From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Howells <dhowells@redhat.com>
Subject: Re: [PATCH 3/17] afs: convert afs_dir_get_page to read_kmap_page
Date: Thu, 12 Apr 2007 20:43:11 +0100
Message-ID: <30839.1176406991@redhat.com>
References: <5c49b0ed0704121227h48c3271bp2511e082704d639f@mail.gmail.com>  <20070412024938.27380.44677.patchbomb.py@localhost> <1273.1176375482@redhat.com> <5c49b0ed0704121123x341d10afxfdf6930b4b26417b@mail.gmail.com> <29898.1176404243@redhat.com>
Cc: "Andrew Morton" <akpm@osdl.org>,
	"Alexander Viro" <viro@zeniv.linux.org.uk>,
	"Christoph Hellwig" <hch@infradead.org>,
	"Roman Zippel" <zippel@linux-m68k.org>,
	"Mikulas Patocka" <mikulas@artax.karlin.mff.cuni.cz>,
	"David Woodhouse" <dwmw2@infradead.org>,
	"Dave Kleikamp" <shaggy@austin.ibm.com>,
	"Anton Altaparmakov" <aia21@cantab.net>,
	"Evgeniy Dushistov" <dushistov@mail.ru>,
	linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	reiserfs-dev@namesys.com
To: "Nate Diller" <nate.diller@gmail.com>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from mx1.redhat.com ([66.187.233.31]:55662 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S932339AbXDLToG (ORCPT <rfc822;linux-fsdevel@vger.kernel.org>);
	Thu, 12 Apr 2007 15:44:06 -0400
In-Reply-To: <5c49b0ed0704121227h48c3271bp2511e082704d639f@mail.gmail.com> 
Sender: linux-fsdevel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

Nate Diller <nate.diller@gmail.com> wrote:

> but that's a lot of code to avoid a single stack allocation.  The
> whole fake file pointer thing still strikes me as a little ugly, and
> you're definitely not the first one who needed this sort of hackery.
> ugh

A better way might be to stick a void * in struct file and pass that through
to readpage() and readpages() instead of the struct file *.  That way, anyone
who wants the traditional arrangement can just point that extra void * at the
struct file.

Of course, I'm in favour of making it a struct key * like this:

	struct address_space_operations {
	...
		int (*readpage)(struct key *, struct page *);
	...
		int (*readpages)(struct key *, struct address_space *,
				struct list_head *, unsigned);
	...
	};

	struct file {
	...
		struct key *f_key;
	};

	struct page *filemap_nopage(struct vm_area_struct *area, ...)
	{
	...
		struct file *file = area->vm_file;
	...
		error = mapping->a_ops->readpage(file->f_key, page);
	...
	}
	
But I'm not sure the NFS crew, for instance, would be happy with that.  Maybe
passing file->private_data through would do.  That's basically what NFS and
FUSE, for instance, want, and it would also do for AFS.

David