Clarify memory model atomic operation language

[ewerness-nv]

The spec has this language in https://registry.khronos.org/vulkan/specs/1.3-extensions/html/vkspec.html#memory-model-atomic-operation:

Two atomic operations A and B are mutually-ordered if and only if they are potentially-mutually-ordered and any of the following are true:

• A and B are both device operations.
• A and B are both host operations.
• A is a device operation, B is a host operation, and the implementation supports concurrent host- and device-atomics.

There's no bit to indicate that a Vulkan implementation supports concurrent atomics in the last bullet, but it seems weird for the spec to imply a support bit that doesn't exist and assume that we should read it as disable because the bit doesn't exist.

[oddhack]

Telecon: @jeffbolznv or someone else from the MM TSG will propose MR.