drivers: scmi-msg: correct voltage domain protocol version & clock rates enumeration #6951
Conversation
etienne-lms
changed the title
There was a problem hiding this comment.
For "drivers: scmi-msg: correct voltage domain protocol version":
Acked-by: Jerome Forissier <[email protected]>
core/drivers/scmi-msg/clock.c
Outdated
| @@ -321,7 +321,7 @@ static void scmi_clock_describe_rates(struct scmi_msg *msg) | |||
| } | |||
|
|
|||
| out_count = rate_index - in_args->rate_index; | |||
| remaining = nb_rates - in_args->rate_index; | |||
| remaining = nb_rates - in_args->rate_index - out_count; | |||
There was a problem hiding this comment.
nb_rates - rate_index would be better, no?
There was a problem hiding this comment.
I would have said nb_rates - out_count as out_count is the number of entries filled in the buffer. Maybe I'm missing something.
There was a problem hiding this comment.
nb_rates - rate_index would be better, no?
Indeed simpler.
I would have said
nb_rates - out_countas out_count is the number of entries filled in the buffer. Maybe I'm missing something.
nb_rates gives the overall number of supported rates. out_count gives the number of rates filled in this very response message but does not count the rates skipped, from index 0 to index in_args->rate_index - 1.
core/drivers/scmi-msg/clock.c
Outdated
| @@ -321,7 +321,7 @@ static void scmi_clock_describe_rates(struct scmi_msg *msg) | |||
| } | |||
|
|
|||
| out_count = rate_index - in_args->rate_index; | |||
| remaining = nb_rates - in_args->rate_index; | |||
| remaining = nb_rates - in_args->rate_index - out_count; | |||
There was a problem hiding this comment.
nb_rates - rate_index would be better, no?
Indeed simpler.
I would have said
nb_rates - out_countas out_count is the number of entries filled in the buffer. Maybe I'm missing something.
nb_rates gives the overall number of supported rates. out_count gives the number of rates filled in this very response message but does not count the rates skipped, from index 0 to index in_args->rate_index - 1.
|
|
|
Sure, |
Fix the version ID of the implemented SCMI voltage domain protocol that is v2.0 (ID 0x20000), not v3.0 (ID 0x30000). Fixes: 006d89b ("drivers: scmi-msg: add SCMI Voltage Domain protocol") Signed-off-by: Etienne Carriere <[email protected]> Acked-by: Jerome Forissier <[email protected]> Acked-by: Gatien Chevallier <[email protected]>
Fix value of remaining clocks to describe on SCMI clock protocol message CLOCK_DESCRIBE_RATES that does not take into account the number of returned clock in the response. Fixes: 90252e2 ("drivers: scmi-msg: clock adapts to output buffer size") Signed-off-by: Etienne Carriere <[email protected]> Acked-by: Jerome Forissier <[email protected]> Acked-by: Gatien Chevallier <[email protected]>
|
Review tags applied. Thanks. |
Fix the version ID of the implemented SCMI voltage domain protocol that is v2.0 (ID 0x20000), not v3.0 (ID 0x30000).
(Fixes: 006d89b ("drivers: scmi-msg: add SCMI Voltage Domain protocol"))
Fix value of remaining clocks to describe on SCMI clock protocol message CLOCK_DESCRIBE_RATES that does not take into account the number of returned clock in the response.
(Fixes: 90252e2 ("drivers: scmi-msg: clock adapts to output buffer size"))