linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Raghavendra D Prabhu <raghu.prabhu13@gmail.com>
To: Fengguang Wu <fengguang.wu@intel.com>
Cc: linux-mm@kvack.org, viro@zeniv.linux.org.uk, akpm@linux-foundation.org
Subject: Re:  Re: Re: [PATCH 2/5] mm/readahead: Change the condition for SetPageReadahead
Date: Tue, 16 Oct 2012 23:12:52 +0530	[thread overview]
Message-ID: <20121016174252.GB2826@Archie> (raw)
In-Reply-To: <20120928115623.GB1525@localhost>

[-- Attachment #1: Type: text/plain, Size: 2127 bytes --]

Hi,


* On Fri, Sep 28, 2012 at 07:56:23PM +0800, Fengguang Wu <fengguang.wu@intel.com> wrote:
>On Wed, Sep 26, 2012 at 06:59:00AM +0530, Raghavendra D Prabhu wrote:
>> Hi,
>>
>>
>> * On Sat, Sep 22, 2012 at 08:49:20PM +0800, Fengguang Wu <fengguang.wu@intel.com> wrote:
>> >On Sat, Sep 22, 2012 at 04:03:11PM +0530, raghu.prabhu13@gmail.com wrote:
>> >>From: Raghavendra D Prabhu <rprabhu@wnohang.net>
>> >>
>> >>If page lookup from radix_tree_lookup is successful and its index page_idx ==
>> >>nr_to_read - lookahead_size, then SetPageReadahead never gets called, so this
>> >>fixes that.
>> >
>> >NAK. Sorry. It's actually an intentional behavior, so that for the
>> >common cases of many cached files that are accessed frequently, no
>> >PG_readahead will be set at all to pointlessly trap into the readahead
>> >routines once and again.
>>
>> ACK, thanks for explaining that. However, regarding this, I would
>> like to know if the implications of the patch
>> 51daa88ebd8e0d437289f589af29d4b39379ea76 will still apply if
>> PG_readahead is not set.
>
>Would you elaborate the implication and the possible problematic case?

Certainly.


An implication of 51daa88ebd8e0d437289f589af29d4b39379ea76 is 
that, a redundant check for PageReadahead(page) is avoided since  
async is piggy-backed into the synchronous readahead itself.


So, in case of 

     page = find_get_page()
     if(!page)
         page_cache_sync_readahead()
     else if (PageReadahead(page))
         page_cache_async_readahead();

isnt' there a possibility that PG_readahead won't be set at all 
if page is not in cache (causing page_cache_sync_readahead) but 
page at index  (nr_to_read - lookahead_size) is already in the 
cache? (due to if (page) continue; in the code)?


Hence, I changed the condition from equality to >= for setting 
SetPageReadahead(page) (and added a 
variable so that it is done only once). 


>
>Thanks,
>Fengguang
>




Regards,
-- 
Raghavendra Prabhu
GPG Id : 0xD72BE977
Fingerprint: B93F EBCB 8E05 7039 CD3C A4B8 A616 DCA1 D72B E977
www: wnohang.net

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

  reply	other threads:[~2012-10-16 17:42 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-09-22 10:33 [PATCH 0/5] Readahead fixes / improvements raghu.prabhu13
     [not found] ` <cover.1348309711.git.rprabhu@wnohang.net>
2012-09-22 10:33   ` [PATCH 1/5] mm/readahead: Check return value of read_pages raghu.prabhu13
2012-09-22 12:43     ` Fengguang Wu
2012-09-26  1:25       ` Raghavendra D Prabhu
2012-09-28 11:54         ` Fengguang Wu
2012-10-16 17:47           ` Raghavendra D Prabhu
2012-10-17  2:53             ` Fengguang Wu
2012-09-22 10:33   ` [PATCH 2/5] mm/readahead: Change the condition for SetPageReadahead raghu.prabhu13
2012-09-22 12:49     ` Fengguang Wu
2012-09-26  1:29       ` Raghavendra D Prabhu
2012-09-28 11:56         ` Fengguang Wu
2012-10-16 17:42           ` Raghavendra D Prabhu [this message]
2012-10-17  2:34             ` Fengguang Wu
2012-09-22 10:33   ` [PATCH 3/5] Remove file_ra_state from arguments of count_history_pages raghu.prabhu13
2012-09-22 12:40     ` Fengguang Wu
2012-10-16 18:21       ` Raghavendra D Prabhu
2012-10-17  3:15         ` Fengguang Wu
2012-09-22 10:33   ` [PATCH 4/5] Move the check for ra_pages after VM_SequentialReadHint() raghu.prabhu13
2012-09-22 12:42     ` Fengguang Wu
2012-09-26  1:39       ` Raghavendra D Prabhu
2012-10-16 18:15       ` Raghavendra D Prabhu
2012-10-17  3:13         ` Fengguang Wu
2012-09-22 10:33   ` [PATCH 5/5] mm/readahead: Use find_get_pages instead of radix_tree_lookup raghu.prabhu13
2012-09-22 13:15     ` Fengguang Wu
2012-09-26  2:58       ` Raghavendra D Prabhu
2012-09-28 12:18         ` Fengguang Wu
2012-10-16 16:59           ` Raghavendra D Prabhu
2012-10-17  2:12             ` 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=20121016174252.GB2826@Archie \
    --to=raghu.prabhu13@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=fengguang.wu@intel.com \
    --cc=linux-mm@kvack.org \
    --cc=viro@zeniv.linux.org.uk \
    /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;
as well as URLs for NNTP newsgroup(s).