WHT SCSI and Genve/Genmod Geneve boot issues

[+jedimatt42]

I've found and read the contents of this archive:  http://ftp.whtech.com/datasheets and manuals/Hardware/WHT/scsi/

 

The contents of the GeneveBootEprom.zip by Tony Knerr suggest everything should be working... but it isn't...

 

I can load from GPL, the SCS1.SCSI/SYS via rompage mode ( but only if GPL Speed is set to 1 )  it will then successfully load SYSTEM/SYS, verify the CRC, and then dump me into MDOS on A:

 

But if I reboot the Geneve, and hold the space-bar to go to the boot device menu, and select SCSI, it fails to load SCSI/SYS.   Similarly if in rompage/GPL at speed 2-5, then attempting to load SCS1.SCSI/SYS fails with a not found error. 

 

The TIMODE or TURBO switches make no difference.

 

Is there an undocumented step I'm missing? 

 

CRUBASE of the card is >1200, I do see the card LED flash briefly when boot rom searches, before it goes off to the floppy.  Then 'TI / Geneve' switch is ON...  

 

Drive is a Compact Flash card, on a SCSI to IDE adapter, that I initialized with DU2K on my 4A, and copied the SYSTEM/SYS and SCSI/SYS files.  SCSI/SYS says it is v2.1,  SYSTEM/SYS is MDOS 6.50.

 

-M@

 

 

 

 

[+jedimatt42]

Additionally, the hardware manual on WHTech seems to be missing something:

 

 

SW2, 1-4 are all ON currently.

 

I think SW2 switch 4 is TIMODE:off  GeneveMode:on

I suspect SW2 switches 1-3 encode the card's SCSI ID,  all on would be 7

 

-M@

[+jedimatt42]

http://www.mainbyte.com/ti99/community/faire_2002.html

 

Quote

There are two types of scsi cards, original wht unmodified cards and new modified or German cards. These are accessed by MDOS 6.5 differently. Use the new CYA that comes with MDOS 6.5 to set up MDOS and setup your autoexec file. MDOS 6.5 is for newer scsi cards. CYA now has option for DMA scsi support. Modify MDOS one way for old scsi cards and another way for the newer cards.

 

I don't think I'm getting this far, as it pertains to modified SYSTEM/SYS, and I'm not even loading that yet, but this is the first I've heard of a modified WHT SCSI card...  How can the modification be identified?

 

-M@

[+InsaneMultitasker]

Does your card have a revision letter, e.g., E,F,G ?     Swim probably has the best working knowledge about the card.   As I recall, the rev E and rev F will have a daughter-board installed and updated chips, typically with Michael's initials "M.B." written on the labels. 

 

The Geneve adjusts GPL "speed" by using a combination of slow/fast memory pages and twiddling with wait state CRU bit.  SCSI/SYS moves itself into one of the fast ram pages after it is loaded.  This would need to be confirmed but I expect you'd find that at speed 1 the memory pages are mapped in such a way that SCSI/SYS can complete the memory map/move without stomping on itself. 

 

Try disconnecting the switchbox, remove the Memex card, then test again.  Be aware that if the OS loads it won't start up because you are missing the 32K sram but the test may shed light on the cause. For giggles you can remove any other cards to eliminate potential DSR interference.

 

 

[+jedimatt42]

It is Revision G. 

 

The GPL speed shows an issue before any code in SCSI/SYS is executed. If the speed is not 1, then the DSR is not even finding the SCSI/SYS file... the LOAD call fails.  

 

I have tried the naked Geneve + SCSI ( no floppy controller, no memory, nothing else in the box ) no change. Boot ROM fails to find SCSI/SYS

 

I haven't tried screwing with CRU addresses (default setting at >1200) yet... or scrutinizing the boot rom code.  But it does poke at the SCSI card (red light flashes on) when using the boot menu and picking SCSI. And it does work if I go deeply into GPL, ROMPAGE mode, speed 1.  But that doesn't help MDOS (6.50) have mass storage. 

 

--- 

 

Also, if it ever did work in MDOS, I can't find any docs on what drive letter a SCSI device would be assigned...  or what device number it would use with the REMAP command.  I would expect to be able to test it from MDOS even though it isn't booted from SCSI. Attempting to do a VOL H: lights up the drive and hangs.. 

