[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20121221140935.GB9729@avionic-0098.adnet.avionic-design.de>
Date: Fri, 21 Dec 2012 15:09:35 +0100
From: Thierry Reding <thierry.reding@...onic-design.de>
To: Lucas Stach <dev@...xeye.de>
Cc: Terje Bergstrom <tbergstrom@...dia.com>, airlied@...ux.ie,
dri-devel@...ts.freedesktop.org, linux-tegra@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCHv4 8/8] gpu: host1x: Register DRM dummy device
On Fri, Dec 21, 2012 at 02:53:51PM +0100, Lucas Stach wrote:
> Am Freitag, den 21.12.2012, 13:39 +0200 schrieb Terje Bergstrom:
[...]
> > +void host1x_unregister_drm_device(struct host1x *host1x)
> > +{
> > + if (host1x->drm_device)
> > + platform_device_unregister(host1x->drm_device);
> > +}
> > +EXPORT_SYMBOL(host1x_unregister_drm_device);
> > +
> > +struct platform_device *host1x_drm_device(struct platform_device *pdev)
> > +{
> > + struct host1x *host = platform_get_drvdata(pdev);
> > + return host->drm_device;
> > +}
> > +EXPORT_SYMBOL(host1x_drm_device);
> This should be called host1x_get_drm_device
Or maybe even host1x_drm_get_device() to make it more obvious that it
doesn't return a struct drm_device. Also I think it would be better to
make it take a struct device * and return a struct device * instead.
Users of the API will probably call this like so:
struct device *dummy = host1x_drm_get_device(pdev->dev.parent);
struct drm_device *drm = dev_get_drvdata(dummy);
So we save ourselves some needless up-casting.
Thierry
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists