On 29/08/15 02:33, Ilia Mirkin wrote:
> On Fri, Aug 28, 2015 at 8:27 PM, Jordan Justen
> <jordan.l.jus...@intel.com> wrote:
>> On 2015-08-05 01:30:17, Iago Toral Quiroga wrote:
>>> From: Samuel Iglesias Gonsalvez <sigles...@igalia.com>
>>>
>>> This commit also adds functions to calculate std430 base alignment and sizes
>>>
>>> Signed-off-by: Samuel Iglesias Gonsalvez <sigles...@igalia.com>
>>> ---
>>> src/glsl/ast.h | 1 +
>>> src/glsl/ast_to_hir.cpp | 20 +++++--
>>> src/glsl/ast_type.cpp | 1 +
>>> src/glsl/glsl_parser.yy | 2 +
>>> src/glsl/glsl_types.cpp | 117
>>> +++++++++++++++++++++++++++++++++++++++
>>> src/glsl/glsl_types.h | 15 ++++-
>>> src/glsl/link_uniform_blocks.cpp | 15 ++++-
>>> src/mesa/main/mtypes.h | 3 +-
>>> 8 files changed, 165 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/src/glsl/ast.h b/src/glsl/ast.h
>>> index d8c6cea..ac3f80f 100644
>>> --- a/src/glsl/ast.h
>>> +++ b/src/glsl/ast.h
>>> @@ -491,6 +491,7 @@ struct ast_type_qualifier {
>>> /** \name Layout qualifiers for GL_ARB_uniform_buffer_object */
>>> /** \{ */
>>> unsigned std140:1;
>>> + unsigned std430:1;
>>> unsigned shared:1;
>>> unsigned packed:1;
>>> unsigned column_major:1;
>>> diff --git a/src/glsl/ast_to_hir.cpp b/src/glsl/ast_to_hir.cpp
>>> index 6e63d86..2dadb03 100644
>>> --- a/src/glsl/ast_to_hir.cpp
>>> +++ b/src/glsl/ast_to_hir.cpp
>>> @@ -2886,11 +2886,12 @@ apply_type_qualifier_to_variable(const struct
>>> ast_type_qualifier *qual,
>>> var->data.depth_layout = ir_depth_layout_none;
>>>
>>> if (qual->flags.q.std140 ||
>>> + qual->flags.q.std430 ||
>>> qual->flags.q.packed ||
>>> qual->flags.q.shared) {
>>> _mesa_glsl_error(loc, state,
>>> - "uniform block layout qualifiers std140, packed,
>>> and "
>>> - "shared can only be applied to uniform blocks, not "
>>> + "uniform block layout qualifiers std140, std430,
>>> packed, "
>>> + "and shared can only be applied to uniform blocks,
>>> not "
>>> "members");
>>> }
>>>
>>> @@ -5655,12 +5656,14 @@ ast_process_structure_or_interface_block(exec_list
>>> *instructions,
>>> const struct ast_type_qualifier *const qual =
>>> & decl_list->type->qualifier;
>>> if (qual->flags.q.std140 ||
>>> + qual->flags.q.std430 ||
>>> qual->flags.q.packed ||
>>> qual->flags.q.shared) {
>>> _mesa_glsl_error(&loc, state,
>>> "uniform/shader storage block layout
>>> qualifiers "
>>> - "std140, packed, and shared can only be
>>> applied "
>>> - "to uniform/shader storage blocks, not
>>> members");
>>> + "std140, std430, packed, and shared can only
>>> be "
>>> + "applied to uniform/shader storage blocks,
>>> not "
>>> + "members");
>>> }
>>>
>>> if (qual->flags.q.constant) {
>>> @@ -5872,6 +5875,13 @@ ast_interface_block::hir(exec_list *instructions,
>>> this->block_name);
>>> }
>>>
>>> + if (!this->layout.flags.q.buffer &&
>>> + this->layout.flags.q.std430) {
>>> + _mesa_glsl_error(&loc, state,
>>> + "std430 storage block layout qualifier is supported
>>> "
>>> + "only for shader storage blocks");
>>> + }
>>> +
>>> /* The ast_interface_block has a list of ast_declarator_lists. We
>>> * need to turn those into ir_variables with an association
>>> * with this uniform block.
>>> @@ -5881,6 +5891,8 @@ ast_interface_block::hir(exec_list *instructions,
>>> packing = GLSL_INTERFACE_PACKING_SHARED;
>>> } else if (this->layout.flags.q.packed) {
>>> packing = GLSL_INTERFACE_PACKING_PACKED;
>>> + } else if (this->layout.flags.q.std430) {
>>> + packing = GLSL_INTERFACE_PACKING_STD430;
>>> } else {
>>> /* The default layout is std140.
>>> */
>>> diff --git a/src/glsl/ast_type.cpp b/src/glsl/ast_type.cpp
>>> index a4671e2..6e69ba2 100644
>>> --- a/src/glsl/ast_type.cpp
>>> +++ b/src/glsl/ast_type.cpp
>>> @@ -123,6 +123,7 @@ ast_type_qualifier::merge_qualifier(YYLTYPE *loc,
>>> ubo_layout_mask.flags.q.std140 = 1;
>>> ubo_layout_mask.flags.q.packed = 1;
>>> ubo_layout_mask.flags.q.shared = 1;
>>> + ubo_layout_mask.flags.q.std430 = 1;
>>>
>>> ast_type_qualifier ubo_binding_mask;
>>> ubo_binding_mask.flags.i = 0;
>>> diff --git a/src/glsl/glsl_parser.yy b/src/glsl/glsl_parser.yy
>>> index 2b0c8bd..e1c0f3d 100644
>>> --- a/src/glsl/glsl_parser.yy
>>> +++ b/src/glsl/glsl_parser.yy
>>> @@ -1197,6 +1197,8 @@ layout_qualifier_id:
>>> $$.flags.q.std140 = 1;
>>> } else if (match_layout_qualifier($1, "shared", state) == 0) {
>>> $$.flags.q.shared = 1;
>>> + } else if (match_layout_qualifier($1, "std430", state) == 0) {
>>> + $$.flags.q.std430 = 1;
>>> } else if (match_layout_qualifier($1, "column_major", state) ==
>>> 0) {
>>> $$.flags.q.column_major = 1;
>>> /* "row_major" is a reserved word in GLSL 1.30+. Its token is
>>> parsed
>>> diff --git a/src/glsl/glsl_types.cpp b/src/glsl/glsl_types.cpp
>>> index 755618a..d29769a 100644
>>> --- a/src/glsl/glsl_types.cpp
>>> +++ b/src/glsl/glsl_types.cpp
>>> @@ -1357,6 +1357,123 @@ glsl_type::std140_size(bool row_major) const
>>> return -1;
>>> }
>>>
>>> +unsigned
>>> +glsl_type::std430_base_alignment(bool row_major) const
>>> +{
>>> + unsigned base_alignment = 0;
>>> + /* When using the "std430" storage layout, shader storage
>>> + * blocks will be laid out in buffer storage identically to uniform and
>>> + * shader storage blocks using the "std140" layout, except that the base
>>> + * alignment of arrays of scalars and vectors in rule (4) and of
>>> structures
>>> + * in rule (9) are not rounded up a multiple of the base alignment of a
>>> vec4.
>>> + */
>>> +
>>> + /* (1) If the member is a scalar consuming <N> basic machine units, the
>>> + * base alignment is <N>.
>>> + *
>>> + * (2) If the member is a two- or four-component vector with components
>>> + * consuming <N> basic machine units, the base alignment is 2<N> or
>>> + * 4<N>, respectively.
>>> + *
>>> + * (3) If the member is a three-component vector with components
>>> consuming
>>> + * <N> basic machine units, the base alignment is 4<N>.
>>> + */
>>> + if (this->is_array()) {
>>> + base_alignment =
>>> this->fields.array->std430_base_alignment(row_major);
>>> + } else {
>>> + /* For the rest of cases, use std140_base_alignment() */
>>> + base_alignment = this->std140_base_alignment(row_major);
>>> + }
>>> + return base_alignment;
>>> +}
>>> +
>>> +unsigned
>>> +glsl_type::std430_size(bool row_major) const
>>
>> I think you should move these two new glsl_type functions into a
>> separate, earlier commit.
>>
>> For the parse/link portions of this patch:
>> Reviewed-by: Jordan Justen <jordan.l.jus...@intel.com>
>>
>>> +{
>>> + unsigned N = is_double() ? 8 : 4;
>>> +
>>> + /* (1) If the member is a scalar consuming <N> basic machine units, the
>>> + * base alignment is <N>.
>>> + *
>>> + * (2) If the member is a two- or four-component vector with components
>>> + * consuming <N> basic machine units, the base alignment is 2<N> or
>>> + * 4<N>, respectively.
>>> + *
>>> + * (3) If the member is a three-component vector with components
>>> consuming
>>> + * <N> basic machine units, the base alignment is 4<N>.
>>> + */
>>> + if (this->is_scalar() || this->is_vector()) {
>>> + if (this->vector_elements > 2)
>>> + return glsl_align(this->vector_elements * N, 16);
>>> + else
>>> + return this->vector_elements * N;
>>> + }
>>> +
>>> + if (this->without_array()->is_matrix()) {
>>> + const struct glsl_type *element_type;
>>> + const struct glsl_type *vec_type;
>>> + unsigned int array_len;
>>> +
>>> + if (this->is_array()) {
>>> + element_type = this->fields.array;
>>> + array_len = this->length;
>>> + } else {
>>> + element_type = this;
>>> + array_len = 1;
>>> + }
>>> +
>>> + if (row_major) {
>>> + vec_type = get_instance(element_type->base_type,
>>> + element_type->matrix_columns, 1);
>>> +
>>> + array_len *= element_type->vector_elements;
>>> + } else {
>>> + vec_type = get_instance(element_type->base_type,
>>> + element_type->vector_elements, 1);
>>> + array_len *= element_type->matrix_columns;
>>> + }
>>> + const glsl_type *array_type = glsl_type::get_array_instance(vec_type,
>>> +
>>> array_len);
>>> +
>>> + return array_type->std430_size(false);
>>> + }
>>> +
>>> + /* When using the "std430" storage layout, shader storage
>>> + * blocks will be laid out in buffer storage identically to uniform and
>>> + * shader storage blocks using the "std140" layout, except that the base
>>> + * alignment of arrays of scalars and vectors in rule (4) and of
>>> structures
>>> + * in rule (9) are not rounded up a multiple of the base alignment of a
>>> vec4.
>>> + */
>>> + if (this->is_array())
>>> + return this->length * this->fields.array->std430_size(row_major);
>>
>> Hmm, you quoted the extension spec, but the 4.30 spec has: "When using
>> the std430 storage layout, shader storage blocks will be laid out in
>> buffer storage identically to uniform and shader storage blocks using
>> the std140 layout, except that the base alignment and stride of arrays
>> of scalars and vectors in rule 4 and of structures in rule 9 are not
>> rounded up a multiple of the base alignment of a vec4 ."
>>
>> Notably, it mentions stride. Wouldn't this be wrong for a vec3 array,
>> where the stride should be 12 bytes?
>>
>> I wonder if we'll need to employ some of idr's std140 tricks to work
>> out all the kinks in std430.
>>
>> http://www.paranormal-entertainment.com/idr/blog/posts/2014-10-08T13:28:09Z-Stochastic_Search-Based_Testing_for_Uniform_Block_Layouts/
>
> I highly recommend it -- helped work out a lot of kinks in fp64.
>
> http://cgit.freedesktop.org/~idr/piglit/log/?h=ubo-lolz
>
> However he doesn't have an implementation of std430_packing_rules in
> there, but it should be moderately straightforward, looking at the
> std140 ones.
>
> If you end up testing fp64 with it, note that it has a few typos in
> that area. IMO we should mainline this and push it to piglit, even if
> it's not run on a regular basis.
>
OK, I will take a look at it. Thanks!
Sam
> -ilia
> _______________________________________________
> 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
