Virtual OS/2 International Consumer Education
VOICE Home Page: http://www.os2voice.org
October 2001

[Newsletter Index]
[Previous Page] [Next Page]
[Feature Index]

editor@os2voice.org


Letters, Addenda, Errata

If you have any comments regarding articles or tips in this or any previous issue of the VOICE Newsletter, please send them to editor@os2voice.org. We are always interested in what our readers have to say.

September 3, 2001 - A followup to two of the OS/2 Tips in the September issue of the newsletter from Al Savage:

Hello, Editor:

Two of September's "OS/2 Tips" reference me (Al Savage).

About the first (availability of older AIC7870.ADD SCSI drivers):
Munson's statement that he tried a succession of Adaptec driver versions (obtained from my site) and found one that worked for him, is a bit misleading.

The NG thread continued after this post: earlier, I'd advised him to try the latest driver (April or July of 1999, variously), v2.4 . Later, after trying all of them he found that, surprise! v2.4 worked for him. I'd later gone on to point that out :)

Summary: use the last available AIC7870.ADD with the 29xx series Adaptec HAs (but not the U2W or newer). It will work just fine :) It will load once for multiple HAs installed (for example, if you have two 2940s or, as I do, a 2940UW Pro, and a 2930b).

And, if you have a "retail BIOS" 2940UW, there is a flash BIOS upgrade available (v2.20 from the website, v2.57.2 from my site) which also works very well.

Also, don't forget!
If you run a HDD larger than about 9GB, add this line to your CONFIG.sys (and don't forget to update your boot floppies, Disk1, as well):

BASEDEV=CHKDSK.SYS
which enables preallocation of RAM for CHKDSK's use for processing large HDDs. This avoids the all-too-familiar "hangs at boot while CHKDSKing my 18GB (or larger) partition". And be prepared for drastically reduced CHKDSK times (my two 18GB partitions CHKDSK in about seven minutes, on a K6-3/450), and an extra reboot after CHKDSK completes.

==============================

About the second tip (concerning HPFS386 boot diskettes):

I've had two people email me already, that the boot diskettes work OK but boot plain HPFS instead of HPFS386. I need to point out that only Disk1 & Disk2 are special for HPFS386 -- Disk0 & Disk3 are the same whether you boot to HPFS or HPFS386.

I've just updated all the diskette image files to include:
Disk0 : Updated kernel files to 0821
Disk1 (both versions): Updated to include "BASEDEV=CHKDSK.SYS"

As usual, download (if using NS, RMB on the file and choose "Save Link As" to download), then run LOADDSKF.exe to create the diskettes:

LOADDSKF HPFS_0.DSK A:
LOADDSKF HPFS3861.DSK A:
LOADDSKF HPFS3862.DSK A:
LOADDSKF HPFS_3.DSK A:
Regards,
Al S.

September 4, 2001 - A letter from Jeffrey S. Worley concerning Michal Necasek's series of articles on OS/2 history in the past 2 issues of the newsletter.:
I've been an OS/2-phile since the release of verion 2.0 and think I may have an answer as to the nasty rumor that early versions were released 'crippled for IBM machines only'.

I was then as now a technician for a volume computer var and saw everything from CP/M to Xenix scuff my bench.

Back when the 386dx was a new processor and folks were paying big bucks for the CPU alone, an embarrassing flaw was discovered which prevented early models from running in protected mode. These chips were for the most part marked and sold to manufacturers as '16 bit only' chips. They were still faster than predecessors at 16 bit execution and since nothing really ran in protected mode, it wasn't seen as a requirement the customer be forewarned. The problem was quickly fixed in production and so...... This bug caused Intel little grief from clients unlike the ol' P60/66 fdiv bug we know and love. Intel did not, unlike the Fdiv bug, offer any sort of trade or recall program.

When OS/2 2.0 and Windows 3.1 were released, the flaw became a more serious issue but not a huge one. I've only seen 20 or so of these 16bit only 386 processors which means that only the first few shipped were so limited.

I know the problem discussed on your site is focusing on pre-2.0 OS/2, I think the above might be the actual root of the rumour.

Regards,

Jeff

Michal's response:
Interesting. I wasn't around in those olden days but I do know about the buggy 'B0 stepping' 386s.

OS/2 knows about them too - these are actual messages from OS2LDR.MSG:

The system detected an Intel 80386 CPU stepping (B0) that is not supported by the OS/2 operating system. Contact your hardware service representative or replace your 80386 CPU.

The system detected an Intel 80386 CPU that does not reliably execute 32-bit multiple operations and cannot be supported by the OS/2 operating system. Contact your hardware service representative or replace your 80386 CPU.

Regarding the comment "I know the problem discussed on your site is focusing on pre-2.0 OS/2, I think the above might be the actual root of the rumour."

IF the bug affected any protected mode operation then it would affect OS/2 1.x too...

Regards,

- Mike


[Feature Index]
editor@os2voice.org
[Previous Page] [Newsletter Index] [Next Page]
VOICE Home Page: http://www.os2voice.org