Comment 2 for bug 1940972

The test script adds a tracing filter:

echo '*lock*' > stack_trace_filter

..that takes a few minutes to apply and during that time we see the RCU and/or softlockup messages. This is kind of expected since we're blocked in kernel space updating ftrace points across the kernel which is slow and expensive on the RISC-V unmatched/unleashed H/W.

One remedy is to do the following:

echo 0 > /proc/sys/kernel/soft_watchdog
echo 1 > /sys/module/rcupdate/parameters/rcu_cpu_stall_suppress

.. run test

echo 0 > /sys/module/rcupdate/parameters/rcu_cpu_stall_suppress
echo 1 > /proc/sys/kernel/soft_watchdog

The long term fix is to make the ftrace stack_trace_filter interface more yieldy