[PATCH 5/9] ipa: rkisp1: params: Check for empty parameters

[Stefan Klug]

Hi Kieran,

Thank you for the review. 

On Mon, Mar 31, 2025 at 05:38:27PM +0100, Kieran Bingham wrote:
> Quoting Stefan Klug (2025-03-31 15:43:44)
> > Add functions to check for empty parameter blocks. Modify the constructor
> > so that a parameter block constructed from an empty span stays empty.
> > 
> > Signed-off-by: Stefan Klug <stefan.klug at ideasonboard.com>
> > ---
> >  src/ipa/rkisp1/params.cpp |  3 +++
> >  src/ipa/rkisp1/params.h   | 13 +++++++++++++
> >  2 files changed, 16 insertions(+)
> > 
> > diff --git a/src/ipa/rkisp1/params.cpp b/src/ipa/rkisp1/params.cpp
> > index 4c0b051ce65d..b4a889e415fc 100644
> > --- a/src/ipa/rkisp1/params.cpp
> > +++ b/src/ipa/rkisp1/params.cpp
> > @@ -82,6 +82,9 @@ RkISP1ParamsBlockBase::RkISP1ParamsBlockBase(RkISP1Params *params, BlockType typ
> >                                              const Span<uint8_t> &data)
> >         : params_(params), type_(type)
> >  {
> > +       if (data.empty())
> > +               return;
> > +
> >         if (params_->format() == V4L2_META_FMT_RK_ISP1_EXT_PARAMS) {
> >                 header_ = data.subspan(0, sizeof(rkisp1_ext_params_block_header));
> >                 data_ = data.subspan(sizeof(rkisp1_ext_params_block_header));
> > diff --git a/src/ipa/rkisp1/params.h b/src/ipa/rkisp1/params.h
> > index 40450e34497a..e0aa04ae0f8d 100644
> > --- a/src/ipa/rkisp1/params.h
> > +++ b/src/ipa/rkisp1/params.h
> > @@ -89,6 +89,9 @@ public:
> >  
> >         void setEnabled(bool enabled);
> >  
> > +       bool isValid() const { return !data_.empty(); }
> > +       explicit operator bool() const { return !data_.empty(); }
> > +
> >  private:
> >         LIBCAMERA_DISABLE_COPY(RkISP1ParamsBlockBase)
> >  
> > @@ -109,6 +112,16 @@ public:
> >         {
> >         }
> >  
> > +       const Type *get() const
> > +       {
> > +               return reinterpret_cast<const Type *>(data().data());
> > +       }
> > +
> > +       Type *get()
> > +       {
> > +               return reinterpret_cast<Type *>(data().data());
> > +       }
> > +
> 
> These seem unrelated ?

Oh, you are right. Dropped it for v2.

Best regards,
Stefan

> 
> --
> Kieran
> 
> 
> >         const Type *operator->() const
> >         {
> >                 return reinterpret_cast<const Type *>(data().data());
> > -- 
> > 2.43.0
> >