xfs
[Top] [All Lists]

Re: [PATCH] default to 64 bit inodes & add feature flag

To: Eric Sandeen <sandeen@xxxxxxxxxxx>
Subject: Re: [PATCH] default to 64 bit inodes & add feature flag
From: "Josef 'Jeff' Sipek" <jeffpc@xxxxxxxxxxxxxx>
Date: Thu, 8 Mar 2012 11:14:18 -0500
Cc: Ben Myers <bpm@xxxxxxx>, xfs-oss <xfs@xxxxxxxxxxx>
In-reply-to: <4F58D35D.7080504@xxxxxxxxxxx>
References: <4F5798F9.2050809@xxxxxxxxxx> <20120308013413.GL3592@dastard> <4F5813D8.8070305@xxxxxxxxxxx> <20120308153721.GS7762@xxxxxxx> <4F58D35D.7080504@xxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Thu, Mar 08, 2012 at 09:42:21AM -0600, Eric Sandeen wrote:
> So, after thinking about this (and talking on irc) some more, I'm
> not convinced that a feature flag is the way to go.
> 
> If we set a feature flag, suddenly old filesystems with 64-bit
> inodes will grow a new feature, and this will force a userspace
> upgrade - but there is no real new feature.  This seems like a bad
> idea.  My original patch (which Dave responded to with this one)
> simply made inode64 default, with no feature flags.
> 
> Unless someone has a really compelling argument for the flag,
> I'm thinking this is the wrong approach after all.
> 
> Perhaps I should resend the just-make-it-default patch.
> 
> Comments?

I was thinking about this sort of scenario.  You are right, there's no
on-disk format change.  My initial thought about how to handle this was to
just make inode64 the default on 64-bit builds.  I think the feature flag
idea is good because it essentially acts as a taint flag - much like the
attr2 feature flag.  The difference is, in the inode64 case...

1) it's the same on-disk format
2) there are years of ambiguous-inode filesystems out there

Out of curiosity...is there a reason we can't do both?  Default to 64-bit,
and slowly introduce the 64-bit inodes feature flag?

Jeff.

-- 
What is the difference between Mechanical Engineers and Civil Engineers?
Mechanical Engineers build weapons, Civil Engineers build targets.

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