usb: dwc3: of-simple: fix use-after-free on remove

[Tony Lindgren <tony@atomide.com>]

* Tero Kristo <t-kristo@ti.com> [180619 12:13]:
> 
> TI SoC arm64 PM is done via genpd attached to each domain, which in turn
> passes control messages to the PM firmware when transitions happen. See
> drivers/soc/ti_sci_pm_domains.c for details.
> 
> Not quite sure about the discussion following later, but there is work
> ongoing to get rid of both omap_hwmod and omap_device, and replace this with
> a proper bus driver for OMAP architectures. It might solve the issues being
> talked about, or then not. Tony, any comments on that?

As long as drivers keep working we should add the callbacks suggested by
Felipe if that allows making drivers generic. Then changing the driver
probes can be done one driver at a time hopefully.

> Anyway, main reason that omap devices are actually suspended on HW level
> during probe is the aggressive power management applied on the device, and
> based on the documentation this is actually a valid thing to do. This makes
> sure that all devices get to idle if they are not used (no driver probed
> etc.), and will allow the SoC to idle core powerdomain.

What Felipe is suggesting is adding the callbacks that won't change the
runtime PM usecount. They just enable the device for probe, then bus code
can disable it again on failed probe. In the device driver probe things
in theory should work then both the old way and the PCI style way Felipe
wants. Needs to be tested of course :)

Regards,

Tony
---
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html