RE: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Dave Wade via cctalk]

> -Original Message-
> From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Jules
> Richardson via cctalk
> Sent: 10 October 2017 12:22
> To: cctalk@classiccmp.org
> Subject: Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and
> IDE ]
> 
> On 10/09/2017 12:52 PM, Chuck Guzis via cctalk wrote:
> > On 10/09/2017 06:52 AM, Jules Richardson via cctalk wrote:
> >
> >> My understanding there is that true SASI supports just a single
> >> target, and so there's no selection phase like there is with SCSI
> >> (and SCSI provides an extra signal on the connector uses during
> >> selection, which simply isn't there with SASI). However, there seemed
> >> to be some significant overlap and blurring of lines between SCSI and
> >> SASI, such that some early devices calling themselves SCSI aren't
> >> quite - and it's possible that some hardware which talks about SASI
> >> actually behaves more like SCSI.
> >
> > I'm not entirely sure about that--the PC Megastore contained both a
> > disk and a tape drive.  So more than a single device.
> 
> Yes, looking at the Xebec S1410A manual it talks about multiple boards on
the
> SASI bus, so that appears to be bit rot on my part - I must be thinking of
> something which predated SASI.
> 
> Having said that, I think that some bridge boards were capable of driving
both
> a ST506/412-type disk and QIC tape, so that particular setup wasn't
unheard
> of, although most bridges handled disk only.
> 
> cheers
> 
> Jules

I am pretty sure I had a XEBEC SASI board that would run as SCSI and had 2 x
MFM drives as 2 x LUNS onto a single SCSi address...
... on my ATARI ...

Dave

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Jules Richardson via cctalk]

On 10/09/2017 12:52 PM, Chuck Guzis via cctalk wrote:

On 10/09/2017 06:52 AM, Jules Richardson via cctalk wrote:


My understanding there is that true SASI supports just a single target,
and so there's no selection phase like there is with SCSI (and SCSI
provides an extra signal on the connector uses during selection, which
simply isn't there with SASI). However, there seemed to be some
significant overlap and blurring of lines between SCSI and SASI, such
that some early devices calling themselves SCSI aren't quite - and it's
possible that some hardware which talks about SASI actually behaves more
like SCSI.


I'm not entirely sure about that--the PC Megastore contained both a disk
and a tape drive.  So more than a single device.


Yes, looking at the Xebec S1410A manual it talks about multiple boards on 
the SASI bus, so that appears to be bit rot on my part - I must be thinking 
of something which predated SASI.


Having said that, I think that some bridge boards were capable of driving 
both a ST506/412-type disk and QIC tape, so that particular setup wasn't 
unheard of, although most bridges handled disk only.


cheers

Jules

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Liam Proven via cctalk]

On 9 October 2017 at 19:52, Chuck Guzis via cctalk
 wrote:

> I'm not entirely sure about that--the PC Megastore contained both a disk
> and a tape drive.  So more than a single device.

Did they have separately-settable IDs?

Otherwise, LUNs, I'd think, but I'd also expect that to flummox many drivers.

-- 
Liam Proven • Profile: https://about.me/liamproven
Email: lpro...@cix.co.uk • Google Mail/Talk/Plus: lpro...@gmail.com
Twitter/Facebook/Flickr: lproven • Skype/LinkedIn/AIM/Yahoo: liamproven
UK: +44 7939-087884 • ČR/WhatsApp/Telegram/Signal: +420 702 829 053

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Warner Losh via cctalk]

On Oct 9, 2017 11:51 AM, "Mark Linimon via cctalk" 
wrote:

On Mon, Oct 09, 2017 at 08:52:26AM -0500, Jules Richardson via cctalk wrote:
> Lots of early SCSI devices have no support for the Inquiry command, which
> trips up modern software which expects it - I don't know if it was simply
> ignored, or if there was a point in time where it wasn't present in the
> spec, and was only added later.

