Re: Future of OS/2 : was: ECS questions

Sir:

David T. Johnson wrote:
Sten Solberg wrote:
On Tue, 11 Apr 2006 17:55:15 UTC, "David T. Johnson" <djohnson@xxxxxxxxxxxx> wrote:


Sten Solberg wrote:

Wish I knew. The OS/2 boot process stops during loading of PMSHELL.EXE when I boot from a cold start. Booting Window first, then making a hot reboot to OS/2, works fine. Subsequent hot reboots from OS/2 also work fine. Strange thing, but more of a nuisance than a problem.

This is interesting. I saw the pmshell.exe hang with another system and never figured out a way to work around it so I switched to a different motherboard that was known to work. I never thought to try booting windows first, though, and then booting OS/2. The problem sounds like some sort of BIOS restriction that OS/2 doesn't know what to do with. I never considered that as a possibility. <Smack on forehead.>

What BIOS and motherboard do you have? A couple of years ago, Microsoft entered into an agreement with Phoenix to integrate their BIOS more closely with Windows and maybe this is some of the dubious fruit from that.

http://news.zdnet.co.uk/software/developer/0,39020387,39116902,00.htm


Interesting, but hardly sensational... :)

Booting to Windows first was suggested to me when I sweated and swore about this
problem last summer. The motherboard on this machine is an oldish, by now, Aopen
n250a-fr, Socket 754 for AMD Athlon 64, ATX, NVIDIA nForce3 250 chipset, HyperTransport @ 800 Mhz

The Aopen n250a-fr has an award bios which is a phoenix bios since award and phoenix are the same company. The interesting thing for me is that cold-booting to Windows with your system somehow changes something on the system so that a subsequent warm boot to OS/2 will succeed. Since the dram is all initialized at boot, the mysterious-something cannot be there and it obviously cannot be on the windows hard drive volume so the only place left for the mysterious-something to be written is to the Bios flash memory. What is being written? Whatever it is, we know that it goes away when the system powers down but if the system remains powered up, it is preserved even when a different os is booting.

What bios version do you have? Apparently Windows XP absolutely will not install on the board with certain older bios versions although windows 98 will install okay which further suggests that windows xp is trying to write something to the bios flash memory that the older bios will not allow but the newer bios updates are okay with. It would be interesting to see if the newer bios versions postdate microsoft's agreement with phoenix. One thing you could try is to install an older bios version (if they are still available somewhere) and then see if os/2 boots okay without a windows warm boot with the 'old' bios. If you are running windows xp, though, probably it would not longer boot with an old bios so you would have to have some other means of reflashing the new bios back. I also notice, though, that the board has 'dual bios' with two physical bios chips mounted on it. I've never had one of those and I'm not sure how they work or what they do but that might make it difficult to flash the bios back to an older version if the system doesn't want you to.


There are a few things in the BIOS I still have not figured out, like APICA. And, even though I have 1 GB physical RAM, SysInfo never reports a higher usage than 480 MB no matter how many apps are loaded/in use. This seems strange to me,
and I have sometimes wondered whether I am overlooking a setting in BIOS that would make OS/2 utilise all the RAM. Well, I don't understand computers.



I am not the expert, but it is known that ACPI requires certain data structures to exist in the BIOS data area. This data is written as the BIOS initializes the hardware. It is further known that XP at install time requires the manufacturer CD to be inserted into the CD drive to get certain drivers from it. It is also known that ACPI has code that can be read by a Forth interpreter. Thus it is not unreasonable to assume that without these mfg drivers written in Forth, that certain parts of the hardware is never initialized. Since OS/2 does not have the Forth interpreter in its boot loader and those who are writing our ACPI driver, which in my humble opinion needs to renamed to APIC driver because that is all it does, don't seem interested in including one, OS/2 may never get a boot time Forth interpreter to read/execute these drivers. Thus, your conjecture that booting XP would be necessary is the correct work-around. Those drivers, BTW, are stored on XP's drive and are loaded during boot time, but can be used by any OS if the OS copied them from the mfg's disk and then so loaded them at boot time. Getting back to the BIOS data structures, they are not complete if those drivers don't run and ACPI will not work correctly without them. Nor as you noticed will OS/2 boot as the snoops crash on the uninitialized hardware. Linux gets around this by including C versions of these Forth drivers in their ACPI code, which is a circle pull in keeping up with the hardware.

You got an AMD64 machine. Did it come with these drivers? The drivers would have *.f extension and be only a few hundred bytes each. I know Intel main boards do to initialized the SBbus hardware and to do that active power management thing.

PS. Did you install Dani's patch for memory reporting?
--
Bill
Thanks a Million!
.

Relevant Pages

  • Re: Future of OS/2 : was: ECS questions
    ... Booting Window first, then making a hot reboot to OS/2, works fine. ... The problem sounds like some sort of BIOS restriction that OS/2 doesn't know what to do with. ... Microsoft entered into an agreement with Phoenix to integrate their BIOS more closely with Windows and maybe this is some of the dubious fruit from that. ... The interesting thing for me is that cold-booting to Windows with your system somehow changes something on the system so that a subsequent warm boot to OS/2 will succeed. ...
    (comp.os.os2.misc)
  • Re: CD-ROM & Floppy Drives Dont Work
    ... many things that could well have been done to a W98 [any ... W98 Windows Explorer in either ... Find out if the 6 year old knows how to get into the BIOS, ... See if it will boot from the CD. ...
    (microsoft.public.windowsxp.hardware)
  • Re: 2 wondows versions detected on pc startup?
    ... Alternative to RegEdit find of Windows ... BIOS examines the very first sector of the disk for a Master Boot Record ... Data Type: REG_MULTI_SZ ...
    (microsoft.public.windowsxp.setup_deployment)
  • Operating System not found - dead HD ? retrive data
    ... windows created a restore point. ... setup I have only 3 tabs - Main, Boot and Exit. ... This morning when I went into BIOS and under BOOT I changed the order ... Microsoft Encarta Online Deluxe Encyclopedia - 1 year subscription, ...
    (microsoft.public.windowsxp.general)
  • Re: cmos battery
    ... How to get into your BIOS depends, not on Windows, but on what ... before Windows even starts to boot. ... When was the last time you thought about the battery in your PC? ...
    (microsoft.public.windowsxp.basics)