Multiple system-wide low power-states
summary
Kernel now exposes a user-space latency knob and a boolean flag so suspend-to-idle can consider per-device PMQoS resume-latency votes (values are microseconds, aggregated by existing max/last-wins rules), and the merged patchset teaches the PM domain code to pass a hint as PSCI arguments so firmware can pick a different SOC idle state. Remaining work covers separating runtime vs system suspend semantics, handling devices outside domain hierarchies (storage), and describing SCMI domains with Nexus node maps for nested domain layout.
AI-written summary · about
chapters (5)
- 0:00Introduction and overview
- 1:26Suspend idle state selection
- 10:01Device QoS constraints for suspend
- 20:11PM domain hierarchy challenges
- 23:59Summary and acknowledgments
related talks
41mLinux Power Management Features & Their Interactions, Part 2
59mDevice power management and idle
36mLinux Power Management Features, Their Relationships and Interactions- 36mLinux Power Management Features, Their Relationships and Interactions
32mAdding Runtime Power Management Capabilities to Device Drivers
33mRemote Core Power Management in AMP Systems
30mHost and Guest Suspend Under Xen - S3 and S0ix
47mLinux Kernel Power Management - 101