Re: XFree86 5.0 TODO
Sven Luther <[email protected]>
| Newsgroups | gmane.comp.xfree86.forum |
|---|---|
| Message-ID | <20030320202955.GA8035@iliana> |
On Thu, Mar 20, 2003 at 03:24:50PM -0500, Mark Vojkovich wrote: > The major problem with the old XAA was that it made too many > hardware assumptions. That was necessary to allow XAA to achieve > one of its goals, that was, to allow people who didn't know anything > about graphics hardware to write drivers. We had fewer professional > graphics people writing XFree86 drivers back then so that was an > important consideration. This is less important now. Mmm, > In the new XAA, surface allocation is relegated to the driver, > which can use any mechanism it wants to allocate these surfaces. But will it not mean less code sharing if each driver has to do memory management ? > I have this working with NVIDIA's binary drivers. Allocations > are offloaded to a system-wide allocator in the kernel. The NVIDIA In the drm driver, or somewhere else ? And how is this related to Ian's DRI memeory manager plans ? > binary drivers use a non-linear framebuffer so this was a good > test. What kind of non-linear framebuffer is used (if you are allowed to tell that is) and do you use a automated conversion unit or something such ? > Multidepth support is of interest to the render extension (all > those 8 bit alpha masks) and in the case of overlays. Ok, ... Friendly, Sven Luther