rcutorture: Also use GP sequence to judge forward progress
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Wed, 18 Jul 2018 22:39:37 +0000 (15:39 -0700)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Wed, 29 Aug 2018 16:20:48 +0000 (09:20 -0700)
commit119248bec9d318ae41da8ab8f400f07e7a610cc3
treea573e308403ac5f16cb38bd49061d09e661580f6
parent1b27291b1ea4f1f2090fb07c3425db474cdb99ba
rcutorture: Also use GP sequence to judge forward progress

Currently, rcutorture relies solely on the progress of
rcu_torture_writer() to judge grace-period forward progress.  In theory,
this is the gold standard of forward progress, but in practice rcutorture
separately detects and reports rcu_torture_writer() stalls.  This commit
therefore adds the grace-period sequence number (when provided) to the
judgment of grace-period forward progress, which makes it easier to
distinguish between failure of actual grace periods to progress on the
one hand and downstream forward-progress failures on the other.

For example, given this change, if rcu_torture_writer() stalls,
but rcu_torture_fwd_prog() does not complain, then the grace-period
computation is working, which is a hint that the failure lies in callback
processing, wakeup of the rcu_torture_writer() kthread, or similar.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
kernel/rcu/rcutorture.c