Hi Hans,
On Thu, Feb 08, 2018 at 09:36:46AM +0100, Hans Verkuil wrote:
> The VIDIOC_DBG_G/S_REGISTER ioctls imply that VIDIOC_DBG_G_CHIP_INFO is also
> present, since without that you cannot use v4l2-dbg.
>
> Just like the implementation in v4l2-ioctl.c this can be implemented in the
> core and no drivers need to be modified.
>
> It also makes it possible for v4l2-compliance to properly test the
> VIDIOC_DBG_G/S_REGISTER ioctls.
>
> Signed-off-by: Hans Verkuil <hans.verk...@cisco.com>
> ---
> drivers/media/v4l2-core/v4l2-subdev.c | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
>
> diff --git a/drivers/media/v4l2-core/v4l2-subdev.c
> b/drivers/media/v4l2-core/v4l2-subdev.c
> index 6cabfa32d2ed..2a5b5a3fa7a3 100644
> --- a/drivers/media/v4l2-core/v4l2-subdev.c
> +++ b/drivers/media/v4l2-core/v4l2-subdev.c
> @@ -255,6 +255,19 @@ static long subdev_do_ioctl(struct file *file, unsigned
> int cmd, void *arg)
> return -EPERM;
> return v4l2_subdev_call(sd, core, s_register, p);
> }
> + case VIDIOC_DBG_G_CHIP_INFO:
> + {
> + struct v4l2_dbg_chip_info *p = arg;
> +
> + if (p->match.type != V4L2_CHIP_MATCH_SUBDEV || p->match.addr)
> + return -EINVAL;
> + if (sd->ops->core && sd->ops->core->s_register)
> + p->flags |= V4L2_CHIP_FL_WRITABLE;
> + if (sd->ops->core && sd->ops->core->g_register)
> + p->flags |= V4L2_CHIP_FL_READABLE;
> + strlcpy(p->name, sd->name, sizeof(p->name));
> + return 0;
> + }
This is effectively doing the same as debugfs except that it's specific to
V4L2. I don't think we should endorse its use, and especially not without a
real use case.
> #endif
>
> case VIDIOC_LOG_STATUS: {
> --
> 2.15.1
>
--
Sakari Ailus
e-mail: sakari.ai...@iki.fi
