[PATCH] drm/i915: Empty the ring before disabling

[PATCH] drm/i915: Empty the ring before disabling

[Chris Wilson]

An interesting snippet from Sandybridge's prm:

"Although a Ring Buffer can be enabled in the non-empty state, it must
not be disabled unless it is empty. Attempting to disable a Ring Buffer
in the non-empty state is UNDEFINED."

Let's avoid the undefined behaviour as we disable the rings prior to
reset and resume.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 4 +++-
 drivers/gpu/drm/i915/intel_uncore.c     | 4 +++-
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 05e01446b00b..3f2073a9d37a 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -480,10 +480,12 @@ static bool stop_ring(struct intel_engine_cs *engine)
 		}
 	}
 
-	I915_WRITE_CTL(engine, 0);
 	I915_WRITE_HEAD(engine, 0);
 	I915_WRITE_TAIL(engine, 0);
 
+	/* The ring must be empty before it is disabled */
+	I915_WRITE_CTL(engine, 0);
+
 	return (I915_READ_HEAD(engine) & HEAD_ADDR) == 0;
 }
 
diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index 20e3c65c0999..ac688ee6fe4e 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -1387,10 +1387,12 @@ static void gen3_stop_engine(struct intel_engine_cs *engine)
 		DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n",
 				 engine->name);
 
-	I915_WRITE_FW(RING_CTL(base), 0);
 	I915_WRITE_FW(RING_HEAD(base), 0);
 	I915_WRITE_FW(RING_TAIL(base), 0);
 
+	/* The ring must be empty before it is disabled */
+	I915_WRITE_FW(RING_CTL(base), 0);
+
 	/* Check acts as a post */
 	if (I915_READ_FW(RING_HEAD(base)) != 0)
 		DRM_DEBUG_DRIVER("%s: ring head not parked\n",
-- 
2.15.0.rc2

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

✗ Fi.CI.BAT: failure for drm/i915: Empty the ring before disabling

[Patchwork]

== Series Details ==

Series: drm/i915: Empty the ring before disabling
URL   : https://patchwork.freedesktop.org/series/32727/
State : failure

== Summary ==

Series 32727v1 drm/i915: Empty the ring before disabling
https://patchwork.freedesktop.org/api/1.0/series/32727/revisions/1/mbox/

Test gem_exec_suspend:
        Subgroup basic-s3:
                pass       -> DMESG-FAIL (fi-kbl-7560u) fdo#103039
        Subgroup basic-s4-devices:
                pass       -> DMESG-FAIL (fi-kbl-7560u) fdo#102846 +1
Test gem_flink_basic:
        Subgroup bad-flink:
                pass       -> DMESG-WARN (fi-kbl-7560u) fdo#103049 +4
Test kms_pipe_crc_basic:
        Subgroup suspend-read-crc-pipe-b:
                pass       -> INCOMPLETE (fi-cnl-y)

fdo#103039 https://bugs.freedesktop.org/show_bug.cgi?id=103039
fdo#102846 https://bugs.freedesktop.org/show_bug.cgi?id=102846
fdo#103049 https://bugs.freedesktop.org/show_bug.cgi?id=103049

fi-bdw-5557u     total:289  pass:268  dwarn:0   dfail:0   fail:0   skip:21  time:443s
fi-bdw-gvtdvm    total:289  pass:265  dwarn:0   dfail:0   fail:0   skip:24  time:454s
fi-blb-e6850     total:289  pass:223  dwarn:1   dfail:0   fail:0   skip:65  time:376s
fi-bsw-n3050     total:289  pass:243  dwarn:0   dfail:0   fail:0   skip:46  time:531s
fi-bwr-2160      total:289  pass:183  dwarn:0   dfail:0   fail:0   skip:106 time:262s
fi-bxt-dsi       total:289  pass:259  dwarn:0   dfail:0   fail:0   skip:30  time:498s
fi-bxt-j4205     total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:495s
fi-byt-j1900     total:289  pass:253  dwarn:1   dfail:0   fail:0   skip:35  time:498s
fi-byt-n2820     total:289  pass:249  dwarn:1   dfail:0   fail:0   skip:39  time:489s
fi-cfl-s         total:289  pass:253  dwarn:4   dfail:0   fail:0   skip:32  time:564s
fi-cnl-y         total:246  pass:221  dwarn:0   dfail:0   fail:0   skip:24 
fi-elk-e7500     total:289  pass:229  dwarn:0   dfail:0   fail:0   skip:60  time:417s
fi-gdg-551       total:289  pass:178  dwarn:1   dfail:0   fail:1   skip:109 time:251s
fi-glk-1         total:289  pass:261  dwarn:0   dfail:0   fail:0   skip:28  time:577s
fi-glk-dsi       total:289  pass:258  dwarn:0   dfail:0   fail:1   skip:30  time:490s
fi-hsw-4770      total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:430s
fi-hsw-4770r     total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:431s
fi-ilk-650       total:289  pass:228  dwarn:0   dfail:0   fail:0   skip:61  time:437s
fi-ivb-3520m     total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:498s
fi-ivb-3770      total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:456s
fi-kbl-7500u     total:289  pass:264  dwarn:1   dfail:0   fail:0   skip:24  time:490s
fi-kbl-7560u     total:125  pass:105  dwarn:5   dfail:2   fail:0   skip:12 
fi-kbl-7567u     total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:480s
fi-kbl-r         total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:588s
fi-pnv-d510      total:289  pass:222  dwarn:1   dfail:0   fail:0   skip:66  time:548s
fi-skl-6260u     total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:455s
fi-skl-6600u     total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:590s
fi-skl-6700hq    total:289  pass:263  dwarn:0   dfail:0   fail:0   skip:26  time:648s
fi-skl-6700k     total:289  pass:265  dwarn:0   dfail:0   fail:0   skip:24  time:520s
fi-skl-6770hq    total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:504s
fi-skl-gvtdvm    total:289  pass:266  dwarn:0   dfail:0   fail:0   skip:23  time:460s
fi-snb-2520m     total:289  pass:250  dwarn:0   dfail:0   fail:0   skip:39  time:562s
fi-snb-2600      total:289  pass:249  dwarn:0   dfail:0   fail:0   skip:40  time:419s

