OSDev.org

This basic release of DynatOS is to help determine if proper memory management is being initialized amongst different IA32 platforms. Basically, it is a user-feedback test with very little interaction capability in order to test the stability of the kernel. What I need the most is feedback from the "mem" command, but if it doesn't even boot up that far... please let me know

I already know how this test version works on emulators such as Bochs, VirtualPC and VMWare. I would greatly appreciate feedback/replies to this thread from anyone who downloads and runs this test version on a Real PC.

Features
-Command Line Interface (extremely basic, QWERTY scancode response)
-Runs in Protected Mode w/ Ring-3 Context Switch capability (base for later)
-Accepts 3 commands, "help", "clear" and "mem"

Requirements for Testing
-Intel 386 Compatible PC or Higher
-Floppy Disk Drive (and a blank floppy)
-Keyboard
-Monitor

Installation
-Download from link below
-Install on a standard 3.44 Floppy Disk using RawWrite (Windows), DD (*nix/Linux/BSD) or some other equivalent program. It can be written to a bootable CD such as Nero, but I wouldn't want anyone to do so unless it is a CD-RW (no loss, this test isn't worth wasting a CD-R over).

Outlook
-Keep releasing this test version until Memory Management Initialization Bugs are fixed
-Implement (enable) Memory Management capability
-Implement basic Process Management capability
-Design a few more hard-coded commands and loadable programs

Download Link
-----> http://www.dynatos.org/project/dynatos-20060914.bin <-----

VMWare Workstation (Linux/5.5.1)
Memory: 32 MB
Hard Disk (SCSI 0:0): 1024MB
Floppy: (using bin downloaded)
Processor: Intel Celeron Mobile 2.6GHZ

I also tried 256 MB of memory, and the same result.

The screen blanked after showing the BIOS message, and stayed blanked. I waited no longer than 15 to 20 seconds.

I was aware of results on all Emulators (including VMWare). The "blank screen" issue results from the old-old-old-old Video Initialization routine that depends on the CMOS values. Since that basic release (yours is not the first complaint), I changed it to rely on the BDA instead.

Thanks for testing anyhow. If you have any test results from a real computer, please let me know

I ran a test on my dual-p2 box "Io", and i was able to boot it.
However, the output of 'mem' doesnt really seem to match the actual hardware. (unless thats for a reason...)

screendump:
[pre]
<sys://> mem

------------------------------
| Memory Information |
------------------------------
Total Memory: 131272 KB
Physical Memory: 130048 KB
Available Memory: 129916 KB

<sys://>[/pre]
Physical memory reads as 127 MB (theres 128 installed, maybe the first 1 MB is ignored?)
Total Memory reads as 128.2 MB which doesnt sound correct either. (i tried some random thoughts but i cant reconstruct this number from anything)

Full specs of Io

Sorry for the delayed response, I've been gone awhile.

Thanks for testing it out.

For Physical Memory, the hole at the 15-16 MB Mark is ignored. Since you don't have any means of "getting it back", it is simply not counted.

There are still a few things I have to do to tweak the memory detection process. The biggest one is detecting for overlapping E820 entries, which I think might be occurring with your example since it is 200KB above 128MB.

At any rate, the detection routine seems to be working in general... so that allows me to proceed on to fixing and adding other things

Once again, thanks for testing it out