Hi Elric, I was looking at the xHCI command cancellation code, and I realized that xhci_cancel_cmd() calls xhci_abort_cmd_ring() with the xHCI spinlock held. That function spins for up to 5 seconds, waiting for the command ring running bit to be clear.
We really can't have the spinlock held for that long, since other things like port status change events can cause the interrupt handler to be called (which would attempt to acquire the spinlock). Would it be OK to drop the spinlock in xhci_abort_cmd_ring() before spinning on the CRR? Or were you trying to protect some other xHCI structures/registers with the lock while you waited to abort the command ring? It's commit ID b92cc66c047ff7cf587b318fe377061a353c120f "xHCI: add aborting command ring function" if you need to refresh your memory. Sarah Sharp -- 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