ab2f0644893988b173763303a96b7213315b8acf drm-tip: 2017y-10m-26d-22h-25m-08s UTC integration manifest
ebb78aa45a24 drm/i915: Empty the ring before disabling

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_6219/
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH] drm/i915: Empty the ring before disabling

[Mika Kuoppala]

Chris Wilson <chris@chris-wilson.co.uk> writes:

> An interesting snippet from Sandybridge's prm:
>
> "Although a Ring Buffer can be enabled in the non-empty state, it must
> not be disabled unless it is empty. Attempting to disable a Ring Buffer
> in the non-empty state is UNDEFINED."
>
> Let's avoid the undefined behaviour as we disable the rings prior to
> reset and resume.
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> ---
>  drivers/gpu/drm/i915/intel_ringbuffer.c | 4 +++-
>  drivers/gpu/drm/i915/intel_uncore.c     | 4 +++-
>  2 files changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
> index 05e01446b00b..3f2073a9d37a 100644
> --- a/drivers/gpu/drm/i915/intel_ringbuffer.c
> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
> @@ -480,10 +480,12 @@ static bool stop_ring(struct intel_engine_cs *engine)
>  		}
>  	}
>  
> -	I915_WRITE_CTL(engine, 0);
>  	I915_WRITE_HEAD(engine, 0);
>  	I915_WRITE_TAIL(engine, 0);
>  
> +	/* The ring must be empty before it is disabled */
> +	I915_WRITE_CTL(engine, 0);
> +

I am thinking that does it need even more finesse.

Like

I915_WRITE_HEAD(engine, I915_READ_TAIL(engine));
I915_WRITE_TAIL(engine, 0);
I915_WRITE_HEAD(engine, 0);
I915_WRITE_CTL(engine, 0);

Regardless, nice find and has the promise of making
those nasty resets on active rings more robust.

Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>


>  	return (I915_READ_HEAD(engine) & HEAD_ADDR) == 0;
>  }
>  
> diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
> index 20e3c65c0999..ac688ee6fe4e 100644
> --- a/drivers/gpu/drm/i915/intel_uncore.c
> +++ b/drivers/gpu/drm/i915/intel_uncore.c
> @@ -1387,10 +1387,12 @@ static void gen3_stop_engine(struct intel_engine_cs *engine)
>  		DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n",
>  				 engine->name);
>  
> -	I915_WRITE_FW(RING_CTL(base), 0);
>  	I915_WRITE_FW(RING_HEAD(base), 0);
>  	I915_WRITE_FW(RING_TAIL(base), 0);
>  
> +	/* The ring must be empty before it is disabled */
> +	I915_WRITE_FW(RING_CTL(base), 0);
> +
>  	/* Check acts as a post */
>  	if (I915_READ_FW(RING_HEAD(base)) != 0)
>  		DRM_DEBUG_DRIVER("%s: ring head not parked\n",
> -- 
> 2.15.0.rc2
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH] drm/i915: Empty the ring before disabling

[Chris Wilson]

Quoting Mika Kuoppala (2017-10-27 07:44:31)
> Chris Wilson <chris@chris-wilson.co.uk> writes:
> 
> > An interesting snippet from Sandybridge's prm:
> >
> > "Although a Ring Buffer can be enabled in the non-empty state, it must
> > not be disabled unless it is empty. Attempting to disable a Ring Buffer
> > in the non-empty state is UNDEFINED."
> >
> > Let's avoid the undefined behaviour as we disable the rings prior to
> > reset and resume.
> >
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_ringbuffer.c | 4 +++-
> >  drivers/gpu/drm/i915/intel_uncore.c     | 4 +++-
> >  2 files changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
> > index 05e01446b00b..3f2073a9d37a 100644
> > --- a/drivers/gpu/drm/i915/intel_ringbuffer.c
> > +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
> > @@ -480,10 +480,12 @@ static bool stop_ring(struct intel_engine_cs *engine)
> >               }
> >       }
> >  
> > -     I915_WRITE_CTL(engine, 0);
> >       I915_WRITE_HEAD(engine, 0);
> >       I915_WRITE_TAIL(engine, 0);
> >  
> > +     /* The ring must be empty before it is disabled */
> > +     I915_WRITE_CTL(engine, 0);
> > +
> 
> I am thinking that does it need even more finesse.

I thought so too, but we already stop the ring, and so gave up thinking.

> Like
> 
> I915_WRITE_HEAD(engine, I915_READ_TAIL(engine));

Nevertheless, I'll give that a spin since that's neat ;)
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH] drm/i915: Empty the ring before disabling

[Ville Syrjälä]

On Fri, Oct 27, 2017 at 12:56:39AM +0100, Chris Wilson wrote:
> An interesting snippet from Sandybridge's prm:
> 
> "Although a Ring Buffer can be enabled in the non-empty state, it must
> not be disabled unless it is empty. Attempting to disable a Ring Buffer
> in the non-empty state is UNDEFINED."

However elsewhere it still says:
"Writing the Head Offset while the RB is enabled is UNDEFINED"
and
"It can be enabled or disabled regardless of whether there
 are valid instructions pending."

And gen2/3 bspec says this:
"Ring buffers can be disabled with valid instructions pending.
When disabled, the ring buffer will be removed from command 
stream arbitration at the next arbitration point (i.e.,
instruction boundary)."

> 
> Let's avoid the undefined behaviour as we disable the rings prior to
> reset and resume.
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> ---
>  drivers/gpu/drm/i915/intel_ringbuffer.c | 4 +++-
>  drivers/gpu/drm/i915/intel_uncore.c     | 4 +++-
>  2 files changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
> index 05e01446b00b..3f2073a9d37a 100644
> --- a/drivers/gpu/drm/i915/intel_ringbuffer.c
> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
> @@ -480,10 +480,12 @@ static bool stop_ring(struct intel_engine_cs *engine)
>  		}
>  	}
>  
> -	I915_WRITE_CTL(engine, 0);
>  	I915_WRITE_HEAD(engine, 0);
>  	I915_WRITE_TAIL(engine, 0);
>  
> +	/* The ring must be empty before it is disabled */
> +	I915_WRITE_CTL(engine, 0);
> +
>  	return (I915_READ_HEAD(engine) & HEAD_ADDR) == 0;
>  }
>  
> diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
> index 20e3c65c0999..ac688ee6fe4e 100644
> --- a/drivers/gpu/drm/i915/intel_uncore.c
> +++ b/drivers/gpu/drm/i915/intel_uncore.c
> @@ -1387,10 +1387,12 @@ static void gen3_stop_engine(struct intel_engine_cs *engine)
>  		DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n",
>  				 engine->name);
>  
> -	I915_WRITE_FW(RING_CTL(base), 0);
>  	I915_WRITE_FW(RING_HEAD(base), 0);
>  	I915_WRITE_FW(RING_TAIL(base), 0);
>  
> +	/* The ring must be empty before it is disabled */
> +	I915_WRITE_FW(RING_CTL(base), 0);
> +
>  	/* Check acts as a post */
>  	if (I915_READ_FW(RING_HEAD(base)) != 0)
>  		DRM_DEBUG_DRIVER("%s: ring head not parked\n",
> -- 
> 2.15.0.rc2
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH] drm/i915: Empty the ring before disabling

[Chris Wilson]

Quoting Ville Syrjälä (2017-10-27 12:27:59)
> On Fri, Oct 27, 2017 at 12:56:39AM +0100, Chris Wilson wrote:
> > An interesting snippet from Sandybridge's prm:
> > 
> > "Although a Ring Buffer can be enabled in the non-empty state, it must
> > not be disabled unless it is empty. Attempting to disable a Ring Buffer
> > in the non-empty state is UNDEFINED."
> 
> However elsewhere it still says:
> "Writing the Head Offset while the RB is enabled is UNDEFINED"
> and
> "It can be enabled or disabled regardless of whether there
>  are valid instructions pending."

Nothing like a catch 22. Let's just reset the GPU first... 
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx