Re: [PATCH rcu 5/6] doc/rcutorture: Add description of rcutorture.stall_cpu_block

[Joel Fernandes]

On Wed, May 10, 2023 at 10:12 AM Paul E. McKenney <paulmck@xxxxxxxxxx> wrote:
>
> From: Zqiang <qiang1.zhang@xxxxxxxxx>
>
> If you build a kernel with CONFIG_PREEMPTION=n and CONFIG_PREEMPT_COUNT=y,
> then run the rcutorture tests specifying stalls as follows:
>
> runqemu kvm slirp nographic qemuparams="-m 1024 -smp 4" \
>         bootparams="console=ttyS0 rcutorture.stall_cpu=30 \
>         rcutorture.stall_no_softlockup=1 rcutorture.stall_cpu_block=1" -d
>
> The tests will produce the following splat:
>
> [   10.841071] rcu-torture: rcu_torture_stall begin CPU stall
> [   10.841073] rcu_torture_stall start on CPU 3.
> [   10.841077] BUG: scheduling while atomic: rcu_torture_sta/66/0x0000000
> ....
> [   10.841108] Call Trace:
> [   10.841110]  <TASK>
> [   10.841112]  dump_stack_lvl+0x64/0xb0
> [   10.841118]  dump_stack+0x10/0x20
> [   10.841121]  __schedule_bug+0x8b/0xb0
> [   10.841126]  __schedule+0x2172/0x2940
> [   10.841157]  schedule+0x9b/0x150
> [   10.841160]  schedule_timeout+0x2e8/0x4f0
> [   10.841192]  schedule_timeout_uninterruptible+0x47/0x50
> [   10.841195]  rcu_torture_stall+0x2e8/0x300
> [   10.841199]  kthread+0x175/0x1a0
> [   10.841206]  ret_from_fork+0x2c/0x50

Another way to get rid of the warning would be to replace the
cur_ops->readlock() with rcu_read_lock(). Though perhaps that will not
test whether the particular RCU flavor under testing is capable of
causing a stall :-).

>         rcutorture.stall_cpu_block= [KNL]
>                         Sleep while stalling if set.  This will result
> -                       in warnings from preemptible RCU in addition
> -                       to any other stall-related activity.
> +                       in warnings from preemptible RCU in addition to
> +                       any other stall-related activity.  Note that
> +                       in kernels built with CONFIG_PREEMPTION=n and
> +                       CONFIG_PREEMPT_COUNT=y, this parameter will
> +                       cause the CPU to pass through a quiescent state.
> +                       Any such quiescent states will suppress RCU CPU
> +                       stall warnings, but the time-based sleep will
> +                       also result in scheduling-while-atomic splats.

Could change last part to "but may also result in
scheduling-while-atomic splats as preemption might be disabled for
certain RCU flavors in order to cause the stall".

> +                       Which might or might not be what you want.
> +

Suggest drop this line ;-).

 - Joel

>         rcutorture.stall_cpu_holdoff= [KNL]
>                         Time to wait (s) after boot before inducing stall.
> --
> 2.40.1
>