From: Christoph Lameter <clameter@sgi.com>
To: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org
Cc: akpm@linux-foundation.org
Cc: dkegel@google.com
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: David Miller <davem@davemloft.net>
Cc: Nick Piggin <npiggin@suse.de>
Subject: [RFC 2/7] Move checks from pageout() to shrink_page_list
Date: Mon, 20 Aug 2007 14:50:42 -0700 [thread overview]
Message-ID: <20070820215316.287244006@sgi.com> (raw)
In-Reply-To: 20070820215040.937296148@sgi.com
[-- Attachment #1: move_checks_to_shrink_page_list --]
[-- Type: text/plain, Size: 4650 bytes --]
This is necessary because we soon will do other things than calling
pageout() from shrink_page_list().
Signed-off-by: Christoph Lameter <clameter@sgi.com>
---
mm/vmscan.c | 90 ++++++++++++++++++++++++++++++------------------------------
1 file changed, 45 insertions(+), 45 deletions(-)
Index: linux-2.6/mm/vmscan.c
===================================================================
--- linux-2.6.orig/mm/vmscan.c 2007-08-19 21:39:55.000000000 -0700
+++ linux-2.6/mm/vmscan.c 2007-08-19 21:47:56.000000000 -0700
@@ -273,8 +273,6 @@ static void handle_write_error(struct ad
/* possible outcome of pageout() */
typedef enum {
- /* failed to write page out, page is locked */
- PAGE_KEEP,
/* move page to the active list, page is locked */
PAGE_ACTIVATE,
/* page has been sent to the disk successfully, page is unlocked */
@@ -289,44 +287,6 @@ typedef enum {
*/
static pageout_t pageout(struct page *page, struct address_space *mapping)
{
- /*
- * If the page is dirty, only perform writeback if that write
- * will be non-blocking. To prevent this allocation from being
- * stalled by pagecache activity. But note that there may be
- * stalls if we need to run get_block(). We could test
- * PagePrivate for that.
- *
- * If this process is currently in generic_file_write() against
- * this page's queue, we can perform writeback even if that
- * will block.
- *
- * If the page is swapcache, write it back even if that would
- * block, for some throttling. This happens by accident, because
- * swap_backing_dev_info is bust: it doesn't reflect the
- * congestion state of the swapdevs. Easy to fix, if needed.
- * See swapfile.c:page_queue_congested().
- */
- if (!is_page_cache_freeable(page))
- return PAGE_KEEP;
- if (!mapping) {
- /*
- * Some data journaling orphaned pages can have
- * page->mapping == NULL while being dirty with clean buffers.
- */
- if (PagePrivate(page)) {
- if (try_to_free_buffers(page)) {
- ClearPageDirty(page);
- printk("%s: orphaned page\n", __FUNCTION__);
- return PAGE_CLEAN;
- }
- }
- return PAGE_KEEP;
- }
- if (mapping->a_ops->writepage == NULL)
- return PAGE_ACTIVATE;
- if (!may_write_to_queue(mapping->backing_dev_info))
- return PAGE_KEEP;
-
if (clear_page_dirty_for_io(page)) {
int res;
struct writeback_control wbc = {
@@ -504,18 +464,58 @@ static unsigned long shrink_page_list(st
if (!sc->may_writepage)
goto keep_locked;
+ /*
+ * If the page is dirty, only perform writeback if
+ * that write will be non-blocking. To prevent this
+ * allocation from being stalled by pagecache
+ * activity. But note that there may be stalls if
+ * we need to run get_block(). We could test
+ * PagePrivate for that.
+ *
+ * If this process is currently in
+ * generic_file_write() against this page's queue,
+ * we can perform writeback even if that will block.
+ *
+ * If the page is swapcache, write it back even if
+ * that would block, for some throttling. This happens
+ * by accident, because swap_backing_dev_info is bust:
+ * it doesn't reflect the congestion state of the
+ * swapdevs. Easy to fix, if needed.
+ * See swapfile.c:page_queue_congested().
