[RFC PATCH 3/3] sched/Documentation: Point out use of preempt_schedule_irq()

[Valentin Schneider]

Since there are a few archs out there that call preempt_schedule_irq()
within a need_resched() loop, point out that it's not needed.

Signed-off-by: Valentin Schneider <valentin.schneider@xxxxxxx>
Cc: Jonathan Corbet <corbet@xxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Julien Thierry <julien.thierry@xxxxxxx>
Cc: linux-doc@xxxxxxxxxxxxxxx
---
 Documentation/scheduler/sched-arch.txt | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/Documentation/scheduler/sched-arch.txt b/Documentation/scheduler/sched-arch.txt
index a2f27bbf2cba..ae41a94da700 100644
--- a/Documentation/scheduler/sched-arch.txt
+++ b/Documentation/scheduler/sched-arch.txt
@@ -59,6 +59,16 @@ Your cpu_idle routines need to obey the following rules:
 arch/x86/kernel/process.c has examples of both polling and
 sleeping idle functions.
 
+Kernel preemption
+=================
+When returning from interrupt context, you should call either of
+preempt_schedule() or preempt_schedule_irq() if preemption is enabled
+and need_resched() is true.
+
+Note that, unlike preempt_schedule(), preempt_schedule_irq() handles the
+enabling and disabling of IRQs. Since it also loops on need_resched(),
+you don't need to have that loop in your arch code. See
+arch/arm64/kernel/entry.S for an example.
 
 Possible arch/ problems
 =======================
--
2.20.1