xfs
[Top] [All Lists]

Re: XFS: attempt to access beyond end of device

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: XFS: attempt to access beyond end of device
From: Amit Sahrawat <amit.sahrawat83@xxxxxxxxx>
Date: Wed, 21 Dec 2011 09:10:13 +0530
Cc: xfs@xxxxxxxxxxx, Christoph Hellwig <hch@xxxxxxxxxxxxx>
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; bh=tys+pctq9sh+M66z/7xZB1EGymTyx8oFgLPwtKxRUk8=; b=LJIclxUch2RTEe0uSQR45i/cAWqzAn+RNsJTp/f47lKdlP4Pq8mIUwncu8TPjnm1Ki 6DPA8qsOHunaZZoEW3H4rEsqcBLDdnDBjgoyFnUImt1Z5GO0FE0KwNIl/bJBRimyKvty fGuO6WSxvrxTpRPGDmhgvQMClES9vT7M+lrI0=
In-reply-to: <20111220220433.GZ23662@dastard>
References: <CADDb1s1cSM0vS3gnO369hwHwe5dJ0Afaf_6wNVDf0eT-jTvKCg@xxxxxxxxxxxxxx> <20111220220433.GZ23662@dastard>
On Wed, Dec 21, 2011 at 3:34 AM, Dave Chinner <david@xxxxxxxxxxxxx> wrote:
> On Tue, Dec 20, 2011 at 07:18:15PM +0530, Amit Sahrawat wrote:
>> Hi,
>>
>> While trying to open a file with a large number of extents in append
>> mode – it resulted in the below error logs continually.
>> I was trying to open the file in append mode, but it showed error, so
>> I checked for the extents – it has around 2238 extents. Although, it
>
> The first question you have to get to the bottom of is if there are
> only 2238 extents in the file, why is the xfs_bmap output showing
> more than that?
Actually the output got mismatched, 2238 was the number when I checked
the issue, while xfs_bmap results were for the last traces observed.
Sorry for the confusion.
>
>> is not repetitive, when I try on next mount and write – it is working
>> fine at times.
>> Also, the system does not seem to be corrupted. Since, running
>> xfs_repair after umount does not show any errors.
>
> Which implies memory corruption. Have you turned on memory poisoning
> and the memory leak detection to see if it detects any problems?
>
> What about running with CONFIG_XFS_DEBUG=y? Does that trip any
> problems sooner than the corruption is detected?
I will try with CONFIG_XFS_DEBUG.

Thanks & Regards,
Amit Sahrawat
>
> Cheers,
>
> Dave.
> --
> Dave Chinner
> david@xxxxxxxxxxxxx

<Prev in Thread] Current Thread [Next in Thread>