From: Matthew Wilcox <willy@debian.org>
To: Alex Tomas <bzzz@tmi.comex.ru>
Cc: Matthew Wilcox <willy@debian.org>,
Andreas Dilger <adilger@clusterfs.com>,
linux-kernel <linux-kernel@vger.kernel.org>,
ext2-devel@lists.sourceforge.net, Andrew Morton <akpm@digeo.com>
Subject: Re: [Ext2-devel] [PATCH] distributed counters for ext2 to avoid group scaning
Date: Mon, 17 Mar 2003 15:27:19 +0000 [thread overview]
Message-ID: <20030317152719.GD28607@parcelfarce.linux.theplanet.co.uk> (raw)
In-Reply-To: <m3ptoqjagt.fsf@lexa.home.net>
On Mon, Mar 17, 2003 at 06:09:54PM +0300, Alex Tomas wrote:
> >>>>> Matthew Wilcox (MW) writes:
> MW> And then we have 3 of these (an additional 3k..). Per
> MW> blockgroup. My 4GB / has 30 blockgroups; my 30GB /home has 232.
> MW> So that's a little under 8 per GB. My _laptop_ has a 40GB drive,
> MW> so that's on the order of 320 blockgroups -- almost an additional
> MW> megabyte of ram consumed for these counters.
>
> no-no!
>
> _one_ dcounter to maintain number of free blocks _per_ fs.
> _one_ dcounter to maintain number of inode blocks _per_ fs.
> _one_ dcounter to maintain number of dirs _per_ fs.
>
> 3 dcounter per fs. no more.
Gah. That's your fault. Use diff -p in future; I saw:
diff -uNr linux/include/linux/ext2_fs_sb.h edited/include/linux/ext2_fs_sb.h
--- linux/include/linux/ext2_fs_sb.h Sun Mar 16 17:21:34 2003
+++ edited/include/linux/ext2_fs_sb.h Mon Mar 17 00:12:00 2003
@@ -16,6 +16,8 @@
#ifndef _LINUX_EXT2_FS_SB
#define _LINUX_EXT2_FS_SB
+#include <linux/dcounter.h>
+
struct ext2_bg_info {
u8 debts;
spinlock_t balloc_lock;
@@ -52,6 +54,9 @@
u32 s_next_generation;
unsigned long s_dir_count;
struct ext2_bg_info *s_bgi;
+ struct dcounter free_blocks_dc;
+ struct dcounter free_inodes_dc;
+ struct dcounter dirs_dc;
};
which makes it look like the dcounters are added to ext2_bg_info.
diff -p would have put the name of the struct after the @@ line. Not to
mention you didn't follow the `s_' prefix style used everywhere else in
that struct.
Anyway, I think dcounters should probably be allocated from kmalloc_percpu()
rather than as part of the dcounter struct.
--
"It's not Hollywood. War is real, war is primarily not about defeat or
victory, it is about death. I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk
next prev parent reply other threads:[~2003-03-17 15:16 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-03-16 15:01 Alex Tomas
2003-03-16 17:44 ` [Ext2-devel] " Andreas Dilger
2003-03-16 21:55 ` Alex Tomas
2003-03-17 15:11 ` Matthew Wilcox
2003-03-17 15:09 ` Alex Tomas
2003-03-17 15:27 ` Matthew Wilcox [this message]
2003-03-17 15:25 ` Alex Tomas
2003-03-17 20:23 ` Andrew Morton
2003-03-17 20:27 ` Matthew Wilcox
2003-03-17 20:40 ` Alex Tomas
2003-03-17 9:37 ` William Lee Irwin III
2003-03-17 9:48 ` William Lee Irwin III
2003-03-17 10:41 ` Alex Tomas
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=20030317152719.GD28607@parcelfarce.linux.theplanet.co.uk \
--to=willy@debian.org \
--cc=adilger@clusterfs.com \
--cc=akpm@digeo.com \
--cc=bzzz@tmi.comex.ru \
--cc=ext2-devel@lists.sourceforge.net \
--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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®