Re: [PATCH 01/22] xfs: query the per-AG reservation counters

From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Dave Chinner <david@fromorbit.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 01/22] xfs: query the per-AG reservation counters
Date: Mon, 24 Jul 2017 12:07:07 -0700	[thread overview]
Message-ID: <20170724190707.GF4352@magnolia> (raw)
In-Reply-To: <20170723222516.GZ17762@dastard>

On Mon, Jul 24, 2017 at 08:25:16AM +1000, Dave Chinner wrote:
> On Thu, Jul 20, 2017 at 09:38:35PM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Establish an ioctl for userspace to query the original and current
> > per-AG reservation counts.  This will be used by xfs_scrub to
> > check that the vfs counters are at least somewhat sane.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >  fs/xfs/libxfs/xfs_fs.h |   10 ++++++++++
> >  fs/xfs/xfs_fsops.c     |   29 +++++++++++++++++++++++++++++
> >  fs/xfs/xfs_fsops.h     |    2 ++
> >  fs/xfs/xfs_ioctl.c     |   16 ++++++++++++++++
> >  fs/xfs/xfs_ioctl32.c   |    1 +
> >  5 files changed, 58 insertions(+)
> > 
> > 
> > diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
> > index 8c61f21..5dedab9 100644
> > --- a/fs/xfs/libxfs/xfs_fs.h
> > +++ b/fs/xfs/libxfs/xfs_fs.h
> > @@ -469,6 +469,15 @@ typedef struct xfs_swapext
> >  #define XFS_FSOP_GOING_FLAGS_NOLOGFLUSH		0x2	/* don't flush log nor data */
> >  
> >  /*
> > + * AG reserved block counters
> > + */
> > +struct xfs_fsop_ag_resblks {
> > +	__u64 resblks;		/* blocks reserved now */
> 
> 		current_reservation
> 
> > +	__u64 resblks_orig;	/* blocks reserved at mount time */
> 
> 		mount_reservation;
> 
> > +	__u64 reserved[2];

So long as I'm going to add a flags field I might as well bump this
whole structure up to a full 64 bytes.

struct xfs_fsop_ag_resblks {
	__u32 flags;			/* output flags, none defined now */
	__u32 reserved;			/* zero */
	__u64 current_resv;		/* blocks reserved now */
	__u64 mount_resv;		/* blocks reserved at mount time */
	__u64 reserved[5];		/* zero */
};

> > +};
> 
> Also, any new structure we pass to userspace should be versioned
> from the start. At minimum, a flags field so we can, in future, tell
> userspace what the reserved space means in future.

Will do.

> > +
> > +/* Query the per-AG reservations to see how many blocks we have reserved. */
> > +int
> > +xfs_fs_get_ag_reserve_blocks(
> > +	struct xfs_mount		*mp,
> > +	struct xfs_fsop_ag_resblks	*out)
> > +{
> > +	struct xfs_ag_resv		*r;
> > +	struct xfs_perag		*pag;
> > +	xfs_agnumber_t			agno;
> > +
> > +	out->resblks = 0;
> > +	out->resblks_orig = 0;
> > +	out->reserved[0] = out->reserved[1] = 0;
> 
> memset() the structure so we don't forget in future to zero it
> properly.

Ok.

--D

> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html