>On Tue, Jun 08, 1999 at 12:47:03AM +0000, Linus Torvalds wrote:
>> I certainly always advocate having some swap-space: it makes the "uhhuh,
>> I have too little memory" case degrade much more gracefully..
>
>yes, with some swap space the box swaps to death, while without swap it
>simply locks hard ;)
Please try to reproduce on:
ftp://e-mind.com/pub/andrea/kernel-patches/2.2.9_andrea-VM3.gz
(it's against clean 2.2.9)
About being slower without swap I think it's because if we don't find
something to swapout we never return from swap_out(). I think we should
simply return 1 even when unmapping page/swap cache pages (if the
reference count of the page is equal 1, and that means the only refernce
belongs to the page cache and so the next shrink_mmap can do progress in
freeing memory). I'll do some test soon.
Andrea Arcangeli
-
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/