On Tue, Jan 19, 2010 at 10:44:23AM +0200, Pekka Enberg wrote:

On Tue, Jan 19, 2010 at 10:26 AM, Gleb Natapov [off-list ref] wrote:

quoted

quoted

me that you're trying to use a big hammer (mlock) when a polite hint
for the VM would probably be sufficient for it do its job.

I what to tell to VM "swap this, don't swap that" and as far as I see
there is no other way to do it currently.

Yeah, which is why I was suggesting that maybe posix_madvise() needs
to be extended to have a MADV_NEED_BUT_LESS_IMPORTANT flag that can be
used as a hint by mm/vmscan.c to first swap the guest address spaces.

If such thing would exist may be I would have used it since swapping out
of a wrong page is not live or death matter in my case, but mlockall()
provides me with exactly what I need and without swapping out wrong
pages. Speaking about adding such madvise call wouldn't it be even
harder to justify? It obviously not good enough for real-time use and my
case, I admit, is unusual. Also if we start prioritise memory why stop
on binary, why not set value like "this memory is more important then
that memory by factor of 5"?

--
			Gleb.

--
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>