I think I may have found a slight problem with IEEE1284 (ECP) direction
changing in ieee1284_ops.c. Consider the following scenario:
Initially the phase is IEEE1284_PH_FWD_IDLE following a write. A read is
posted that causes ecp_forward_to_reverse to be entered. If this fails (in
this case due to faulty firmware on the peripheral) the read returns with 0
bytes read (no error?). At this point the hardware may be in an indeterminate
state. A subsequent write does not attempt to change direction because the
state is still IEEE1284_PH_FWD_IDLE. See code snippet:
if (port->ieee1284.phase != IEEE1284_PH_FWD_IDLE)
if (ecp_reverse_to_forward (port))
I would propose adding another state IEEE1284_PH_FWD_UNKNOWN that would be
raised on such error conditions, thus forcing subsequent operations to retry.
Also, why does the read operation apparently succeed?
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2b29 : Tue Jul 31 2001 - 10:29:23 EDT