Re: [PATCH v3] PM / core: conditionally skip system pm in device/driver model

[Guan-Yu Lin]

On Fri, Feb 23, 2024 at 11:18 PM Andy Shevchenko
<andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
>
> On Fri, Feb 23, 2024 at 02:38:29PM +0000, Guan-Yu Lin wrote:
> > In systems with a main processor and a co-processor, asynchronous
> > controller management can lead to conflicts.  One example is the main
> > processor attempting to suspend a device while the co-processor is
> > actively using it. To address this, we introduce a new sysfs entry
> > called "conditional_skip". This entry allows the system to selectively
> > skip certain device power management state transitions. To use this
> > feature, set the value in "conditional_skip" to indicate the type of
> > state transition you want to avoid.  Please review /Documentation/ABI/
> > testing/sysfs-devices-power for more detailed information.
>
> ...
>
> > +static ssize_t conditional_skip_store(struct device *dev,
> > +                                   struct device_attribute *attr,
> > +                                   const char *buf, size_t n)
> > +{
>
> > +     int ret;
>
> > +     if (kstrtoint(buf, 0, &ret))
>
> Why is it int? It seems like flags, should not be unsigned as u32 or so?
>

The ".event" member in struct pm_message is an int, but the values
assigned to it are used like bit flags (e.g. PM_EVENT_FREEZE=0x1,
PM_EVENT_SUSPEND=0x2, PM_EVENT_HIBERNATE=0x4). Is this an intentional
design choice? We might need to change the design accordingly.

> > +             return -EINVAL;
>
> Do not shadow the real error code without justification.
>

Thanks for suggesting the desired implementation. I'll refactor it in
the next version.

> > +     ret &= (PM_EVENT_FREEZE|PM_EVENT_SUSPEND|PM_EVENT_HIBERNATE);
> > +
> > +     dev->power.conditional_skip_pm = ret;
> > +
> > +     return n;
> > +}
>
> > +
>
> Redundant blank line.
>

Thanks for the heads-up.

> > +static DEVICE_ATTR_RW(conditional_skip);
>
> --
> With Best Regards,
> Andy Shevchenko
>
>