When I worked on SCSI device drivers in the late 1980s, we found that
support for the full spec was rare.  "Get it to work well enough and
ship it" seemed to be the explanation.  (cf: any BIOS implementation)

We wound up using a subset of the spec to get things to interoperate
correctly with our cards.


Open source kernels from the time had big tables of quirks to cope with
dodgy standards compliance. Things haven't changed much: there are still
non compliant devices today... don't want to use a feature that has a bug
which will eat data...

Warner

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Chuck Guzis via cctalk]

On 10/09/2017 06:52 AM, Jules Richardson via cctalk wrote:

> My understanding there is that true SASI supports just a single target,
> and so there's no selection phase like there is with SCSI (and SCSI
> provides an extra signal on the connector uses during selection, which
> simply isn't there with SASI). However, there seemed to be some
> significant overlap and blurring of lines between SCSI and SASI, such
> that some early devices calling themselves SCSI aren't quite - and it's
> possible that some hardware which talks about SASI actually behaves more
> like SCSI.

I'm not entirely sure about that--the PC Megastore contained both a disk
and a tape drive.  So more than a single device.

--Chuck

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Mark Linimon via cctalk]

On Mon, Oct 09, 2017 at 08:52:26AM -0500, Jules Richardson via cctalk wrote:
> Lots of early SCSI devices have no support for the Inquiry command, which
> trips up modern software which expects it - I don't know if it was simply
> ignored, or if there was a point in time where it wasn't present in the
> spec, and was only added later.

When I worked on SCSI device drivers in the late 1980s, we found that
support for the full spec was rare.  "Get it to work well enough and
ship it" seemed to be the explanation.  (cf: any BIOS implementation)

We wound up using a subset of the spec to get things to interoperate
correctly with our cards.

mcl

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Jules Richardson via cctalk]

On 10/09/2017 10:42 AM, Jon Elson wrote:

On 10/09/2017 08:52 AM, Jules Richardson via cctalk wrote:


My understanding there is that true SASI supports just a single target,
and so there's no selection phase like there is with SCSI

No, not true.  Each of the 8 data lines selected one target, so you could
have 8 targets on one controller.


Yes, you're right... I don't know where I was remembering that from (did 
SASI grow out of something which was single target?) but I was just looking 
at the Xebec S1410A manual and it does indeed talk about multiple boards on 
the SASI bus.


cheers

Jules

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Jon Elson via cctalk]

On 10/09/2017 08:52 AM, Jules Richardson via cctalk wrote:


My understanding there is that true SASI supports just a 
single target, and so there's no selection phase like 
there is with SCSI
No, not true.  Each of the 8 data lines selected one target, 
so you could have 8 targets on one controller.


(and SCSI provides an extra signal on the connector uses 
during selection, which simply isn't there with SASI). 
However, there seemed to be some significant overlap and 
blurring of lines between SCSI and SASI, such that some 
early devices calling themselves SCSI aren't quite - and 
it's possible that some hardware which talks about SASI 
actually behaves more like SCSI.


Yes, I'm sure this was true, as vendors moved their code 
over to comply with the new standard.
I don't think SASI had any parity support, either - but I 
think that in a lot of cases early HBA's relied on parity 
checking in software, which meant that it could simply be 
ignored.


I looked up my SASI interface, and I didn't see any parity 
circuit there.


Jon

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Jules Richardson via cctalk]

On 10/05/2017 01:50 PM, Chuck Guzis via cctalk wrote:

You could well be right--I do recall that there was "Mac SCSI" and then
the slightly different "Everyone else's SCSI".  I ran into this when
talking with some SMS/OMTI engineers about an ST506-to-SCSI bridge
board that I have.  Their reaction was "Oh, that's Mac SCSI--you want
real SCSI".


Hmm, I was under the impression that at the connector level the Mac flavor 
still carries all the signals of real SCSI, but leaves out a lot of the 
grounds so that it'll fit into 25 pins. At the protocol level, it's the 
same. At the low-level software side of things, Macs could be picky and 
only talk to a SCSI device which identified itself with Apple branding - 
which would certainly cause problems in hooking up something like an OMTI 
board.



