[v2] kernel/sched: Modify initial boot task idle setup

Commit Message

Liam R. Howlett Sept. 15, 2023, 5:44 p.m. UTC

Initial booting is setting the task flag to idle (PF_IDLE) by the call
path sched_init() -> init_idle().  Having the task idle and calling
call_rcu() in kernel/rcu/tiny.c means that TIF_NEED_RESCHED will be
set.  Subsequent calls to any cond_resched() will enable IRQs,
potentially earlier than the IRQ setup has completed.  Recent changes
have caused just this scenario and IRQs have been enabled early.

This causes a warning later in start_kernel() as interrupts are enabled
before they are fully set up.

Fix this issue by setting the PF_IDLE flag later in the boot sequence.

Although the boot task was marked as idle since (at least) d80e4fda576d,
I am not sure that it is wrong to do so.  The forced context-switch on
idle task was introduced in the tiny_rcu update, so I'm going to claim
this fixes 5f6130fa52ee.

Link: https://lore.kernel.org/linux-mm/[e-mail]/
Link: https://lore.kernel.org/linux-mm/CAMuHMdWpvpWoDa=[e-mail]/
Fixes: 5f6130fa52ee ("tiny_rcu: Directly force QS when call_rcu_[bh|sched]() on idle_task")
Cc: [e-mail]
Cc: Geert Uytterhoeven <[e-mail]>
Cc: "Paul E. McKenney" <[e-mail]>
Cc: Christophe Leroy <[e-mail]>
Cc: Andreas Schwab <[e-mail]>
Cc: Matthew Wilcox <[e-mail]>
Cc: Peng Zhang <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Ingo Molnar <[e-mail]>
Cc: Juri Lelli <[e-mail]>
Cc: Vincent Guittot <[e-mail]>
Cc: Andrew Morton <[e-mail]>
Cc: "Mike Rapoport (IBM)" <[e-mail]>
Cc: Vlastimil Babka <[e-mail]>
Signed-off-by: Liam R. Howlett <[e-mail]>
---

v1: https://lore.kernel.org/linux-mm/[e-mail]/

 kernel/sched/core.c | 2 +-
 kernel/sched/idle.c | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

Comments

Peter Zijlstra Sept. 19, 2023, 8:49 a.m. UTC | #1

On Fri, Sep 15, 2023 at 01:44:44PM -0400, Liam R. Howlett wrote:
> Initial booting is setting the task flag to idle (PF_IDLE) by the call
> path sched_init() -> init_idle().  Having the task idle and calling
> call_rcu() in kernel/rcu/tiny.c means that TIF_NEED_RESCHED will be
> set.  Subsequent calls to any cond_resched() will enable IRQs,
> potentially earlier than the IRQ setup has completed.  Recent changes
> have caused just this scenario and IRQs have been enabled early.
> 
> This causes a warning later in start_kernel() as interrupts are enabled
> before they are fully set up.
> 
> Fix this issue by setting the PF_IDLE flag later in the boot sequence.
> 
> Although the boot task was marked as idle since (at least) d80e4fda576d,
> I am not sure that it is wrong to do so.  The forced context-switch on
> idle task was introduced in the tiny_rcu update, so I'm going to claim
> this fixes 5f6130fa52ee.
> 
> Link: https://lore.kernel.org/linux-mm/[e-mail]/
> Link: https://lore.kernel.org/linux-mm/CAMuHMdWpvpWoDa=[e-mail]/
> Fixes: 5f6130fa52ee ("tiny_rcu: Directly force QS when call_rcu_[bh|sched]() on idle_task")
> Signed-off-by: Liam R. Howlett <[e-mail]>
> ---

Thanks!, I've queued this up for sched/urgent but will let the robots at
it before I push it out to -tip.

Geert Uytterhoeven Sept. 19, 2023, 10:11 a.m. UTC | #2

Hi Liam,

On Fri, Sep 15, 2023 at 7:45 PM Liam R. Howlett <[e-mail]> wrote:
> Initial booting is setting the task flag to idle (PF_IDLE) by the call
> path sched_init() -> init_idle().  Having the task idle and calling
> call_rcu() in kernel/rcu/tiny.c means that TIF_NEED_RESCHED will be
> set.  Subsequent calls to any cond_resched() will enable IRQs,
> potentially earlier than the IRQ setup has completed.  Recent changes
> have caused just this scenario and IRQs have been enabled early.
>
> This causes a warning later in start_kernel() as interrupts are enabled
> before they are fully set up.
>
> Fix this issue by setting the PF_IDLE flag later in the boot sequence.
>
> Although the boot task was marked as idle since (at least) d80e4fda576d,
> I am not sure that it is wrong to do so.  The forced context-switch on
> idle task was introduced in the tiny_rcu update, so I'm going to claim
> this fixes 5f6130fa52ee.
>
> Link: https://lore.kernel.org/linux-mm/[e-mail]/
> Link: https://lore.kernel.org/linux-mm/CAMuHMdWpvpWoDa=[e-mail]/
> Fixes: 5f6130fa52ee ("tiny_rcu: Directly force QS when call_rcu_[bh|sched]() on idle_task")
> Cc: [e-mail]
> Cc: Geert Uytterhoeven <[e-mail]>
> Cc: "Paul E. McKenney" <[e-mail]>
> Cc: Christophe Leroy <[e-mail]>
> Cc: Andreas Schwab <[e-mail]>
> Cc: Matthew Wilcox <[e-mail]>
> Cc: Peng Zhang <[e-mail]>
> Cc: Peter Zijlstra <[e-mail]>
> Cc: Ingo Molnar <[e-mail]>
> Cc: Juri Lelli <[e-mail]>
> Cc: Vincent Guittot <[e-mail]>
> Cc: Andrew Morton <[e-mail]>
> Cc: "Mike Rapoport (IBM)" <[e-mail]>
> Cc: Vlastimil Babka <[e-mail]>
> Signed-off-by: Liam R. Howlett <[e-mail]>

Thanks for your patch!

This fixes the

      WARNING: CPU: 0 PID: 0 at init/main.c:992 start_kernel+0x2f0/0x480

I was seeing during boot on Renesas RZ/A1 and RZ/A2 since commit
cfeb6ae8bcb96ccf ("maple_tree: disable mas_wr_append() when other
readers are possible") in v6.5.

And unlike v1, this does not cause lots of new warnings on e.g. R-Car M2-W.

Tested-by: Geert Uytterhoeven <[e-mail]>

Gr{oetje,eeting}s,

                        Geert

Patch

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index c52c2eba7c73..e8f73ff12126 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9271,7 +9271,7 @@  void __init init_idle(struct task_struct *idle, int cpu)
 	 * PF_KTHREAD should already be set at this point; regardless, make it
 	 * look like a proper per-CPU kthread.
 	 */
-	idle->flags |= PF_IDLE | PF_KTHREAD | PF_NO_SETAFFINITY;
+	idle->flags |= PF_KTHREAD | PF_NO_SETAFFINITY;
 	kthread_set_per_cpu(idle, cpu);
 
 #ifdef CONFIG_SMP
diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
index 342f58a329f5..5007b25c5bc6 100644
--- a/kernel/sched/idle.c
+++ b/kernel/sched/idle.c
@@ -373,6 +373,7 @@  EXPORT_SYMBOL_GPL(play_idle_precise);
 
 void cpu_startup_entry(enum cpuhp_state state)
 {
+	current->flags |= PF_IDLE;
 	arch_cpu_idle_prepare();
 	cpuhp_online_idle(state);
 	while (1)