Messages in this thread | | | Date | Tue, 10 Nov 2020 19:19:33 +0000 | From | Cristian Marussi <> | Subject | Re: [PATCH v2 2/6] firmware: arm_scmi: add SCMIv3.0 Sensors descriptors extensions |
| |
On Tue, Nov 10, 2020 at 06:50:04PM +0100, Peter Hilber wrote: > Hi Cristian, > > sorry, I mistakenly used the wrong sender ("Mailing Lists") for the > original comment mail. Please see below for my reply. > > On 10.11.20 18:21, Cristian Marussi wrote: > > On Tue, Nov 10, 2020 at 05:00:05PM +0100, Mailing Lists wrote: > >> On 26.10.20 21:10, Cristian Marussi wrote: > >>> Add support for new SCMIv3.0 Sensors extensions related to new sensors' > >>> features, like multiple axis and update intervals, while keeping > >>> compatibility with SCMIv2.0 features. > >>> While at that, refactor and simplify all the internal helpers macros and > >>> move struct scmi_sensor_info to use only non-fixed-size typing. > >>> > >>> Signed-off-by: Cristian Marussi <cristian.marussi@arm.com> > >>> --- > >>> v1 --> v2 > >>> - restrict segmented intervals descriptors to single triplet > >>> - add proper usage of scmi_reset_rx_to_maxsz > >>> --- > >>> drivers/firmware/arm_scmi/sensors.c | 391 ++++++++++++++++++++++++++-- > >>> include/linux/scmi_protocol.h | 219 +++++++++++++++- > >>> 2 files changed, 584 insertions(+), 26 deletions(-) > >>> > >>> diff --git a/drivers/firmware/arm_scmi/sensors.c b/drivers/firmware/arm_scmi/sensors.c > >>> index 6aaff478d032..5a18f8c84bef 100644 > >>> --- a/drivers/firmware/arm_scmi/sensors.c > >>> +++ b/drivers/firmware/arm_scmi/sensors.c > >>> @@ -7,16 +7,21 @@ > >>> > >>> #define pr_fmt(fmt) "SCMI Notifications SENSOR - " fmt > >>> > >>> +#include <linux/bitfield.h> > >>> #include <linux/scmi_protocol.h> > >>> > >>> #include "common.h" > >>> #include "notify.h" > >>> > >>> +#define SCMI_MAX_NUM_SENSOR_AXIS 64 > >> > >> IMHO the related 6 bit wide fields, like SENSOR_DESCRIPTION_GET "Number > >> of axes", should determine the maximum value, so 64 -> 63. > >> > > > > Yes, bits [21:16] 'Number of Axes' in sensor_attributes_high, but this > > #define was meant to represent the maximum number of sensor axis (64...ranging > > from 0 to 63) not the highest possible numbered (63). > > > > But in my understanding the actual maximum number of sensor axes is 63 > due to the maximum value 63 of 'Number of Axes', 64 would overflow > already. The ids would range from 0 to 62.
Ah damn, you're right ... maximum that I can set in 5 bits is anyway 63. I'll fix.
Thanks
Cristian
> > That said, in my understanding there is no problem with retaining a > higher value ATM. > > Best regards, > > Peter
| |