Instead of periodically resetting watchdogs from thread context,
this patch simply forces resched and checks rq->sched_count.
Watchdog is reset if the sched count increases. If the same thread
is picked by pick_next_task during resched, there is no context
switch.
With the new method we lose coverage on: a migration/n thread
actually gets picked and we actually context switch to the
migration/n thread. These steps are unlikely to silently fail.
The change would provide nearly the same level of protection with
less latency / jitter.
v3:
- Removed the old method and boot option
- Still need to check resched
v2:
- Use sched_count instead of having sched calling into watchdog code
- Remove the sysctl and add a boot option, which can be removed later
- Changed the subject line
Xi Wang (1):
sched: watchdog: Touch kernel watchdog with sched count
include/linux/sched.h | 4 ++++
kernel/sched/core.c | 23 +++++++++++++++++++--
kernel/sched/sched.h | 6 +++++-
kernel/watchdog.c | 47 +++++++++++++------------------------------
4 files changed, 44 insertions(+), 36 deletions(-)
--
2.29.0.rc2.309.g374f81d7ae-goog