Tweak siop scripts assembler for new oosiop.

First cut at a esiop driver (enhanced siop). Doesn't implement tagged queuing
yet.
If is restricted to SIOP which implement the load/store instruction, and
has 10 scratch registers (basically, 825 and newer, possibly 770).
It implements a different interface between host and script, using a real
ring for command starts, and improved support for reconnect which will allow
256 tag per device. It uses interrupt on the fly to signal complete command,
which allows several commands to be serviced per interrupt and doesn't require
the script to stop to signal command completion.

MKfoo=no -> NOfoo

Now generate oosiop.out, the compiled 53c700 script.

branches: 1.2.2;
Add a MI driver for the Symbios/NCR 53c710 SCSI controller.
This is based on amiga's siop driver, but converted to use
bus_space(9) functions and modified to fit bus_dma(9) framework.

Currently tested on NetBSD/arc with jazzio 53c710 SCSI,
which really requires bus_dma(9) functions :-)
Sync transfers and disconnect/reconnect are also working.

TODO:
- Test under more heavy load
- Clean up osiop_checkintr() hander
- Reorganize command queue and sync negotiation handling more suitable
for thorpej-scsipi mid-layer
- Re-think defered interrupt handling for amiga

branches: 1.1.6; 1.1.8;
Snapshot of work in progress: new driver for the NCR 53c8xx SCSI controller
(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'.

Regen: save offset at disconnect after a data phase without save data pointer.

branches: 1.13.10; 1.13.12; 1.13.18;
Regen from esiop.ss 1.18

Regen from esiop.ss 1.17.

branches: 1.11.4;
Regen from esiop.ss 1.15.

branches: 1.10.4; 1.10.6; 1.10.8; 1.10.14;
Regen: move a few things around to save a few JUMP.

Regen: use a u_int32_t in script RAM to pass flags between script and driver.

Regen: implement done command ring.

Regen: 2 DSA entry per lun, load SCRATCHC before select, avoid race condition
when setting f_cmd_free.

Regen: get rid or ID in scheduler slot, to save RAM.

Regen: bump the number of slots in the cmd ring to 256.

Regen: bump message size to 16 octets.

Regen: block on wait MSG_OUT after a SELECT.

Regen: tagged queuing support, and led on/off fragments.

Generated from esiop.ss, for the new esiop driver.

s/beggining/beginning/ in comments.

Remove closes 3 & 4 from my licence. Lots of thanks to Soren Jacobsen
for the booring work !

branches: 1.20.74;
Some drives disconnect after the last data phase without a save data pointer
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.

branches: 1.19.2; 1.19.4; 1.19.10;
nuke trailing whitespace

branches: 1.18.4; 1.18.6;
when an unexpected disconnect occurs only compute the resid; do the
real save data pointers when we get the message (or rather, at disconnect
time following the message).
Factor out code to do this, and to deal with xs->resid, in siop_common.c.

Properly compute xs->resid, instead of assuming it'll always be 0 when
a command is done.

branches: 1.16.4;
Remove references to University of California from my copyright notices.

Properly handle messages that we get at reconnect time (other than
SIMPLE_Q_TAG). Some drives start sync/wide negotiation at reconnect time.
Problem reported and fix tested by Andreas Wrede.

branches: 1.14.6;
Fix typo in comment.

branches: 1.13.2; 1.13.4; 1.13.6; 1.13.8;
Move a few things around, to save some JUMP.

Use a u_int32_t in script RAM instead of the SEM bit in ISTAT to pass flags
between script and driver. This allows more than one flag, and is easier to
manage (we almost can't read/write istat outside of the irq handler).

Implement a ring for the completed commands. This avoid a (problematic only
if error occurs after status is collected) race condition
when using the status byte to detect completed commands (a command descriptor
could be recycled before the device disconnected), and make the
interrupt routine handling completed commands more efficient (no need to
scan target * lun * tag array any more).

- We can't share the per-lun DSA entry for untagged and tag table DSA;
there may be tagged commands still running when we queue a request sense
command.
Solve this by using 2 DSA entry per LUN
- Now that we have the command DSA before select, we can load T/L/Q in
SCRATCHC. This makes the selection timeout handler simpler.
- Avoid a race condition when setting the free flag in the cmd ring (see
comment in the script)
- don't forget to update the ID in the head of LUN table after a sync/wide
negotiation. This fixes the command timeout at the first data command
after negotiation (the bus reset handler did update the ID properly,
so subsequent commands were OK).
- for DMA interrupts, clear fifo if it's not empty. Leaving the fifo dirty
would prevent subsequent interrupts from coming in.
- Various improvements in debug messages
- misc cleanups.

