scsi "killer" ?

[Amiga1992]

Originally Posted by Pat the Cat

the poster asked because they wanted to reduce boot times, not increase them as you advise.

I advised what? you really are making things up real hardcore.
I'll leave you to yourself and the moderation team, not posting on this thread anymore.

[kolla]

I would advice people to not take advice from Pat the Cat, the stuff he posts is mostly wrong and irrelevant. Which is damn annoying, because I think he really wants to be helpful.

[matthey]

It would be necessary to enter a floppy with a kickstart which supports SCSI every cold start if there was no kickstart with SCSI support in ROM. The super kickstart 3000s have a kickstart with SCSI in ROM. It is an early version of 1.4/2.x AmigaOS which looks like a hybrid between them (blue screen but 3D look). As I recall, it is possible to click in the upper left hand corner of the super kickstart screen to select this old kickstart and even boot with it.

[whaka]

Originally Posted by Daedalus

H
In theory it would be possible to create a device that responds to queries with a last unit flag, but such a device wouldn't be a passive item like the IDE or floppy killers - it is more likely to require some sort of intelligent controller to deal with the queries correctly. A far simpler idea however would be to get hold of an old SCSI hard drive, or if power/heat/noise/space are a priority, a SCSI2SD adaptor, set that to unit 0 and set the last unit flag on it. That will give the minimum boot time possible without going to drastic lengths, however will need reconfiguring to a higher ID when connecting the CD-ROM drive.

thanks
if i get a "died" hard drive, and just keep the pcb, do you think this can do the trick ?

if i have time, i was thinking build a little pcb wich can fit i a db 25 connector. this way it could be possible to do it externally

Originally Posted by Pat the Cat

Interesting. The A3000 was released with Kickstart on Hard disk - so without a SCSI device in the chain, how could such a machine start up?

this concern only very early motherboard revision, such as rev 6 and maybe very first rev 7. and worse, this revisions cannot boot at all if no hard drive is attached, even with a full kickstart rom.
for those, they require a wire jump between a pin on scsi chip and a resistor, to get rid on this issue.

rev 7.3, 8 and 9 (the most common) are in standard with full kickstart in rom, and can fully boot without hard drive.

[hese]

Originally Posted by whaka

this concern only very early motherboard revision, such as rev 6 and maybe very first rev 7. and worse, this revisions cannot boot at all if no hard drive is attached, even with a full kickstart rom.
for those, they require a wire jump between a pin on scsi chip and a resistor, to get rid on this issue.

rev 7.3, 8 and 9 (the most common) are in standard with full kickstart in rom, and can fully boot without hard drive.

I have a rev 6.1 A3000 motherboard here with Kickstart 3.1 ROMs and it boots fine from floppy disk without any SCSI drive attached. And there are no wires between the SCSI chip and resistor.

[whaka]

this is because it's not a "real" 6.1. it's a rev 7.x but marked as 6.1 on motherboard.they exist.
my source is dave haynie.

you can find all this info on the dave haynie archives website

[Daedalus]

Originally Posted by whaka

thanks
if i get a "died" hard drive, and just keep the pcb, do you think this can do the trick ?

Probably not, hard drives don't usually respond until they have started up, which probably can't happen if the disc is missing.

if i have time, i was thinking build a little pcb wich can fit i a db 25 connector. this way it could be possible to do it externally

That would be a nice solution indeed.

[Shadowfire]

Originally Posted by whaka

thanks
if i get a "died" hard drive, and just keep the pcb, do you think this can do the trick ?

if i have time, i was thinking build a little pcb wich can fit i a db 25 connector. this way it could be possible to do it externally



this concern only very early motherboard revision, such as rev 6 and maybe very first rev 7. and worse, this revisions cannot boot at all if no hard drive is attached, even with a full kickstart rom.

This is technically not true. The superkickstart A3000's were shipped prior to OS 2.0 being finished; these particular A3000's were supplied with an earlier version of 2.0 (actually, at the time it was called Kickstart 1.4) in ROM. When you turn on the power of these units, a control program will look for the partition with the device name "wb_2.x:", mount it, then check in the /devs directory for a file called "kickstart" (going off of memory here). If its found, it is softkicked, reset, and a normal boot follow, except any partition named "wb_1.3:" is not mounted. (A similar path is followed if booting 1.3, except the partition named "wb_1.3:" is searched for kickstart, and "wb_2.x:" is not mounted during bootup)

On these machines, if you hold down both mouse buttons during that initial powerup, you are presented with a boot screen which will allow you to boot "Workbench 1.3" or "Workbench 2.x" from either floppy or hard disk. If the appropriate kickstart image was not found on the partition, that particular hard disk boot option will be ghosted out. In a standard shipping system, both Os 1.3 and 2.0 were included on the hard disk, and you would need to go here to initiate booting Workbench 1.3.

Additionally, you had the option to boot 1.3 or 2.x from floppy - this is where the Superkickstart disk comes in. This would read the kickstart from floppy, similar to an A1000. The Superkickstart disk contains both OS 1.3, and whatever flavor of 2.x, images.

A third option would be to boot into the 1.4 OS in ROM. On the OS selection screen, if you clicked on the invisible window-close gadget in the upper left corner of the screen, the OS selection menu would go away and it would attempt an OS 1.4 boot.

Note that the A3000 superkickstart images had the scsi.device included on the disk (although not part of the main ROM image, it was loaded separately by the 1.4 OS bootloader); also, there was a program included on those pre-release A3000 installation disks that would read a superkickstart disk and generate the devs:kickstart files.

To address the OP, the posters talking about the device scanning times are who you want to listen to. If you have no hard disk attached, the OS will run through the ropes scanning devices until it finds a hard disk that was marked as the last device. SCSI termination has nothing at all to do with the delay, and you can't build an inexpensive "terminator" to change this (well, actually, if you had the technical engineering prowess you could build up something that responded to SCSI reads and feed the A3000 bogus RigidDiskBlocks, but its far easier to just put a scsi hard disk on the bus, run HDToolbox, detect the device, then write out the RigidDiskBlock with no partitions defined and call it a day).

[Jope]

In this case I'd probably swap in a custom kick with no scsi.device and loadmodule it if I need it. :-)

[Hedeon]

I'm a bit puzzled, though. The cycle is maybe half a second for each ID. So a few seconds. Or maybe the 060 speeds it up here. Just make sure LUN checking is off otherwise it will take 8 times longer.

[whaka]

yep, LUN are not enabled.

Originally Posted by Shadowfire

This is technically not true.

if it's for the non booting early A3000 without scsi drive attached, go say that to dave haynie...

http://www.thule.no/haynie/systems/a...cs/pcb.rev.txt

[Shadowfire]

That same sheet also says the only pcb revision which had this problem (7.0) did not ship out to customers. So unless you have a prototype 7.0 A3000, which is highly unlikely if you didn't work in West Chester, it isn't true.

[Chucky]

just name the disk/partiton wb_2.x: (cannot remember exact name. buit I think it is) and it will boot the kickstart from that. diskette, IDE from extra controller etc doesn't mind. it will boot that kickstart and it will work. SCSI not needed.