I do know that many SASI devices work as SCSI-1 devices.  Somewhere, I
still have an early PC ISA SASI (not SCSI) adapter for an Ampex
Megastore unit.


My understanding there is that true SASI supports just a single target, and 
so there's no selection phase like there is with SCSI (and SCSI provides an 
extra signal on the connector uses during selection, which simply isn't 
there with SASI). However, there seemed to be some significant overlap and 
blurring of lines between SCSI and SASI, such that some early devices 
calling themselves SCSI aren't quite - and it's possible that some hardware 
which talks about SASI actually behaves more like SCSI.


I don't think SASI had any parity support, either - but I think that in a 
lot of cases early HBA's relied on parity checking in software, which meant 
that it could simply be ignored.


Lots of early SCSI devices have no support for the Inquiry command, which 
trips up modern software which expects it - I don't know if it was simply 
ignored, or if there was a point in time where it wasn't present in the 
spec, and was only added later.


cheers

Jules

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Fred Cisin via cctalk]

On Fri, 6 Oct 2017, Geoffrey Oltmans via cctalk wrote:

Supposedly the Mac Plus SCSI implementation is slightly broken/non-standard
(or at least draft standard), either in its drivers or the SCSI controller
chip, so maybe that's what they were referring to.


Even TRIVIAL differences can bite you.
Such as Apple/Future Domain? cabling of "TermPwr"

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Geoffrey Oltmans via cctalk]

On Thu, Oct 5, 2017 at 1:50 PM, Chuck Guzis via cctalk <
cctalk@classiccmp.org> wrote:

>
> You could well be right--I do recall that there was "Mac SCSI" and then
> the slightly different "Everyone else's SCSI".  I ran into this when
> talking with some SMS/OMTI engineers about an ST506-to-SCSI bridge board
> that I have.  Their reaction was "Oh, that's Mac SCSI--you want real SCSI".
>
>
Supposedly the Mac Plus SCSI implementation is slightly broken/non-standard
(or at least draft standard), either in its drivers or the SCSI controller
chip, so maybe that's what they were referring to.

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Liam Proven via cctalk]

On 6 October 2017 at 06:11, r.stricklin via cctalk
 wrote:
>
> Notwithstanding the first Macintosh models lacked SCSI entirely. The first 
> Macintosh with SCSI was the Plus, in 1986.

Well, this is true, but then again, there weren't many models of Mac
before the Mac Plus, were there?

AFAIK there was only the original Macintosh (128 kB RAM) and then the
Fat Mac (same machine, but with 512 kB RAM). That's it. No?

I have a Plus but it died while I was experimenting and I have the
electronics knowledge of an orang utan. How I'm going to fix it, I do
not know.

-- 
Liam Proven • Profile: https://about.me/liamproven
Email: lpro...@cix.co.uk • Google Mail/Talk/Plus: lpro...@gmail.com
Twitter/Facebook/Flickr: lproven • Skype/LinkedIn/AIM/Yahoo: liamproven
UK: +44 7939-087884 • ČR/WhatsApp/Telegram/Signal: +420 702 829 053

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[r.stricklin via cctalk]

On Oct 5, 2017, at 11:18 AM, Tom Gardner via cctech wrote:

> I was at Shugart at that time and to the best of my recollection Apple was 
> not a driver of the ANSI activity.
> The Macintosh shipped in January 1984 well after the ANSI SCSI work started 
> and its major distinguishing feature was the non-standard connector

Notwithstanding the first Macintosh models lacked SCSI entirely. The first 
Macintosh with SCSI was the Plus, in 1986.

ok
bear.

-- 
until further notice

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[allison via cctalk]

On 10/05/2017 09:39 PM, Jerry Weiss via cctalk wrote:
> The DK Driver for VMS versions around 5.x definitely had a problem with 
> non-DEC disks.  6.X and greater were slightly more forgiving.

