Re: [PATCH 19/21] swap_prefetch: Conversion of nr_unstable to ZVC

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tuesday 13 June 2006 07:14, Christoph Lameter wrote:
> Subject: swap_prefetch: conversion of nr_unstable to per zone counter
> From: Christoph Lameter <[email protected]>
>
> The determination of the vm state is now not that expensive
> anymore after we remove the use of the page state.
> Remove the logic to avoid the expensive checks.
>
> Signed-off-by: Christoph Lameter <[email protected]>
> Signed-off-by: Andrew Morton <[email protected]>

Nack. You're changing some other code unintentionally.

>
> Index: linux-2.6.17-rc6-cl/mm/swap_prefetch.c
> ===================================================================
> --- linux-2.6.17-rc6-cl.orig/mm/swap_prefetch.c	2006-06-12
> 12:55:39.368358636 -0700 +++
> linux-2.6.17-rc6-cl/mm/swap_prefetch.c	2006-06-12 12:55:41.321362692 -0700
> @@ -298,7 +298,7 @@ static int prefetch_suitable(void)
>  {
>  	unsigned long limit;
>  	struct zone *z;
> -	int node, ret = 0, test_pagestate = 0;
> +	int node, ret = 0;
>
>  	/* Purposefully racy */
>  	if (test_bit(0, &swapped.busy)) {
> @@ -306,19 +306,6 @@ static int prefetch_suitable(void)
>  		goto out;
>  	}
>
> -	/*
> -	 * get_page_state and above_background_load are expensive so we only
> -	 * perform them every SWAP_CLUSTER_MAX prefetched_pages.
> -	 * We test to see if we're above_background_load as disk activity
> -	 * even at low priority can cause interrupt induced scheduling
> -	 * latencies.
> -	 */
> -	if (!(sp_stat.prefetched_pages % SWAP_CLUSTER_MAX)) {
> -		if (above_background_load())
> -			goto out;

We still need the above_background_load check, just remove the test_pagestate 
check and it can be checked every time. Drop the comment about get_page_state 
only.

> -		test_pagestate = 1;
> -	}
> -
>  	clear_current_prefetch_free();
>
>  	/*
> @@ -357,7 +344,6 @@ static int prefetch_suitable(void)
>  	 */
>  	for_each_node_mask(node, sp_stat.prefetch_nodes) {
>  		struct node_stats *ns = &sp_stat.node[node];
> -		struct page_state ps;
>
>  		/*
>  		 * We check to see that pages are not being allocated
> @@ -375,11 +361,6 @@ static int prefetch_suitable(void)
>  		} else
>  			ns->last_free = ns->current_free;
>
> -		if (!test_pagestate)
> -			continue;
> -
> -		get_page_state_node(&ps, node);
> -
>  		/* We shouldn't prefetch when we are doing writeback */
>  		if (node_page_state(node, NR_WRITEBACK)) {
>  			node_clear(node, sp_stat.prefetch_nodes);
> @@ -394,7 +375,8 @@ static int prefetch_suitable(void)
>  			node_page_state(node, NR_ANON) +
>  			node_page_state(node, NR_SLAB) +
>  			node_page_state(node, NR_DIRTY) +
> -			ps.nr_unstable + total_swapcache_pages;
> +			node_page_state(node, NR_UNSTABLE) +
> +			total_swapcache_pages;
>  		if (limit > ns->prefetch_watermark) {
>  			node_clear(node, sp_stat.prefetch_nodes);
>  			continue;

-- 
-ck
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]
  Powered by Linux