Skip to content

Conversation

@bardliao
Copy link
Collaborator

CLOCK_STOP_MODE1 is used when the Peripheral might have entered a deeper power-saving mode that does not retain state while the Clock is stopped. It is useful when the device is more power consumption sensitive. Add it back to allow the Peripheral use CLOCK_STOP_MODE1.

CLOCK_STOP_MODE1 is used when the Peripheral might have entered a deeper
power-saving mode that does not retain state while the Clock is stopped.
It is useful when the device is more power consumption sensitive. Add it
back to allow the Peripheral use CLOCK_STOP_MODE1.

Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
@plbossart
Copy link
Member

@bardliao how would you test this? In practice the intel host does a bus reset after restarting the clock which forces the device to lose context anyways.
IOW it doesn't matter if the device supports clock stop0 or 1, context always needs to be restored.

@bardliao
Copy link
Collaborator Author

@plbossart Indeed, we will do bus reset after restarting the clock. Using mode 1 is just for power consumption during clock stop. On the other hands, the peripheral will need to restore the context anyway, why don't we use mode 1 If the peripheral supports it?

@plbossart
Copy link
Member

I don't disagree @bardliao, if the host does a bus reset we would benefit from the MODE1 on the device side.
There are still several opens
a) how was MODE0 tested, if we do a bus reset then by construction we cannot test the instant restart
b) if a device supports MODE1, do we always select MODE1? it's a trade-off between power and latency, not sure the property definition is that clear.
c) if the host does a bus reset, should it also override all the properties to set the MODE1?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants