Message-ID: <406361F2.6060308@stesmi.com>
Date: Thu, 25 Mar 2004 23:49:22 +0100
From: Stefan Smietanowski <stesmi@stesmi.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7b) Gecko/20040316
MIME-Version: 1.0
To: Steve Lord <lord@xfs.org>
CC: Christoph Hellwig <hch@infradead.org>, Chris Wedgwood <cw@f00f.org>,
        "IKARASHI, Seiichi" <ikarashi@miraclelinux.com>, linux-xfs@oss.sgi.com
Subject: Re: synchronization of XFS
References: <4060F7FC.8090602@miraclelinux.com> <20040325063902.GA9697@dingdong.cryptoapps.com> <4062C97A.6030702@miraclelinux.com> <20040325124152.GA12078@dingdong.cryptoapps.com> <4062D7E5.6070501@stesmi.com> <20040325132200.GA12333@dingdong.cryptoapps.com> <4062E19A.90207@xfs.org> <20040325140723.GA12558@dingdong.cryptoapps.com> <20040325144519.A23764@infradead.org> <40635F04.6010109@xfs.org> <40636032.3000402@stesmi.com> <4063612E.4030109@xfs.org>
In-Reply-To: <4063612E.4030109@xfs.org>
Content-type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 8bit
Sender: linux-xfs-bounce@oss.sgi.com
Errors-to: linux-xfs-bounce@oss.sgi.com
Precedence: bulk
Content-Length: 1357
Lines: 37

Hi Steve.

>>>>>> Now if grub is opening the block device and reading out of that, it
>>>>>> is looking at the same pages for metadata that xfs is looking at in
>>>>>> memory. There is a bug where you can get corruption if you access
>>>>>> the block device in parallel with the filesystem. Possibly this is
>>>>>> behind the problem.
>>>>>
>>>>> This will cause an oops on 2.6.x won't it --- so I suspect if this is
>>>>> behind the problem the report will be have been different.
>>>>
>>>> I don't think they're hitting the problem, the symptoms look very 
>>>> different.
>>>
>>> And thinking about it some more, having grub make the filesystem 
>>> remount readonly would force everything down to disk unlike just 
>>> doing a sync
>>> call.
>>
>> Tried and failed :(
>>
>> Tried that before but it didn't help unfortunately.
> 
> So what exactly is the sequence of events here, some files are created
> in the boot directory via the kernel, then grub wants to look at them
> via the block device api and its emulation of the filesystem? If we
> can nail this down, then maybe we can really work out what is going
> on here.
> 
> Steve

Yup. That's what's happening. It first does one run with --just-copy
where it writes the files using the filesystem then reads the same
files using the blockdevice and it's own filesystem code basically.

// Stefan