From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+w=401wt.eu-S1755425AbZBCWCE@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1755425AbZBCWCE (ORCPT <rfc822;w@1wt.eu>);
	Tue, 3 Feb 2009 17:02:04 -0500
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751932AbZBCWBv
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Tue, 3 Feb 2009 17:01:51 -0500
Received: from smtp1.linux-foundation.org ([140.211.169.13]:41057 "EHLO
	smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK)
	by vger.kernel.org with ESMTP id S1751049AbZBCWBu (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Tue, 3 Feb 2009 17:01:50 -0500
Date: Tue, 3 Feb 2009 14:01:35 -0800
From: Andrew Morton <akpm@linux-foundation.org>
To: Eric Sandeen <sandeen@redhat.com>
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
       t-sato@yk.jp.nec.com
Subject: Re: [PATCH V4] Allow SysRq emergency thaw to thaw frozen
 filesystems
Message-Id: <20090203140135.fe8360f6.akpm@linux-foundation.org>
In-Reply-To: <498779D6.7010304@redhat.com>
References: <496EB639.6090800@redhat.com>
	<4970E087.8020308@redhat.com>
	<4970E4FB.7040407@redhat.com>
	<20090130134009.02d4a576.akpm@linux-foundation.org>
	<498779D6.7010304@redhat.com>
X-Mailer: Sylpheed version 2.2.4 (GTK+ 2.8.20; i486-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Mon, 02 Feb 2009 16:55:18 -0600
Eric Sandeen <sandeen@redhat.com> wrote:

> (this version fixes a couple small issues raised by Randy Dunlap)
> (and adds sb_lock locking I forgot, as akpm pointed out)
> (Randy's CONFIG_BLOCK fix should still apply over this)

<converts it to a delta so I can see what you did>

--- a/fs/buffer.c~allow-sysrq-emergency-thaw-to-thaw-frozen-filesystems-v4
+++ a/fs/buffer.c
@@ -263,11 +263,13 @@ void do_thaw_all(unsigned long unused)
 	struct super_block *sb;
 	char b[BDEVNAME_SIZE];
 
+	spin_lock(&sb_lock);
 	list_for_each_entry(sb, &super_blocks, s_list) {
 		while (sb->s_bdev && !thaw_bdev(sb->s_bdev, sb))
 			printk(KERN_WARNING "Emergency Thaw on %s\n",
 			       bdevname(sb->s_bdev, b));
 	}
+	spin_unlock(&sb_lock);
 	printk(KERN_WARNING "Emergency Thaw complete\n");
 }
 
_


Can't call thaw_bdev() under spinlock.


If we're going to do this, I think it will need the whole
sb_lock/s_count/s_umount song-n-dance.

It's a pretty common operation.  What you want is, I think, identical
to sync_supers(), only with one line changed.

so we could do

void apply_to_all_supers(void (fn)(struct super_block *))
{
	struct super_block *sb;

	spin_lock(&sb_lock);
restart:
	list_for_each_entry(sb, &super_blocks, s_list) {
		if (sb->s_dirt) {
			sb->s_count++;
			spin_unlock(&sb_lock);
			down_read(&sb->s_umount);
			(*fn)(sb);
			up_read(&sb->s_umount);
			spin_lock(&sb_lock);
			if (__put_super_and_need_restart(sb))
				goto restart;
		}
	}
	spin_unlock(&sb_lock);

}

That isn't quite sufficient to use for get_super(), but I think it
could be made so.

Ditto user_get_super().

Ditto do_emergency_remount()

But that's a separate little project for someone.  For the purposes of
this patch I guess you could do yet another copy-n-paste.