 

---

 

So which BBS does Richard hang out on? @Swim

 

-M@

[atrax27407]

Richard is the sysop of "The Hidden Reef"

[+9640News]

The SCSI drive will not need to use the REMAP command.  The REMAP command is only used for floppy like devices.

 

If you just type ASSIGN at the MDOS prompt, it will give you a list of drive letters and device assignments.

 

You can use in AUTOEXEC, the command listed below if you want to change drive assignments which is good if you are dealing with many drive letters.

 

ASSIGN E=SCS1:

 

This will make drive letter E represent SCS1 drive if you want to change the default settings from the ASSIGN command.

 

Or if you want to reassign floppy drives:

 

ASSIGN H=DSK1:

 

Another comment.  I think what I read suggested you have a MEMEX in the system.  Does the MEMEX have a switch on it besides the normal dip switches, perhaps along the top edge of the card?  I ask, because I have a MEMEX with that mod that came from a fully expanded MEMEX system.  I noted that if the switch is in one position, I could not boot to MDOS with a Myarc HFDC.  However, if it was in the opposite position, it would.  This particular configuration was a fully expanded MEMEX but NOT a GenMOD Geneve, so I was only able to use 504K of the MEMEX.

 

I've not seen a switch like this on a MEMEX before and I have not seen any notes about it, so I do not know what it specifically does.

 

Beery

 

 

 

[+mizapf]

I think we have some article on that in Ninerpedia ... yes: https://www.ninerpedia.org/wiki/Geneve_drive_remapping

[+InsaneMultitasker]

SCSI card must be set to CRU 1200 for MDOS to recognize it.

 

I am reminded of a similar problem where I was unable to load the Horizon ramdisk ROS program image file from the SCSI card under ROMPAGE.   I made a cryptic note about the FAC and/or error status byte. Not sure where I was headed with that one and since that time, my system has been running Heatwave 24x7.

 

See Beery's comments and also type "PDMA" at the MDOS prompt.  PDMA should be set to 'on' before accessing the SCSI device.  The default may be OFF as at the time of release, there were still people with the non-PDMA cards. 

 

 

[+jedimatt42]

OK, cool info... so the drive mapping table in the documentation and echoed on ninerpedia seems to be conditionally correAfct.  H isn't always a ramdisk... it could be a SCS1... 

 

After issuing: 

PDMA ON

ASSIGN E=SCS1:

 

I can see the scsi drive in MDOS... 

 

So, it requires PDMA ON after in MDOS... that doesn't get it to booting.. but at least we know it works in MDOS.

 

I would have assumed the v1.0 boot rom delegates to the DSR... Tony's doc asserts that if an ea5 loader can run SCS1.SCSI/SYS then the Geneve will boot it.  There must be another requirement. Maybe this one:

 

Quote

It's also possible that a SCSI hard drive may not power up quickly enough to present itself on the SCSI bus in time to be recognized by the boot sequence. In this case the user will receive the "SCSI/SYS NOT FOUND" error message or the boot code will proceed to the next device. It is recommended that users of such "slow" SCSI drives do one of the following:

• A: Obtain a newer, faster-to-start SCSI drive.
• B: Depress the spacebar to access the boot menu and wait until the drive is ready before making a selection.
• ? Power-up the drive first before powering the rest of the system.

 

Although I've tried option B numerous times.  And I'd think the flash adapter would be fast, but maybe it doesn't init until poked?  First access in MDOS is relatively slow.

 

The latest DSR on whtech appears to be 1.6b, and there is no changelog info like there is from inception to 1.5... This board has 1.6a on it, with a note that it doesn't do 'WDS'

 

Sadly, I think the video ram on my geneve is at the end of its life.  Sitting at the MDOS prompt the SPACE character loses a few bits, filling the screen with garbage...  I need to try a different 9938 first, since it is responsible for memory refresh. A new problem today

[+jedimatt42]

OK, 9938 replacement solved the video corruption... Geneve is still alive! 

 

Still doesn't boot from SCSI, but progress...

 

-M@

 

[+InsaneMultitasker]

Please elaborate - what do you mean by "First access in MDOS is relatively slow"?  

 

Find (WHT?) and download the MDOS 6.70RC2 disk image.  It hasn't been tested with a Genmod so I cannot guarantee that it will detect Genmod and PDMA properly but it should do so based on the boot eprom and scsi dsr eprom, respectively.  The SCSI access routines are approx. 2x faster. 