Having many of the era DEC VAX I can say the only SCSI issue I had was
the boot disk greater than 1.07 GB.
for second or third disks that was not an issue.  It was a system
diagnostic/boot rom issue.

I've used most anything I could find that was SCSI or SCSI-2. Most were
Seagate or WD, and a few with Compaq/fujitsu
labels.  With VMS 5, but not 5.0.

Allison
> The specifics are summarized in a note from Ralph Weber in 
> https://groups.google.com/forum/#!search/SCSI$20Mode$20Page$20Requirements$20$20axp/comp.os.vms/RAaUpP_XXEw/BWn64YZYwBQJ
>  .  
>
> I don’t think there is list of non-DEC disks in the driver as it instead 
> checked the SCSI Mode bits and other disk configuration settings.   There is 
> a list (table) for DEC Drives (idiosyncrasies?) and another SCSI2 Tagged 
> Queuing devices requirements used for Clusters in the driver.
>
> Regards,
> Jerry
>
>
>
>> On Oct 5, 2017, at 6:23 PM, Peter Coghlan via cctalk  
>> wrote:
>>
>>
>>
>>> The biggest problem you had was the requirement to assert ATN when selected 
>>> properly.� Later the tag queuing caused huge headaches as manufacturers 
>>> implemented that feature.
>>>
>>> It eventually was made mandatory for the most part by linux, and perhaps 
>>> Windows requiring the tag queuing drilled own to the lowest level of the 
>>> system's use of the disk.  The capability to do that, or fake it is 
>>> required to allow the kernel to queue commands to run, and have the OS 
>>> continue to run till command completion.
>>>
>> I recall VMS having issues with SCSI disks which claimed to do tag queueing
>> (and bad block replacement) but didn't do it right, before I'd even heard
>> of linux.
>>
>> Customers complained that VMS refused to work with commodity SCSI disks
>> and thought that it was a conspiracy to get them to buy expensive DEC branded
>> disks.  DEC claimed that only the disks with their firmware did tag queueing
>> and bad block replacement correctly.  The VMS SCSI driver supposedly had 
>> (has?)
>> a list of specific disks known to mess up which it would refuse to bring
>> online.
>>
>> I wasn't well up on Sun but I expect the same issue existed there too.
>>
>> Regards,
>> Peter Coghlan.
>
>

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Jon Elson via cctalk]

On 10/05/2017 01:50 PM, Chuck Guzis via cctalk wrote:



What I found curious was the CDC manual that called SCSI "SASI subset".
To me that says that SASI was the more elaborate protocol and SCSI
initially picked and chose from it.


Oh, no!  SASI was VERY simple.  You could read and write a 
number of blocks, and sense the error status.
That was just about the total command set.  Even SCSI-I had 
a much wider set, to accommodate tapes,
scanners, printers,and other things.  It allowed for several 
different types of status sensing, could report
a lot of information about the drive (make & model, physical 
configuration, etc.)


At least the Memorex SASI drive/controller I put on my CP/M 
system was really basic, as far as I remember.


Jon

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Jerry Weiss via cctalk]

The DK Driver for VMS versions around 5.x definitely had a problem with non-DEC 
disks.  6.X and greater were slightly more forgiving.

The specifics are summarized in a note from Ralph Weber in 
https://groups.google.com/forum/#!search/SCSI$20Mode$20Page$20Requirements$20$20axp/comp.os.vms/RAaUpP_XXEw/BWn64YZYwBQJ
 .  

I don’t think there is list of non-DEC disks in the driver as it instead 
checked the SCSI Mode bits and other disk configuration settings.   There is a 
list (table) for DEC Drives (idiosyncrasies?) and another SCSI2 Tagged Queuing 
devices requirements used for Clusters in the driver.

