2015-05-17 05:19:58

by NeilBrown

[permalink] [raw]
Subject: [PATCH - RE-RESEND] IRQ: don't suspend nested_thread irqs over system suspend.



Nested IRQs can only fire when the parent irq fires.
So when the parent is suspended, there is no need to suspend
the child irq.

Suspending nested irqs can cause a problem is they are suspended or
resumed in the wrong order.
If an interrupt fires while the parent is active but the child is
suspended, then the interrupt will not be acknowledged properly
and so an interrupt storm can result.
This is particularly likely if the parent is resumed before
the child, and the interrupt was raised during suspend.

Ensuring correct ordering would be possible, but it is simpler
to just never suspend nested interrupts.

Signed-off-by: NeilBrown <[email protected]>

---
At Rafael's request I'm resending this. No response last time (except from Rafael),
no change in the code.

This change will fix a bug so that the work-around introduced by
8b41669ceba0 ("mfd: twl4030: Fix chained irq handling on resume from suspend")
can be reverted. This work-around misuses the IRQF_EARLY_RESUME flag, and triggers
a warning on resume:

[ 56.095825] WARNING: CPU: 0 PID: 3 at ../kernel/irq/manage.c:661 irq_nested_primary_handler+0x18/0x28()

(though the line number might have changed).

NeilBrown


diff --git a/kernel/irq/pm.c b/kernel/irq/pm.c
index 5204a6d1b985..d22786a6dbde 100644
--- a/kernel/irq/pm.c
+++ b/kernel/irq/pm.c
@@ -123,6 +123,8 @@ void suspend_device_irqs(void)
unsigned long flags;
bool sync;

+ if (irq_settings_is_nested_thread(desc))
+ continue;
raw_spin_lock_irqsave(&desc->lock, flags);
sync = suspend_device_irq(desc, irq);
raw_spin_unlock_irqrestore(&desc->lock, flags);
@@ -163,6 +165,8 @@ static void resume_irqs(bool want_early)

if (!is_early && want_early)
continue;
+ if (irq_settings_is_nested_thread(desc))
+ continue;

raw_spin_lock_irqsave(&desc->lock, flags);
resume_irq(desc, irq);


Attachments:
(No filename) (811.00 B)
OpenPGP digital signature

2015-05-18 12:16:50

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH - RE-RESEND] IRQ: don't suspend nested_thread irqs over system suspend.

On Sun, May 17, 2015 at 03:19:34PM +1000, NeilBrown wrote:
>
>
> Nested IRQs can only fire when the parent irq fires.
> So when the parent is suspended, there is no need to suspend
> the child irq.
>
> Suspending nested irqs can cause a problem is they are suspended or
> resumed in the wrong order.
> If an interrupt fires while the parent is active but the child is
> suspended, then the interrupt will not be acknowledged properly
> and so an interrupt storm can result.
> This is particularly likely if the parent is resumed before
> the child, and the interrupt was raised during suspend.
>
> Ensuring correct ordering would be possible, but it is simpler
> to just never suspend nested interrupts.

Looks sane to me, but it's Thomas' call.

FWIW:

Acked-by: Peter Zijlstra (Intel) <[email protected]>

>
> Signed-off-by: NeilBrown <[email protected]>
> ---
> diff --git a/kernel/irq/pm.c b/kernel/irq/pm.c
> index 5204a6d1b985..d22786a6dbde 100644
> --- a/kernel/irq/pm.c
> +++ b/kernel/irq/pm.c
> @@ -123,6 +123,8 @@ void suspend_device_irqs(void)
> unsigned long flags;
> bool sync;
>
> + if (irq_settings_is_nested_thread(desc))
> + continue;
> raw_spin_lock_irqsave(&desc->lock, flags);
> sync = suspend_device_irq(desc, irq);
> raw_spin_unlock_irqrestore(&desc->lock, flags);
> @@ -163,6 +165,8 @@ static void resume_irqs(bool want_early)
>
> if (!is_early && want_early)
> continue;
> + if (irq_settings_is_nested_thread(desc))
> + continue;
>
> raw_spin_lock_irqsave(&desc->lock, flags);
> resume_irq(desc, irq);

Subject: [tip:irq/core] genirq: Don' t suspend nested_thread irqs over system suspend

Commit-ID: 3c646f2c6aa9e918d7fc77867df7f430059f9ccc
Gitweb: http://git.kernel.org/tip/3c646f2c6aa9e918d7fc77867df7f430059f9ccc
Author: NeilBrown <[email protected]>
AuthorDate: Sun, 17 May 2015 15:19:34 +1000
Committer: Thomas Gleixner <[email protected]>
CommitDate: Mon, 18 May 2015 17:23:47 +0200

genirq: Don't suspend nested_thread irqs over system suspend

Nested IRQs can only fire when the parent irq fires. So when the
parent is suspended, there is no need to suspend the child irq.

Suspending nested irqs can cause a problem is they are suspended or
resumed in the wrong order. If an interrupt fires while the parent is
active but the child is suspended, then the interrupt will not be
acknowledged properly and so an interrupt storm can result. This is
particularly likely if the parent is resumed before the child, and the
interrupt was raised during suspend.

Ensuring correct ordering would be possible, but it is simpler to just
never suspend nested interrupts.

Signed-off-by: NeilBrown <[email protected]>
Cc: GTA04 owners <[email protected]>
Cc: Kalle Jokiniemi <[email protected]>
Acked-by: Peter Zijlstra <[email protected]>
Cc: Rafael J. Wysocki <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Thomas Gleixner <[email protected]>

---
kernel/irq/pm.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/kernel/irq/pm.c b/kernel/irq/pm.c
index 5204a6d..d22786a 100644
--- a/kernel/irq/pm.c
+++ b/kernel/irq/pm.c
@@ -123,6 +123,8 @@ void suspend_device_irqs(void)
unsigned long flags;
bool sync;

+ if (irq_settings_is_nested_thread(desc))
+ continue;
raw_spin_lock_irqsave(&desc->lock, flags);
sync = suspend_device_irq(desc, irq);
raw_spin_unlock_irqrestore(&desc->lock, flags);
@@ -163,6 +165,8 @@ static void resume_irqs(bool want_early)

if (!is_early && want_early)
continue;
+ if (irq_settings_is_nested_thread(desc))
+ continue;

raw_spin_lock_irqsave(&desc->lock, flags);
resume_irq(desc, irq);