f



IBM SCSI Microcode versions - a little touchup

http://ps-2.kev009.com/ohlandl/IBM_SCSI/IBM_SCSI_microcode_versions.html

Just added some Anchors to David Beem's opus magnus on SCSI uCode Versions.

Still the same micro channel goodness in every byte.
0
Louis
8/17/2016 4:23:22 PM
comp.sys.ibm.ps2.hardware 10363 articles. 0 followers. Post Follow

11 Replies
258 Views

Similar Articles

[PageSpeed] 24

> Just added some Anchors to David Beem's opus magnus on SCSI uCode Versions.

Who?
0
davidb
9/11/2016 10:58:25 PM
On Wednesday, 17 August 2016 17:23:22 UTC+1, Louis Ohland  wrote:
> http://ps-2.kev009.com/ohlandl/IBM_SCSI/IBM_SCSI_microcode_versions.html
> 
> Just added some Anchors to David Beem's opus magnus on SCSI uCode Versions.
> 
> Still the same micro channel goodness in every byte.

A big thumbs-up, Louis.

Cheers,
WBST
0
WBSTClarke
9/12/2016 5:32:43 AM
In the dimly lit era of titans, there was the mythic being only known as 
David Beem. David collected knowledge on the many faces of God as 
expounded in the IBM SCSI code versions, and then provided that 
information to mere mortals.

But I wonder if he suffered the fate of Prometheus.

Part of David's punishment was to have the gift corrupted by agents of 
pure evil. I have poured many libations of coffee and raised my 
supplications to the heavens while using the Internet Archive to salvage 
that page.

In short, David found most versions of PS/2 SCSI adapters and put them 
on his page.

  http://ps-2.kev009.com/ohlandl/IBM_SCSI/IBM_SCSI_microcode_versions.html



On 9/11/2016 17:58, davidb@wnmt.com wrote:
>> Just added some Anchors to David Beem's opus magnus on SCSI uCode Versions.
>
> Who?
>

0
Louis
9/12/2016 3:28:39 PM
> In the dimly lit era of titans, there was the mythic being only known as=
=20
> David Beem. David collected knowledge on the many faces of God as=20
> expounded in the IBM SCSI code versions, and then provided that=20
> information to mere mortals.

The really sad thing is that some of that information might be truly gone. =
I had a former co-worker deliberately delete the content from my domain (ev=
erything but the logs, or I wouldn't have figured out what happened to it).=
 I'm hoping that some of the SCSILVL work might be on the Model 56 I used t=
o expand upon it, I'll need to dig.
0
davidb
9/13/2016 4:42:25 PM
Bill Gates' chaotic minions are everywhere.

IBM is too tied up in corporate internal scheming to actually do things 
like that. Or so I would imagine.

On 9/13/2016 11:42, davidb@wnmt.com wrote:
>> In the dimly lit era of titans, there was the mythic being only known as
>> David Beem. David collected knowledge on the many faces of God as
>> expounded in the IBM SCSI code versions, and then provided that
>> information to mere mortals.
>
> The really sad thing is that some of that information might be truly gone.
I had a former co-worker deliberately delete the content from my domain
(everything but the logs, or I wouldn't have figured out what happened 
to it).
  I'm hoping that some of the SCSILVL work might be on the Model 56 I 
used to
expand upon it, I'll need to dig.
>

0
Louis
9/13/2016 5:18:37 PM
On Monday, 12 September 2016 16:28:40 UTC+1, Louis Ohland  wrote:
> In the dimly lit era of titans, there was the mythic being only known as 
> David Beem. David collected knowledge on the many faces of God as 
> expounded in the IBM SCSI code versions, and then provided that 
> information to mere mortals.
> 
> But I wonder if he suffered the fate of Prometheus.
> 
> Part of David's punishment was to have the gift corrupted by agents of 
> pure evil. I have poured many libations of coffee and raised my 
> supplications to the heavens while using the Internet Archive to salvage 
> that page.
> 
> In short, David found most versions of PS/2 SCSI adapters and put them 
> on his page.
> 
>   http://ps-2.kev009.com/ohlandl/IBM_SCSI/IBM_SCSI_microcode_versions.html
> 
> 
> 
> On 9/11/2016 17:58, d....b@wnmt.com wrote:
> >> Just added some Anchors to David Beem's opus magnus on SCSI uCode Versions.
> >
> > Who?
> >

Let me weigh in here, I believe the revision level will be 'hidden' in some of the extra POS registers for the adapter(s)/planar(s) or the extended CMOS RAM copies both for those with on-planar SCSI and/or the individual adapters.

Cheers,
WBST
0
WBSTClarke
9/14/2016 3:04:47 AM
> Let me weigh in here, I believe the revision level will be 'hidden' in so=
me of the extra POS registers for the adapter(s)/planar(s) or the extended =
CMOS RAM copies both for those with on-planar SCSI and/or the individual ad=
apters.

On the Tribbles I was actually pulling the ROM chip to read the microcode v=
ersion in an EPROM programmer - I accidentally burned up my only copy of th=
e late Tribble version when I put it backwards back on the adapter socket! =
The software routine may very well pull it out of a more dynamic area - Una=
l did much more investigation than I did.