Regards,
Jerry



> On Oct 5, 2017, at 6:23 PM, Peter Coghlan via cctalk  
> wrote:
> 
> 
> 
>> 
>> The biggest problem you had was the requirement to assert ATN when selected 
>> properly.� Later the tag queuing caused huge headaches as manufacturers 
>> implemented that feature.
>> 
>> It eventually was made mandatory for the most part by linux, and perhaps 
>> Windows requiring the tag queuing drilled own to the lowest level of the 
>> system's use of the disk.  The capability to do that, or fake it is required 
>> to allow the kernel to queue commands to run, and have the OS continue to 
>> run till command completion.
>> 
> 
> I recall VMS having issues with SCSI disks which claimed to do tag queueing
> (and bad block replacement) but didn't do it right, before I'd even heard
> of linux.
> 
> Customers complained that VMS refused to work with commodity SCSI disks
> and thought that it was a conspiracy to get them to buy expensive DEC branded
> disks.  DEC claimed that only the disks with their firmware did tag queueing
> and bad block replacement correctly.  The VMS SCSI driver supposedly had 
> (has?)
> a list of specific disks known to mess up which it would refuse to bring
> online.
> 
> I wasn't well up on Sun but I expect the same issue existed there too.
> 
> Regards,
> Peter Coghlan.

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Peter Coghlan via cctalk]

The biggest problem you had was the requirement to assert ATN when 
selected properly.  Later the tag queuing caused huge headaches as 
manufacturers implemented that feature.


It eventually was made mandatory for the most part by linux, and perhaps 
Windows requiring the tag queuing drilled own to the lowest level of the 
system's use of the disk.  The capability to do that, or fake it is 
required to allow the kernel to queue commands to run, and have the OS 
continue to run till command completion.




I recall VMS having issues with SCSI disks which claimed to do tag queueing
(and bad block replacement) but didn't do it right, before I'd even heard
of linux.

Customers complained that VMS refused to work with commodity SCSI disks
and thought that it was a conspiracy to get them to buy expensive DEC branded
disks.  DEC claimed that only the disks with their firmware did tag queueing
and bad block replacement correctly.  The VMS SCSI driver supposedly had (has?)
a list of specific disks known to mess up which it would refuse to bring
online.

I wasn't well up on Sun but I expect the same issue existed there too.

Regards,
Peter Coghlan.

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Phil Blundell via cctalk]

On Thu, 2017-10-05 at 11:50 -0700, Chuck Guzis via cctalk wrote:
> 
> What I found curious was the CDC manual that called SCSI "SASI
> subset".
> To me that says that SASI was the more elaborate protocol and SCSI
> initially picked and chose from it.

I think that's just bad/ambiguous wording and the intended meaning was
"SCSI (but only the parts that are in SASI)" as distinct from "SCSI
(all parts)".

p.

Re: The origin of SCSI [WAS:RE: The origin of the phrases ATA and IDE ]

[Chuck Guzis via cctalk]

On 10/05/2017 11:18 AM, Tom Gardner via cctalk wrote:
> I suspect this might start another discussion, but as I understand it Apple 
> had little to do with the evolution of SASI into SCSI.
> Shugart Associates published SASI in 1981 and took it to ANSI in 1982 where 
> they renamed it SCSI to avoid using a vendors name.  

You could well be right--I do recall that there was "Mac SCSI" and then
the slightly different "Everyone else's SCSI".  I ran into this when
talking with some SMS/OMTI engineers about an ST506-to-SCSI bridge board
that I have.  Their reaction was "Oh, that's Mac SCSI--you want real SCSI".

What I found curious was the CDC manual that called SCSI "SASI subset".
To me that says that SASI was the more elaborate protocol and SCSI
initially picked and chose from it.

I do know that many SASI devices work as SCSI-1 devices.  Somewhere, I
still have an early PC ISA SASI (not SCSI) adapter for an Ampex
Megastore unit.

I'm also well-acquainted with what Andy Johnson-Laird called "SCSI
Voodoo" in trying to get several different SCSI devices to work off the
same SCSI-2 bus.

--Chuck