Re: [PATCH] mm: prevent mmap_cache race in find_vma()

[Linus Torvalds]

On Thu, Apr 4, 2013 at 12:01 PM, Hugh Dickins <hughd@xxxxxxxxxx> wrote:
>
> When Paul reminded us of it yesterday, I came to wonder if actually
> every use of ACCESS_ONCE in the read form should strictly be matched
> by ACCESS_ONCE whenever modifying the location.
>
> My uneducated guess is that strictly it ought to, in the sense of
> insurance policy; but that (apart from that strange split writing
> issue which came up a couple of months ago) in practice our compilers
> have not "advanced" to the point of making this an issue yet.

I don't see how a compiler could reasonably really ever do anything
different, but I do think the ACCESS_ONCE() modification version might
be a good thing just as a "documentation".

This is a good example of this issue, exactly because we have a mix of
both speculative cases (the find_vma() lookup and modification)
together with strictly exclusive locked accesses to the same field
(the ones that invalidate the cache under the write lock). So
documenting that the write in find_vma() is this kind of "optimistic
unlocked access" is actually a potentially interesting piece of
information for programmers, completely independently of whether the
compiler will then treat it really differently or not.

Of course, a plain comment would do the same, but would be less greppable.

And despite the verbiage here, I don't really have a very strong
opinion on this. I'm going to let it go, and if somebody sends me a
patch with a good explanation in the next merge window, I'll probably
apply it.

                    Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/