[Mesa-dev] [PATCH 1/2] nir: Handle <bits>=32 case in bitfield_insert lowering.

[Connor Abbott]

Both are

Reviewed-by: Connor Abbott <cwabbott0 at gmail.com>

On Wed, Jan 13, 2016 at 2:25 PM, Matt Turner <mattst88 at gmail.com> wrote:
> The OpenGL specifications for bitfieldInsert() says:
>
>    The result will be undefined if <offset> or <bits> is negative, or if
>    the sum of <offset> and <bits> is greater than the number of bits
>    used to store the operand.
>
> Therefore passing bits=32, offset=0 is legal and defined in GLSL.
>
> But the earlier SM5 bfi opcode is specified to accept a bitfield width
> ranging from 0-31. As such, Intel and AMD instructions read only the low
> 5 bits of the width operand, making them not able to implement the
> GLSL-specified behavior directly.
>
> This commit fixes the lowering of bitfield_insert to handle the trivial
> case of <bits> = 32 as
>
>    bitfieldInsert:
>       bits > 31 ? insert : bfi(bfm(bits, offset), insert, base)
>
> Fixes:
>    ES31-CTS.shader_bitfield_operation.bitfieldInsert.uint_2
>    ES31-CTS.shader_bitfield_operation.bitfieldInsert.uvec4_3
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92595
> ---
> These two patches replace 8/9 and 9/9 of the previous series.
> The first 7 patches from it have been reviewed and committed.
>
>  src/glsl/nir/nir_opcodes.py       | 1 +
>  src/glsl/nir/nir_opt_algebraic.py | 6 +++++-
>  2 files changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/src/glsl/nir/nir_opcodes.py b/src/glsl/nir/nir_opcodes.py
> index 1c65def..3e43438 100644
> --- a/src/glsl/nir/nir_opcodes.py
> +++ b/src/glsl/nir/nir_opcodes.py
> @@ -558,6 +558,7 @@ triop("fcsel", tfloat, "(src0 != 0.0f) ? src1 : src2")
>  opcode("bcsel", 0, tuint, [0, 0, 0],
>        [tbool, tuint, tuint], "", "src0 ? src1 : src2")
>
> +# SM5 bfi assembly
>  triop("bfi", tuint, """
>  unsigned mask = src0, insert = src1, base = src2;
>  if (mask == 0) {
> diff --git a/src/glsl/nir/nir_opt_algebraic.py b/src/glsl/nir/nir_opt_algebraic.py
> index 1eb044a..0d31e39 100644
> --- a/src/glsl/nir/nir_opt_algebraic.py
> +++ b/src/glsl/nir/nir_opt_algebraic.py
> @@ -225,9 +225,13 @@ optimizations = [
>
>     # Misc. lowering
>     (('fmod', a, b), ('fsub', a, ('fmul', b, ('ffloor', ('fdiv', a, b)))), 'options->lower_fmod'),
> -   (('bitfield_insert', a, b, c, d), ('bfi', ('bfm', d, c), b, a), 'options->lower_bitfield_insert'),
>     (('uadd_carry', a, b), ('b2i', ('ult', ('iadd', a, b), a)), 'options->lower_uadd_carry'),
>     (('usub_borrow', a, b), ('b2i', ('ult', a, b)), 'options->lower_usub_borrow'),
> +
> +   (('bitfield_insert', 'base', 'insert', 'offset', 'bits'),
> +    ('bcsel', ('ilt', 31, 'bits'), 'insert',
> +              ('bfi', ('bfm', 'bits', 'offset'), 'insert', 'base')),
> +    'options->lower_bitfield_insert'),
>  ]
>
>  # Add optimizations to handle the case where the result of a ternary is
> --
> 2.4.9
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev