[PATCH 2/5] wined3d: Handle CS op emission from CS op handlers.

[Sebastian Lackner]

On 17.02.2017 20:07, Stefan Dösinger wrote:
> Hi Henri,
> 
> What is your long term plan regarding this? I'm sure you are aware that
> my last version avoided emitting CS command from inside the CS and the
> ringbuffer code is designed around having a single producer and single
> consumer.

I don't know if its planned as a permanent solution, but emitting commands
on the command stream thread (and dispatching them immediately) does not
really mean that we can't use a ringbuffer (or any other queue) for
commands from other threads. We only have to be careful that wait_idle()
doesn't cause a deadlock. ;)

> 
> Is this patch intended as a temporary measure, or are you planning to
> retain the ability to emit CS commands from the CS thread?
> 
> Stefan
> 
>