On 2/26/2022 12:31 AM, Jerome Shidel wrote:
On Feb 25, 2022, at 8:00 PM, Ralf Quint <[email protected]> wrote:
On 2/25/2022 4:02 AM, Jim Hall wrote:
https://www.heise.de/news/Betriebssystem-Oldtimer-FreeDOS-1-3-mit-neuen-Programmen-und-Features-6515775.html
Just had a quick read over that one and noticed one strange thing mentioned.
Apparently the CPU detection in the installer would only recognize an 80186
even though apparently installation on an 80286 was attempted... :?
Ralf
Probably do to translation, this is not entirely accurate.
Which translation?
Although not a bug, the information used to perform CPU detection is not 100%
reliable with all processors. 8086/88 is obviously assumed. 80186 is fine.
However, 286 tests fail on some 486 machines. I don’t know if they function on
586 (no Pentium to test). But they do work on 686 (Pentium Pro) and higher
CPU’s. Being the 286 test is for a 286, it should work there. However, I’ve
only got so many test machines and it is unknown if the test passes on machines
with less then a 686.
So at present, results of the test algorithm for specific CPU’s are 8086/88
(good), 80186 (good), 286 (possibly good), 386 (possibly good), 486 (rarely
good), 586 (possible good), 686+ (good), Virtual Platforms (good, handled
differently).
So on some 486 systems (probably most, including my 486DX2/66), the 286 test
fails and it is detected as a 80186.
I don't have access to my source code, but I never had problems with all
of those "back in the days". Certainly not to differentiate between an
80186 and an 80286. Only for later CPUs of various clone manufacturers
there were issues regarding proper detection of 386+ features....
But that should not apply to what was written in that Heise article....
Ralf
--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
_______________________________________________
Freedos-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-devel
