This is a completely rewritten scsipi_xfer execution engine, and the
associated changes to HBA drivers. Overview of changes & features:
- All xfers are queued in the mid-layer, rather than doing so in an
ad-hoc fashion in individual adapter drivers.
- Adapter/channel resource management in the mid-layer, avoids even trying
to start running an xfer if the adapter/channel doesn't have the resources.
- Better communication between the mid-layer and the adapters.
- Asynchronous event notification mechanism from adapter to mid-layer and
peripherals.
- Better peripheral queue management: freeze/thaw, sorted requeueing during
recovery, etc.
- Clean separation of peripherals, adapters, and adapter channels (no more
scsipi_link).
- Kernel thread for each scsipi_channel makes error recovery much easier
(no more dealing with interrupt context when recovering from an error).
- Mid-layer support for tagged queueing: commands can have the tag type
set explicitly, tag IDs are allocated in the mid-layer (thus eliminating
the need to use buggy tag ID allocation schemes in many adapter drivers).
- support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command
will be requeued, or a REQUEST SENSE will be sent as appropriate.
Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
This bug is only visible on some disks, and I have verified that it works
correctly on VAX. However, du to the author, it is a problem on other
machines/disks as well. Here's the author's comments:
> The MI code for the NCR5380 has a small bug in it The RZ56 wants to
> negotiate for Synchronous Data Transfers and when it does, the current code
> looks like it will send a REJECT message as the spec calls for when we cant
> handle the request. However, a couple of lines of code in the
> ncr5380_msg_in routine are missing wrt the PARITY message and anything that
> wants to send a REJECT message. The ATN setup that is done in
> ncr_sched_msgout is immediately negated by the final byte ACK sequence in
> this routine. This causes bad things to happen to an RZ56.
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.
use PIO (or equivalent) code to do asynchronous transfers: In
ncr5380_scsi_cmd(), test to see if the request has been completed after
the call to ncr5380_sched(), and return COMPLETE if so. This avoids
going into an infinite loop in scsipi_execute_xs() while waiting for an
interrupt to trigger completion of the transfer... which, of course,
never happens, since it's already done.
(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.
particularly older ones, don't do a SAVE DATA POINTER before disconnecting.
When the driver reconnects, it does an implicit RESTORE POINTERS, and
restores the data pointer from *before* the last transfer. If the driver
calculates the residual using the data pointer and doesn't account for
this, the residual will be wrong.
What we do is test for the SDEV_AUTOSAVE quirk when we see a disconnect
message, falling through to do a SAVE DATA POINTER if the quirk applies
to this target.
The workaround was suggested by Mycroft, who also provided the explanation
of the problem. Actual code was (mostly) ripped off from the 53C9x driver.
bits that tell the driver not to expect parity from those targets.
XXX: For now, it is up to MD code to set this mask. Eventually,
XXX: we should identify targets first (without demanding parity)
XXX: and then use a new "SCSI quirk" to set the right bit.
detatch everywhere else. Reverted to the english spelling.
Also fixed the rest of the prototype warnings while I was at it.
ic/ncr5380sbc.c: Don't declare Debugger()... I have to clean this
everywhere :-(
simonb