From: Christoph Lameter <clameter@sgi.com>
To: akpm@osdl.org
Cc: linux-mm@kvack.org,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Lee Schermerhorn <lee.schermerhorn@hp.com>,
Christoph Lameter <clameter@sgi.com>,
Hugh Dickins <hugh@veritas.com>
Subject: [PATCH 7/7] page migration: Add new fallback function
Date: Thu, 27 Apr 2006 23:03:33 -0700 (PDT) [thread overview]
Message-ID: <20060428060333.30257.43096.sendpatchset@schroedinger.engr.sgi.com> (raw)
In-Reply-To: <20060428060302.30257.76871.sendpatchset@schroedinger.engr.sgi.com>
page migration: Add new fallback function that checks properly for dirty pages
Add a new migration function that checks for PageDirty after unmapping
the ptes. It then directly writes out the page without relying on pageout().
Add some logic to deal with writepage() potentially unlocking the page.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Index: linux-2.6.17-rc2-mm1/mm/migrate.c
===================================================================
--- linux-2.6.17-rc2-mm1.orig/mm/migrate.c 2006-04-27 21:34:12.171976982 -0700
+++ linux-2.6.17-rc2-mm1/mm/migrate.c 2006-04-27 22:52:35.363141399 -0700
@@ -24,6 +24,7 @@
#include <linux/topology.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
+#include <linux/writeback.h>
#include "internal.h"
@@ -445,6 +446,66 @@
}
EXPORT_SYMBOL(buffer_migrate_page);
+static int fallback_migrate_page(struct address_space *mapping,
+ struct page *newpage, struct page *page)
+{
+ int rc;
+
+ if (try_to_unmap(page, 1) == SWAP_FAIL)
+ /* A vma has VM_LOCKED set -> permanent failure */
+ return -EPERM;
+
+ /*
+ * Removing the ptes may have dirtied the page
+ */
+ if (PageDirty(page)) {
+ struct writeback_control wbc = {
+ .sync_mode = WB_SYNC_NONE,
+ .nr_to_write = 1,
+ .range_start = 0,
+ .range_end = LLONG_MAX,
+ .nonblocking = 1,
+ .for_reclaim = 1
+ };
+
+ if (!mapping->a_ops->writepage)
+ /* No write method for the address space */
+ return -EINVAL;
+
+ if (!clear_page_dirty_for_io(page))
+ /* Someone else already triggered a write */
+ return -EAGAIN;
+
+ if (mapping->a_ops->writepage(page, &wbc) < 0)
+ /* I/O Error writing */
+ return -EIO;
+
+ /*
+ * Retry if writepage() removed the lock or the page
+ * is still dirty or undergoing writeback.
+ */
+ if (!PageLocked(page) ||
+ PageWriteback(page) || PageDirty(page))
+ return -EAGAIN;
+ }
+
+ /*
+ * Buffers are managed in a filesystem specific way.
+ * We must have no buffers or drop them.
+ */
+ if (page_has_buffers(page) &&
+ !try_to_release_page(page, GFP_KERNEL))
+ return -EAGAIN;
+
+ rc = migrate_page_move_mapping(mapping, newpage, page);
+
+ if (rc)
+ return rc;
+
+ migrate_page_copy(newpage, page);
+ return 0;
+}
+
/*
* migrate_pages
*
@@ -527,59 +588,19 @@
* Try to migrate the page.
*/
mapping = page_mapping(page);
- if (!mapping) {
+ if (!mapping)
rc = migrate_page(mapping, newpage, page);
- goto unlock_both;
- } else
- if (mapping->a_ops->migratepage) {
- /*
- * Most pages have a mapping and most filesystems
- * should provide a migration function. Anonymous
- * pages are part of swap space which also has its
- * own migration function. This is the most common
- * path for page migration.
- */
+ else if (mapping->a_ops->migratepage)
rc = mapping->a_ops->migratepage(mapping,
newpage, page);
- goto unlock_both;
- }
-
- /*
- * Default handling if a filesystem does not provide
- * a migration function. We can only migrate clean
- * pages so try to write out any dirty pages first.
- */
- if (PageDirty(page)) {
- switch (pageout(page, mapping)) {
- case PAGE_KEEP:
- case PAGE_ACTIVATE:
- goto unlock_both;
-
- case PAGE_SUCCESS:
- unlock_page(newpage);
- goto next;
-
- case PAGE_CLEAN:
- ; /* try to migrate the page below */
- }
- }
-
- /*
- * Buffers are managed in a filesystem specific way.
- * We must have no buffers or drop them.
- */
- if (!page_has_buffers(page) ||
- try_to_release_page(page, GFP_KERNEL)) {
- rc = migrate_page(mapping, newpage, page);
- goto unlock_both;
- }
+ else
+ rc = fallback_migrate_page(mapping, newpage, page);
-unlock_both:
unlock_page(newpage);
-
unlock_page:
- unlock_page(page);
+ if (PageLocked(page)) /* writepage() may unlock */
+ unlock_page(page);
next:
if (rc) {
--
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>
next prev parent reply other threads:[~2006-04-28 6:03 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-04-28 6:03 [PATCH 1/7] page migration: Reorder functions in migrate.c Christoph Lameter
2006-04-28 6:03 ` [PATCH 2/7] page migration: Remove unnecessarily exported functions Christoph Lameter
2006-04-28 6:03 ` [PATCH 3/7] page migration: Change handling of address spaces Christoph Lameter
2006-04-28 6:03 ` [PATCH 4/7] page migration: Drop nr_refs parameter Christoph Lameter
2006-04-28 7:30 ` KAMEZAWA Hiroyuki
2006-04-28 13:57 ` Christoph Lameter
2006-04-28 6:03 ` [PATCH 5/7] page migration: synchronize from and to lists Christoph Lameter
2006-04-28 7:46 ` KAMEZAWA Hiroyuki
2006-04-28 15:31 ` Christoph Lameter
2006-04-29 0:27 ` KAMEZAWA Hiroyuki
2006-04-29 0:33 ` Christoph Lameter
2006-04-29 0:44 ` KAMEZAWA Hiroyuki
2006-04-29 0:46 ` Christoph Lameter
2006-04-28 6:03 ` [PATCH 6/7] page migration: Extract try_to_unmap Christoph Lameter
2006-04-28 6:03 ` Christoph Lameter [this message]
2006-04-28 6:42 ` [PATCH 7/7] page migration: Add new fallback function Christoph Lameter
2006-04-28 22:08 ` [PATCH 1/7] page migration: Reorder functions in migrate.c Andrew Morton
2006-04-28 23:01 ` Christoph Lameter
2006-04-28 23:18 ` Andrew Morton
2006-04-29 0:14 ` Christoph Lameter
2006-04-29 0:36 ` Andrew Morton
2006-04-29 0:46 ` Christoph Lameter
2006-04-29 2:27 ` Andrew Morton
2006-04-29 2:48 ` Christoph Lameter
2006-04-29 0:54 ` Christoph Lameter
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=20060428060333.30257.43096.sendpatchset@schroedinger.engr.sgi.com \
--to=clameter@sgi.com \
--cc=akpm@osdl.org \
--cc=hugh@veritas.com \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=lee.schermerhorn@hp.com \
--cc=linux-mm@kvack.org \
/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