On 2/16/12 10:57 AM, Andrea Righi wrote:
Maybe we should try to push ...something... in the memcg code for the
short-term future, make it as much generic as possible, and for the
long-term try to reuse the same feature (totally or in part) in the
per-fd approach via fadvise().
Yes - the two approaches are complementary and we should probably pursue
both.
There are a number of apps which are already using fadvise though:
https://issues.apache.org/jira/browse/MAPREDUCE-3289
http://highscalability.com/blog/2012/1/12/peregrine-a-map-reduce-framework-for-iterative-and-pipelined.html
and probably many other similar cases that are not open source.
Some of these apps may be better off using NOREUSE instead of DONTNEED,
since they may not have a clue on what else is going on in the system.
The way I think about it: NOREUSE is a statement about what my process
is doing and DONTNEED is a statement about the entire system.
-Arun
--
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/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>