Hi Niklas,
Thank you for the patch.
On Wed, Mar 18, 2020 at 10:30:47PM +0100, Niklas Söderlund wrote:
> Add a flags field to the media_device_info structure by taking one
> of the reserved u32 fields. The use-case is to have a way to
> (optionally) report to user-space if the media graph is complete or not.
>
> Also define two flags to carry information about if the graph is
> complete or not. If neither of the two flags are set the
> media device does not support reporting its graph status. The other bits
> in the flags field are unused for now, but could be claimed to carry
> other type of information in the future.
>
> Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@xxxxxxxxxxxx>
> ---
> include/uapi/linux/media.h | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/include/uapi/linux/media.h b/include/uapi/linux/media.h
> index 383ac7b7d8f07eca..9b37ed8b41d0d866 100644
> --- a/include/uapi/linux/media.h
> +++ b/include/uapi/linux/media.h
> @@ -34,9 +34,16 @@ struct media_device_info {
> __u32 media_version;
> __u32 hw_revision;
> __u32 driver_version;
> - __u32 reserved[31];
> + __u32 flags;
> + __u32 reserved[30];
I think this information should be added to media_v2_topology, not
media_device_info, otherwise you'll have a race condition between
retrieving the media device information and the topology.
media_device_info is really supposed to be static.
> };
>
> +/*
> + * Graph flags
> + */
> +#define MEDIA_INFO_FLAG_INCOMPLETE (1 << 0)
> +#define MEDIA_INFO_FLAG_COMPLETE (1 << 1)
> +
> /*
> * Base number ranges for entity functions
> *
--
Regards,
Laurent Pinchart
