USB: Don't rebind before "complete" callback

This patch (as1130) fixes an incompatibility between the new PM
infrastructure and USB power management.  We are not allowed to call
drivers' probe routines during a system sleep transition between the
"prepare" and "complete" callbacks, but that's exactly what we do when
a driver doesn't have full suspend/resume support.  Such drivers are
unbound during the "suspend" call and reprobed during the "resume" call.

The patch causes the reprobe step to be skipped if the "complete"
callback hasn't been issued yet, i.e., if the interface's
dev.power.status field is not equal to DPM_ON.  Thus during the
"resume" callback nothing bad will happen, and during the final
"complete" callback the reprobing will occur as desired.

This fixes the problem reported in Bugzilla #11263.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

1 files changed, 9 insertions, 21 deletions

diff --git a/drivers/usb/core/driver.c b/drivers/usb/core/driver.c
index 2da70b4d33f..5a7fa6f0995 100644
--- a/drivers/usb/core/driver.c
+++ b/drivers/usb/core/driver.c
@@ -814,7 +814,8 @@ void usb_forced_unbind_intf(struct usb_interface *intf)
  * The caller must hold @intf's device's lock, but not its pm_mutex
  * and not @intf->dev.sem.
  *
- * FIXME: The caller must block system sleep transitions.
+ * Note: Rebinds will be skipped if a system sleep transition is in
+ * progress and the PM "complete" callback hasn't occurred yet.
  */
 void usb_rebind_intf(struct usb_interface *intf)
 {
@@ -830,10 +831,12 @@ void usb_rebind_intf(struct usb_interface *intf)
         }
 
         /* Try to rebind the interface */
-        intf->needs_binding = 0;
-        rc = device_attach(&intf->dev);
-        if (rc < 0)
-                dev_warn(&intf->dev, "rebind failed: %d\n", rc);
+        if (intf->dev.power.status == DPM_ON) {
+                intf->needs_binding = 0;
+                rc = device_attach(&intf->dev);
+                if (rc < 0)
+                        dev_warn(&intf->dev, "rebind failed: %d\n", rc);
+        }
 }
 
 #ifdef CONFIG_PM
@@ -845,7 +848,6 @@ void usb_rebind_intf(struct usb_interface *intf)
  * or rebind interfaces that have been unbound, according to @action.
  *
  * The caller must hold @udev's device lock.
- * FIXME: For rebinds, the caller must block system sleep transitions.
  */
 static void do_unbind_rebind(struct usb_device *udev, int action)
 {
@@ -867,22 +869,8 @@ static void do_unbind_rebind(struct usb_device *udev, int action)
                                 }
                                 break;
                         case DO_REBIND:
-                                if (intf->needs_binding) {
-
-        /* FIXME: The next line is needed because we are going to probe
-         * the interface, but as far as the PM core is concerned the
-         * interface is still suspended.  The problem wouldn't exist
-         * if we could rebind the interface during the interface's own
-         * resume() call, but at the time the usb_device isn't locked!
-         *
-         * The real solution will be to carry this out during the device's
-         * complete() callback.  Until that is implemented, we have to
-         * use this hack.
-         */
-//                                      intf->dev.power.sleeping = 0;
-
+                                if (intf->needs_binding)
                                         usb_rebind_intf(intf);
-                                }
                                 break;
                         }
                 }