For a new command, use the id in the command table and get rid of the ID in the
scheduler slot. This costs a few more instructions but divide the size of the
scheduler ring by 2, saving 1k of onboard RAM (a bus with 15 devices would
overflow the on-board RAM by 128 bytes).

Bump the number of slots in the command ring to 256.

More copyright fixes, pointed out by Thomas. Thanks !

Fix copyrigth.

Bump the space for SCSI messages from 8 to 16 octets, as 8 may not be
enouth for IDENTIFY + TAG + PPR.
Get rid of constants in C code by use of a offsetof macro.

It seems that the script processor continue to exec a few intructions
after the selection timeout is posted but the number executed isn't
reliable. So wait for MSG_OUT rigth after the select so that the state of
the script when the interruption is handled is known.

Add the led_on/led_off script fragments.

Add support for tagged queuing to esiop (256 tags per device).
For this add another indirecton: the DSA in the LUN table points to
a table of DSA indexed by the tag number when tagged command is in use.
For non tagged command, the LUN DSA still points to the tables describing the
xfer directly.

First cut at a esiop driver (enhanced siop). Doesn't implement tagged queuing
yet.
If is restricted to SIOP which implement the load/store instruction, and
has 10 scratch registers (basically, 825 and newer, possibly 770).
It implements a different interface between host and script, using a real
ring for command starts, and improved support for reconnect which will allow
256 tag per device. It uses interrupt on the fly to signal complete command,
which allows several commands to be serviced per interrupt and doesn't require
the script to stop to signal command completion.

s/substract/subtract/ in comments and error message.
s/obtainted/obtained/ in one comment.

Drop 3rd and 4th clauses (except on files where copyright is shared
with UC, in which case only the ad clause has been removed). Approved
by mhitch@ (copyright holder).

branches: 1.15.74;
merge ktrace-lwp.

White space nit- don't put a space before/after increment/decrement
operators.

branches: 1.13.4; 1.13.6;
Make it compile with recent compilers.

branches: 1.12.2;
add missing __RCSID()

branches: 1.11.2;
Tweak siop scripts assembler for new oosiop.

branches: 1.10.4;
- handle INTFLY (interrupt on the fly) command
- hanlde set/clear carry
- handle shl, shr (shift left and rigth) and xor operators

Fix scratchd2 address.

Really use a JUMP for NOP, not a block move.

Added support for the Symbios/NCR 53c700 SCSI controller.
This chip is very similar to the 53c710, except that it's
missing a few registers, doesn't do relative jumps and
doesn't have the table-indirect addressing mode.

I have checked that this still generates the same output
on the existing scripts for the other chips.

branches: 1.6.2;
Correct register address map for 53c710.

branches: 1.5.2; 1.5.4;
Allow conditionnal branch like "JUMP foo, IF FALSE"
The branch will never be taken as is, but the driver can patch the script.

Fix off by one error when using RETURN, WHEN/IF ...

Emit a warning when a symbol is undefined.

branches: 1.2.4;
- Add support for NOP instruction
- fix dsp value for second operand of memory move, so that patch list will be
correct.

Script assembler for the NCR 53c7xx/8xx scsi controller. Initial work
done by Michael L. Hitch, support for the 8xx and some improvements
by me.

Regen scripts for new oosiop.

Regen.

branches: 1.1.2; 1.1.4; 1.1.12; 1.1.14;
Generated from oosiop.ss rev. 1.1:
Initial compiled script for 53c700.

Add MI NCR/Symbios 53c700 SCSI driver.
This "oosiop" driver was originally written by Shuichiro URATA
for arc port, and then it was modified by me to make it work
also on hp700.

