[MPlayer-dev-eng] Vidix Status

[Rich Felker]

On Mon, Feb 05, 2007 at 10:11:34AM +0100, Alex Beregszaszi wrote:
> Hi,
> 
> > > > Merging both vidix implementations is most welcome.  We've been talking
> > > > about this for years now, but there was never a volunteer to take on the
> > > > work.  If you do this it will be a valuable contribution.
> > > 
> > > While we're at it, is anyone interested in making wrappers so that
> > > vidix drivers build as kernelspace modules? Running them in userspace
> > > is exceedingly stupid, requiring your movie player to run as root or
> > > else use kernel modules with root vulns (svgalib_helper).
> > 
> > Alex wrote the kernelhelper module long long ago. But it never
> > really worked. I never checked why it does not. I don't even 
> > know whether the vidix API is suitable for a kernel module.
> 
> It worked, but kernel API's changed every second day, thus the memory
> mapping part was not always working.

Just use the mga_vid code as the wrapper, and make it call the linked
vidix module instead of directly writing to mga registers...

> However, vidix.sf.net's verison contains and extended dhahelper with DMA
> and IRQ support, which is working.

And which provides root to anyone who can open the device! There's no
way to make a "generic" kernelspace helper for userspace drivers. The
code that does the mmio, etc. really needs to be in kernelspace so
that the userspace code is limited to making requests and not doing
its own (arbitrary, possibly destructive) io.

Rich