> From: Nicolin Chen <nicolinc@xxxxxxxxxx>
> Sent: Sunday, September 11, 2022 7:36 AM
>
> On Thu, Sep 08, 2022 at 09:08:38AM -0300, Jason Gunthorpe wrote:
> > On Thu, Sep 08, 2022 at 09:30:57AM +0000, Tian, Kevin wrote:
>
> > > There are also cases where common kAPIs are called in the attach
> > > path which may return -EINVAL and random errno, e.g.:
> > >
> > > omap_iommu_attach_dev()
> > > omap_iommu_attach()
> > > iommu_enable()
> > > pm_runtime_get_sync()
> > > __pm_runtime_resume()
> > > rpm_resume()
> > > if (dev->power.runtime_error) {
> > > retval = -EINVAL;
>
> > Yes, this is was also on my mind with choosing an unpopular return
> > code, it has a higher chance of not coming out of some other kernel
> > API
>
> I wonder if it would be safe to just treat a pm_runtime_get_sync()
> failure as -ENODEV, since a PM resume() mostly occurs to the IOMMU
> that an IOMMU client/master device is behind, while an iommu_domain
> rarely intervenes.
Yes, this is a condition preventing the device from being attached by
a domain hence converting -EINVAL to -ENODEV probably makes sense.
But as replied in another we might want to keep other errno's as is.
