linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: David Rientjes <rientjes@google.com>
To: Lee Schermerhorn <lee.schermerhorn@hp.com>,
	Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm@kvack.org, linux-numa@vger.kernel.org,
	Mel Gorman <mel@csn.ul.ie>,
	Randy Dunlap <randy.dunlap@oracle.com>,
	Nishanth Aravamudan <nacc@us.ibm.com>,
	Adam Litke <agl@us.ibm.com>, Andy Whitcroft <apw@canonical.com>,
	Christoph Lameter <cl@linux-foundation.org>,
	eric.whitney@hp.com, Yasunori Goto <y-goto@jp.fujitsu.com>
Subject: [patch] mm: clear node in N_HIGH_MEMORY and stop kswapd when all memory is offlined
Date: Wed, 7 Oct 2009 01:24:59 -0700 (PDT)	[thread overview]
Message-ID: <alpine.DEB.1.00.0910070043140.16136@chino.kir.corp.google.com> (raw)
In-Reply-To: <20091006031924.22576.35018.sendpatchset@localhost.localdomain>

On Mon, 5 Oct 2009, Lee Schermerhorn wrote:

> [PATCH 11/11] hugetlb:  offload [un]registration of sysfs attr to worker thread
> 
> Against:  2.6.31-mmotm-090925-1435
> 
> New in V6
> 
> V7:  + remove redundant check for memory{ful|less} node from 
>        node_hugetlb_work().  Rely on [added] return from
>        hugetlb_register_node() to differentiate between transitions
>        to/from memoryless state.
> 

That doesn't work because the memory hotplug code doesn't clear the 
N_HIGH_MEMORY bit for status_change_nid on MEM_OFFLINE, so 
hugetlb_register_node() will always return true under such conditions.

The following should fix it.  Christoph?



mm: clear node in N_HIGH_MEMORY and stop kswapd when all memory is offlined

When memory is hot-removed, its node must be cleared in N_HIGH_MEMORY if
there are no present pages left.

In such a situation, kswapd must also be stopped since it has nothing
left to do.

Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: David Rientjes <rientjes@google.com>
---
 include/linux/swap.h |    1 +
 mm/memory_hotplug.c  |    4 ++++
 mm/vmscan.c          |   28 ++++++++++++++++++++++------
 3 files changed, 27 insertions(+), 6 deletions(-)

diff --git a/include/linux/swap.h b/include/linux/swap.h
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -273,6 +273,7 @@ extern int scan_unevictable_register_node(struct node *node);
 extern void scan_unevictable_unregister_node(struct node *node);
 
 extern int kswapd_run(int nid);
+extern void kswapd_stop(int nid);
 
 #ifdef CONFIG_MMU
 /* linux/mm/shmem.c */
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -838,6 +838,10 @@ repeat:
 
 	setup_per_zone_wmarks();
 	calculate_zone_inactive_ratio(zone);
+	if (!node_present_pages(node)) {
+		node_clear_state(node, N_HIGH_MEMORY);
+		kswapd_stop(node);
+	}
 
 	vm_total_pages = nr_free_pagecache_pages();
 	writeback_set_ratelimit();
diff --git a/mm/vmscan.c b/mm/vmscan.c
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2163,6 +2163,7 @@ static int kswapd(void *p)
 	order = 0;
 	for ( ; ; ) {
 		unsigned long new_order;
+		int ret;
 
 		prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
 		new_order = pgdat->kswapd_max_order;
@@ -2174,19 +2175,23 @@ static int kswapd(void *p)
 			 */
 			order = new_order;
 		} else {
-			if (!freezing(current))
+			if (!freezing(current) && !kthread_should_stop())
 				schedule();
 
 			order = pgdat->kswapd_max_order;
 		}
 		finish_wait(&pgdat->kswapd_wait, &wait);
 
-		if (!try_to_freeze()) {
-			/* We can speed up thawing tasks if we don't call
-			 * balance_pgdat after returning from the refrigerator
-			 */
+		ret = try_to_freeze();
+		if (kthread_should_stop())
+			break;
+
+		/*
+		 * We can speed up thawing tasks if we don't call balance_pgdat
+		 * after returning from the refrigerator
+		 */
+		if (!ret)
 			balance_pgdat(pgdat, order);
-		}
 	}
 	return 0;
 }
@@ -2441,6 +2446,17 @@ int kswapd_run(int nid)
 	return ret;
 }
 
