Start our preparations for guaranteeing endless execution.
First, we just want to estimate the 'ulta-low latency' dispatch overhead
by running an endless chain of batch buffers. The legacy binding process
here will be replaced by async VM_BIND, but for the moment this
suffices to construct the GTT
Start our preparations for guaranteeing endless execution.
First, we just want to estimate the direct userspace dispatch overhead
of running an endless chain of batch buffers. The legacy binding process
here will be replaced by async VM_BIND, but for the moment this
suffices to construct the GTT a
Verify that an endless batch is not forcibly reset across suspend, but
allowed to finish its task on resume.
Just an initial stub for the most concerning problem.
Signed-off-by: Chris Wilson
---
tests/Makefile.sources| 3 ++
tests/i915/gem_exec_endless.c | 94 ++
Chris Wilson writes:
> Start our preparations for guaranteeing endless execution.
>
> First, we just want to estimate the 'ulta-low latency' dispatch overhead
> by running an endless chain of batch buffers. The legacy binding process
> here will be replaced by async VM_BIND, but for the moment th