+8
−8
Loading
The rcu_torture_writer() polling currently uses timeouts ranging from zero to 16 milliseconds to wait for the polled grace period to end. This works, but it would be better to have a higher probability of exercising races with the code that cleans up after a grace period. This commit therefore switches from these millisecond-scale timeouts to timeouts ranging from zero to 128 microseconds, and with a full microsecond's worth of timeout fuzz. Signed-off-by:Paul E. McKenney <paulmck@kernel.org> Signed-off-by:
Uladzislau Rezki (Sony) <urezki@gmail.com>