Is there any documentation available which describes how Linux manages the
OMAP chip resources, such as resets, power domains, and memory spaces? I'm
currently designing an OMAP 3515 based board that shares many of the same
features with the Beagle Board, but has stringent battery life requirements.
I'm not a Linux expert so I'm not familiar with what's available as far as
low level system documentation. Does Linux allow for dynamic power control,
shutting off parts of the OMAP when not in use? For example, if the ISP
camera interface is used intermittently, can that power domain be shut off
from a Linux driver? Can Linux gracefully move up and down the power states
when the OMAP is essentially idle? And most important, is any of this
written down somewhere?
Jack Peacock
Jack Peacock wrote:
Is there any documentation available which describes how Linux manages the
OMAP chip resources, such as resets, power domains, and memory spaces? I'm
currently designing an OMAP 3515 based board that shares many of the same
features with the Beagle Board, but has stringent battery life requirements.
I'm not a Linux expert so I'm not familiar with what's available as far as
low level system documentation. Does Linux allow for dynamic power control,
shutting off parts of the OMAP when not in use? For example, if the ISP
camera interface is used intermittently, can that power domain be shut off
from a Linux driver? Can Linux gracefully move up and down the power states
when the OMAP is essentially idle? And most important, is any of this
written down somewhere?
Yes, current Linux on OMAP3 does a lot of these things, if you make your board SmartReflex aware/capable it will also dynamically adjust the core voltage depending on the current CPU needs.
there is this presentation that can help:
http://www.celinux.org/elc08_presentations/TI_OMAP3430_Linux_PM_reference.ppt