Skip to content

[WIP] kernel behavior for Privacy shutter - #103

Draft
deep0294 wants to merge 7 commits into
intel:mainfrom
deep0294:copilot/deep0294-change-pm-runtime-behavior
Draft

deep0294 wants to merge 7 commits into
intel:mainfrom
deep0294:copilot/deep0294-change-pm-runtime-behavior

Conversation

@deep0294

Copy link
Copy Markdown

pm runtime behavior

Copilot AI and others added 3 commits September 16, 2026 08:13
Co-authored-by: deep0294 <154404789+deep0294@users.noreply.github.com>
Reorder ipu7_isys_video_set_streaming()'s stop path so the sensor
subdevice is disabled before stop_streaming_firmware()/
close_streaming_firmware() are called. This closes the window where
new SOF/frame-capture commands from a still-streaming sensor race the
firmware's flush/abort snapshot, which was observed to cause the FW
stream-stop to time out. Also drop the early return on
v4l2_subdev_disable_streams() failure so the FW stream is always
flushed and closed even if the sensor disable fails.

Co-authored-by: deep0294 <154404789+deep0294@users.noreply.github.com>
@deep0294 deep0294 changed the title pm runtime behavior [WIP] kernel behavior for Privacy shutter Sep 17, 2026
Add dev_info prints in ipu_buttress_powerup() and ipu_buttress_powerdown()
to log power UP and DOWN status for ISYS sub-domains, matching the
logging format of PSYS sub-domains. Also log pm_runtime_put_sync_suspend
in ipu7_isys_fw_close().

Signed-off-by: Amarnath Kumar, Deepak <deepak.amarnath.kumar@intel.com>
…ut_sync_suspend

Log return value and atomic usage_count of adev->auxdev.dev when
pm_runtime_put_sync_suspend() is called to verify if the device
actually transitions to suspended state or is blocked by non-zero
usage counter.

Signed-off-by: Amarnath Kumar, Deepak <deepak.amarnath.kumar@intel.com>
@deep0294
deep0294 force-pushed the copilot/deep0294-change-pm-runtime-behavior branch 2 times, most recently from 40c7214 to 0daa5f5 Compare October 1, 2026 06:30
ipu7_psys_release() calls pm_runtime_put_sync() but discarded its
return value and the resulting usage_count, making it impossible to
tell from logs whether PSYS actually suspended (and powered down) on
device close versus staying busy. Log both, mirroring the existing
isys fw_close diagnostic.

Signed-off-by: Amarnath Kumar, Deepak <deepak.amarnath.kumar@intel.com>
Pairs with the existing release()-side logging. Shows whether the
usage_count baseline is already off by one right after our own get,
or grows later during streaming, to help find the reference that
keeps PSYS from suspending between shutter toggles.

Signed-off-by: Amarnath Kumar, Deepak <deepak.amarnath.kumar@intel.com>

This branch has not been deployed

No deployments
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