Re: DRI documentation ...was Re: [Dri-devel] Announcement

[Daryll S. <da...@va...>]

The documentation makes a good base and is mostly still correct. The
problem is that this stuff is growing very quickly. For example, we've
come up with fairly standardized ways of writing a device driver. From
an architectural point of view it's still just a device driver, but from
an implementation point of view it would be nice to have more
documentation that describes how everything fits together.

As with any engineering project it is tough to create current
documentation. Documentation quickly becomes outdated. There's no
documentation for the Linux kernel either for pretty much the same
reason. Your best bet is going to be to dive into the code, and ask
questions if you get stuck.

					      - |Daryll


On Fri, Jun 08, 2001 at 02:58:15PM -0500, ad...@hu... wrote:
>=20
> I don't want to beat it to death, but following the "Announcement" sparke=
d=20
> discussion I wanted to mention my take:
>=20
> One of my problems (time being the biggest one of course :) with getting=
=20
> involved with the code-part of the DRI project was the developer document=
ation.=20
> I am greatful for having some of it  (I have read these a few times:  DRI=
intro.html=20
>     dr.html     drm.html      security.html DRIdataflow.jpg  DRIcomponent=
s.jpg=20
>  Free86-DRI.html  drill.html  locking.html ), but I still found myself ne=
eding=20
> more to grasp the inner workings of the DRI better.  (some were basically=
=20
> function prototypes/descriptions which didn't tell me much about how and=
=20
> where these functions where used)
>=20
> So since I am speaking of the PI docs, I would like a response from the=
=20
> PI guys in the list:  how current are these docs to the present-day DRI=
=20
> ?
> While reading them I saw many mentions about features that were to be imp=
lemented=20
> in the future,  (seems the docs were done in time for some LinuxShow/Conf=
erence=20
> demo) and got the feeling that the docs were quite out of touch with pres=
ent-
> day DRI. Of course I know the dates of the docs, but how applicable are=
=20
> they today?  =20
>=20
> Just want to get an idea, because if I use them as my base of understandi=
ng=20
> for the DRI (as I do today), then I want to know how off-base I am.
>=20
> I realize that going through the code would answer this for me, but I am=
=20
> not that familiar with it ; I am still working on grasping the high-level=
=20
> layout of the project.  I know how to bang registers and stuff like that=
=20
> makes sense, but the overall flow of things is my struggling point.
>=20
> thanks
>=20
> tony
> Free, encrypted, secure Web-based email at www.hushmail.com