 

A few minutes ago I restarted the heatwave system, loaded MDOS 6.50, GPL 6.50, speed 1, ROMPAGE.  SCSI/SYS loaded just fine but after displaying the 'loading system/sys' message it locked up tight.   Come to think of it, I don't know that I've ever had success completing a scsi-based system/sys load in this manner.   Are you using EA cart or a program to load SCSI/SYS?   

 

Yes, the Geneve boot EPROM uses the SCSI DSR to load SCSI/SYS.  SCSI/SYS uses the SCSI DSR to load the OS.

 

The Ninerpedia drive mapping table does NOT represent drive letter assignments in the left-most column.  Any "disk" device may be mapped to DSK1 through DSK9.  Since there are more than 10 device types, letters are used to select the appropriate device for each disk number.   

- ASSIGN A=DSK1:  assigns DSK1 to the drive letter A: however... 

- REMAP 5A assigns DSK5. to the first physical disk drive connected to a standard floppy controller at CRU >1100

- REMAP 8N assigned DSK8. to the 16-bit RAMDISK at CRU >1400  (16-bit refers to the memory addressing capacity in this case)

 

 

 

 

 

[+jedimatt42]

Elaborating: 

 

After booting from floppy, running PDMA ON, ASSIGN E=SCS1:, then DIR E:

 

It takes a while, then another DIR E: if very quick. 

 

I read somewhere in a guide on removable media and this controller, that some data is cached, thus you must reboot if you change zip disks for instance. I am assuming the first access is getting that data, and the second is using the cache. 

 

When I load SCS1.SCSI/SYS in GPL to test, I used 9640 menu's Run Program option. It loads fine, loads SYSTEM/SYS fine, validates CRC, then dumps me into MDOS. 

 

-M@

[+jedimatt42]

15 hours ago, InsaneMultitasker said:

\Find (WHT?) and download the MDOS 6.70RC2 disk image.  It hasn't been tested with a Genmod so I cannot guarantee that it will detect Genmod and PDMA properly but it should do so based on the boot eprom and scsi dsr eprom, respectively.  The SCSI access routines are approx. 2x faster. 

 

http://ftp.whtech.com/sitelist.txt   <--- NO MDOS greater than 650 on here.   

 

-M@

[+9640News]

6 minutes ago, jedimatt42 said:

 

http://ftp.whtech.com/sitelist.txt   <--- NO MDOS greater than 650 on here.   

 

-M@

6.70 Beta can be found here if you are on YahooGroups:

 

https://groups.yahoo.com/neo/groups/Geneve9640/files/MDOS/

 

Please note file areas are quickly disappearing from YahooGroups, so it is possible in another day or two, the files could be gone.  I've also included it here in this message.  It looks like it is a 720K DSK image file.

mdos670betaRC2-720k.DSK

[+mizapf]

I just uploaded it to the folder Geneve on WHTech (as GeneveOS_670betaRC2-720k.dsk).

[+jedimatt42]

I never saw file archives on the Yahoo groups, during my brief attendance. 

 

I believe I had myself removed after just a couple weeks.

 

-M@

[+InsaneMultitasker]

On 11/21/2019 at 10:36 AM, jedimatt42 said:

It takes a while, then another DIR E: if very quick. 

 

I read somewhere in a guide on removable media and this controller, that some data is cached, thus you must reboot if you change zip disks for instance. I am assuming the first access is getting that data, and the second is using the cache.

MDOS reads sector 0 and caches a few bytes of information including the total sectors per AU; there should not be a substantial delay caused by this operation under normal circumstances.  Is there some performance hit due to Genmod?  Maybe.  Someone with a genmod would need to test the operations and determine the cause.  (I no longer have a Genmod at my disposal. )

[Swim]

Hmmm? I have to start at the top of the page to see what's going on here. lol

[+jedimatt42]

3 hours ago, Swim said:

Hmmm? I have to start at the top of the page to see what's going on here. lol

 

I have a rev G card, with DSR ROM 1.6a.  I can find DSR ROM 1.6b on whtech.com - haven't had a chance to find a 64k eprom to try yet... but basically the SCSI configuration is working fine on my GenMod Geneve, except booting... I have SCSI/SYS on the scsi drive (compact-flash) but it doesn't load from the Geneve Boot ROM ( v1.0 genmod )  - I just get a SCSI/SYS not found error.  
I can load and run SCSI/SYS off of the SCS1. volume from GPL speed 1, ROMPAGE mode, but I'm missing something to be able to boot from it.

