On Fri, 28 Aug 2015, Roland Weber wrote:
> Hmm, looks like my last two mails didn't get sent to the list.
> So, with some delay...
>
> > > If you want, you could try an even finer bisection. The commit you
> > > identified adds a mutex_lock and a mutex_unlock, and it also changes an
> > > alloc_ordered_workqueue to alloc_workqueue. You could leave the mutex
> > > stuff out and just include the alloc_workqueue change, or vice versa.
>
> I've tried both. The kernel with just the mutex change is good. The one
> which only changes alloc_ordered_workqueue to alloc_workqueue is bad.
Interesting. It's possible (though unlikely) that your problem arises
from a bug in the workqueue code rather than in the USB stack.
> Does that mean I'll have to dig into hardware specs next, to see if
> there's some restriction on the order in which ports must be probed
> or reset on that specific USB hub?
No hubs have any restrictions of that sort. Besides, changing
alloc_ordered_workqueue to alloc_workqueue could affect the order in
which different hubs are managed, but it can't affect the order in
which ports within a single hub are handled.
> Or maybe I can gather some more
> debug information about the initialization? As long as I'm the only one
> bitten by this change, I wouldn't ask to simply revert it.
You can try the debugging patch below. It will tell us when the items
on the workqueue get executed, which could be a useful clue. You
should test this from a VT console with the logging level set high
enough so that the log messages show up on the screen. That way you'll
be able to see them even after a hang.
Alan Stern
Index: usb-4.2/drivers/usb/core/hub.c
===================================================================
--- usb-4.2.orig/drivers/usb/core/hub.c
+++ usb-4.2/drivers/usb/core/hub.c
@@ -4990,6 +4990,8 @@ static void hub_event(struct work_struct
hub_dev = hub->intfdev;
intf = to_usb_interface(hub_dev);
+ dev_info(hub_dev, "entry\n");
+
dev_dbg(hub_dev, "state %d ports %d chg %04x evt %04x\n",
hdev->state, hdev->maxchild,
/* NOTE: expects max 15 ports... */
@@ -5093,6 +5095,8 @@ out_autopm:
out_hdev_lock:
usb_unlock_device(hdev);
+ dev_info(hub_dev, "exit\n");
+
/* Balance the stuff in kick_hub_wq() and allow autosuspend */
usb_autopm_put_interface(intf);
kref_put(&hub->kref, hub_release);
--
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
