(6/8/12 2:54 AM), Pekka Enberg wrote:
On Fri, Jun 8, 2012 at 6:55 AM, KOSAKI Motohiro
<kosaki.motohiro@xxxxxxxxx> wrote:
For some cases, yes, but probably not for all.
For example, if userspace doesn't know about "about to swap real soon"
condition, it can continue to grow its caches making
fallocate(VOLATILE) pretty much useless.
Fair enough. But, Please consider your scenario don't need vmevent(2),
cat /proc/meminfo works enough. Only dropping activity need a quick work
and just in time notification. cache growing limitation don't need.
Why do you keep bringing this up? Yes, you might get away with polling
/proc/meminfo or /proc/vmstat for some specific cases. It does not
make it a proper generic solution to the problem "vmevent" ABI tries
to address.
Guys, current vmevent _is_ a polling interface. It only is wrapped kernel
timer.
--
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>