Re: [PATCH 3/7] scsi: Drop runtime PM usage count after host is added

[Mika Westerberg]

On Fri, Feb 19, 2016 at 09:50:50AM +1100, Julian Calaby wrote:
> Hi Mika,
> 
> On Thu, Feb 18, 2016 at 7:54 PM, Mika Westerberg
> <mika.westerberg@xxxxxxxxxxxxxxx> wrote:
> > Runtime PM of the SCSI host is already handled by calls to
> > scsi_autopm_get_host() and scsi_autopm_put_host() from appropriate places
> > whenever the host needs to be powered on. This works fine when there is
> > device connected to the host as once it runtime suspends the host will too.
> >
> > However, if there is no device connected the host is never runtime
> > suspended (the usage counter is always 0).
> >
> > Allow runtime suspend of host even if it has no devices connected by
> > calling scsi_autopm_put_host() at the end of scsi_add_host_with_dma(). We
> > temporarily increase runtime PM usage counter first so call to
> > scsi_autopm_put_host() will result idle request to be scheduled for the
> > device.
> >
> > Signed-off-by: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
> > ---
> >  drivers/scsi/hosts.c | 7 +++++++
> >  1 file changed, 7 insertions(+)
> >
> > diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
> > index 82ac1cd818ac..e46bf4d152a0 100644
> > --- a/drivers/scsi/hosts.c
> > +++ b/drivers/scsi/hosts.c
> > @@ -250,6 +250,12 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
> >         if (error)
> >                 goto out_destroy_freelist;
> >
> > +       /*
> > +        * Increase usage count temporarily here so that calling
> > +        * scsi_autopm_put_host() will trigger runtime idle if there is
> > +        * nothing else preventing suspending the device.
> > +        */
> > +       pm_runtime_get_noresume(&shost->shost_gendev);
> >         pm_runtime_set_active(&shost->shost_gendev);
> >         pm_runtime_enable(&shost->shost_gendev);
> >         device_enable_async_suspend(&shost->shost_gendev);
> > @@ -290,6 +296,7 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
> >                 goto out_destroy_host;
> >
> >         scsi_proc_host_add(shost);
> > +       scsi_autopm_put_host(shost);
> 
> Would it be cleaner to export the code that runs when the usage
> counter decrements and call it here?

There actually is no code to run. This just ensures that the device
runtime_idle gets called which allows the parent device to runtime
suspend (as it returns 0). Alternative way would be just to call
pm_request_idle() directly here.