usb_hcd_flush_endpoint() has a retry loop that starts with a spin_lock_irq(), but only gives up the spinlock, not the irq_disable before jumping to the rescan label.
Split the spin_lock_irq into the retryable part and the local_irq_disable() that is only done once as a micro-optimization and slight cleanup. Signed-off-by: Arnd Bergmann <[EMAIL PROTECTED]> --- On Tuesday 23 October 2007, I wrote: > I tried the trivial annotation below and (with lockdep enabled) got a few > warnings at boot time, but only one that I could still find in the log > buffer: One more such example that was not found by lockdep. I guess this counts as a false positive, as it is clearly harmless, but working around it is a small optimization for the case where local_irq_disable() is a hypervisor call. Should we try to fix this class of (non-)problem in other places? Will this patch cause a different warning with lockdep since now we are pairing spin_lock() with spin_unlock_irq()? --- a/drivers/usb/core/hcd.c +++ b/drivers/usb/core/hcd.c @@ -1312,8 +1312,9 @@ void usb_hcd_flush_endpoint(struct usb_device *udev, hcd = bus_to_hcd(udev->bus); /* No more submits can occur */ + local_irq_disable(); rescan: - spin_lock_irq(&hcd_urb_list_lock); + spin_lock(&hcd_urb_list_lock); list_for_each_entry (urb, &ep->urb_list, urb_list) { int is_in; - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/