 
> yet). 
I assume not even in a drm git branch ?

> Xserver with EXA 2.2, 
So xserver >= 1.4

> libwsbm.
> libwsbm is in principle ready-to-go, but we're discussing some interface 
> changes to simplify it a little.
> libwsbm is basically a buffer manager abstraction that is intended to 
> provide buffer management with- or without drm.
>
Ok

> So for now, consider the branch a work-in-progress thing that also is 
> used to keep track with latest openchrome development, and for reviews 
> and comments.
> 
> What's missing at this point is the device initialization code and 
> command submission code needed for non-drm operation,
> and XvMC fixes. Perhaps rename the ochr_ files to via_ considering 
> recent events.
Ok, let us know when at least part of the work is in a
testable/compilable state (failing on libwsbm currently).

Regards,
Xavier