+/*
+ * Called by memory hotplug when all memory in a node is offlined.
+ */
+void kswapd_stop(int nid)
+{
+	struct task_struct *kswapd = NODE_DATA(nid)->kswapd;
+
+	if (kswapd)
+		kthread_stop(kswapd);
+}
+
 static int __init kswapd_init(void)
 {
 	int nid;

--
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>

  parent reply	other threads:[~2009-10-07  8:25 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-10-06  3:17 [PATCH 0/11] hugetlb: V9 numa control of persistent huge pages alloc/free Lee Schermerhorn
2009-10-06  3:17 ` [PATCH 1/11] hugetlb: rework hstate_next_node_* functions Lee Schermerhorn
2009-10-06  3:17 ` [PATCH 2/11] hugetlb: add nodemask arg to huge page alloc, free and surplus adjust fcns Lee Schermerhorn
2009-10-06  9:09   ` David Rientjes
2009-10-07  3:26   ` David Rientjes
2009-10-07 14:13     ` Lee Schermerhorn
2009-10-06  3:17 ` [PATCH 3/11] hugetlb: factor init_nodemask_of_node Lee Schermerhorn
2009-10-07  3:21   ` David Rientjes
2009-10-06  3:18 ` [PATCH 4/11] hugetlb: derive huge pages nodes allowed from task mempolicy Lee Schermerhorn
2009-10-07  3:26   ` David Rientjes
2009-10-07 16:30     ` Lee Schermerhorn
2009-10-07 20:09       ` David Rientjes
2009-10-06  3:18 ` [PATCH 5/11] hugetlb: accomodate reworked NODEMASK_ALLOC Lee Schermerhorn
2009-10-06  3:18 ` [PATCH 6/11] hugetlb: add generic definition of NUMA_NO_NODE Lee Schermerhorn
2009-10-06  9:28   ` David Rientjes
2009-10-06  3:18 ` [PATCH 7/11] hugetlb: add per node hstate attributes Lee Schermerhorn
2009-10-07  4:04   ` David Rientjes
2009-10-06  3:18 ` [PATCH 8/11] hugetlb: update hugetlb documentation for NUMA controls Lee Schermerhorn
2009-10-06  3:18 ` [PATCH 9/11] hugetlb: use only nodes with memory for huge pages Lee Schermerhorn
2009-10-06  3:18 ` [PATCH 10/11] hugetlb: handle memory hot-plug events Lee Schermerhorn
2009-10-07  4:12   ` David Rientjes
2009-10-06  3:19 ` [PATCH 11/11] hugetlb: offload per node attribute registrations Lee Schermerhorn
2009-10-06 16:01   ` Andi Kleen
2009-10-06 16:28     ` Lee Schermerhorn
2009-10-06 16:46       ` Andi Kleen
2009-10-06 17:57         ` Lee Schermerhorn
2009-10-07  8:24   ` David Rientjes [this message]
2009-10-07 14:25     ` [patch] mm: clear node in N_HIGH_MEMORY and stop kswapd when all memory is offlined Christoph Lameter
2009-10-07 16:48     ` Lee Schermerhorn
2009-10-07 19:53       ` David Rientjes
2009-10-06 16:02 ` [PATCH 0/11] hugetlb: V9 numa control of persistent huge pages alloc/free Andi Kleen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=alpine.DEB.1.00.0910070043140.16136@chino.kir.corp.google.com \
    --to=rientjes@google.com \
    --cc=agl@us.ibm.com \
    --cc=akpm@linux-foundation.org \
    --cc=apw@canonical.com \
    --cc=cl@linux-foundation.org \
    --cc=eric.whitney@hp.com \
    --cc=lee.schermerhorn@hp.com \
    --cc=linux-mm@kvack.org \
    --cc=linux-numa@vger.kernel.org \
    --cc=mel@csn.ul.ie \
    --cc=nacc@us.ibm.com \
    --cc=randy.dunlap@oracle.com \
    --cc=y-goto@jp.fujitsu.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox