Adaptec 1522A SCSI Support (was re: New TestFDC Results Registry)

[systems_glitch]

I often have a 1522A in my disk imaging machine. Linux supports it, which
means I can use `ddrescue` with it. It'll also talk to a lot of
older/slower drives that my 2940UW doesn't like (e.g. very old DEC drives
from VAXen, it does better with some old tape drives than the 2940UW). It's
slow, but I've also found that it's less likely to cause a kernel panic if
you're working on a disk that is very near the end of its life. I suspect
that's due to being a PIO device.

Thanks,
Jonathan

On Fri, Jan 19, 2018 at 6:31 AM, jim stephens via cctalk <
cctalk at classiccmp.org> wrote:

>
>
> On 1/18/2018 5:44 PM, Jason T via cctalk wrote:
>
>> On Thu, Jan 18, 2018 at 6:58 PM, Adrian Graham via cctalk
>> <cctalk at classiccmp.org> wrote:
>>
>>> I could, but I guess by the time I’ve sourced a replacement I might as
>>> well have bought an AHA-1522A instead, I have a couple of scouts out
>>> looking for them as we speak :) The 1522A is a full pass for TESTFDC.
>>>
>> Has anyone using one of these cards made use of the SCSI function?  It
>> has a Centronics 50 connector, which isn't terribly useful unless
>> you've got the right cable, but if you're building an all-in-one
>> imaging machine, it might be handy to have SCSI capability as well.
>> It seems the driver hasn't been in Linux for quite a few versions.
>> Not sure about the BSDs.
>>
>>
>> The 1522 was based on a lower cost chipset which Adaptec introduced to
> complement the 154x boards.  SCSI was suffering from being undercut by
> other interfaces in the market due to the nonstandard bios interface that
> the systems equipped with the 154x boards had to run with as bios got more
> complicated.
>
> Also I think this was the first chip that could be in the system w/o
> having the bios initiate the interface.  The 154x (1542 for example)
> required having the bus initiated and reset at boot by the bios, since
> there was a lot of logic involved in that chip.  The 152x and the like
> could be fully integrated into a reasonable sized driver and only have to
> go the the time consuming process of resetting and enumerating devices and
> initiators, etc. when needed.
>
> As Richard said for use with scanners and other peripheral media. The
> systems became very annoying if you didn't have boot media attached to the
> system controller to justify all the time it took to do that on every boot.
>
> And this and the 151x controllers were priced lower.
>
> I suspect the 154x and the PCI interface controllers Adaptec had should
> still be in the kernel.  You might be able to find the drivers and enable
> them in the Linux Kernel, unless some kernel driver rewrite had some
> feature that couldn't be implemented for this controller.
>
> I don't have a lot of references for the above, and if anyone has updates
> or comments, please add them.  I'm still using the HP DL360 and DL380 and
> the like which have the higher end controllers integrated in for the only
> SCSI work I do, and have not followed these or other support for a long
> time for parallel scsi interfaces.
>
> DL360's are so cheap that buying one with SCSI drives in the Generation 5
> or 6 or later is the best way I know of to run SCSI. Not great for power,
> but they are excellent and solid systems with a lot of OS support.
> thanks
> Jim
>