SMP TWD (Timer & Watchdog) driver need to be reentrant
Registered by
Linus Walleij
When CPUs and clusters of CPUs come and go the .setup() and .stop() hooks of the SMP TWD timer will be called repeatedly. As of kernel 3.7 this is not possible: the .setup()+.stop() pair is not reentrant and will produce warnings and leak resources.
Blueprint information
- Status:
- Complete
- Approver:
- Amit Kucheria
- Priority:
- Undefined
- Drafter:
- None
- Direction:
- Needs approval
- Assignee:
- Linus Walleij
- Definition:
- Pending Approval
- Series goal:
- None
- Implementation:
- Implemented
- Milestone target:
- None
- Started by
- Linus Walleij
- Completed by
- Linus Walleij
Related branches
Related bugs
Sprints
Whiteboard
A patch that fixes the reentrance problem has been iterated on LAKML and reached consensus. It is currently pending in Russell King's patch tracker:
http://
(?)
Work Items
Work items:
Prepare a patch: DONE
Iterate the patch: DONE
Submit to Russell's patch tracker: DONE
Wait for Russell to pick it: TODO
Patch merged upstream to Torvalds: TODO