On 11/02/2012 06:41 PM, Michel Lespinasse wrote:

On Fri, Nov 2, 2012 at 9:52 AM, Rik van Riel [off-list ref] wrote:

quoted

On 10/31/2012 06:33 AM, Michel Lespinasse wrote:

I guess the suggestion is OK in the sense that I can't see a case
where it'd hurt. However, it still won't find all cases where we just
unmapped a region of size N with the correct alignment - it could be
that the first region we find has insufficient alignment, and then the
search with an increased length could fail, even though there exists
an aligned gap (just not the first) of the desired size. So, this is
only a partial solution.

Unfortunately, I don't think there is an efficient solution to the
general problem, and the partial solutions discussed above (both yours
and mine) don't seem to cover enough cases to warrant the complexity
IMO...

The common case (anonymous memory) is that no alignment is
required, so your solution and mine would be equivalent :)

You are right that your solution and mine are pretty
much the same for the (rarer) cases where alignment
is needed.  Lets stick with your simpler version.

ACK to your version

-- 
All rights reversed

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>