From Newsgroup: comp.sys.ibm.ps2.hardware

Folks, I looked at the T0 and T1 ver 1.02 refdisks.

http://ps-2.kev009.com/ohland/8590/8590_386-102_Refdisks.html

Note that each expanded refdisk is linked to the column header.

I >think< the files that differ by a few bytes have a different system
ID [like "5C" versus "5D"] only need a few bytes twiddled.

Files that need to be compared:

$BOOT.IML
$IMAGE.IML
MEMORY.DGS
PLANAR.DGS

Why the unnatural interest? The only refdisk for the T0 can only see
8MB, while the T1 ver 1.02 supports 32MB. -MAYBE- we can do
complex-affirming surgery by fixing T0-level files to the 1.02 level?
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.sys.ibm.ps2.hardware

Processor Complex Sub-Model Matrix
https://ardent-tool.com/complex/matrix.html

Submodels for T0- and T0+ in a Model 90
T0- 15
T0+ 17

Submodels for 0Gx, 0Jx, and 0Kx in a Model 90
0Gx 2D [486SX]
0Gx 2F [487SX]
0Jx 11
0Kx 13

NOTE: the 0Gx can use a 486SX or 487SX, submodel changes with J2

WBST noticed the submodels were in Command.com or simular DOS system
files. When the refdisk runs, it czechs for the submodel bytes.

My SWAG is that the system files in the 8590-386 refdisk has only two
submodel bytes, 15/17. The T1 ver 1.02 probably tests for 2D/2F/11/13
and lacks 15/17. Files that differ by only a few bytes between 402 and
T1 might have different submodel bytes accounting for the difference.

Larger size differences may be due to changed functions?

Louis Ohland wrote:

Folks, I looked at the T0 and T1 ver 1.02 refdisks.

http://ps-2.kev009.com/ohland/8590/8590_386-102_Refdisks.html

Note that each expanded refdisk is linked to the column header.

I >think< the files that differ by a few bytes have a different system
ID [like "5C" versus "5D"] only need a few bytes twiddled.

Files that need to be compared:

$BOOT.IML
$IMAGE.IML
MEMORY.DGS
PLANAR.DGS

Why the unnatural interest? The only refdisk for the T0 can only see
8MB, while the T1 ver 1.02 supports 32MB. -MAYBE- we can do complex-affirming surgery by fixing T0-level files to the 1.02 level?

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.sys.ibm.ps2.hardware

https://ardent-tool.com/config/refdisk.html#Functions_COMMAND_CMD

Functions and Relationship of COMMAND.COM and CMD.COM (from WBST)

Each RefDisk has its own, unique COMMAND.COM which:
Checks for correct system unit type (via Int 15h AH=C0h call and test of
the returned values, primarily the Model-Submodel-Revision bytes, but
possibly others for specific feature support).

Displays the announcement "splash" screen

Implements any delay -or- key-press processing before entering Main Menu display

Invokes CMD.COM when CTRL-A is pressed [Advanced Diagnostics].
Note: CMD.COM must accept a parameter to distinguish between
"Advanced" and "Test the Computer" invocations.

Louis Ohland wrote:

Processor Complex Sub-Model Matrix https://ardent-tool.com/complex/matrix.html

Submodels for T0- and T0+ in a Model 90
T0- 15
T0+ 17

Submodels for 0Gx, 0Jx, and 0Kx in a Model 90
0Gx 2D [486SX]
0Gx 2F [487SX]
0Jx 11
0Kx 13

NOTE: the 0Gx can use a 486SX or 487SX, submodel changes with J2

WBST noticed the submodels were in Command.com or simular DOS system
files. When the refdisk runs, it czechs for the submodel bytes.

My SWAG is that the system files in the 8590-386 refdisk has only two submodel bytes, 15/17. The T1 ver 1.02 probably tests for 2D/2F/11/13
and lacks 15/17. Files that differ by only a few bytes between 402 and
T1 might have different submodel bytes accounting for the difference.

Larger size differences may be due to changed functions?

Louis Ohland wrote:

Folks, I looked at the T0 and T1 ver 1.02 refdisks.

http://ps-2.kev009.com/ohland/8590/8590_386-102_Refdisks.html

Note that each expanded refdisk is linked to the column header.

I >think< the files that differ by a few bytes have a different system
ID [like "5C" versus "5D"] only need a few bytes twiddled.

Files that need to be compared:

$BOOT.IML
$IMAGE.IML
MEMORY.DGS
PLANAR.DGS

Why the unnatural interest? The only refdisk for the T0 can only see
8MB, while the T1 ver 1.02 supports 32MB. -MAYBE- we can do
complex-affirming surgery by fixing T0-level files to the 1.02 level?

--- Synchronet 3.22a-Linux NewsLink 1.2