[RFC]get_swap_page():delay update swap_list.next

[RFC]get_swap_page():delay update swap_list.next

[Bo Liu]

If scan_swap_map() successed in current si, there is no
need to update swap_list.next.So get_swap_page next time 
called can start search from the last swap_info(which still
have free slots).
 
 
Signed-off-by: Bo Liu <bo-liu@hotmail.com>
---
 
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 63ce10f..b3ba2c5 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -480,13 +480,13 @@ swp_entry_t get_swap_page(void)
   if (!(si->flags & SWP_WRITEOK))
    continue;
 
-  swap_list.next = next;
   /* This is called for allocating swap entry for cache */
   offset = scan_swap_map(si, SWAP_CACHE);
   if (offset) {
    spin_unlock(&swap_lock);
    return swp_entry(type, offset);
   }
+  swap_list.next = next;
   next = swap_list.next;
  }
  		 	   		  
_________________________________________________________________
Windows Live: Friends get your Flickr, Yelp, and Digg updates when they e-mail you.
http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/social-network-basics.aspx?ocid=PID23461::T:WLMTAGL:ON:WL:en-xm:SI_SB_3:092010

--
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>

Re: [RFC]get_swap_page():delay update swap_list.next

[hugh.dickins]

On Mon, 19 Oct 2009 you wrote:

> If scan_swap_map() successed in current si, there is no
> need to update swap_list.next.So get_swap_page next time
> called can start search from the last swap_info(which still
> have free slots).

No, I think you'll find that's mistaken.  It is intended behaviour that
allocating from two swap areas of the same priority will alternate
between them page by page (or cycle around three or more of
the same priority) - a kind of poor man's striping.

You may think that's silly, and not want that behaviour: in which
case assign different priorities to those swap areas, and it will
then behave like you were changing the same priority case to
behave.  So there is no reason to make your change.

Hugh

[I apologize for breaking the threading on your mail: I've been
caught by surprise to find myself unable to reply in the usual way.]




Get 50% off Norton Security- http://www.tiscali.co.uk/securepc

--
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>