Metal: Fix dynamic uniform buffer offset corruption when rebinding sets

[stuartcarnie]

Closes #114069

When the same uniform set is bound multiple times within a render pass (e.g., OPAQUE pass then ALPHA pass with different instance buffers), the dynamic offset bits were being OR'd together, corrupting the packed frame indices.

For example, if OPAQUE binds set 1 with frame_idx=1 (0x10) and ALPHA binds set 1 with frame_idx=2 (0x20), the OR produces 0x30 instead of the correct 0x20, causing the shader to read from the wrong buffer offset.

This fix clears the relevant bits for each set being bound before OR'ing the new values, ensuring only the latest frame indices are used.

Fixes rendering artefacts in the mobile renderer which uses two dynamic uniforms (uniform buffer + storage buffer) in set 1, unlike the clustered renderer which only has one.

🤖 AI disclosure

I used AI to identify the bug. I have validated the code and verified the fix.

I used the following prompt:

A rendering regression for Metal was introduced by commit 230adb7, which I expect is related to UNIFORM_TYPE_UNIFORM_BUFFER_DYNAMIC or UNIFORM_TYPE_STORAGE_BUFFER_DYNAMIC. It appears there is corruption rendering triangles. The artefacts only appear with the mobile renderer @servers/rendering/renderer_rd/shaders/forward_mobile/ Can you see any issues with how Metal handles dynamic uniforms. Essentially, Godot requests a buffer, and a MTLBuffer is allocated by multiplying the requested size * the number of in-flight frames. Each time the buffer is used for a frame, an index is increased (modulus the number of in-flight frames) to write to that section of the buffer.

[akien-mga]

Thanks!