Re: [PATCH] cowlinks v2

[Jamie Lokier]

Pavel Machek wrote:
> > I know a writable mmap needs to trigger a copy in that case.
> > And then are fun cases with MAP_FIXED which may mean invalidation
> > is not allowed.
> 
> How is "invalidation not allowed" for MAP_FIXED? Application will
> never see the fault...

I think Eric secretly meant MAP_LOCKED and/or mlock().

Even if the file isn't copied, when you have two different mappings of
different cowlinks both mapped with MAP_LOCKED, the pages need to be
different in RAM or we break locked page expectations.

-- Jamie
-
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/