message. In such case we would not update resid with the proper value
(eventually resid would not be updated at all if there was only one data
phase). To fix this, have the script save the offset in the data tables at
disconnect time if there was a transfer, and use this to compute the resid
if the current offset is 0.
Problem reported and patch tested by edwin, Roy Bixler and YAMAMOTO Takashi.
Fix kern/31990 by YAMAMOTO Takashi.
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
- move status handling in siop_scsicmd_end(), it's better than in siop_intr()
- define 2 internal SIOP status, for "no status reported by device" and
reset condition
- add a list of "urgent" command, to be executed before the list of command
queued the normal way; this is used for command which got aborted
by a QUEUE FULL and have to be requeued in order.
- Don't accept to send a Q_TAG message not immediatly folowing a IDENTIFY
Implement SCBUSACCEL ioctl.
Snapshot of work in progress on tagged queuing: we can send/receive
queue tag messages. Infrastructure to manage multiple commands per
devices not here yet.
'command slots' in which the host can put command and wait for the script
to start them
- Change siop.c to do full disconnect/reslelect, allowing as much as one
command per target/lun to run in parallel.
- Fix bug in registers init where a board without BIOS would end at
ID 0 (now the driver works on alpha too).
- better handling of messages, sending back a MSG_EXT_SDTR in response to an
incoming MSG_EXT_SDTR, and MSG_MESSAGE_REJECT for unhandled messages.
- fix use of bus_dmamap_sync() and htole32().
- supports shared interrups
- change some int8 and int16 to int, for alpha and mips benefits ( suggested by
Toru Nishimura)
(the name 'siop' is still being discussed, may change).
Only basic disconnect/reselect for now, no sync/wide negotiation.
Tested with 810A, 875 and 895 on i386 only.
The bus-independant part should also be able to handle the 53c720 and 53c770.
A new driver with enhanced script should appear for the 825/875/895 'soon'.
