On Fri, 21 Mar 2014, Kelley Nielsen wrote:
> The function try_to_unuse() is of quadratic complexity, with a lot of
> wasted effort. It unuses swap entries one by one, potentially iterating
> over all the page tables for all the processes in the system for each
> one.
>
> This new proposed implementation of try_to_unuse simplifies its
> complexity to linear. It iterates over the system's mms once, unusing
> all the affected entries as it walks each set of page tables. It also
> makes similar changes to shmem_unuse.
>
> Improvement
>
> Time took by swapoff on a swap partition containing about 240M of data,
> with about 1.1G free memory and about 520M swap available. Swap
> partition was on a laptop with a hard disk drive (not SSD).
>
> Present implementation....about 13.8s
> Prototype.................about 5.5s
I haven't studied the patch yet (and won't manage to do so in the week
ahead), nor examined its performance; but I have taken it out for a
spin, and I'm impressed by its robustness - swap being as racy as it
is, I had expected plenty of trouble, but very little - well done.
Just three little self-explanatory fixes needed so far, all down
at the out_put end: please fold in to your next version whenever.
Hugh
--- 3.14-rc7-kn/mm/swapfile.c 2014-03-22 12:22:46.420136358 -0700
+++ linux/mm/swapfile.c 2014-03-23 07:29:59.852002968 -0700
@@ -1479,12 +1479,14 @@ out_put:
* that we must not delete, since it may not have been written
* out to swap yet.
*/
+ lock_page(page);
if (PageSwapCache(page) &&
- likely(page_private(page) == entry.val)){
- lock_page(page);
+ likely(page_private(page) == entry.val)) {
+ wait_on_page_writeback(page);
delete_from_swap_cache(page);
- unlock_page(page);
}
+ unlock_page(page);
+ page_cache_release(page);
}
out:
return retval;
--
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>