Fix uninitialized variable ending up sent to Vulkan

[darksylinc]

The first time a shader is compiled Godot performs the following:

for (uint32_t i = 0; i < SHADER_STAGE_MAX; i++) {
	if(spirv_data.push_constant_stages_mask.has_flag((ShaderStage)(1 << i))) {
		binary_data.push_constant_vk_stages_mask |= shader_stage_masks[i];
	}
}

However binary_data.push_constant_vk_stages_mask is never initialized to 0 and thus contains garbage data or'ed with the good data.

This value is used by push constants (and many other things) thus it can be a big deal.

Fortunately because the relevant flags are always guaranteed to be set (but not guaranteed to be unset), the damage is restricted to:

1. Performance (unnecessary flushing & over-excessive barriers)
2. Overwriting push descriptors already set (this would be serious, doesn't seem to be an issue)
3. Driver implementations going crazy when they see bits set they don't expect (unknown if this is an issue)

This uninitialized value is later saved into the binary cache.

Valgrind is able to detect this bug on the first run, but not on the subsequent ones because the data comes from a file.

cache_file_version has been bumped to force rebuild of all cached shaders. Because the ones generated so far are compromised.

Can this fix TDRs? (aka DEVICE_LOST)

Possibly.

Should this be backported to older versions?

Probably, I didn't check

[Calinou]

I wonder if this can help resolve #77427.

[darksylinc]

I wonder if this can help resolve #77427.

That bug is still present. But I thought it was on purpose / a limitation (zero-initializing 3D volume data takes time).

I can look into it if that's an issue, but it looks like it's just the 3D buffer not being initialized.

[BastiaanOlij]

Auch...

[clayjohn]

Looks great! Congratulations on your first (of many) merged PR.

Note for maintainers, this should probably not be cherry picked to 4.1. technically it's fine, the only thing it will do is invalidate the previous cache.