All of lore.kernel.org
 help / color / mirror / Atom feed
From: Fengguang Wu <wfg@mail.ustc.edu.cn>
To: Andrew Morton <akpm@osdl.org>
Cc: linux-kernel@vger.kernel.org
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Jens Axboe <jens.axboe@oracle.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Steven Pratt <slpratt@austin.ibm.com>
Cc: Ram Pai <linuxram@us.ibm.com>
Subject: [PATCH 6/9] readahead: convert filemap invocations
Date: Thu, 17 May 2007 06:47:58 +0800	[thread overview]
Message-ID: <379355696.29993@ustc.edu.cn> (raw)
Message-ID: <20070516224818.963553696@mail.ustc.edu.cn> (raw)
In-Reply-To: 20070516224752.500812933@mail.ustc.edu.cn

[-- Attachment #1: readahead-invocations-filemap.patch --]
[-- Type: text/plain, Size: 2767 bytes --]

Convert filemap reads to use on-demand readahead.

The new call scheme is to
- call readahead on non-cached page
- call readahead on look-ahead page
- update prev_index when finished with the read request

Signed-off-by: Fengguang Wu <wfg@mail.ustc.edu.cn>
---
 mm/filemap.c |   51 +++++++++++++++++++++++++++++--------------------
 1 file changed, 31 insertions(+), 20 deletions(-)

--- linux-2.6.22-rc1-mm1.orig/mm/filemap.c
+++ linux-2.6.22-rc1-mm1/mm/filemap.c
@@ -916,15 +916,20 @@ void do_generic_mapping_read(struct addr
 		nr = nr - offset;
 
 		cond_resched();
-		if (index == next_index)
-			next_index = page_cache_readahead(mapping, &ra, filp,
-					index, last_index - index);
-
 find_page:
 		page = find_get_page(mapping, index);
-		if (unlikely(page == NULL)) {
-			handle_ra_miss(mapping, &ra, index);
-			goto no_cached_page;
+		if (!page) {
+			page_cache_readahead_ondemand(mapping,
+					&ra, filp, page,
+					index, last_index - index);
+			page = find_get_page(mapping, index);
+			if (unlikely(page == NULL))
+				goto no_cached_page;
+		}
+		if (PageReadahead(page)) {
+			page_cache_readahead_ondemand(mapping,
+					&ra, filp, page,
+					index, last_index - index);
 		}
 		if (!PageUptodate(page))
 			goto page_not_up_to_date;
@@ -1075,6 +1080,7 @@ no_cached_page:
 
 out:
 	*_ra = ra;
+	_ra->prev_index = prev_index;
 
 	*ppos = ((loff_t) index << PAGE_CACHE_SHIFT) + offset;
 	if (cached_page)
@@ -1378,26 +1384,30 @@ struct page *filemap_fault(struct vm_are
 		goto no_cached_page;
 
 	/*
-	 * The readahead code wants to be told about each and every page
-	 * so it can build and shrink its windows appropriately
-	 *
-	 * For sequential accesses, we use the generic readahead logic.
-	 */
-	if (VM_SequentialReadHint(vma))
-		page_cache_readahead(mapping, ra, file, fdata->pgoff, 1);
-
-	/*
 	 * Do we have something in the page cache already?
 	 */
 retry_find:
 	page = find_lock_page(mapping, fdata->pgoff);
+	/*
+	 * For sequential accesses, we use the generic readahead logic.
+	 */
+	if (VM_SequentialReadHint(vma)) {
+		if (!page) {
+			page_cache_readahead_ondemand(mapping, ra, file, page,
+							   fdata->pgoff, 1);
+			page = find_lock_page(mapping, fdata->pgoff);
+			if (!page)
+				goto no_cached_page;
+		}
+		if (PageReadahead(page)) {
+			page_cache_readahead_ondemand(mapping, ra, file, page,
+							   fdata->pgoff, 1);
+		}
+	}
+
 	if (!page) {
 		unsigned long ra_pages;
 
-		if (VM_SequentialReadHint(vma)) {
-			handle_ra_miss(mapping, ra, fdata->pgoff);
-			goto no_cached_page;
-		}
 		ra->mmap_miss++;
 
