From mboxrd@z Thu Jan  1 00:00:00 1970
From: Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH 12/14] quota: Split off quota tree handling into a
 separate file
Date: Mon, 27 Oct 2008 22:52:42 -0700
Message-ID: <20081027225242.cad4ee18.akpm@linux-foundation.org>
References: <1225109512852-git-send-email-jack@suse.cz>
	<12251095123763-git-send-email-jack@suse.cz>
	<12251095124166-git-send-email-jack@suse.cz>
	<1225109512758-git-send-email-jack@suse.cz>
	<12251095122803-git-send-email-jack@suse.cz>
	<12251095124165-git-send-email-jack@suse.cz>
	<1225109513735-git-send-email-jack@suse.cz>
	<12251095131095-git-send-email-jack@suse.cz>
	<12251095133646-git-send-email-jack@suse.cz>
	<12251095133281-git-send-email-jack@suse.cz>
	<12251095131431-git-send-email-jack@suse.cz>
	<1225109513299-git-send-email-jack@suse.cz>
	<12251095133700-git-send-email-jack@suse.cz>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Cc: linux-fsdevel@vger.kernel.org
To: Jan Kara <jack@suse.cz>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from smtp1.linux-foundation.org ([140.211.169.13]:48662 "EHLO
	smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK)
	by vger.kernel.org with ESMTP id S1752124AbYJ1FxS (ORCPT
	<rfc822;linux-fsdevel@vger.kernel.org>);
	Tue, 28 Oct 2008 01:53:18 -0400
In-Reply-To: <12251095133700-git-send-email-jack@suse.cz>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Mon, 27 Oct 2008 13:11:50 +0100 Jan Kara <jack@suse.cz> wrote:

> There is going to be a new version of quota format having 64-bit
> quota limits and a new quota format for OCFS2. They are both
> going to use the same tree structure as VFSv0 quota format. So
> split out tree handling into a separate file and make size of
> leaf blocks, amount of space usable in each block (needed for
> checksumming) and structures contained in them configurable
> so that the code can be shared.
> 
>
> ...
>
> +
> +typedef char *dqbuf_t;

This thing is fairly unpleasant.  It would have been better to have
open-coded `char *' everywhere - the typedef just obscures what's going
on.  Or void*.

Also, in most cases the return value of getdqbuf() is cast to a `struct
qt_disk_dqdbheader *', so why not make getdqbuf() return a pointer of
that type?

Anyway, that's all old code and this patch just moves it around. 
Please sometime take a look at whether this code is as tip-top,
super-duper as we'd like?

> +static inline ssize_t read_blk(struct qtree_mem_dqinfo *info, uint blk, dqbuf_t buf)
> +{
> +	struct super_block *sb = info->dqi_sb;
> +
> +	memset(buf, 0, info->dqi_usable_bs);
> +	return sb->s_op->quota_read(sb, info->dqi_type, (char *)buf,
> +	       info->dqi_usable_bs, blk << info->dqi_blocksize_bits);
> +}
> +
> +static inline ssize_t write_blk(struct qtree_mem_dqinfo *info, uint blk, dqbuf_t buf)
> +{
> +	struct super_block *sb = info->dqi_sb;
> +
> +	return sb->s_op->quota_write(sb, info->dqi_type, (char *)buf,
> +	       info->dqi_usable_bs, blk << info->dqi_blocksize_bits);
> +}

I expect you'll get some nice code-size and cache-footprint
improvements by uninlining these.