I remember it as an offset from the base I/O address - the routine I mimick=
ed had to make sure of the I/O range of the circuitry. If it had been easie=
r to find elsewhere I would have changed my method. As I said, unless I pul=
l that machine (a Model 56SLC3 I still have tucked away) I won't be able to=
 resume work in that area at all.
0
davidb
9/15/2016 5:02:48 AM
On Thursday, 15 September 2016 06:02:49 UTC+1, dav...@wnmt.com  wrote:
> > Let me weigh in here, I believe the revision level will be 'hidden' in =
some of the extra POS registers for the adapter(s)/planar(s) or the extende=
d CMOS RAM copies both for those with on-planar SCSI and/or the individual =
adapters.
>=20
> On the Tribbles I was actually pulling the ROM chip to read the microcode=
 version in an EPROM programmer - I accidentally burned up my only copy of =
the late Tribble version when I put it backwards back on the adapter socket=
! The software routine may very well pull it out of a more dynamic area - U=
nal did much more investigation than I did.
>=20
> I remember it as an offset from the base I/O address - the routine I mimi=
cked had to make sure of the I/O range of the circuitry. If it had been eas=
ier to find elsewhere I would have changed my method. As I said, unless I p=
ull that machine (a Model 56SLC3 I still have tucked away) I won't be able =
to resume work in that area at all.

I seem to remember that for those adapters/planars that support them, POS r=
egisters 104h and 105h can be indexed through additional registers/values. =
It's in Peter's later HITR/BITR manuals, IIRC.

Cheers,
WBST
0
WBSTClarke
9/21/2016 10:34:15 AM
On Wednesday, 21 September 2016 11:34:16 UTC+1, WBSTClarke  wrote:
> On Thursday, 15 September 2016 06:02:49 UTC+1, dav...@wnmt.com  wrote:
> > > Let me weigh in here, I believe the revision level will be 'hidden' i=
n some of the extra POS registers for the adapter(s)/planar(s) or the exten=
ded CMOS RAM copies both for those with on-planar SCSI and/or the individua=
l adapters.
> >=20
> > On the Tribbles I was actually pulling the ROM chip to read the microco=
de version in an EPROM programmer - I accidentally burned up my only copy o=
f the late Tribble version when I put it backwards back on the adapter sock=
et! The software routine may very well pull it out of a more dynamic area -=
 Unal did much more investigation than I did.
> >=20
> > I remember it as an offset from the base I/O address - the routine I mi=
micked had to make sure of the I/O range of the circuitry. If it had been e=
asier to find elsewhere I would have changed my method. As I said, unless I=
 pull that machine (a Model 56SLC3 I still have tucked away) I won't be abl=
e to resume work in that area at all.
>=20
> I seem to remember that for those adapters/planars that support them, POS=
 registers 104h and 105h can be indexed through additional registers/values=
.. It's in Peter's later HITR/BITR manuals, IIRC.
>=20
> Cheers,
> WBST

http://www.mcamafia.de/pdf/ibm_hitra02.pdf
Page 108: POS Registers (10)6h & (10)7h are set to 01 & 00 then read (10)4h=
 to autoincrement through the Subaddress Space, if (10)6h reads back as 01 =
(indexing supported).

That's how I've interpreted it, anyway.

Cheers,
WBST
0
WBSTClarke
9/21/2016 10:59:34 AM
On Wednesday, 21 September 2016 11:59:35 UTC+1, WBSTClarke  wrote:
> On Wednesday, 21 September 2016 11:34:16 UTC+1, WBSTClarke  wrote:
> > On Thursday, 15 September 2016 06:02:49 UTC+1, dav...@wnmt.com  wrote:
> > > > Let me weigh in here, I believe the revision level will be 'hidden'=
 in some of the extra POS registers for the adapter(s)/planar(s) or the ext=
ended CMOS RAM copies both for those with on-planar SCSI and/or the individ=
ual adapters.
> > >=20
> > > On the Tribbles I was actually pulling the ROM chip to read the micro=
code version in an EPROM programmer - I accidentally burned up my only copy=
 of the late Tribble version when I put it backwards back on the adapter so=
cket! The software routine may very well pull it out of a more dynamic area=
 - Unal did much more investigation than I did.
> > >=20
> > > I remember it as an offset from the base I/O address - the routine I =
mimicked had to make sure of the I/O range of the circuitry. If it had been=
 easier to find elsewhere I would have changed my method. As I said, unless=
 I pull that machine (a Model 56SLC3 I still have tucked away) I won't be a=
ble to resume work in that area at all.
> >=20
> > I seem to remember that for those adapters/planars that support them, P=
OS registers 104h and 105h can be indexed through additional registers/valu=
es. It's in Peter's later HITR/BITR manuals, IIRC.
> >=20
> > Cheers,
> > WBST
>=20
> http://www.mcamafia.de/pdf/ibm_hitra02.pdf
> Page 108: POS Registers (10)6h & (10)7h are set to 01 & 00 then read (10)=
4h to autoincrement through the Subaddress Space, if (10)6h reads back as 0=
1 (indexing supported).
>=20
> That's how I've interpreted it, anyway.
>=20
> Cheers,
> WBST

End of Subaddress Space will be indicated by the read value of the indexing=
 pair (106h) & (10)7h wrapping back to zeros, I believe.

End of extended memory dump.
WBST
0
WBSTClarke
9/21/2016 11:03:14 AM
This is what I wrote, more than ten years ago (!):

> Very convoluted to me. You OUT the memory locations to the controller
> ports where you want the data (including the microcode value), then 
> give the command to another controller port & bango/presto, the values 
> appear, transfered with DMA. I'm going to start working on my code 
> fragments tomorrow.
0
davidb
9/22/2016 4:36:10 AM
Reply: