Work around a remote wakeup vs port poweroff request race. A wakeup request
sets the timer, but by the time it expires the port has been turned off, so we
hit:
if ((raw_port_status & PORT_RESET) ||
!(raw_port_status & PORT_PE))
return 0xffffffff;
...in xhci_get_port_status
When userspace has set the port policy to allow poweroff the assumption is that
it no longer cares about remote wakeup requests for that port. If a request
happens to collide with autosuspend of the port we need to make sure to cancel
an in-flight request otherwise we can end up in an infinite loop.
Signed-off-by: Dan Williams <dan.j.willi...@intel.com>
---
drivers/usb/host/xhci-hub.c | 12 ++++++++++--
1 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/usb/host/xhci-hub.c b/drivers/usb/host/xhci-hub.c
index 9992fbfec85f..2333ec573594 100644
--- a/drivers/usb/host/xhci-hub.c
+++ b/drivers/usb/host/xhci-hub.c
@@ -1004,9 +1004,16 @@ int xhci_hub_control(struct usb_hcd *hcd, u16 typeReq,
u16 wValue,
xhci_disable_port(hcd, xhci, wIndex,
port_array[wIndex], temp);
break;
- case USB_PORT_FEAT_POWER:
- writel(temp & ~PORT_POWER, port_array[wIndex]);
+ case USB_PORT_FEAT_POWER: {
+ struct xhci_bus_state *bus_state;
+ bus_state = &xhci->bus_state[hcd_index(hcd)];
+ writel(temp & ~PORT_POWER, port_array[wIndex]);
+ if (test_and_clear_bit(wIndex,
&bus_state->resuming_ports)) {
+ bus_state->resume_done[wIndex] = 0;
+ xhci_dbg(xhci, "port%d: resume cancelled\n",
+ wIndex);
+ }
spin_unlock_irqrestore(&xhci->lock, flags);
temp = usb_acpi_power_manageable(hcd->self.root_hub,
wIndex);
@@ -1015,6 +1022,7 @@ int xhci_hub_control(struct usb_hcd *hcd, u16 typeReq,
u16 wValue,
wIndex, false);
spin_lock_irqsave(&xhci->lock, flags);
break;
+ }
default:
goto error;
}
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