This driver has been tested on my NEC Express5800/240 with 53c700-66
for several months, and also tested on HP9000 735/125 with 53c700
(though current hp700 port has been broken since SA merge).
Both sync transfer and disconnect/reselect work fine,
but tagged queuing is not implemented yet.

branches: 1.1.2; 1.1.4;
New script for the Symbios/NCR 53c700 SCSI controller.
This is closely related to osiop.ss, the script for the
53c710.

branches: 1.1.2; 1.1.18;
Generated from osiop.ss rev. 1.1:
Initial compiled script for 53c710.

acknowleg -> acknowledg, proceedure -> procedure.
only comments were changed.

from miod

branches: 1.2.22; 1.2.40; 1.2.44;
Drop 3rd and 4th clauses (except on files where copyright is shared
with UC, in which case only the ad clause has been removed). Approved
by mhitch@ (copyright holder).

branches: 1.1.2; 1.1.126;
Add a MI driver for the Symbios/NCR 53c710 SCSI controller.
This is based on amiga's siop driver, but converted to use
bus_space(9) functions and modified to fit bus_dma(9) framework.

Currently tested on NetBSD/arc with jazzio 53c710 SCSI,
which really requires bus_dma(9) functions :-)
Sync transfers and disconnect/reconnect are also working.

TODO:
- Test under more heavy load
- Clean up osiop_checkintr() hander
- Reorganize command queue and sync negotiation handling more suitable
for thorpej-scsipi mid-layer
- Re-think defered interrupt handling for amiga

Regen: save offset at disconnect after a data phase without save data pointer.

branches: 1.15.10; 1.15.12; 1.15.18;
Regen from siop.ss 1.18

branches: 1.14.4; 1.14.10; 1.14.14;
Regen: bump message size to 16 octets.

Regen: add hooks to drive LED via GPIO.

Regen.

branches: 1.11.2; 1.11.4; 1.11.6;
Regen

Regen

Regen

Regen

Regen

branches: 1.6.4;
Regen.

Regen

Ops, regen siop.out from the last siop.ss.

Add a flag hold in scratcha0, set when we're doing a data transfer. Used
for save data pointers.

- Change the script to start new commands in an asyncronous way, using
'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)

Snapshot of work in progress: new driver for the NCR 53c8xx SCSI controller
(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'.

Remove closes 3 & 4 from my licence. Lots of thanks to Soren Jacobsen
for the booring work !

branches: 1.20.74;
Some drives disconnect after the last data phase without a save data pointer
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.

branches: 1.19.2; 1.19.4; 1.19.10;
nuke trailing whitespace

branches: 1.18.4; 1.18.6;
when an unexpected disconnect occurs only compute the resid; do the
real save data pointers when we get the message (or rather, at disconnect
time following the message).
Factor out code to do this, and to deal with xs->resid, in siop_common.c.

branches: 1.17.6; 1.17.10;
Spell '[Rr]ight' correctly. From Jim Bernard.

enouth -> enough.

branches: 1.15.2; 1.15.4;
More copyright fixes, pointed out by Thomas. Thanks !

Bump the space for SCSI messages from 8 to 16 octets, as 8 may not be
enouth for IDENTIFY + TAG + PPR.
Get rid of constants in C code by use of a offsetof macro.

Add hooks to drive the activity led via the GPIO pin 1, as required by some
tekram and symbios adapters.

branches: 1.12.2; 1.12.4; 1.12.6;
Rearrange things a bit to have less jumps
Install the core of the scheduler in main script (so it's in RAM when there is
one), and avoid jump in the common case. The command part of the scheduler now
lives in host memory, with tables.
Add template for a tag switch.

Move restore_scntl3 before the lun switch, to ease freeing of unused lun
slots.

Redo the target/lun sw in a better way, and handle queue tag messages.

Fix recurent typo: shed->sched

Handle reselect entirely in script; for this provide a script area,
modified by the driver, which will load the appropriate DSA for a given
target/lun.

branches: 1.7.2;
sheduler->scheduler, as pointed out by Klaus Klein

branches: 1.6.2;
Separate the sheduler from the main script, allocate another DMA-safe
memory page for the sheduler. Put the main script in the on-chip RAM when
available.
Avoid a null-pointer dereference when DSA is invalid.

- use my standart copyrigth notice
- After reselect, don't clear ack after the IDENTIFY message, this breaks
when the reselecting target has different sync parameters than the previous
one.

Add a flag hold in scratcha0, set when we're doing a data transfer. Used
for save data pointers.

Add NetBSD RCS Id and copyrigth notice.

- Change the script to start new commands in an asyncronous way, using
'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)

Snapshot of work in progress: new driver for the NCR 53c8xx SCSI controller
(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'.