On Mon, Jan 26, 2015 at 04:19:06PM +0100, Andrea Arcangeli wrote:
> Overall I don't see how we could collapse in readonly vma and where
> the bug is for this case, but I may be overlooking something obvious.
I just realized what the problem was... that the "ro" is not the total
number of readonly ptes mapped by the pmd,.. because we don't count
the none ones as readonly too.
It misses a ro increase or equivalent adjustment:
if (pte_none(pteval)) {
+ ro++;
if (++none <= khugepaged_max_ptes_none)
[..]
Andrea
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>