From: Shubhrajyoti Datta <omaplinuxkernel@xxxxxxxxx> Sent: Thursday, June 11, 2015 6:11 PM
> To: Duan Fugang-B38611
> Cc: Gao Pan-B54642; Wolfram Sang; Linux-I2C; Li Frank-B20596
> Subject: Re: [Patch v1] i2c: imx: add runtime pm support to improve the
> performance
>
> <snip>
> >> > static irqreturn_t i2c_imx_isr(int irq, void *dev_id) @@ -583,6
> >> > +582,9 @@ static irqreturn_t i2c_imx_isr(int irq, void *dev_id)
> >> > struct imx_i2c_struct *i2c_imx = dev_id;
> >> > unsigned int temp;
> >> >
> >> > + if (pm_runtime_suspended(i2c_imx->adapter.dev.parent))
> >> > + return IRQ_NONE;
> >> > +
> >>
> >> Didn't quite get this one.
> >
> > Yes, there don't need to add pm_runtime_suspended() check in isr
> > handler. But in some worse worse case, like system is very busy and
> > irq is blocked by others
>
> you mean other irqs?
Maybe.
Or spin_lock_irq_save() is called for some bad routine for long time ....
>
> > that irq response coming is very late while i2c clock is gated off, the
> check can avoid system hang.
> >
> > So I think it can be reasonable. How do you think ?
> >
> > Regards,
> > Andy
��.n��������+%������w��{.n�����{��-��)�� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
