I guess what I'm asking is, how are you running the tests? Manually? Automatically? Do you intend to have all the tests pass? Are you interested in collaborating on tests?
For instance, if your end goal is to have all the tests pass, I'm not immediately sure what this kind of commit gets you, since there's still a failure to fix. If you're running CI, it's not going to be useful if you have a bunch of failing tests. I don't see any value in this kind of thing in general—a failure is a failure.
On the other hand, if you're not just interested in using the existing tests, but also in collaborating on tests and helping to contribute to the test suite, I'm certainly open to adding e.g. todo_dxvk annotations and specifically accounting for DXVK-specific failures.