xfs
[Top] [All Lists]

Re: [PATCH 1/2] xfs_quota: allow operation on foreign filesystem types

To: xfs@xxxxxxxxxxx
Subject: Re: [PATCH 1/2] xfs_quota: allow operation on foreign filesystem types
From: Eric Sandeen <sandeen@xxxxxxxxxxx>
Date: Thu, 4 Feb 2016 23:27:39 -0600
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <56B3E6F3.802@xxxxxxxxxxx>
References: <1454627718-19583-1-git-send-email-david@xxxxxxxxxxxxx> <1454627718-19583-2-git-send-email-david@xxxxxxxxxxxxx> <56B3E6F3.802@xxxxxxxxxxx>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.11; rv:38.0) Gecko/20100101 Thunderbird/38.5.1

On 2/4/16 6:04 PM, Eric Sandeen wrote:
> Looks ok, but now with the new option:
> 
> 1) needs a manpage update
> 2) usage() should be updated to include -f

So, I haven't quite worked out what this *is* doing, but on further reflection,
it seems like "-f" should definitely relate to behavior which iterates over
all filesystems.  i.e. without -f, non-xfs filesystems are skipped; with -f,
"foreign" filesystems are included.  That was my main concern.

But if an xfs_quota command is pointed directly at a non-xfs filesystem,
I'm not sure what's best.  Assume the user intended it, and operate on
that fs w/o needing -f?  Or require "-f" for consistency?  What do you think?

And, we can specify multiple mount points to operate on, i.e.

xfs_quota -c "foo" /mnt/ext4 /mnt/xfs

so ... I guess I don't know if that should require -f or not.  principle of
least surprise?  Keep old behavior of ignoring the non-xfs mount?

-Eric

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