[Mesa-stable] [Mesa-dev] [PATCH] i965: Respect stride and subreg_offset for ATTR registers

[Matt Turner]

On Wed, Sep 23, 2015 at 5:11 PM, Kristian Høgsberg Kristensen
<krh at bitplanet.net> wrote:
> When we assign hw regs to attributes, we don't incorporate the stride
> and subreg_offset from the fs_reg. It's rarely used, but the integer
> multiplication lowering uses unusual stride and subreg_offset
> combination breaks when one source is an attribute.
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91970
> Cc: "11.0" <mesa-stable at lists.freedesktop.org>

The lower_integer_multiplication() pass was added before 10.6 so it
should get Cc'd there as well.

> Signed-off-by: Kristian Høgsberg Kristensen <krh at bitplanet.net>
> ---
>  src/mesa/drivers/dri/i965/brw_fs.cpp | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
> index 225a312..618bbd2 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
> @@ -1564,7 +1564,10 @@ fs_visitor::assign_vs_urb_setup()
>
>              inst->src[i].file = HW_REG;
>              inst->src[i].fixed_hw_reg =
> -               retype(brw_vec8_grf(grf, 0), inst->src[i].type);
> +               stride(byte_offset(retype(brw_vec8_grf(grf, 0), inst->src[i].type),
> +                                  inst->src[i].subreg_offset),
> +                      inst->exec_size * inst->src[i].stride,
> +                      inst->exec_size, inst->src[i].stride);

I had to think a little bit about the V, W, H parameters to stride(),
but I believe they are correct.

Nice work. This explains why I couldn't spot anything wrong when I
looked at it a little bit during XDC -- I was debugging the lowering
pass. Everything was right there, we were just throwing away the
subreg offset later. :)

Reviewed-by: Matt Turner <mattst88 at gmail.com>