xfs
[Top] [All Lists]

Re: [PATCH 1/2] 286: Test FITRIM where length is smaller than FSB

To: Lukas Czerner <lczerner@xxxxxxxxxx>
Subject: Re: [PATCH 1/2] 286: Test FITRIM where length is smaller than FSB
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Wed, 17 Oct 2012 07:16:45 +1100
Cc: xfs@xxxxxxxxxxx
In-reply-to: <1350380494-17482-1-git-send-email-lczerner@xxxxxxxxxx>
References: <1350380494-17482-1-git-send-email-lczerner@xxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Tue, Oct 16, 2012 at 11:41:33AM +0200, Lukas Czerner wrote:
> This tests corner case in FITRIM implementation where range size is
> smaller than file system block or zero. In this case FITRIM should fail
> with EINVAL.
> 
> The problem was spotted in xfs and ext4 where in case of length = 0 the
> 'end' variable underflowed. In case of length smaller than 1 FSB FITRIM
> finished successfully, but we really should rather return EINVAL in both
> cases.
> 
> (This patch has to be applied after 'Use upstream version of fstrim
> instead of the local one')
> 
> Signed-off-by: Lukas Czerner <lczerner@xxxxxxxxxx>
> ---
>  286     |   58 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  286.out |    6 ++++++

We've already got 286 (and 287), so this should probably be 288...

Otherwise, it looks good.

Reviewed-by: Dave Chinner <dchinner@xxxxxxxxxx>
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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