[Top] [All Lists]

Re: [PATCH 0/9] xfs: metadata CRCs, kernel, first batch

To: Ben Myers <bpm@xxxxxxx>
Subject: Re: [PATCH 0/9] xfs: metadata CRCs, kernel, first batch
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Sat, 23 Feb 2013 13:38:18 +1100
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20130222235029.GA30652@xxxxxxx>
References: <1358774760-21841-1-git-send-email-david@xxxxxxxxxxxxx> <20130222151931.GU22182@xxxxxxx> <20130222231217.GA26081@dastard> <20130222235029.GA30652@xxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Fri, Feb 22, 2013 at 05:50:29PM -0600, Ben Myers wrote:
> On Sat, Feb 23, 2013 at 10:12:17AM +1100, Dave Chinner wrote:
> > On Fri, Feb 22, 2013 at 09:19:31AM -0600, Ben Myers wrote:
> > This is mentioned in the patch zero description:
> > 
> > "This series makes is through to 001-092 in xfstests - there is a
> > problem in the dquot verifier that causes log recovery of dquot
> > buffers to follow a NULL pointer."
> >
> > Basically, mp->m_quotainfo is not initialised until after log
> > recovery occurs, so this has to be detected in the verify/crc
> > routines otherwise it goes splat like above. My current patch series
> > has this fixed.
> Cool.  Sorry for the extra noise.

No, that's fine. It tells me that you're actaully looking at the
code and seeing what it does ;)

I'm close to having a new version of the kernel patchset out. I've
just got to finish debugging the attribute changes I've made and
move the superblock support patch to the end of the series and I'll
post it out.

All I'm aiming for with the next version of the patch set is that
existing filesystems (i.e. no CRCs) are regression free. I've
actaully done very little CRC enabled testing while doing all the
directory and attribute code changes, mainly because I can't test
them properly until the userspace support is there. However, the
patch set up to the final patch (i.e. everything but the attribute
changes) seems to work just fine with the existing toolchain and

Put simply, my focus for testing the patch set is currently "no
regressions for existing users", not "CRCs work perfectly".


Dave Chinner

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