[Intel-gfx] [PATCH 12/34] drm/i915: Issue engine resets onto idle engines

Chris Wilson chris at chris-wilson.co.uk
Wed Jan 23 01:31:59 UTC 2019


Quoting John Harrison (2019-01-23 01:18:36)
> On 1/21/2019 14:20, Chris Wilson wrote:
> > @@ -479,8 +477,6 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active)
> >                                       break;
> >                               }
> >   
> > -                             GEM_BUG_ON(!rq->global_seqno);
> > -                             seqno = rq->global_seqno - 1;
> AFAICT this saved seqno value was never used anyway? It only exists 
> inside the loop, was only used in a pr_err earlier in the loop, and the 
> start of the loop always (re-)initialises it. Or am I missing some 
> hidden macro magic somewhere?

I can't remember what I intended it to be in the first place. History
says that we had to set engine->hangcheck.seqno correctly to force a
per-engine reset. Lost in

commit bba0869b18e44ff2f713c98575ddad8c7c5e9b10
Author: Chris Wilson <chris at chris-wilson.co.uk>
Date:   Fri Apr 6 23:03:53 2018 +0100

    drm/i915: Treat i915_reset_engine() as guilty until proven innocent

-Chris


More information about the Intel-gfx mailing list