Re: Spinlocks waiting with interrupts disabled / preempt disabled.

[Ingo Molnar]

* Christoph Lameter <clameter@xxxxxxx> wrote:

> @@ -132,10 +132,14 @@ unsigned long __lockfunc _write_lock_irq
>  {
>  	unsigned long flags;
>  
> +retry:
>  	local_irq_save(flags);
> -	preempt_disable();
> -	_raw_write_lock(lock);
> -	return flags;
> +	if (_write_trylock(lock))
> +		return flags;
> +	local_irq_restore(flags);
> +	while (!write_can_lock(lock))
> +		cpu_relax();
> +	goto retry;
>  }
>  EXPORT_SYMBOL(_write_lock_irqsave);

hm, this is done on a too high level and will turn off some debugging 
code. I.e. if we dont just loop long but truly deadlock here we wont 
call lib/spinlock_debug.c's _raw_write_lock() code that does some sanity 
checks in the debug case.

so how about doing this on a deeper level and adding a new 
__raw_write_lock_flags() primitive that would look at the flags value 
and could enable interrupts in the lowlevel code?

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/