Re: linux-next: Tree for Jul 7 (scsi/lpfc/lpfc_init.c)

[Viresh Kumar]

Hi James,

On Tue, Jul 7, 2020 at 11:43 PM James Smart <james.smart@xxxxxxxxxxxx> wrote:
>
> On 7/7/2020 10:09 AM, Randy Dunlap wrote:
> > On 7/7/20 1:08 AM, Stephen Rothwell wrote:
> >> Hi all,
> >>
> >> Changes since 20200706:
> >>
> > on i386:
> >
> > when CONFIG_ACPI is not set/enabled:
> >
> >
> > ../drivers/scsi/lpfc/lpfc_init.c:1265:15: error: implicit declaration of function 'get_cpu_idle_time'; did you mean 'get_cpu_device'? [-Werror=implicit-function-declaration]
> >
> >
> > The cpufreq people want justification for using

I am one of cpufreq people :)

> > get_cpu_idle_time().  Please see
> > https://lore.kernel.org/linux-scsi/20200707030943.xkocccy6qy2c3hrx@vireshk-i7/
> >
>
> The driver is using cpu utilization in order to choose between softirq
> or work queues in handling an interrupt. Less-utilized, softirq is used.
> higher utilized, work queue is used.  The utilization is checked
> periodically via a heartbeat.

I understand that you need to use this functionality, what I was
rather asking was
why do you need to use a routine that is for cpufreq related stuff only.

I also see that drivers/macintosh/rack-meter.c has its own
implementation for this.

What I would suggest is that, if required, we should add/move a
generic implementation
of this to another file (which is available to all) and then let
everyone use it.

--
viresh