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

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Sun, 2014-07-06 at 17:05 +1000, Benjamin Herrenschmidt wrote:
> On Sun, 2014-07-06 at 16:46 +1000, Benjamin Herrenschmidt wrote:
> > At this point, I"m tempted to just revert that commit. What do you
> > think Gerd ?
> 
> I mean that hunk of the commit... I missed that the commit itself
> added a whole lot more bound checking.

So we have an number of other problems :-)

 - With GTK the qemu cursor constantly gets removed... it works when
established at boot but as soon as we take the grab, or change anything,
the GTK ui removes the custom cursor and it generally remains hidden
form then on until we release the grab. I am not familiar with that
code, I'll try to debug it but help is welcome. Basically most of the
time we only see the "emulated" HW cursor (or nothing with a shared
surface).

 - With SDL, the screen colors are all wrong :-) It looks like a
component is partially missing. Not sure what's up there, again,
something else to debug. Everything has a blue tint (this is full emu on
x86_64 host).

 - With SDL, the cursor quickly goes bonkers, starts jumping around all
over the place, I'm not sure what exactly is going on here. It starts ok
but as soon as one does a too fast movement, it's dead.

 - With VNC, we see both cursors as expected but ... with an offset,
more or less variable. Ie I would expect 3 cursors: The VNC host cursor,
the little round thing that VNC puts where it thinks the guest cursor
is, and the emulated guest cursor, but the little round thing, instead
of following the guest cursor (emulated), follows the host one (though
the clicks are recorded by windows at the position of the guest one). it
looks like the wrong positions are sent to VNC.

Now, it's all code I know nothing about :-) I'll spend some time
investigating in my little spare time, but hints are welcome.

Cheers,
Ben.