[Intel-gfx] [igt-dev] [PATCH i-g-t] lib/i915: Bump conservative threshold for ring size

[Chris Wilson]

Quoting Tvrtko Ursulin (2019-10-07 17:56:20)
> 
> On 07/10/2019 09:27, Chris Wilson wrote:
> > We are still hitting the occasional stall upon submission, so be extra
> > caution and leave one more spare.
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >   lib/i915/gem_ring.c | 4 ++--
> >   1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/lib/i915/gem_ring.c b/lib/i915/gem_ring.c
> > index 272264659..9f099edff 100644
> > --- a/lib/i915/gem_ring.c
> > +++ b/lib/i915/gem_ring.c
> > @@ -104,7 +104,7 @@ __gem_measure_ring_inflight(int fd, unsigned int engine, enum measure_ring_flags
> >       } while (1);
> >   
> >       igt_assert_eq(__execbuf(fd, &execbuf), -EINTR);
> > -     igt_assert(count > 1);
> > +     igt_assert(count > 2);
> >   
> >       memset(&itv, 0, sizeof(itv));
> >       setitimer(ITIMER_REAL, &itv, NULL);
> > @@ -120,7 +120,7 @@ __gem_measure_ring_inflight(int fd, unsigned int engine, enum measure_ring_flags
> >       gem_quiescent_gpu(fd);
> >   
> >       /* Be conservative in case we must wrap later */
> > -     return count - 1;
> > +     return count - 2;
> >   }
> >   
> >   /**
> > 
> 
> Will one make a difference?

I'm willing to find out!

All I am going on is

rcs0
*** WEDGED ***
	Awake? 66
	Hangcheck: 11959 ms ago
	Reset count: 0 (global 8)
	Requests:
	MMIO base:  0x00002000
	CCID: 0x00000000
	RING_START: 0x00000000
	RING_HEAD:  0x00000000
	RING_TAIL:  0x00000000
	RING_CTL:   0x00000000
	RING_MODE:  0x00000200 [idle]
	RING_IMR: ffffffff
	ACTHD:  0x00000000_00000000
	BBADDR: 0x00000000_00000000
	DMA_FADDR: 0x00000000_00000000
	IPEIR: 0x00000000
	IPEHR: 0x00000000
		E  3:139cd!  @ 12914ms: gem_eio[1345]
		E  3:139ce!  @ 12914ms: gem_eio[1345]
		E  3:139cf!  @ 12914ms: gem_eio[1345]
		E  3:139d0!  @ 12914ms: gem_eio[1345]
		E  3:139d1!  @ 12914ms: gem_eio[1345]
		E  3:139d2!  @ 12913ms: gem_eio[1345]
		E  3:139d3!  @ 12913ms: gem_eio[1345]
		...skipping 25 executing requests...
		E  3:139ed!  @ 0ms: gem_eio[1345]

which tells me it was blocked in submission until hangcheck kicked in
and declared the system wedged.
-Chris