[PATCH] USB: wakeup flag updates (2/3) uhci-hcd
authorDavid Brownell <david-b@pacbell.net>
Mon, 7 Nov 2005 23:34:41 +0000 (15:34 -0800)
committerGreg Kroah-Hartman <gregkh@suse.de>
Wed, 4 Jan 2006 21:48:30 +0000 (13:48 -0800)
This makes UHCI stop using the HCD glue wakeup flags to report whether
the controller can wake the system.  The existing code was wrong anyway;
having a PCI PM capability doesn't imply it reports PME# is supported.

I skimmed Intel's ICH7 datasheet and that basically says the wakeup
signaling gets routed only through ACPI registers.  (On the other hand,
many VIA chips provide the PCI PM capabilities...)  I think that doing
this correctly with UHCI is going to require the ACPI folk to associate
the /proc/acpi/wakeup identifiers (and wakeup enable/disable flags)
with the relevant /sys/devices/pci*/...  devices.

From: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
drivers/usb/host/uhci-hcd.c

index 79efaf7d86a398a9806b44efb8e12097c8cff501..5589d4010e36cdc73c3870817f7bc6299af879d9 100644 (file)
@@ -478,8 +478,6 @@ static int uhci_start(struct usb_hcd *hcd)
        struct dentry *dentry;
 
        hcd->uses_new_polling = 1;
-       if (pci_find_capability(to_pci_dev(uhci_dev(uhci)), PCI_CAP_ID_PM))
-               hcd->can_wakeup = 1;            /* Assume it supports PME# */
 
        dentry = debugfs_create_file(hcd->self.bus_name,
                        S_IFREG|S_IRUGO|S_IWUSR, uhci_debugfs_root, uhci,