Re: [PATCH v2 7/7] drm: Document how to register devices without struct drm_bus

From: David Herrmann
Date: Tue May 13 2014 - 11:56:05 EST


Hi

On Tue, May 13, 2014 at 5:30 PM, Thierry Reding
<thierry.reding@xxxxxxxxx> wrote:
> From: Thierry Reding <treding@xxxxxxxxxx>
>
> With the recent addition of the drm_set_unique() function, devices can
> now be registered without requiring a drm_bus. Add a brief description
> to the DRM docbook to show how that can be achieved.
>
> Signed-off-by: Thierry Reding <treding@xxxxxxxxxx>
> ---
> Documentation/DocBook/drm.tmpl | 29 +++++++++++++++++++++++++++++
> 1 file changed, 29 insertions(+)
>
> diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
> index f26a8e4fbe47..f48227caf41a 100644
> --- a/Documentation/DocBook/drm.tmpl
> +++ b/Documentation/DocBook/drm.tmpl
> @@ -142,6 +142,12 @@
> to register it with the DRM subsystem.
> </para>
> <para>
> + Newer drivers that no longer require a <structname>drm_bus</structname>
> + structure can alternatively use the low-level device initialization and
> + registration functions such as <function>drm_dev_alloc()</function> and
> + <function>drm_dev_register()</function> directly.
> + </para>
> + <para>
> The <structname>drm_driver</structname> structure contains static
> information that describes the driver and features it supports, and
> pointers to methods that the DRM core will call to implement the DRM API.
> @@ -290,6 +296,29 @@ char *date;</synopsis>
> !Fdrivers/gpu/drm/drm_pci.c drm_pci_init drm_pci_exit
> !Fdrivers/gpu/drm/drm_usb.c drm_usb_init drm_usb_exit
> !Fdrivers/gpu/drm/drm_platform.c drm_platform_init
> + <para>
> + New drivers that no longer rely on the services provided by the
> + <structname>drm_bus</structname> structure can call the low-level
> + device registration functions directly. The
> + <function>drm_dev_alloc()</function> function can be used to allocate
> + and initialize a new <structname>drm_device</structname> structure.
> + Drivers will typically want to perform some additional setup on this
> + structure, such as allocating driver-specific data and storing a
> + pointer to it in the DRM device's <structfield>dev_private</structfield>
> + field. Drivers should also set the device's unique name using the
> + <function>drm_set_unique()</function> function. After it has been set up
> + a device can be registered with the DRM subsystem by calling
> + <function>drm_dev_register()</function>. This will cause the device to
> + be exposed to userspace and will call the driver's
> + <structfield>.load()</structfield> implementation. When a device is
> + removed, the DRM device can safely be unregistered and freed using a

That's not true. drm_put_dev() is anything but safe.. but that's an
implementation bug and I'm working on it. Not that easy, though.
Furthermore, please use drm_dev_unregister() and drm_dev_unref()
directly. No reason to use drm_put_dev() in new drivers. Once I fixed
the runtime issues, a call to drm_dev_unregister() will _immediately_
remove a device and detach all active user-space connections.

> + call to <function>drm_put_dev()</function>.
> + </para>
> +!Fdrivers/gpu/drm/drm_stub.c drm_dev_alloc

Please change this to:

+!Fdrivers/gpu/drm/drm_stub.c drm_dev_alloc
+!Fdrivers/gpu/drm/drm_stub.c drm_dev_ref
+!Fdrivers/gpu/drm/drm_stub.c drm_dev_unref
+!Fdrivers/gpu/drm/drm_stub.c drm_dev_register
+!Fdrivers/gpu/drm/drm_stub.c drm_dev_unregister
+!Fdrivers/gpu/drm/drm_stub.c drm_dev_set_unique

All these helpers are needed and exported to drivers.

Otherwise, looks all good! Thanks!
David

> + </sect2>
> <sect2>
> <title>Driver Load</title>
> <para>
> --
> 1.9.2
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/