Re: [PATCH 04/11] media: iris: Add helper to create a context bank device on iris vpu bus

[Dmitry Baryshkov]

On Fri, Apr 17, 2026 at 08:49:44PM +0530, Vishnu Reddy wrote:
> 
> On 4/14/2026 8:48 PM, Dmitry Baryshkov wrote:
> > On Tue, Apr 14, 2026 at 10:30:00AM +0530, Vishnu Reddy wrote:
> > > From: Vikash Garodia<vikash.garodia@xxxxxxxxxxxxxxxx>
> > > 
> > > Add a helper function to allocate and register context bank (CB) device
> > > on the iris vpu bus. The function ID associated with the CB is specified
> > > from the platform data, allowing the bus dma_configure callback to apply
> > > correct stream ID mapping when device is registered.
> > > 
> > > Signed-off-by: Vikash Garodia<vikash.garodia@xxxxxxxxxxxxxxxx>
> > > Signed-off-by: Vishnu Reddy<busanna.reddy@xxxxxxxxxxxxxxxx>
> > > ---
> > >   drivers/media/platform/qcom/iris/iris_resources.c | 33 +++++++++++++++++++++++
> > >   drivers/media/platform/qcom/iris/iris_resources.h |  1 +
> > >   2 files changed, 34 insertions(+)
> > > 
> > > diff --git a/drivers/media/platform/qcom/iris/iris_resources.c b/drivers/media/platform/qcom/iris/iris_resources.c
> > > index 773f6548370a..a25e0f2e9d26 100644
> > > --- a/drivers/media/platform/qcom/iris/iris_resources.c
> > > +++ b/drivers/media/platform/qcom/iris/iris_resources.c
> > > @@ -6,6 +6,7 @@
> > >   #include <linux/clk.h>
> > >   #include <linux/devfreq.h>
> > >   #include <linux/interconnect.h>
> > > +#include <linux/iris_vpu_bus.h>
> > >   #include <linux/pm_domain.h>
> > >   #include <linux/pm_opp.h>
> > >   #include <linux/pm_runtime.h>
> > > @@ -141,3 +142,35 @@ int iris_disable_unprepare_clock(struct iris_core *core, enum platform_clk_type
> > >   	return 0;
> > >   }
> > > +
> > > +static void iris_release_cb_dev(struct device *dev)
> > > +{
> > > +	kfree(dev);
> > > +}
> > > +
> > > +struct device *iris_create_cb_dev(struct iris_core *core, const char *name, const u32 *f_id)
> > Please move into the bus code and make it generic enough.
> Do you suggest to add a wrapper to pass the varying inputs to the generic
> bus, something like this
> struct device* create_and_register_device(dma_mask, parent_dev, *release,
> dev_name,...)

Definitely not the release function. The devname is also not that
important. The rest, yes, you are correct.

> > > +{
> > > +	struct device *dev;
> > > +	int ret;
> > > +
> > > +	dev = kzalloc_obj(*dev);
> > > +	if (!dev)
> > > +		return ERR_PTR(-ENOMEM);
> > > +
> > > +	dev->release = iris_release_cb_dev;
> > > +	dev->bus = &iris_vpu_bus_type;
> > > +	dev->parent = core->dev;
> > > +	dev->coherent_dma_mask = core->iris_platform_data->dma_mask;
> > > +	dev->dma_mask = &dev->coherent_dma_mask;
> > Would you also need to set the of_node? See
> > device_set_of_node_from_dev()
> 
> It might be needed for FastRPC as they are following sub node approach, Iris
> does not need.

Wouldn't it save you from passing it to of_dma_configure_id()?

> > > +
> > > +	dev_set_name(dev, "%s", name);
> > > +	dev_set_drvdata(dev, (void *)f_id);
> > > +
> > > +	ret = device_register(dev);
> > > +	if (ret) {
> > > +		put_device(dev);
> > > +		return ERR_PTR(ret);
> > > +	}
> > > +
> > > +	return dev;
> > > +}
> > > diff --git a/drivers/media/platform/qcom/iris/iris_resources.h b/drivers/media/platform/qcom/iris/iris_resources.h
> > > index 6bfbd2dc6db0..4a494627ff23 100644
> > > --- a/drivers/media/platform/qcom/iris/iris_resources.h
> > > +++ b/drivers/media/platform/qcom/iris/iris_resources.h
> > > @@ -15,5 +15,6 @@ int iris_unset_icc_bw(struct iris_core *core);
> > >   int iris_set_icc_bw(struct iris_core *core, unsigned long icc_bw);
> > >   int iris_disable_unprepare_clock(struct iris_core *core, enum platform_clk_type clk_type);
> > >   int iris_prepare_enable_clock(struct iris_core *core, enum platform_clk_type clk_type);
> > > +struct device *iris_create_cb_dev(struct iris_core *core, const char *name, const u32 *f_id);
> > >   #endif
> > > 
> > > -- 
> > > 2.34.1
> > > 

-- 
With best wishes
Dmitry