Re: [PATCH 2/2] xhci: EHCI/XHCI ports switching on Intense-PC.

[Sarah Sharp <sarah.a.sharp@linux.intel.com>]

Hi Denis,

I found a couple issues with your second patch.  There are non-PCI xHCI
host controllers, so we can't assume the xHCI host is a PCI host.  So
this code can't run in the generic xHCI shutdown method:

>  void xhci_shutdown(struct usb_hcd *hcd)
>  {
> +	struct pci_dev *pdev;
> +	const char *brd_name;
>  	struct xhci_hcd *xhci = hcd_to_xhci(hcd);
> 
> +	pdev = to_pci_dev(hcd->self.controller);
> +
> +	if (usb_is_intel_switchable_xhci(pdev)) {
> +		brd_name = dmi_get_system_info(DMI_BOARD_NAME);
> +
> +		/* quirk for Compulab's Intense-PC board */
> +		if (brd_name && strstr(brd_name, "Intense-PC"))
> +			usb_disable_xhci_ports(pdev);
> +	}
> +
>  	spin_lock_irq(&xhci->lock);
>  	xhci_halt(xhci);
>  	spin_unlock_irq(&xhci->lock);

I think the to_pci_dev call will oops if this isn't a PCI host.  Plus,
I've found out from the BIOS and chipset folks that this quirk needs to
be more broadly applied.  I'll send you a revised patch.  Can you test
it and make sure it works for you?

Sarah Sharp