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

[Johan Hovold <johan@kernel.org>]

On Thu, Jun 21, 2018 at 12:55:26AM +0200, Rafael J. Wysocki wrote:
> On Thu, Jun 21, 2018 at 12:32 AM, Rafael J. Wysocki <rafael@kernel.org> wrote:
> > On Wed, Jun 20, 2018 at 5:46 PM, Johan Hovold <johan@kernel.org> wrote:
> >> On Wed, Jun 20, 2018 at 02:54:10PM +0200, Rafael J. Wysocki wrote:
> >>> On Wednesday, June 20, 2018 2:23:46 PM CEST Johan Hovold wrote:
> >>> > On Wed, Jun 20, 2018 at 02:16:59AM -0700, Tony Lindgren wrote:
> >>> > > Hi,
> >>> > >
> >>> > > Adding Rafael and linux-pm to Cc as well.
> >>> > >
> >>> > > * Felipe Balbi <balbi@kernel.org> [180619 01:23]:
> >>> > > > This is a direct consequence of not paying attention to the order of
> >>> > > > things. If driver were to assume that pm_domain->activate() would do the
> >>> > > > right thing for the device -- meaning that probe would run with an
> >>> > > > active device --, then we wouldn't need that pm_runtime_get() call on
> >>> > > > probe at all. Rather we would follow the sequence:
> >>> > > >
> >>> > > >         pm_runtime_forbid()
> >>> > > >         pm_runtime_set_active()
> >>> > > >         pm_runtime_enable()
> >>> > > >
> >>> > > >         /* do your probe routine */
> >>> > > >
> >>> > > >         pm_runtime_put_noidle()
> >>> > > >
> >>> > > > Then you remove you would need to call pm_runtime_get_noresume() to
> >>> > > > balance out the pm_runtime_put_noidle() there.
> >>> >
> >>> > > > (If you need to know why the pm_runtime_put_noidle(), remember that
> >>> > > > pm_runtime_set_active() increments the usage counter, so
> >>> > > > pm_runtime_put_noidle is basically allowing pm_runtime to happen as soon
> >>> > > > as userspace writes "auto" to /sys/..../power/control)
> >>> >
> >>> > That's not correct; pm_runtime_set_active() only increments the usage
> >>> > counter of a parent (under some circumstances), so unless you have bus
> >>> > code incrementing the usage counter before probe, the above
> >>> > pm_runtime_put_noidle() would actually introduce an imbalance.

> The confusion regarding the pm_runtime_put_noidle() at the end may
> come from the special requirement of the PCI bus type as per the
> comment in local_pci_probe().

That seems to be the case, but I'm not sure of much of what PCI is doing
that can be applied here (e.g. OMAP platform devices), where unbound
devices should always be powered off and were I assume we want to have
runtime pm allowed by default, for example.

Johan
---
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