Re: [Qemu-devel] [RFC] qemu VGA endian swap low level drawing changes

[Gerd Hoffmann]

On Mi, 2014-07-02 at 19:19 +1000, Benjamin Herrenschmidt wrote:
> On Tue, 2014-07-01 at 21:23 +1000, Benjamin Herrenschmidt wrote:
> > On Tue, 2014-07-01 at 13:15 +0200, Gerd Hoffmann wrote:
> > > TCG works.  Any workaround to make it fly with kvm too?
> > > 
> > > And win98 actually uses the hardware cursor.
> > 
> > Ah ok. I was trying to install NT4 here but was having problems,
> > the screen would go yellow and cirrus would stop displaying anything
> > when trying to "test" the resolution (without my patches, actually with
> > a slightly older qemu too).
> > 
> > I was planning on digging more tomorrow, but I can try Win98 if that's
> > easier.
> 
> NT4 works a lot better without -enable-kvm :-) (file corruption, general
> bad behaviour...).
> 
> I can confirm the cursor is missing in 16bpp with cirrus on NT4 as well,
> I'll play with it later this week or week-end, unless you beat me to it.
> 
> [ Let me know if you start working on it so we don't duplicate work ]

console/pixman bits are here:

https://www.kraxel.org/cgit/qemu/log/?h=rebase/console-wip

Added a patch for hardware cursor support via dpy_cursor_define().
Old hardware cursor code is still in, so in theory this gives you two
pointers.  In practice it only shows that cursor support in relative
mouse mode is (a) flaky and (b) our uis are very inconsistent ...

[ not hacking on it at the moment, busy with other stuff ]

cheers,
  Gerd