Incorrect index validation

[atynagano]

rust-gpu/crates/spirv-std/src/byte_addressable_buffer.rs

Lines 73 to 78 in e87c324

	pub unsafe fn load<T>(&self, byte_index: u32) -> T {
	if byte_index + mem::size_of::<T>() as u32 > self.data.len() as u32 {
	panic!("Index out of range");
	}
	buffer_load_intrinsic(self.data, byte_index)
	}

if byte_index + mem::size_of::<T>() as u32 > (mem::size_of::<u32> * self.data.len()) as u32

Isn't the right-hand side multiplied by 4 because data is &[u32]?

[Patryk27]

ByteAddressableBuffer requires for the data to be aligned to 4 bytes, so the current condition is fine IMO; the naming could be better (word_index: u32), but the comment above explains it's inherited from HLSL.

Note that buffer_load_intrinsic() ends up doing 4 * byte_index:

rust-gpu/crates/rustc_codegen_spirv/src/builder/byte_addressable_buffer.rs

Line 202 in e87c324

let two = self.constant_u32(DUMMY_SP, 2);

[atynagano]

So does the left side size_of<T> needs to be divided by 4? Or does size_of returns dword size?

[Patryk27]

Oh, yeah - I think it should say mem::size_of::<T>() / 4 then 🤔