Re: [PATCH v2 1/4] firmware: arm_scmi: Bypass setting fwnode for scmi cpufreq

[Cristian Marussi]

On Thu, Feb 06, 2025 at 02:31:19PM +0300, Dan Carpenter wrote:
> On Thu, Feb 06, 2025 at 06:52:20PM +0800, Peng Fan wrote:
> > On Wed, Feb 05, 2025 at 03:45:00PM +0300, Dan Carpenter wrote:
> > >On Mon, Jan 20, 2025 at 03:13:29PM +0800, Peng Fan (OSS) wrote:
> > >> diff --git a/drivers/firmware/arm_scmi/bus.c b/drivers/firmware/arm_scmi/bus.c
> > >> index 2c853c84b58f530898057e4ab274ba76070de05e..7850eb7710f499888d32aebf5d99df63db8bfa26 100644
> > >> --- a/drivers/firmware/arm_scmi/bus.c
> > >> +++ b/drivers/firmware/arm_scmi/bus.c
> > >> @@ -344,6 +344,21 @@ static void __scmi_device_destroy(struct scmi_device *scmi_dev)
> > >>  	device_unregister(&scmi_dev->dev);
> > >>  }
> > >>  
> > >> +static int
> > >> +__scmi_device_set_node(struct scmi_device *scmi_dev, struct device_node *np,
> > >> +		       int protocol, const char *name)
> > >> +{
> > >> +	/* cpufreq device does not need to be supplier from devlink perspective */
> > >> +	if ((protocol == SCMI_PROTOCOL_PERF) && !strcmp(name, "cpufreq")) {
> > >
> > >I don't love this...  It seems like an hack.  Could we put a flag
> > >somewhere instead?  Perhaps in scmi_device?  (I'm just saying that
> > >because that's what we're passing to this function).
> > 
> > This means when creating scmi_device, a flag needs to be set which requires
> > to extend scmi_device_id to include a flag entry or else.
> > 
> > As below in scmi-cpufreq.c
> > { SCMI_PROTOCOL_PERF, "cpufreq", SCMI_FWNODE_NO }
> > 
> 
> Yeah, I like that.
> 
> -	if ((protocol == SCMI_PROTOCOL_PERF) && !strcmp(name, "cpufreq")) {
> +	if (scmi_dev->flags & SCMI_FWNODE_NO) {
> 
> Or we could do something like "if (scmi_dev->no_fwnode) {"

I proposed a flag a few review ago about this, it shoule come somehow
from the device_table above like Peng was proposing, so that a driver
can just declare that does NOT need fw_devlink.

Thanks,
Cristian