> oom-9 is out. I added a very useful anti-starvation test at the top of
> swap_out() and shm_swap():
>
> if (!nr_swap_pages)
> return 0;
This is broken, because swap_out() also is used to:
- clear out clean pages which have a mapping in swap space
- unmap mmap()ed pages from the application's address space
> I also replaced the nr_free_pages < freepages.min with
> !free_memory_available() in mmput() to use better kswapd when some
> process release memory.
That is a very good change, because it will allow us to
change the free-memory criterium in _one_ place when we
need to change it someplace in the future.
> Now I' d like to get reports of failed MM allocation even if there's
> some physiscal space available (as Rik pointed out). Once somebody
> will report such problem and how to trigger it I' ll think something
> to improve kswapd.
I'm pretty sure that these reports already are in the linux-kernel
archives of several months back. In any case, I don't think we
should take the risk of going with a less sure behaviour of kswapd,
especially not now we're in a code freeze...
Rik.
+-------------------------------------------------------------------+
| Linux memory management tour guide. H.H.vanRiel@phys.uu.nl |
| Scouting Vries cubscout leader. http://www.phys.uu.nl/~riel/ |
+-------------------------------------------------------------------+
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/