xfs
[Top] [All Lists]

Re: [REVIEW #3] bad_features2 support in user-space

To: Eric Sandeen <sandeen@xxxxxxxxxxx>
Subject: Re: [REVIEW #3] bad_features2 support in user-space
From: "Josef 'Jeff' Sipek" <jeffpc@xxxxxxxxxxxxxx>
Date: Fri, 29 Feb 2008 01:11:18 -0500
Cc: Barry Naujok <bnaujok@xxxxxxx>, "xfs@xxxxxxxxxxx" <xfs@xxxxxxxxxxx>
In-reply-to: <47C79BB8.4010005@sandeen.net>
References: <op.t69rt3bp3jf8g2@pc-bnaujok.melbourne.sgi.com> <47C79BB8.4010005@sandeen.net>
Sender: xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.5.16 (2007-06-11)
On Thu, Feb 28, 2008 at 11:44:24PM -0600, Eric Sandeen wrote:
> Barry Naujok wrote:
> > Ok, xfs_repair will leave bad_features2 in place just in case it is
> > being run with an older kernel that expects features2 in the bad
> > location. But, it will make sure the correct and bad features2 are
> > consistent if bad_features2 in non-zero.
> > 
> > If bad_features2 is zero, it is left alone (eg. new mkfs or new
> > kernel with fixes it during mount time).
> > 
> > This seems to be the best solution to the problem.
> 
> Will look at details later, but I agree w/ the general direction... I
> think maybe the kernel should go this way too?  (i.e. kernel maybe
> shouldn't be zeroing features2 either...

I think so.

> but this is probably more important for userspace than kernelspace; how
> often do you revert to an older kernel...)

All you need is another subsystem in Linux to be broken (*cough*ACPI*cough*)
and you'll want to go back to your previous kernel.

Josef 'Jeff' Sipek.

-- 
A computer without Microsoft is like chocolate cake without mustard.


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