Iñaky Pérez-González
2008-10-20 23:31:44 UTC
The current WUSB device reset code uses a hard reset via the Reset
Device IE which causes the device to enter the UnConnected state and
(potentially) reconnect to a different fake port. This breaks USB
device drivers with .pre_reset and .post_reset methods.
Instead, I propose doing a soft reset of authenticated devices (the only
state where this is valid)
Makes sense to me -- don't have the spec at hand, but if it is mentionedDevice IE which causes the device to enter the UnConnected state and
(potentially) reconnect to a different fake port. This breaks USB
device drivers with .pre_reset and .post_reset methods.
Instead, I propose doing a soft reset of authenticated devices (the only
state where this is valid)
setaddr(0) / setaddr(addr) equates a soft reset, it should be fine.
One thing to bear in mind is how well devices will respond to this. In
theory reset means "do a full device reset". How true is this when doing
a soft reset might confuse a few drivers. Being pesimistic here, but you
might have a to have a plan B.
and simply ignoring reset requests for
unauthenticated devices (these will eventually time out after
TrustTimeout and disconnect themselves so they cannot get stuck in a
funny state).
I'd keep the reset IE for those guys; as mentioned before, when we sayunauthenticated devices (these will eventually time out after
TrustTimeout and disconnect themselves so they cannot get stuck in a
funny state).
reset, we expect an inmediate reset.
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-***@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html