Do you think there's any chance of getting the
root cause fixed?

As I recall, this was a workaround for a bad fix.
Some descriptor (hub?) had two adjacent bitfields
with one bit per port (?), and the bad fix removed
a moderately clean representation thereof.  (Maybe
this is what a comment from your patch describes:

 
-    /* two bitmaps:  ports removable, and usb 1.0 legacy PortPwrCtrlMask */
+    /* ports removable, and usb 1.0 legacy PortPwrCtrlMask */

Yes, that #define has always been ugly, but at the
time it seemed a bit less ugly thanjust ignoring
therelevant spec and disregarding the bitmaps.
--

[ message continues ]