Swap reclaiming

Swap reclaiming

[Pavel Machek]

Hi!

Browsing through 2.4.8 patch, I found this:

+/*
+ * When swap space gets filled up, we will set this flag.
+ * This will make do_swap_page(), in the page fault path,
+ * free swap entries on swapin so we'll reclaim swap space
+ * in order to be able to swap something out.
+ *
+ * At the moment we start reclaiming when swap usage goes
+ * over 80% of swap space.
+ *
+ * XXX: Random numbers, fixme.
+ */
+#define SWAP_FULL_PCT 80
+int vm_swap_full (void)
+{
+       int swap_used = total_swap_pages - nr_swap_pages;
+
+       return swap_used * 100 > total_swap_pages * SWAP_FULL_PCT;
+}
+

Hmm... This kind-of defeats purposes on swap priorities: with this,
you are going to fill slow swap even through there is lots of fast
swap that could be reclaimed.

I'm not sure what fix should be.
								Pavel
-- 
I'm pavel@ucw.cz. "In my country we have almost anarchy and I don't care."
Panos Katsaloulis describing me w.r.t. patents at discuss@linmodems.org

Re: Swap reclaiming

[Rik van Riel]

On Sun, 26 Aug 2001, Pavel Machek wrote:

	[snip swap full code]

> Hmm... This kind-of defeats purposes on swap priorities: with this,
> you are going to fill slow swap even through there is lots of fast
> swap that could be reclaimed.

When swap fills up, we will start freeing swap space on swapin.

This means that we'll free up space in both fast and slow swap.

> I'm not sure what fix should be.

I'm not sure what the problem would be ;)

(but would love a more detailed explanation if you
really have found a problem)

regards,

Rik
-- 
IA64: a worthy successor to i860.

http://www.surriel.com/		http://distro.conectiva.com/

Send all your spam to aardvark@nl.linux.org (spam digging piggy)