Hi Laurent,
On Thu, 2021-01-14 at 04:06 +0200, Laurent Pinchart wrote:
> Hi Ezequiel,
>
> Thank you for the patch.
>
> On Tue, Jan 12, 2021 at 10:23:29AM -0300, Ezequiel Garcia wrote:
> > The use of v4l2_async_notifier_add_subdev is discouraged.
>
> It's not yet, that only happens at the end of the series :-)
>
> s/is discouraged/will be discouraged/
>
> > Drivers are instead encouraged to use a helper such as
> > v4l2_async_notifier_add_i2c_subdev.
> >
> > This fixes a misuse of the API, as v4l2_async_notifier_add_subdev
> > should get a kmalloc'ed struct v4l2_async_subdev,
> > removing some boilerplate code while at it.
> >
> > Use the appropriate helper: v4l2_async_notifier_add_i2c_subdev
> > or v4l2_async_notifier_add_fwnode_remote_subdev, which handles
> > the needed setup, instead of open-coding it.
> >
> > Using v4l2-async to allocate the driver-specific structs,
> > requires to change struct ceu_subdev so the embedded
> > struct v4l2_async_subdev is now the first element.
> >
> > Signed-off-by: Ezequiel Garcia <ezequiel@xxxxxxxxxxxxx>
> > ---
> > drivers/media/platform/renesas-ceu.c | 89 ++++++++++------------------
> > 1 file changed, 31 insertions(+), 58 deletions(-)
> >
> > diff --git a/drivers/media/platform/renesas-ceu.c b/drivers/media/platform/renesas-ceu.c
> > index 4a633ad0e8fa..18485812a21e 100644
> > --- a/drivers/media/platform/renesas-ceu.c
> > +++ b/drivers/media/platform/renesas-ceu.c
> > @@ -152,8 +152,8 @@ static inline struct ceu_buffer *vb2_to_ceu(struct vb2_v4l2_buffer *vbuf)
> > * ceu_subdev - Wraps v4l2 sub-device and provides async subdevice.
> > */
> > struct ceu_subdev {
> > - struct v4l2_subdev *v4l2_sd;
> > struct v4l2_async_subdev asd;
> > + struct v4l2_subdev *v4l2_sd;
> >
> > /* per-subdevice mbus configuration options */
> > unsigned int mbus_flags;
> > @@ -174,7 +174,7 @@ struct ceu_device {
> > struct v4l2_device v4l2_dev;
> >
> > /* subdevices descriptors */
> > - struct ceu_subdev *subdevs;
> > + struct ceu_subdev **subdevs;
> > /* the subdevice currently in use */
> > struct ceu_subdev *sd;
> > unsigned int sd_index;
> > @@ -1195,7 +1195,7 @@ static int ceu_enum_input(struct file *file, void *priv,
> > if (inp->index >= ceudev->num_sd)
> > return -EINVAL;
> >
> > - ceusd = &ceudev->subdevs[inp->index];
> > + ceusd = ceudev->subdevs[inp->index];
> >
> > inp->type = V4L2_INPUT_TYPE_CAMERA;
> > inp->std = 0;
> > @@ -1230,7 +1230,7 @@ static int ceu_s_input(struct file *file, void *priv, unsigned int i)
> > return 0;
> >
> > ceu_sd_old = ceudev->sd;
> > - ceudev->sd = &ceudev->subdevs[i];
> > + ceudev->sd = ceudev->subdevs[i];
> >
> > /*
> > * Make sure we can generate output image formats and apply
> > @@ -1423,7 +1423,7 @@ static int ceu_notify_complete(struct v4l2_async_notifier *notifier)
> > * ceu formats.
> > */
> > if (!ceudev->sd) {
> > - ceudev->sd = &ceudev->subdevs[0];
> > + ceudev->sd = ceudev->subdevs[0];
> > ceudev->sd_index = 0;
> > }
> >
> > @@ -1465,28 +1465,6 @@ static const struct v4l2_async_notifier_operations ceu_notify_ops = {
> > .complete = ceu_notify_complete,
> > };
> >
> > -/*
> > - * ceu_init_async_subdevs() - Initialize CEU subdevices and async_subdevs in
> > - * ceu device. Both DT and platform data parsing use
> > - * this routine.
> > - *
> > - * Returns 0 for success, -ENOMEM for failure.
> > - */
> > -static int ceu_init_async_subdevs(struct ceu_device *ceudev, unsigned int n_sd)
> > -{
> > - /* Reserve memory for 'n_sd' ceu_subdev descriptors. */
> > - ceudev->subdevs = devm_kcalloc(ceudev->dev, n_sd,
> > - sizeof(*ceudev->subdevs), GFP_KERNEL);
>
> I may be missing something, but it looks like the subdevs array isn't
> allocated anymore. It turned to an array of pointers, but it still need
> allocation.
>
After some thought, I'd say it would be better to leave
ceu_init_async_subdevs() untouched, trying to be as
uninvasive as possible.
Thanks for the catch.
Ezequiel
