Re: zap_page_range(): TLB flush race

[Jamie Lokier]

Manfred Spraul wrote:
> Can we ignore the munmap+access case?
> I'd say that if 2 threads race with munmap+access, then the behaviour is
> undefined.
> Tlb flushes are expensive, I'd like to avoid the second tlb flush as in
> Kanoj's patch.

No, you can't ignore it.  A variation called mprotect+access is used by
garbage collection systems that expect to receive SEGVs when access is
to a protected region.

At very least, you'd have to document the race very clearly, and provide
a workaround.

-- Jamie