[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZD1HMkbdhQpAYs7P@ziepe.ca>
Date: Mon, 17 Apr 2023 10:18:42 -0300
From: Jason Gunthorpe <jgg@...pe.ca>
To: Alex Williamson <alex.williamson@...hat.com>
Cc: Nipun Gupta <nipun.gupta@....com>, linux-kernel@...r.kernel.org,
kvm@...r.kernel.org, masahiroy@...nel.org, nathan@...nel.org,
ndesaulniers@...gle.com, nicolas@...sle.eu, git@....com,
harpreet.anand@....com, pieter.jansen-van-vuuren@....com,
nikhil.agarwal@....com, michal.simek@....com
Subject: Re: [PATCH v2] vfio/cdx: add support for CDX bus
On Fri, Apr 14, 2023 at 03:36:14PM -0600, Alex Williamson wrote:
> > +static int vfio_cdx_init_device(struct vfio_device *core_vdev)
> > +{
> > + struct vfio_cdx_device *vdev =
> > + container_of(core_vdev, struct vfio_cdx_device, vdev);
> > + struct cdx_device *cdx_dev = to_cdx_device(core_vdev->dev);
> > +
> > + vdev->cdx_dev = cdx_dev;
> > + vdev->dev = &cdx_dev->dev;
>
> Both of these seem trivial to factor out of this patch, cdx_device is
> always available via to_cdx_device(core_vdev->dev) and the struct
> device is always available via core_vdev->dev. vdev->dev isn't even
> used anywhere yet. Both the init and release functions here could be
> dropped afaict.
Yes please, I have a series someplace that gets rid of all these
redundent ->devs we keep around everwhere.
to_cdx_device(core_vdev->dev) is a good solution, maybe with a static
inline.
Jason
Powered by blists - more mailing lists