On Thu, 4 Aug 2016, Viresh Kumar wrote: > > > What about another patch on top of my patch to fix the deadlock? > > > > Or another patch in place of yours to fix both problems. Has your > > patch been merged yet? I don't see it in any of the branches in > > https://git.kernel.org/cgit/linux/kernel/git/gregkh/usb.git, and it's > > not in the current mainline. > > I don't think Greg has applied it yet. Do you want me to send the new patch? I'm concerned about the locking. hub_activate needs to be mutually exclusive with hub_quiesce. But hub_activate acquire the lock for the usb_interface, whereas callers of hub_quiesce (like hub_pre_reset and hub_event) acquire the lock for the usb_device. I can't remember why I did it that way. It looks like hub_activate really should acquire the usb_device lock. Also, although this isn't strictly necessary, the early-exit path should call usb_autopm_put_interface_async. Ideally, it should jump directly to the end of the routine, where these things are already being done. In short, it appears that fixing this properly requires a series of 3 patches. I'd better write them. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
