[LTP] [PATCH 2/2] [RFC] syscalls/mmap12: Do not fail on non-present pages

Jan Stancek jstancek@redhat.com
Thu Sep 14 16:08:20 CEST 2017

----- Original Message -----
> Hi!
> > > @Michal: Can you think of alternative way to differentiate between:
> > > "MAP_POPULATE did its best" vs. "MAP_POPULATE is broken"?
> > 
> > What about reading /proc/self/smaps for the range and check for RSS
> > field? Seeing 0 could be reported as an unexpected _warning_ but I am
> > not really sure there is any correctness test we could do here. Maybe
> > just compare that MAP_POPULATE will give you the same content as the
> > manual fault in.
> Jan will you try to implement something along these lines or should we
> merge the original patch before the release and look into this later in
> a case we decide to?


Let's go with original for now.

I think RSS != 0 is likely to work, but if we can't guarantee that,
I'd prefer simpler test over sporadic/rare false positives.


More information about the ltp mailing list