[Intel-gfx] [igt-dev] [PATCH i-g-t 3/3] tests/i915/gem_watchdog: Exercise long rendering chains

[Matthew Auld]

On Thu, 18 Mar 2021 at 16:24, Tvrtko Ursulin
<tvrtko.ursulin at linux.intel.com> wrote:
>
> From: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>
> Test to demonstrate a problem with the proposed default fence expiry
> semantics where long rendering chain get silently broken.
>
> If we had fence error propagation (no clear agreement whether to do it or
> not) maybe userspace would see if, assuming fence status is looked at, but
> overall potential rendering corruption is the story in any case.
>
> Note that this is not a single long batch but just a long queue of work
> which. Could be viewed as heavy system load as well (like virtualisation
> or other types of resource sharing).
>
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
Reviewed-by: Matthew Auld <matthew.auld at intel.com>