+ */
+ if (!is_page_cache_freeable(page))
+ goto keep_locked;
+ if (!mapping) {
+ /*
+ * Some data journaling orphaned pages can
+ * have page->mapping == NULL while being
+ * dirty with clean buffers.
+ */
+ if (PagePrivate(page)) {
+ if (try_to_free_buffers(page)) {
+ ClearPageDirty(page);
+ printk("%s: orphaned page\n",
+ __FUNCTION__);
+ goto release_page;
+ }
+ }
+ goto keep_locked;
+ }
+ if (mapping->a_ops->writepage == NULL)
+ goto activate_locked;
+ if (!may_write_to_queue(mapping->backing_dev_info))
+ goto keep_locked;
+
/* Page is dirty, try to write it out here */
switch(pageout(page, mapping)) {
- case PAGE_KEEP:
- goto keep_locked;
case PAGE_ACTIVATE:
goto activate_locked;
case PAGE_SUCCESS:
if (PageWriteback(page) || PageDirty(page))
goto keep;
/*
- * A synchronous write - probably a ramdisk. Go
- * ahead and try to reclaim the page.
+ * A synchronous write - probably a ramdisk.
+ * Go ahead and try to reclaim the page.
*/
if (TestSetPageLocked(page))
goto keep;
@@ -526,7 +526,7 @@ static unsigned long shrink_page_list(st
; /* try to free the page below */
}
}
-
+release_page:
/*
* If the page has buffers, try to free the buffer mappings
* associated with this page. If we succeed we try to free
--
WARNING: multiple messages have this Message-ID (diff)
From: Christoph Lameter <clameter@sgi.com>
To: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org, akpm@linux-foundation.org,
dkegel@google.com, Peter Zijlstra <a.p.zijlstra@chello.nl>,
David Miller <davem@davemloft.net>, Nick Piggin <npiggin@suse.de>
Subject: [RFC 2/7] Move checks from pageout() to shrink_page_list
Date: Mon, 20 Aug 2007 14:50:42 -0700 [thread overview]
Message-ID: <20070820215316.287244006@sgi.com> (raw)
In-Reply-To: 20070820215040.937296148@sgi.com
[-- Attachment #1: move_checks_to_shrink_page_list --]
[-- Type: text/plain, Size: 4876 bytes --]
This is necessary because we soon will do other things than calling
pageout() from shrink_page_list().
Signed-off-by: Christoph Lameter <clameter@sgi.com>
---
mm/vmscan.c | 90 ++++++++++++++++++++++++++++++------------------------------
1 file changed, 45 insertions(+), 45 deletions(-)
Index: linux-2.6/mm/vmscan.c
===================================================================
--- linux-2.6.orig/mm/vmscan.c 2007-08-19 21:39:55.000000000 -0700
+++ linux-2.6/mm/vmscan.c 2007-08-19 21:47:56.000000000 -0700
@@ -273,8 +273,6 @@ static void handle_write_error(struct ad
/* possible outcome of pageout() */
typedef enum {
- /* failed to write page out, page is locked */
- PAGE_KEEP,
/* move page to the active list, page is locked */
PAGE_ACTIVATE,
/* page has been sent to the disk successfully, page is unlocked */
@@ -289,44 +287,6 @@ typedef enum {
*/
static pageout_t pageout(struct page *page, struct address_space *mapping)
{
- /*
- * If the page is dirty, only perform writeback if that write
- * will be non-blocking. To prevent this allocation from being
- * stalled by pagecache activity. But note that there may be
- * stalls if we need to run get_block(). We could test
- * PagePrivate for that.
- *
- * If this process is currently in generic_file_write() against
- * this page's queue, we can perform writeback even if that
- * will block.
- *
- * If the page is swapcache, write it back even if that would
- * block, for some throttling. This happens by accident, because
- * swap_backing_dev_info is bust: it doesn't reflect the
- * congestion state of the swapdevs. Easy to fix, if needed.
- * See swapfile.c:page_queue_congested().
- */
- if (!is_page_cache_freeable(page))
- return PAGE_KEEP;
- if (!mapping) {
- /*
- * Some data journaling orphaned pages can have
- * page->mapping == NULL while being dirty with clean buffers.
