On Tue, 20 Aug 2013, Steve Cotton wrote:
> Last working release: 3.10
> Bisected earliest broken: c1117afb85
> Broken in: 3.11-rc1, 3.11-rc6, bd479f2933 (the current usb-next)
>
> Once my PC has been hibernated and resumed, some devices plugged
> in to the on-motherboard ports stop working, and don't show up in
> lsusb. The pattern seems to be that ports with USB 1.1 devices
> are affected, but ports with USB 2.0 devices are not.
>
> It's an x86-amd64 desktop, with an MSI motherboard based on the
> AMD 970 chipset. Ports in USB 1.1 mode use the ohci-pci driver.
>
> I've bisected the problem to c1117afb85, included in v3.11-rc1.
> c1117afb85 USB: OHCI: make ohci-pci a separate driver
>
> If the PC is booted without a resume image, everything works.
>
> It's running Debian unstable with a local kernel build, and to
> hibernate I'm using the "pm-hibernate" script from pm-utils.
>
> For a sanity check of my local builds, I've seen the same
> behaviour in the Debian archive's build of 3.11-rc4.
My god, I can't believe this got left out! The lines in the patch
below should have been part of that c1117afb85 commit all along. It's
amazing that nobody has complained before this.
Does this patch fix the problem?
Alan Stern
Index: usb-3.11/drivers/usb/host/ohci-pci.c
===================================================================
--- usb-3.11.orig/drivers/usb/host/ohci-pci.c
+++ usb-3.11/drivers/usb/host/ohci-pci.c
@@ -304,6 +304,11 @@ static int __init ohci_pci_init(void)
pr_info("%s: " DRIVER_DESC "\n", hcd_name);
ohci_init_driver(&ohci_pci_hc_driver, &pci_overrides);
+
+ /* Entries for the PCI suspend/resume callbacks are special */
+ ohci_pci_hc_driver.pci_suspend = ohci_suspend;
+ ohci_pci_hc_driver.pci_resume = ohci_resume;
+
return pci_register_driver(&ohci_pci_driver);
}
module_init(ohci_pci_init);
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
