Also, adjust message/command construction in ncr53c9x_select() to work no
matter how the structure alignment works out (needed at least for m68k).
Tested by me on mac68k & alpha, and sanity-checked by eeh.
which is revealed with larger HZ systems like NetBSD/pmax (256Hz)
and NetBSD/alpha (1024Hz) as reported by PR#8645. Polled tape
drive access is done with maximum 6 hour timeout which ended up
with negative time and then confused SCSI bus severely.
and scsipi_device respectively, with size reduction of ncr53c9x_softc.
Specifying NULL instructs the driver to use default adapter and default
device codes. Every target port has ncr53c9x_attach(sc, NULL, NULL) anyway.
timeout()/untimeout() API:
- Clients supply callout handle storage, thus eliminating problems of
resource allocation.
- Insertion and removal of callouts is constant time, important as
this facility is used quite a lot in the kernel.
The old timeout()/untimeout() API has been removed from the kernel.
it implemented under the label `shortcut:' and only use it in these
cases: (1) after successful re-relection, (2) after receiving command-complete
status, and (3) during message-in handshake.
for front-ends to override the allocation to avoid alignment
handling in their DMA engines. Note that that ncr53c9x_msgout()
can request a 1 byte DMA transfer that would be difficult to break up.
an initiator/target nexus and thus mark the correct queue (if any) a
command is on.
* If a disconnected command times out, just leave it on the nexus queue
and do nothing (for now). I need yet to decide on the strategy to
follow in this case. Note: we used to move the command to the `ready'
queue and then do nothing, which is worse.
the target still is in MSG OUT phase. We still send a message (a NO_OP)
in this case and the chip will remove ATN at the appropriate time.
Using the RSTATN command here induces a "illegal command" in some
chip revisions.
This situation only occurs if the target rejects a previous (multi-byte)
message early (by switching to MESSAGE IN and sending a MESSAGE REJECT)
before the chip has completed the entire MSG OUT transfer. ATN will
remain asserted, and the target returns to MESSAGE OUT phase.
* Account for the events above when reporting "DMA not completed"
diagnostic messages.
* Stream-line the selection code a bit, and make the DMA setup code
more like the MSG OUT & DATA XFER setup.
"A sequence step of 0 after a select with ATN can be a selection
timeout, or it can also indicate the target did not respond with
a message out phase. The latter will occur on very old SCSI
devices which do not respond to the ATN signal and go directly to
the command phase".
(currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2
busses to which devices can attach (scsibus and atapibus). This needed to
change some include files and structure names in the low level scsi drivers.
zero. We use this later when the COMPLETE message comes in to set the
final residual count to zero. The flag is reset if the target resumes
a data phase. This obsoletes the `AUTOSAVE' quirk (for this driver).
Also, avoid overwriting the residual count if a SENSE was appended to
the current transaction.
(a) The interrupt is a RESEL interrupt, and
(b) our state is SELECTING.
This condition can occur in perfectly normal operation if we are using
DMA to select the target and we are interrupted by another target
reselecting us. Per discussion with Paul Krannenburg.
we return from the driver expecting to come back due to an
interrupt, because the interrupt might not happen...
Do the untimeout in ncr53c9x_done instead of just before
almost every call to ncr53c9x_done as was done previously.
Make ncr53c9x_sense schedule its own timeout for the new
command it is starting (request sense), separate from the
timeout for the command that just completed.
Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.
For the detailed change history, look at the commit log entries for
the is-newarp branch.