- */
- if (PagePrivate(page)) {
- if (try_to_free_buffers(page)) {
- ClearPageDirty(page);
- printk("%s: orphaned page\n", __FUNCTION__);
- return PAGE_CLEAN;
- }
- }
- return PAGE_KEEP;
- }
- if (mapping->a_ops->writepage == NULL)
- return PAGE_ACTIVATE;
- if (!may_write_to_queue(mapping->backing_dev_info))
- return PAGE_KEEP;
-
if (clear_page_dirty_for_io(page)) {
int res;
struct writeback_control wbc = {
@@ -504,18 +464,58 @@ static unsigned long shrink_page_list(st
if (!sc->may_writepage)
goto keep_locked;
+ /*
+ * If the page is dirty, only perform writeback if
+ * that write will be non-blocking. To prevent this
+ * allocation from being stalled by pagecache
+ * activity. But note that there may be stalls if
+ * we need to run get_block(). We could test
+ * PagePrivate for that.
+ *
+ * If this process is currently in
+ * generic_file_write() against this page's queue,
+ * we can perform writeback even if that will block.
+ *
+ * If the page is swapcache, write it back even if
+ * that would block, for some throttling. This happens
+ * by accident, because swap_backing_dev_info is bust:
+ * it doesn't reflect the congestion state of the
+ * swapdevs. Easy to fix, if needed.
+ * See swapfile.c:page_queue_congested().
+ */
+ if (!is_page_cache_freeable(page))
+ goto keep_locked;
+ if (!mapping) {
+ /*
+ * Some data journaling orphaned pages can
+ * have page->mapping == NULL while being
+ * dirty with clean buffers.
+ */
+ if (PagePrivate(page)) {
+ if (try_to_free_buffers(page)) {
+ ClearPageDirty(page);
+ printk("%s: orphaned page\n",
+ __FUNCTION__);
+ goto release_page;
+ }
+ }
+ goto keep_locked;
+ }
+ if (mapping->a_ops->writepage == NULL)
+ goto activate_locked;
+ if (!may_write_to_queue(mapping->backing_dev_info))
+ goto keep_locked;
+
/* Page is dirty, try to write it out here */
switch(pageout(page, mapping)) {
- case PAGE_KEEP:
- goto keep_locked;
case PAGE_ACTIVATE:
goto activate_locked;
case PAGE_SUCCESS:
if (PageWriteback(page) || PageDirty(page))
goto keep;
/*
- * A synchronous write - probably a ramdisk. Go
- * ahead and try to reclaim the page.
+ * A synchronous write - probably a ramdisk.
+ * Go ahead and try to reclaim the page.
*/
if (TestSetPageLocked(page))
goto keep;
@@ -526,7 +526,7 @@ static unsigned long shrink_page_list(st
; /* try to free the page below */
}
}
-
+release_page:
/*
* If the page has buffers, try to free the buffer mappings
* associated with this page. If we succeed we try to free
--
--
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:[~2007-08-20 21:54 UTC|newest]
Thread overview: 91+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-08-20 21:50 [RFC 0/7] Postphone reclaim laundry to write at high water marks Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-20 21:50 ` [RFC 1/7] release_lru_pages(): Generic release of pages to the LRU Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-21 14:52 ` Mel Gorman
2007-08-21 14:52 ` Mel Gorman
2007-08-21 20:51 ` Christoph Lameter
2007-08-21 20:51 ` Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter [this message]
2007-08-20 21:50 ` [RFC 2/7] Move checks from pageout() to shrink_page_list Christoph Lameter
2007-08-20 21:50 ` [RFC 3/7] shrink_page_list: Support isolating dirty pages on laundry list Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-21 15:04 ` Mel Gorman
