[1/3] kernel: local_irq_{save,restore}_nmi()

Provide local_irq_{save,restore}_nmi() which will allow us to help
architectures that implement NMIs using IRQ priorities like SPARC64
does.

Sparc uses IRQ prio 15 for NMIs and implements local_irq_disable() as
disable <= 14. However if you do that while inside an NMI you re-
enable the NMI priority again, causing all kinds of fun.

A more solid implementation would first check the disable level and
never lower it, however that is more costly and would slow down the
rest of the kernel for no particular reason.

Therefore introduce local_irq_save_nmi() which can implement this
slower but more solid scheme and dis-allow local_irq_save() from NMI
context.

Suggested-by: David Miller <davem@davemloft.net>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
 include/linux/irqflags.h     |   51 ++++++++++++++++++++++++++++++++++++++++---
 kernel/lockdep.c             |    7 +++++
 kernel/trace/trace_irqsoff.c |    8 ++++++
 3 files changed, 63 insertions(+), 3 deletions(-)



--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Tue, 2010-04-06 at 15:28 +0200, Peter Zijlstra wrote:

> Index: linux-2.6/kernel/trace/trace_irqsoff.c
> ===================================================================
> --- linux-2.6.orig/kernel/trace/trace_irqsoff.c
> +++ linux-2.6/kernel/trace/trace_irqsoff.c
> @@ -316,6 +316,14 @@ void trace_hardirqs_off(void)
>  }
>  EXPORT_SYMBOL(trace_hardirqs_off);
>  
> +void trace_hardirqs_off_no_nmi(void)
> +{
> +	WARN_ON_ONCE(in_nmi());

Should we do this for all archs? I can imagine a lot of warning reports
coming in the near future. And they will be passing it towards me.

-- Steve



> +	if (!preempt_trace() && irq_trace())
> +		start_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
> +}
> +EXPORT_SYMBOL(trace_hardirqs_off_no_nmi);
> +
>  void trace_hardirqs_on_caller(unsigned long caller_addr)
>  {
>  	if (!preempt_trace() && irq_trace())
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/


--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

From: Steven Rostedt <rostedt@goodmis.org>
Date: Tue, 06 Apr 2010 21:13:10 -0400

> On Tue, 2010-04-06 at 15:28 +0200, Peter Zijlstra wrote:
> 
>> Index: linux-2.6/kernel/trace/trace_irqsoff.c
>> ===================================================================
>> --- linux-2.6.orig/kernel/trace/trace_irqsoff.c
>> +++ linux-2.6/kernel/trace/trace_irqsoff.c
>> @@ -316,6 +316,14 @@ void trace_hardirqs_off(void)
>>  }
>>  EXPORT_SYMBOL(trace_hardirqs_off);
>>  
>> +void trace_hardirqs_off_no_nmi(void)
>> +{
>> +	WARN_ON_ONCE(in_nmi());
> 
> Should we do this for all archs? I can imagine a lot of warning reports
> coming in the near future. And they will be passing it towards me.

That's the whole point, so that the problem is more easily noticed and
it gets fixed long before I end up accidently testing the code on my
machines :-)

To be honest, the fix is so trivial, you just need to add '_nmi' to
the local_irq_{save,restore}() calls that warn like this.

I'm even willing to have you forward all of those reports to me and
I'll be responsible for fixing them.

How's that? :-)
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Tue, 2010-04-06 at 18:19 -0700, David Miller wrote:

> That's the whole point, so that the problem is more easily noticed and
> it gets fixed long before I end up accidently testing the code on my
> machines :-)
> 
> To be honest, the fix is so trivial, you just need to add '_nmi' to
> the local_irq_{save,restore}() calls that warn like this.
> 
> I'm even willing to have you forward all of those reports to me and
> I'll be responsible for fixing them.
> 
> How's that? :-)

Sure.

/me sets up his procmailrc to search for the WARN_ON line in
trace_irqsoff.c and have it forward to davem.

-- Steve


--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html