BPF program verification is surely irrelevant to userspace patching

BPF program verification is surely irrelevant to userspace patching

BPF programs can be verified to ensure that they do not cause a crash.

BPF programs can be verified to ensure that the programs themselves do not crash, but for this application the BPF programs themselves should obviously be running with minimal privilege, i.e. in user space, so their not crashing is more or less irrelevant to overall system stability. What matters is that installing the patch, and the patch itself, should not crash the userspace program: and verifying that in the general case is, of course, impossible (and even in the non-completely-general real-world case is ridiculously hard).

This really does feel to me like 'kernel developer sees problem, has hammer running in kernel space, concludes that hammer should be employed on userspace nail even though it is the wrong tool and has a raison d'etre which relates to problems that are irrelevant to this use case'.