On Tue, 3 Oct 2006 16:18:35 +0100 (BST)
"Maciej W. Rozycki" [off-list ref] wrote:

2. The driver uses schedule_work() for handling interrupts, but does not 
   make sure any pending work scheduled thus has been completed before 
   driver's structures get freed from memory.  This is especially 
   important as interrupts may keep arriving if the line is shared with 
   another PHY.

   The solution is to ignore phy_interrupt() calls if the reported device 
   has already been halted and calling flush_scheduled_work() from 
   phy_stop_interrupts() (but guarded with current_is_keventd() in case 
   the function has been called through keventd from the MAC device's 
   close call to avoid a deadlock on the netlink lock).

eww, hack.

Also not module-friendly:

WARNING: "current_is_keventd" [drivers/net/phy/libphy.ko] undefined!

Does this

static void flush_cpu_workqueue(struct cpu_workqueue_struct *cwq)
{
	if (cwq->thread == current) {
		/*
		 * Probably keventd trying to flush its own queue. So simply run
		 * it by hand rather than deadlocking.
		 */
		run_workqueue(cwq);

not work???