 		/*
@@ -1450,6 +1460,7 @@ retry_find:
 	 * Found the page and have a reference on it.
 	 */
 	mark_page_accessed(page);
+	ra->prev_index = page->index;
 	return page;
 
 outside_data_content:

--

  parent reply	other threads:[~2007-05-16 22:48 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-05-16 22:47 [PATCH 0/9] on-demand readahead Fengguang Wu
2007-05-16 22:47 ` Fengguang Wu
2007-05-16 22:47 ` [PATCH 1/9] readahead: introduce PG_readahead Fengguang Wu
2007-05-16 22:47   ` Fengguang Wu
2007-05-19  6:28     ` Andrew Morton
2007-05-19 11:35       ` Andi Kleen
2007-05-19 15:19         ` Andrew Morton
2007-05-19 12:30       ` Fengguang Wu
2007-05-19 12:30         ` Fengguang Wu
2007-05-19 15:25           ` Andrew Morton
2007-05-20  3:09             ` Fengguang Wu
2007-05-20  3:09               ` Fengguang Wu
2007-05-20  7:10                 ` Christoph Lameter
2007-06-12  1:04   ` Rusty Russell
2007-06-12  2:52     ` Fengguang Wu
2007-06-12  2:52       ` Fengguang Wu
2007-05-16 22:47 ` [PATCH 2/9] readahead: add look-ahead support to __do_page_cache_readahead() Fengguang Wu
2007-05-16 22:47   ` Fengguang Wu
2007-05-16 22:47 ` [PATCH 3/9] readahead: MIN_RA_PAGES/MAX_RA_PAGES macros Fengguang Wu
2007-05-16 22:47   ` Fengguang Wu
2007-05-16 22:47 ` [PATCH 4/9] readahead: data structure and routines Fengguang Wu
2007-05-16 22:47   ` Fengguang Wu
2007-06-12  3:30   ` Rusty Russell
2007-06-12 12:07     ` Fengguang Wu
2007-06-12 12:07       ` Fengguang Wu
2007-06-13  0:27       ` Rusty Russell
2007-06-13  3:07         ` Fengguang Wu
2007-06-13  3:07           ` Fengguang Wu
2007-05-16 22:47 ` [PATCH 5/9] readahead: on-demand readahead logic Fengguang Wu
2007-05-16 22:47   ` Fengguang Wu
2007-05-19  6:23     ` Andrew Morton
2007-05-19 13:02       ` Fengguang Wu
2007-05-19 13:02         ` Fengguang Wu
2007-06-12  4:36   ` Rusty Russell
2007-06-12 10:35     ` Fengguang Wu
2007-06-12 10:35       ` Fengguang Wu
2007-06-13  1:40       ` Rusty Russell
2007-06-13  4:00         ` Fengguang Wu
2007-06-13  4:00           ` Fengguang Wu
2007-06-13  5:51           ` Rusty Russell
2007-06-13  7:07             ` Fengguang Wu
2007-06-13  7:07               ` Fengguang Wu
2007-05-16 22:47 ` Fengguang Wu [this message]
2007-05-16 22:47   ` [PATCH 6/9] readahead: convert filemap invocations Fengguang Wu
2007-05-16 22:47 ` [PATCH 7/9] readahead: convert splice invocations Fengguang Wu
2007-05-16 22:47   ` Fengguang Wu
2007-05-16 22:48 ` [PATCH 8/9] readahead: convert ext3/ext4 invocations Fengguang Wu
2007-05-16 22:48   ` Fengguang Wu
2007-05-19 12:19   ` Andi Kleen
2007-05-16 22:48 ` [PATCH 9/9] readahead: remove the old algorithm Fengguang Wu
2007-05-16 22:48   ` Fengguang Wu
2007-05-19 12:18   ` Andi Kleen
2007-05-19 13:17     ` Fengguang Wu
2007-05-19 13:17       ` Fengguang Wu

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=379355696.29993@ustc.edu.cn \
    --to=wfg@mail.ustc.edu.cn \
    --cc=akpm@osdl.org \
    --cc=linux-kernel@vger.kernel.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.