On Sat, Apr 04, 2015 at 03:46:32AM -0700, Kenneth Graunke wrote:
> Fixes Piglit's arb_gpu_shader5-xfb-streams-without-invocations.
>
> Signed-off-by: Kenneth Graunke <kenn...@whitecape.org>
> Cc: mesa-sta...@lists.freedesktop.org
> ---
> src/mesa/drivers/dri/i965/gen8_sol_state.c | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/src/mesa/drivers/dri/i965/gen8_sol_state.c
> b/src/mesa/drivers/dri/i965/gen8_sol_state.c
> index 1f122ec..d98a226 100644
> --- a/src/mesa/drivers/dri/i965/gen8_sol_state.c
> +++ b/src/mesa/drivers/dri/i965/gen8_sol_state.c
> @@ -128,6 +128,15 @@ gen8_upload_3dstate_streamout(struct brw_context *brw,
> bool active,
> dw2 |= urb_entry_read_offset << SO_STREAM_0_VERTEX_READ_OFFSET_SHIFT;
> dw2 |= (urb_entry_read_length - 1) <<
> SO_STREAM_0_VERTEX_READ_LENGTH_SHIFT;
>
> + dw2 |= urb_entry_read_offset << SO_STREAM_1_VERTEX_READ_OFFSET_SHIFT;
Can we use SET_FIELD() instead? I know the original code isn't using it
either but I thought we aimed in some point moving to using it instead:
dw2 |= SET_FIELD(urb_entry_read_offset,
SO_STREAM_0_VERTEX_READ_OFFSET);
> + dw2 |= (urb_entry_read_length - 1) <<
> SO_STREAM_1_VERTEX_READ_LENGTH_SHIFT;
> +
> + dw2 |= urb_entry_read_offset << SO_STREAM_2_VERTEX_READ_OFFSET_SHIFT;
> + dw2 |= (urb_entry_read_length - 1) <<
> SO_STREAM_2_VERTEX_READ_LENGTH_SHIFT;
> +
> + dw2 |= urb_entry_read_offset << SO_STREAM_3_VERTEX_READ_OFFSET_SHIFT;
> + dw2 |= (urb_entry_read_length - 1) <<
> SO_STREAM_3_VERTEX_READ_LENGTH_SHIFT;
> +
> /* Set buffer pitches; 0 means unbound. */
> if (xfb_obj->Buffers[0])
> dw3 |= linked_xfb_info->BufferStride[0] * 4;
> --
> 2.3.4
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
