On 07/02/2004 04:20 PM Philippe Gerum wrote:
> On Fri, 2004-07-02 at 14:44, Wolfgang Grandegger wrote:
>> Hi Philippe,
>> 
>> I realized that __adeos_handle_irq() might be called with hard
>> interrupts enabled when called from adeos_tigger_irq(). Is this OK?
>> 
> 
> No, it's a bug, at least on SMP configs because we don't want unexpected
> migrations to occur there. Oops... I'll fix it for x86 at least and
> check arm/armnommu and ia64, just in case. Many thanks for the hint.

I actually checked the interrupt flag (MSR_EE on PowerPC) in
__adeos_set_irq_bit() and similar macros while hunting another problem.
I will apply a similar fix for the ppc.

Thanks.

Wolfgang.






Reply via email to