driver	where dmemem is allocated in	variations of mutex	if_init lock	if_start lock	if_stop lock	if_ioctl lock	softint(9)based TX and RX?	TX and TX are locked with mutex	callout lock	notes
wm(knakahara)	in attach()	tx, rx, core	CORE_LOCK(spin mutex)	TX_LOCK(spin mutex) + sc_stopping flag	CORE_LOCK(spin mutex)+sc_stopping flag	use CORE_LOCK() and TX_LOCK() partially	no	TX and RX differently	TX_LOCK(mutex) and check sc_stopping in the beginning
vioif	in attach()	tx, rx, ctrl_wait	not taken?	TX_LOCK(mutex)+sc_stopping check	TX, RX + sc_stopping check	splnet	only RX is softint'ed	TX and RX differently	callout isn't used
vmxnet
pq3etsec	in attach()	lock, hwlock, mdio_lock	not taken?	not taken. Set flag with atomic_or_uint() and call softintr_schedule()	not taken?	splnet()	Yes	sc_hwlock is for hard interrput. sc_lock is for software interrupt. TX and RX use the same lock	mutex_enter(sc_lock)
bcm53xx_eth	in attach()	lock, hwlock	not taken?	not taken. Set flag with atomic_or_uint() and call softintr_schedule()	not taken?	splnet()	Yes	sc_hwlock is for hard interrput. sc_lock is for software interrupt. TX and RX use the same lock	callout isn't used
ixg	in init() for rx jumbo, in attach() for others	tx, rx, core	CORE_LOCK(adaptive mutex)	TX_LOCK(spin mutex)	CORE_LOCK(spin mutex)	use core lock for init() and set_multi()	Yes. Only one time is execulted in the interupt context. If more request exist, softint is issued.	TX and RX differently	CORE_LOCK() in the beginning
bnx	in attach()	tx	splnet() in the beginning	not taken?	not taken?	splnet() in the beginning	No	only TX uses mutex	splnet() in the beginning
bge	in attach()	none	splnet() in the beginning	not taken?	not taken?	splnet() in the beginning	No	No	splnet() in the beginning	mutex is not used at all