Re: [RFC PATCH v3 1/5] tracing: Introduce faultable tracepoints (v3)

From: Paul E. McKenney
Date: Tue Oct 03 2023 - 13:33:38 EST


On Tue, Oct 03, 2023 at 10:08:54AM -0400, Steven Rostedt wrote:
> On Tue, 3 Oct 2023 06:44:50 -0700
> "Paul E. McKenney" <paulmck@xxxxxxxxxx> wrote:
>
> > > That way it is clear what uses what, as I read the original paragraph a
> > > couple of times and could have sworn that rcu_read_lock_trace() required
> > > tasks to not block.
> >
> > That would work for me. Would you like to send a patch, or would you
> > rather we made the adjustments?
>
> Which ever.

OK, how about like this?

Thanx, Paul

------------------------------------------------------------------------

commit 973eb79ec46c16f13bb5b47ad14d44a1f1c79dc9
Author: Paul E. McKenney <paulmck@xxxxxxxxxx>
Date: Tue Oct 3 10:30:01 2023 -0700

doc: Clarify RCU Tasks reader/updater checklist

Currently, the reader/updater compatibility rules for the three RCU
Tasks flavors are squished together in a single paragraph, which can
result in confusion. This commit therefore splits them out into a list,
clearly showing the distinction between these flavors.

Reported-by: Steven Rostedt <rostedt@xxxxxxxxxxx>
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>

diff --git a/Documentation/RCU/checklist.rst b/Documentation/RCU/checklist.rst
index bd3c58c44bef..c432899aff22 100644
--- a/Documentation/RCU/checklist.rst
+++ b/Documentation/RCU/checklist.rst
@@ -241,15 +241,22 @@ over a rather long period of time, but improvements are always welcome!
srcu_struct. The rules for the expedited RCU grace-period-wait
primitives are the same as for their non-expedited counterparts.

- If the updater uses call_rcu_tasks() or synchronize_rcu_tasks(),
- then the readers must refrain from executing voluntary
- context switches, that is, from blocking. If the updater uses
- call_rcu_tasks_trace() or synchronize_rcu_tasks_trace(), then
- the corresponding readers must use rcu_read_lock_trace() and
- rcu_read_unlock_trace(). If an updater uses call_rcu_tasks_rude()
- or synchronize_rcu_tasks_rude(), then the corresponding readers
- must use anything that disables preemption, for example,
- preempt_disable() and preempt_enable().
+ Similarly, it is necssary to correctly use the RCU Tasks flavors:
+
+ a. If the updater uses synchronize_rcu_tasks() or
+ call_rcu_tasks(), then the readers must refrain from
+ executing voluntary context switches, that is, from
+ blocking.
+
+ b. If the updater uses call_rcu_tasks_trace()
+ or synchronize_rcu_tasks_trace(), then the
+ corresponding readers must use rcu_read_lock_trace()
+ and rcu_read_unlock_trace().
+
+ c. If an updater uses call_rcu_tasks_rude() or
+ synchronize_rcu_tasks_rude(), then the corresponding
+ readers must use anything that disables preemption,
+ for example, preempt_disable() and preempt_enable().

Mixing things up will result in confusion and broken kernels, and
has even resulted in an exploitable security issue. Therefore,