Re: [PATCH V4 1/2] PCI: Omit pci_disable_device() in .shutdown()

[Bjorn Helgaas <helgaas@kernel.org>]

On Fri, Feb 03, 2023 at 12:00:37PM +0800, Huacai Chen wrote:
> Hi, Bjorn,
> 
> On Fri, Feb 3, 2023 at 4:30 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
> >
> > On Thu, Feb 02, 2023 at 09:27:03PM +0800, Huacai Chen wrote:
> > > On Thu, Feb 2, 2023 at 2:17 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > > > On Wed, Feb 01, 2023 at 12:30:17PM +0800, Huacai Chen wrote:
> >
> > > > > +static void pcie_portdrv_shutdown(struct pci_dev *dev)
> > > > > +{
> > > > > +     if (pci_bridge_d3_possible(dev)) {
> > > > > +             pm_runtime_forbid(&dev->dev);
> > > > > +             pm_runtime_get_noresume(&dev->dev);
> > > > > +             pm_runtime_dont_use_autosuspend(&dev->dev);
> > > > > +     }
> > > > > +
> > > > > +     pcie_port_device_remove(dev);
> > > >
> > > > Thanks!  I guess you verified that this actually *does* call all the
> > > > port service .remove() methods, right?  aer_remove(), dpc_remove(),
> > > > etc?
> > >
> > > I have tested, but aer_probe(), dpc_probe() doesn't get called at
> > > boot, so does aer_remove(), dpc_remove() when poweroff. I haven't got
> > > the root cause but I will continue to investigate.
> >
> > We'll only call aer_probe() and dpc_probe() if the port supports those
> > services and the platform has granted us control of them.  I don't
> > know if your platform does.  It may support PCIe native hotplug
> > (pcie_hp_init()) or PME (pcie_pme_init()).
>
> When I use pcie_ports=native to boot kernel, I verified that
> aer_remove() and pcie_pme_remove() are both called, while DPC and
> HOTPLUG are both not supported.

Great, thank you!