[PATCH v3 4/6] firmware: arm_scmi: add SCMIv3.0 Sensors timestamped reads

[Cristian Marussi]

On Thu, Nov 19, 2020 at 11:42:16AM +0000, Sudeep Holla wrote:
> On Wed, Nov 18, 2020 at 04:29:03PM +0000, Cristian Marussi wrote:
> > Add new .reading_get_timestamped() method to sensor_ops to support SCMIv3.0
> > timestamped reads.
> > 
> > Signed-off-by: Cristian Marussi <cristian.marussi at arm.com>
> > ---
> > V2 --> v3
> > - setting rx_size to 0 in sensor_reading_get to allow fw to send
> >   both v2 and v3 replies...even if sensor_reading_get() only handles
> >   v2 spec and returns one single value
> > - using get_unaligned_le64 in lieu of le64_to_cpu
> > - removed refs to v2.1
> > ---
> >  drivers/firmware/arm_scmi/sensors.c | 137 ++++++++++++++++++++++++++--
> >  include/linux/scmi_protocol.h       |  22 +++++
> >  2 files changed, 152 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/firmware/arm_scmi/sensors.c b/drivers/firmware/arm_scmi/sensors.c
> > index 1c83aaae0012..0adc545116a4 100644
> > --- a/drivers/firmware/arm_scmi/sensors.c
> > +++ b/drivers/firmware/arm_scmi/sensors.c
> > @@ -156,6 +156,27 @@ struct scmi_msg_sensor_reading_get {
> >  #define SENSOR_READ_ASYNC	BIT(0)
> >  };
> >  
> > +struct scmi_resp_sensor_reading_get {
> > +	__le64 readings;
> 
> Generally I have avoided such single element structures so far. Any
> particular reasons for having it ?
> 
 Just because there are a few of them so I found it easy and preferable
 to have some typing to distinguish them, but I can drop it.

 Cristian

> -- 
> Regards,
> Sudeep