usb: gadget: f_fs: Add unbind event before functionfs_unbind

commit efb6b53 upstream.

While exercising the unbind path, with the current implementation
the functionfs_unbind would be calling which waits for the ffs->mutex
to be available, however within the same time ffs_ep0_read is invoked
& if no setup packets are pending, it will invoke function
wait_event_interruptible_exclusive_locked_irq which by definition waits
for the ev.count to be increased inside the same mutex for which
functionfs_unbind is waiting.
This creates deadlock situation because the functionfs_unbind won't
get the lock until ev.count is increased which can only happen if
the caller ffs_func_unbind can proceed further.

Following is the illustration:

	CPU1				CPU2

ffs_func_unbind()		ffs_ep0_read()
				mutex_lock(ffs->mutex)
				wait_event(ffs->ev.count)
functionfs_unbind()
  mutex_lock(ffs->mutex)
  mutex_unlock(ffs->mutex)

ffs_event_add()

<deadlock>

Fix this by moving the event unbind before functionfs_unbind
to ensure the ev.count is incrased properly.

Fixes: 6a19da1 ("usb: gadget: f_fs: Prevent race during ffs_ep0_queue_wait")
Cc: stable <stable@kernel.org>
Signed-off-by: Uttkarsh Aggarwal <quic_uaggarwa@quicinc.com>
Link: https://lore.kernel.org/r/20230525092854.7992-1-quic_uaggarwa@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/usb/gadget/function/f_fs.c

@@ -3620,6 +3620,7 @@ static void ffs_func_unbind(struct usb_configuration *c,
 	/* Drain any pending AIO completions */
 	drain_workqueue(ffs->io_completion_wq);
 
+	ffs_event_add(ffs, FUNCTIONFS_UNBIND);
 	if (!--opts->refcnt)
 		functionfs_unbind(ffs);
 
@@ -3644,7 +3645,6 @@ static void ffs_func_unbind(struct usb_configuration *c,
 	func->function.ssp_descriptors = NULL;
 	func->interfaces_nums = NULL;
 
-	ffs_event_add(ffs, FUNCTIONFS_UNBIND);
 }
 
 static struct usb_function *ffs_alloc(struct usb_function_instance *fi)