xfs
[Top] [All Lists]

Re: [PATCH v2 1/3] xfs_quota: add capabilities for use on ext4

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH v2 1/3] xfs_quota: add capabilities for use on ext4
From: "Bill O'Donnell" <billodo@xxxxxxxxxx>
Date: Mon, 22 Aug 2016 11:02:07 -0500
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20160822014742.GQ19025@dastard>
References: <1471356998-2876-1-git-send-email-billodo@xxxxxxxxxx> <1471356998-2876-2-git-send-email-billodo@xxxxxxxxxx> <20160822014742.GQ19025@dastard>
User-agent: Mutt/1.6.2 (2016-07-01)
On Mon, Aug 22, 2016 at 11:47:42AM +1000, Dave Chinner wrote:
> On Tue, Aug 16, 2016 at 09:16:36AM -0500, Bill O'Donnell wrote:
> > This allows xfs_quota to be used on ext4 for project quota testing
> > in xfstests.
> > 
> > This patch was originally submitted by Dave Chinner
> > (http://oss.sgi.com/archives/xfs/2016-02/msg00131.html)
> > 
> > Resubmitting with the following change:
> > quota/init.c: correct logic error in loop contained in init_args_command()
> > function (lines 85-91).
> 
> What logic error?

In your original patch, in init_args_command():

        do {
                fs_path = &fs_table[index++];
-       } while ((fs_path->fs_flags & FS_PROJECT_PATH) && index < fs_count);
+               if (fs_path->fs_flags & FS_PROJECT_PATH)
+                       continue;
+               if (!foreign_allowed && (fs_path->fs_flags & FS_FOREIGN))
+                       continue;
+       } while (index < fs_count);

The loop should break out, when (fs_path->fs_flags & FS_PROJECT_PATH) is false,
but instead moves onto the next test (and then back to the top). See in the
original while statement, the loop stops when the false condition occurs, that 
is,
((fs_path->fs_flags & FS_PROJECT_PATH) && index < fs_count) == False.

My commit message was completely terse, sorry. I'll clarify it in v3.

Thanks-
Bill

> 
> Commit messages like this really don't tell the reader anything
> about what is different the original patch.  I've had to go archive
> spelunking to work out what is different, and I'm still not sure
> what the logic error you fixed is....
> 
> And, FWIW, whilst spelunking, I noticed that Eric's last review
> comments on my original patch:
> 
>       Looks ok, but now with the new option:
> 
>       1) needs a manpage update
>       2) usage() should be updated to include -f
> 
>       3) and I just noticed,
> 
>       _("foreign mount active, %s command is for XFS filesystems only\n"),
> 
>       seems kind of unclear; maybe just
> 
>       _("%s command is for XFS filesystems only\n"),
> 
> have not been addressed by this update.
> 
> Can you please add these these changes, update the commit message
> and resubmit?
> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@xxxxxxxxxxxxx

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