Re: [PATCH] kernel: x86: apic: Replace GFP_ATOMIC with GFP_KERNEL in __add_pin_to_irq_node

[Thomas Gleixner]

On Tue, 13 Feb 2018, Ingo Molnar wrote:
> * Jia-Ju Bai <baijiaju1990@xxxxxxxxx> wrote:
> 
> > The function __add_pin_to_irq_node is not called in atomic context.
> > Thus GFP_ATOMIC is not necessary, and it can be replaced with GFP_KERNEL.
> > 
> > Signed-off-by: Jia-Ju Bai <baijiaju1990@xxxxxxxxx>
> > ---
> >  arch/x86/kernel/apic/io_apic.c |    2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
> > index 201579d..665c013 100644
> > --- a/arch/x86/kernel/apic/io_apic.c
> > +++ b/arch/x86/kernel/apic/io_apic.c
> > @@ -379,7 +379,7 @@ static int __add_pin_to_irq_node(struct mp_chip_data *data,
> >  		if (entry->apic == apic && entry->pin == pin)
> >  			return 0;
> >  
> > -	entry = kzalloc_node(sizeof(struct irq_pin_list), GFP_ATOMIC, node);
> > +	entry = kzalloc_node(sizeof(struct irq_pin_list), GFP_KERNEL, node);
> >  	if (!entry) {
> >  		pr_err("can not alloc irq_pin_list (%d,%d,%d)\n",
> >  		       node, apic, pin);
> 
> NAK: this is called in an atomic section: with IRQs disabled ...

The only invocation where this happens with IRQs disabled is the early
check_timer() code which runs _before_ the scheduler is working. GFP_KERNEL
is valid during that phase. All other call sites have interrupts enabled.

Thanks,

	tglx