Repository navigation
Conversation
There was a problem hiding this comment.
Does this come from host -> DSP ? Wondering if flags would be better as we may permit/deny other operations in the future (not just audio streams) ?
There was a problem hiding this comment.
I've updated the structure just to include only generic fields to disable certain power optimizations. FW will be checking if only pipelines related to keyword detection feature are running (no active streaming to host effectively) and then will opt in for Low Power Sequence (LPS) flow.
There was a problem hiding this comment.
@lbetlej what this intentional? it doesnt seem related to this PR?
There was a problem hiding this comment.
I've reverted the change, this can be cleaned up in another PR
a6b48c6 to
c2588c0
Compare
plbossart
left a comment
plbossart
left a comment
There was a problem hiding this comment.
Comments below. some points aren't clear on what the kernel needs to pass over IPC.
There was a problem hiding this comment.
nit-pick: is this a state or a transition?
There was a problem hiding this comment.
Sorry, I don't get the last two fields. If you set D3 but prevent power gating, this is not D3...
How would the kernel set these two values, based on what?
There was a problem hiding this comment.
I've added diagrams to better illustrate the idea at https://github.com/orgs/thesofproject/teams/sof-developers/discussions/ Could we switch there for a discussion ?
There was a problem hiding this comment.
does this deprecate PM_CTX_SAVE/RESTORE then?
There was a problem hiding this comment.
for now PM_CTX_SAVE/RESTORE would be backward compatible to for D3 entry transition effectively, mid-term they would become deprecated
Could we switch to https://github.com/orgs/thesofproject/teams/sof-developers/discussions/ for a discussion on pm flows ?
mmaka1
changed the title
c2588c0 to
c35f4b7
Compare
The commit extends IPC interface between a driver and FW to support new power flows to be used with keyword detection feature. A new command is added to set request for power state transition request explicitly & definition of D0ix in addtion to D0 , D3 states is added. Signed-off-by: Lech Betlej <lech.betlej@linux.intel.com>
c35f4b7 to
ee32227
Compare
lgirdwood
left a comment
lgirdwood
left a comment
There was a problem hiding this comment.
Will need some sync with ABI minor version, as the kernel and FW still have a remaining ABI minor diff to align on.
lbetlej
changed the title
| case SOF_IPC_PM_STATE_SET: | ||
| return ipc_pm_state_set(header); | ||
| case SOF_IPC_PM_STATE_GET: | ||
| return ipc_pm_state_get(header); |
There was a problem hiding this comment.
@lbetlej just for my understanding. When would the driver use the ipc_pm_state_get()? Clearly when the DSP is in D3, this cannot be used.
|
@lbetlej btw, is there an equivalent kernel ABI update PR ? |
|
@lbetlej CI failing due to conflicts. Will need kernel PR too. |
|
@lbetlej What are we doing with this one? |
|
@lbetlej I think we are good once conflicts/CI issues resolved and and kernel PR is done too.. |
The commit extends IPC interface between a driver and FW to support
new power flows to be used with keyword detection feature. A new
command is added to set request for power state transition request
explicitly & definition of D0ix in addtion to D0 , D3 states is added.
Change to use registers instead of shared memory for cavs specific power
flow will be covered by a separate PR. CHT specific linking issue is tracked as #1330
Signed-off-by: Lech Betlej lech.betlej@linux.intel.com