[Intel-gfx] [PATCH 2/6] drm/i915/vlv: Added a rendering specific Hw WA 'WaSendDummy3dPrimitveAfterSetContext'
Chris Wilson
chris at chris-wilson.co.uk
Mon Mar 24 10:39:11 CET 2014
On Mon, Mar 24, 2014 at 12:19:20PM +0530, sourab.gupta at intel.com wrote:
> From: Akash Goel <akash.goel at intel.com>
>
> This workaround is needed on VLV for the HW context feature.
> It is used after adding the mi_set_context command in ring buffer
> for Hw context switch. As per the spec
> "The software must send a pipe_control with a CS stall and a post sync
> operation and then a dummy DRAW after every MI_SET_CONTEXT and after any
> PIPELINE_SELECT that is enabling 3D mode".
>
> Signed-off-by: Akash Goel <akash.goel at intel.com>
> Signed-off-by: Sourab Gupta <sourab.gupta at intel.com>
> ---
> drivers/gpu/drm/i915/i915_gem_context.c | 65 ++++++++++++++++++++++++++++++++-
> drivers/gpu/drm/i915/i915_reg.h | 3 ++
> drivers/gpu/drm/i915/intel_ringbuffer.c | 9 +++++
> drivers/gpu/drm/i915/intel_ringbuffer.h | 1 +
> 4 files changed, 76 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
> index 6043062..544fc2d 100644
> --- a/drivers/gpu/drm/i915/i915_gem_context.c
> +++ b/drivers/gpu/drm/i915/i915_gem_context.c
> @@ -584,6 +584,58 @@ i915_gem_context_get(struct drm_i915_file_private *file_priv, u32 id)
> return ctx;
> }
>
> +static inline void
> +mi_set_context_dummy3d_prim_wa(struct intel_ring_buffer *ring)
> +{
> + u32 scratch_addr;
> + u32 flags = 0;
> +
> + /*
> + * Check if we have the scratch page allocated needed
> + * for the Pipe Control command, otherwise don't apply
> + * the dummmy 3d primitive workaround & add NOOPs instead
> + */
> + if (get_pipe_control_scratch_addr(ring)) {
0 is a valid address. Plus using an unprefixed global static inline is
frown upon. Also we cannot get here without allocating a scratch_obj, so
we can forgo the test completely (and then kill the
get_pipe_control_scratch_addr).
> + /* Actual scratch location is at 128 bytes offset */
> + scratch_addr = get_pipe_control_scratch_addr(ring) + 128;
> +
> + /*
> + * WaSendDummy3dPrimitveAfterSetContext:vlv
> + * Software must send a pipe_control with a CS stall
> + * and a post sync operation and then a dummy DRAW after
> + * every MI_SET_CONTEXT and after any PIPELINE_SELECT that
> + * is enabling 3D mode. A dummy draw is a 3DPRIMITIVE command
> + * with Indirect Parameter Enable set to 0, UAV Coherency
> + * Required set to 0, Predicate Enable set to 0,
> + * End Offset Enable set to 0, and Vertex Count Per Instance
> + * set to 0, All other parameters are a don't care.
> + */
> +
> + /*
> + * Add a pipe control with CS Stall and postsync op
> + * before dummy 3D_PRIMITIVE
> + */
> + flags |= PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_CS_STALL;
> + intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4));
> + intel_ring_emit(ring, flags);
> + intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT);
> + intel_ring_emit(ring, 0);
> +
> + /* Add a dummy 3D_PRIMITVE */
> + intel_ring_emit(ring, GFX_OP_3DPRIMITIVE());
GFX_OP_3DPRIMITIVE emulates a function, why? I thought it was a
constant.
> + intel_ring_emit(ring, 4); /* PrimTopoType*/
> + intel_ring_emit(ring, 0); /* VertexCountPerInstance */
> + intel_ring_emit(ring, 0); /* StartVertexLocation */
> + intel_ring_emit(ring, 0); /* InstanceCount */
> + intel_ring_emit(ring, 0); /* StartInstanceLocation */
> + intel_ring_emit(ring, 0); /* BaseVertexLocation */
> + } else {
> + int i;
> + for (i = 0; i < 11; i++)
> + intel_ring_emit(ring, MI_NOOP);
> + }
> +}
> +
> static inline int
> mi_set_context(struct intel_ring_buffer *ring,
> struct i915_hw_context *new_context,
> @@ -602,7 +654,10 @@ mi_set_context(struct intel_ring_buffer *ring,
> return ret;
> }
>
> - ret = intel_ring_begin(ring, 6);
> + if (IS_VALLEYVIEW(ring->dev))
> + ret = intel_ring_begin(ring, 6+4+8);
> + else
> + ret = intel_ring_begin(ring, 6);
> if (ret)
> return ret;
>
> @@ -626,7 +681,13 @@ mi_set_context(struct intel_ring_buffer *ring,
> intel_ring_emit(ring, MI_NOOP);
>
> if (IS_GEN7(ring->dev))
> - intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_ENABLE);
> + if (IS_VALLEYVIEW(ring->dev)) {
> + /* FIXME, should also apply to ivb */
> + mi_set_context_dummy3d_prim_wa(ring);
> + intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_ENABLE);
> + intel_ring_emit(ring, MI_NOOP);
> + } else
> + intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_ENABLE);
> else
> intel_ring_emit(ring, MI_NOOP);
>
> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> index adcb9c7..fd25e8e4 100644
> --- a/drivers/gpu/drm/i915/i915_reg.h
> +++ b/drivers/gpu/drm/i915/i915_reg.h
> @@ -348,6 +348,9 @@
> #define PIPE_CONTROL_DEPTH_CACHE_FLUSH (1<<0)
> #define PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */
>
> +#define GFX_OP_3DPRIMITIVE() \
> + ((0x3<<29)|(0x3<<27)|(0x3<<24)| \
> + (0x0<<16)|(0x0<<10)|(0x0<<8)|(7-2))
Please emulate the style we use else where for splitting this up into
client, opcode, subop, len etc.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
More information about the Intel-gfx
mailing list