[PATCH vkd3d 10/12] vkd3d-shader/hlsl: Detect missing loads on rhs when splitting copies of non-numeric types.

[Francisco Casas]

Hello,

April 19, 2022 6:44 AM, "Giovanni Mascellani" <gmascellani at codeweavers.com> wrote:

> Hi,
> 
> Il 18/04/22 08:34, Giovanni Mascellani ha scritto:
> 
>> Otherwise we can get failed assertions:
>> assert(node->type == HLSL_IR_LOAD);
>> because broadcasts to these types are not implemented yet.
>> Signed-off-by: Francisco Casas<fcasas at codeweavers.com>
>> ---
>> libs/vkd3d-shader/hlsl_codegen.c | 14 ++++++++++++++
>> 1 file changed, 14 insertions(+)
>> diff --git a/libs/vkd3d-shader/hlsl_codegen.c b/libs/vkd3d-shader/hlsl_codegen.c
>> index 73e3b73f..2104c48b 100644
>> --- a/libs/vkd3d-shader/hlsl_codegen.c
>> +++ b/libs/vkd3d-shader/hlsl_codegen.c
>> @@ -628,6 +628,13 @@ static bool split_array_copies(struct hlsl_ctx *ctx, struct hlsl_ir_node
>> *instr,
>> element_type = type->e.array.type;
>> element_size = hlsl_type_get_array_element_reg_size(element_type);
>>> + if (rhs->type != HLSL_IR_LOAD)
>> + {
>> + hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NOT_IMPLEMENTED,
>> + "array store rhs is not HLSL_IR_LOAD, missing broadcast?");
>> + return false;
>> + }
> 
> If the problem here is that the compiler is missing a feature, then I don't think hlsl_error() is
> appropriate, because that would be for a user error. I guess hlsl_fixme() is what should be used
> here.
> 

I see, I think you are right.

> Also, notice that the error message is usually formatted with a leading upper case letter and with
> a trailing dot (but no trailing newline).
>

Okay, I will try to remember that. I think I didn't pay too much attention to this patch
because I just wrote it so that the tests in the following patch of my branch:

725ae863 fcasas tests: Test complex broadcasts.

didn't fail the aforementioned assertion.

> Last, I am not completely sure of what the issue is here, but if eventually properly implementing
> other parts of the compiler will make this code useless, then I'd add a comment saying that.
> 

Yes, this issue stops happening with:

6d2a14a0 fcasas vkd3d-shader/hlsl: Lower complex broadcasts.