Hi Peter,
You may ignore my previous minute patch and give a try with the below one.
It makes more sense to ignore the period (synchronous period factor) when
the offset (REQ/ACK offset) is zero, facing synchronous data transfer
negotiation in SCSI. Even if the period should be ignored by SCSI devices
when the offset is zero, I would prefer the driver to behave correctly by
design here (i.e.: ignore the period value when offset is zero).
The below patch just does so in the 2 places the driver checks against a
needed sync. negotiation.
Btw, I would consider as broken any SCSI code that makes decision about
data transfer being asynchronous on period value = 0. ;-)
--- sym_hipd.c.021201 Sun Dec 1 16:09:55 2002
+++ sym_hipd.c Sun Dec 1 16:13:23 2002
@@ -3074,8 +3074,9 @@
/*
* negotiate synchronous transfers?
*/
- else if (tp->tinfo.current.period != tp->tinfo.goal.period ||
- tp->tinfo.current.offset != tp->tinfo.goal.offset)
+ else if (tp->tinfo.current.offset != tp->tinfo.goal.offset ||
+ (tp->tinfo.goal.offset &&
+ tp->tinfo.current.period != tp->tinfo.goal.period))
nego = NS_SYNC;
switch (nego) {
@@ -7838,8 +7839,9 @@
*/
cp->nego_status = 0;
if (tp->tinfo.current.width != tp->tinfo.goal.width ||
- tp->tinfo.current.period != tp->tinfo.goal.period ||
tp->tinfo.current.offset != tp->tinfo.goal.offset ||
+ (tp->tinfo.goal.offset &&
+ tp->tinfo.current.period != tp->tinfo.goal.period)||
tp->tinfo.current.options != tp->tinfo.goal.options) {
if (!tp->nego_cp && lp)
msglen += sym_prepare_nego(np, cp, 0, msgptr + msglen);
------------------------ CUT HERE ------------------------
Let me know if this patch makes the driver work with your
asynchronous-only SCSI scanner.
Gérard.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message