 

What else is required to be able to boot? 
 

I can't find any info on the difference between DSR 1.6a and 1.6b. Is 1.6b known to be required to boot? 

 

-M@

[Swim]

9 hours ago, jedimatt42 said:

 

I have a rev G card, with DSR ROM 1.6a.  I can find DSR ROM 1.6b on whtech.com - haven't had a chance to find a 64k eprom to try yet... but basically the SCSI configuration is working fine on my GenMod Geneve, except booting... I have SCSI/SYS on the scsi drive (compact-flash) but it doesn't load from the Geneve Boot ROM ( v1.0 genmod )  - I just get a SCSI/SYS not found error.  
I can load and run SCSI/SYS off of the SCS1. volume from GPL speed 1, ROMPAGE mode, but I'm missing something to be able to boot from it.

 

What else is required to be able to boot? 
 

I can't find any info on the difference between DSR 1.6a and 1.6b. Is 1.6b known to be required to boot? 

 

-M@

The old grey matter isn't what it used to be and it's been years since I sent all the SCSI stuff back to Don when he sold the WHT SCSI rights to competition computer but going by what I've read above my gut tells me PDMA (Pseudo Direct Memory Access) is not turned on in MDOS. I'm assuming this "G" model has the 1.6 upgrade with a daughter board that changes the on board narrow package 32k to the wide package 32k? We had to make that change because the narrow package speed caused problems that the slower wide package solved. I'm forgetting versions of MDOS but it used to be that you had to turn PDMA on until Gazoo came up with his SCSI Boot EPROM and an MDOS change that had PDMA always on. I'll try to pull out my notes later today and go over Gazoo's notes on setting up the Geneve for SCSI Boot. I don't seem to recall any notes about GenMod requiring different parameters and basically the 1.6 WHT upgrade didn't care what Geneve you had as long as it had the 32k fast RAM upgrade. BTW, after the 1.6 upgrade the "G" model WHT SCSI controller is the thoroughbred of the series with the "F" being the most stable with all manufacturer drive combinations in my opinion, and the "E" being the slowest.

[Swim]

BTW jedimatt, it comes to mind one of the hardware upgrades that was part of the 1.6 upgrade was a Schottky diode at location D1 for those using ZIP drives so keep that in mind if you are booting from a ZIP drive.

 

[+jedimatt42]

The SCSI system all works fine after boot from floppy and set PDMA ON, it's the booting that doesn't work. 

 

I'm using an ACARD SCSI to IDE, and an IDE to CF adapter. 

 

The board does have a diode in D1, and the wide DIP adapter.

 

--- 

 

This morning I tried the 1.6b DSR ROM from here: http://ftp.whtech.com/datasheets and manuals/Hardware/WHT/scsi/  but the result was that the Geneve wouldn't even boot off floppy.   It would however take longer to fail to find SCSI/SYS. 

 

---

 

I suppose all the lore was also on Yahoo? It seems like this was all developed 1998-2002? 

 

-M@

[+jedimatt42]

On 11/19/2019 at 8:59 PM, jedimatt42 said:

FROM THE DOC

 

It's also possible that a SCSI hard drive may not power up quickly enough to present itself on the SCSI bus in time to be recognized by the boot sequence. In this case the user will receive the "SCSI/SYS NOT FOUND" error message or the boot code will proceed to the next device. It is recommended that users of such "slow" SCSI drives do one of the following:

• A: Obtain a newer, faster-to-start SCSI drive.
• B: Depress the spacebar to access the boot menu and wait until the drive is ready before making a selection.
• ? Power-up the drive first before powering the rest of the system.

 

I think my issue may be more related to the SCSI device... during one of the reboots today, the thing actually booted off the SCSI card... was reproducable a couple times after a warm boot... CTRL-SHIFT-SHIFT    But then trying to reproduce from cold, then warm, wasn't possible.  

-M@

[+Ksarul]

Timing issue? Maybe it is just a hair too slow (or a hair too fast) for the system to register that things are right?