Re: [PATCH v7 4/5] media: platform: mediatek: isp: add mediatek ISP3.0 camsv

[Laurent Pinchart <laurent.pinchart@ideasonboard.com>]

On Mon, Nov 25, 2024 at 03:40:18PM +0100, Julien Stephan wrote:
> Le lun. 25 nov. 2024 à 09:14, CK Hu (胡俊光) <ck.hu@mediatek.com> a écrit :
> > On Thu, 2024-11-21 at 09:53 +0100, Julien Stephan wrote:
> > >
> > > From: Phi-bang Nguyen <pnguyen@baylibre.com>
> > >
> > > This driver provides a path to bypass the SoC ISP so that image data
> > > coming from the SENINF can go directly into memory without any image
> > > processing. This allows the use of an external ISP.
> > >
> > > Signed-off-by: Phi-bang Nguyen <pnguyen@baylibre.com>
> > > Signed-off-by: Florian Sylvestre <fsylvestre@baylibre.com>
> > > [Paul Elder fix irq locking]
> > > Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
> > > Co-developed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > > Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > > Co-developed-by: Julien Stephan <jstephan@baylibre.com>
> > > Signed-off-by: Julien Stephan <jstephan@baylibre.com>
> > > ---
> >
> > [snip]
> >
> > > +static void mtk_camsv30_update_buffers_add(struct mtk_cam_dev *cam_dev,
> > > +                                          struct mtk_cam_dev_buffer *buf)
> > > +{
> > > +       mtk_camsv30_img0_write(cam_dev, CAMSV_IMGO_SV_BASE_ADDR, buf->daddr);
> > > +}
> > > +
> >
> > [snip]
> >
> > > +static void mtk_camsv30_fbc_inc(struct mtk_cam_dev *cam_dev)
> > > +{
> > > +       unsigned int fbc_val;
> > > +
> > > +       if (pm_runtime_resume_and_get(cam_dev->dev) < 0) {
> >
> > I think this pm_runtime_resume_and_get() is not necessary.
> > mtk_camsv30_fbc_inc() is called only in mtk_cam_vb2_buf_queue().
> > But when buf_list is empty, mtk_camsv30_update_buffers_add() is called before this function.
> > But mtk_camsv30_update_buffers_add() does not call pm_runtime_resume_and_get() and it works normally.
> > So this function is not necessary to call pm_runtime_resume_and_get().
> 
> Hi CK,
> 
> This one is actually needed because .buf_queue can be called before
> .start_streaming in case a user requests to prepare buffers before
> streaming.
> But you are right, if a user requests to alloc buffer before streaming
> mtk_camsv_update_buffers_address will be called without pm. Streaming
> still works because in start streaming we call
> mtk_camsv_update_buffers_address again.
> So maybe I should put the pm stuff in mtk_cam_vb2_buf_queue ?

mtk_cam_vb2_buf_queue() should only touch the hardware if streaming has
been started, otherwise it shhould just put the buffers in a queue and
return immediately, and the initial hardware setup should be done when
starting streaming. There should be no need to handle runtime PM in this
function or in mtk_cam_vb2_buf_queue().

> > In other register setting function, please also check this pm function is necessary or not.
> >
> > Regards,
> > CK
> >
> > > +               dev_err(cam_dev->dev, "failed to get pm_runtime\n");
> > > +               return;
> > > +       }
> > > +
> > > +       fbc_val = mtk_camsv30_read(cam_dev, CAMSV_IMGO_FBC);
> > > +       fbc_val |= CAMSV_IMGO_FBC_RCNT_INC;
> > > +       mtk_camsv30_write(cam_dev, CAMSV_IMGO_FBC, fbc_val);
> > > +       fbc_val &= ~CAMSV_IMGO_FBC_RCNT_INC;
> > > +       mtk_camsv30_write(cam_dev, CAMSV_IMGO_FBC, fbc_val);
> > > +
> > > +       pm_runtime_put_autosuspend(cam_dev->dev);
> > > +}
> > > +

-- 
Regards,

Laurent Pinchart