2007-08-21 15:04 ` Mel Gorman
2007-08-21 20:53 ` Christoph Lameter
2007-08-21 20:53 ` Christoph Lameter
2007-08-20 21:50 ` [RFC 4/7] Pass laundry through shrink_inactive_list() and shrink_zone() Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-20 21:50 ` [RFC 5/7] Laundry handling for direct reclaim Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-21 15:06 ` Mel Gorman
2007-08-21 15:06 ` Mel Gorman
2007-08-21 20:55 ` Christoph Lameter
2007-08-21 20:55 ` Christoph Lameter
2007-08-21 15:19 ` Mel Gorman
2007-08-21 15:19 ` Mel Gorman
2007-08-21 21:00 ` Christoph Lameter
2007-08-21 21:00 ` Christoph Lameter
2007-08-20 21:50 ` [RFC 6/7] kswapd: Do laundry after reclaim Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-20 21:50 ` [RFC 7/7] Switch of PF_MEMALLOC during writeout Christoph Lameter
2007-08-20 21:50 ` Christoph Lameter
2007-08-20 23:08 ` Andi Kleen
2007-08-20 23:08 ` Andi Kleen
2007-08-20 23:19 ` Christoph Lameter
2007-08-20 23:19 ` Christoph Lameter
2007-08-21 1:13 ` Andi Kleen
2007-08-21 1:13 ` Andi Kleen
2007-08-21 10:36 ` [RFC 0/7] Postphone reclaim laundry to write at high water marks Peter Zijlstra
2007-08-21 10:36 ` Peter Zijlstra
2007-08-21 20:48 ` Christoph Lameter
2007-08-21 20:48 ` Christoph Lameter
2007-08-21 21:13 ` Peter Zijlstra
2007-08-21 21:13 ` Peter Zijlstra
2007-08-21 21:29 ` Christoph Lameter
2007-08-21 21:29 ` Christoph Lameter
2007-08-21 21:43 ` Rik van Riel
2007-08-21 21:43 ` Rik van Riel
2007-08-21 22:32 ` Christoph Lameter
2007-08-21 22:32 ` Christoph Lameter
2007-08-23 12:05 ` Andrea Arcangeli
2007-08-23 12:05 ` Andrea Arcangeli
2007-08-23 20:23 ` Christoph Lameter
2007-08-23 20:23 ` Christoph Lameter
2007-08-21 22:09 ` Peter Zijlstra
2007-08-21 22:09 ` Peter Zijlstra
2007-08-21 22:43 ` Christoph Lameter
2007-08-21 22:43 ` Christoph Lameter
2007-08-22 7:02 ` Peter Zijlstra
2007-08-22 7:02 ` Peter Zijlstra
2007-08-22 19:04 ` Christoph Lameter
2007-08-22 19:04 ` Christoph Lameter
2007-08-22 20:03 ` Peter Zijlstra
2007-08-22 20:03 ` Peter Zijlstra
2007-08-22 20:16 ` Christoph Lameter
2007-08-22 20:16 ` Christoph Lameter
2007-08-23 7:39 ` Peter Zijlstra
2007-08-23 7:39 ` Peter Zijlstra
2007-08-26 4:52 ` Rik van Riel
2007-08-26 4:52 ` Rik van Riel
2007-08-23 12:16 ` Andrea Arcangeli
2007-08-23 12:16 ` Andrea Arcangeli
2007-08-22 7:45 ` Ingo Molnar
2007-08-22 7:45 ` Ingo Molnar
2007-08-22 19:19 ` Christoph Lameter
2007-08-22 19:19 ` Christoph Lameter
2007-08-23 12:08 ` Andrea Arcangeli
2007-08-23 12:08 ` Andrea Arcangeli
2007-08-23 12:59 ` Peter Zijlstra
2007-08-23 12:59 ` Peter Zijlstra
2007-08-21 15:16 ` Rik van Riel
2007-08-21 15:16 ` Rik van Riel
2007-08-21 20:59 ` Christoph Lameter
2007-08-21 20:59 ` Christoph Lameter
2007-08-21 21:14 ` Rik van Riel
2007-08-21 21:14 ` Rik van Riel
2007-08-21 21:30 ` Christoph Lameter
2007-08-21 21:30 ` Christoph Lameter
2007-08-21 15:51 ` Dave McCracken
2007-08-21 15:51 ` Dave McCracken
2007-08-21 21:03 ` 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=20070820215316.287244006@sgi.com \
--to=clameter@sgi.com \
--cc=linux-kernel@vger.kernel.org \
--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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.