OSDev.org

Finally managed to restore the ancient power of Clicker: reading your disks' MBR to display the partition type tag (plus loading IDELBA driver on demand). That concludes the parenthesis opened with the rewrite of paging support one year ago !

Don't hesitate any longer: get your copy today

Hmm, and btw, better keep the floppy away from dad's precious computer, right ? ...

Any feedback is welcome... sources release will follow within a few days.

Tested on a

Pentium-S 100 (i think its a P54C)
8mb ram
i430FX chipset
printed 100010001000DEAD

Pentium III-450
192mb ram
i440BX chipset
printed 100010001000DEAD

Tested on QEmu
first go printed 800080008000DEAD.
seemed to think it was booting off a HD and not a floppy.

worked fine second go when i forced it to think it was a floppy.

hmm ... weird enough ... those codes are bios error codes, iirc. 80h is "timeout (not ready)" and 10h is "uncorrectable CRC or ECC error on read" according to RBIL, so if i were you, i'd toss that floppy away and try with a fresh/confirmed one ...

If it still doesn't work, i'll go back to my sources with humility ...

humm.. ok i will try with a different floppy and let you know.

must have been the floppy. i ran it on the pentium-s 100 and it booted fine. (strange how the image wrote to the 1st disk fine.. ohwell).

seemed to boot up ok and give me all the screens of info.

tnx. For those who want, the sources are online, together with the tools required to compile them ... A Guide for compiling under linux is there too.

As for the ATA test (on test machines only. Remember, no warranty, cannot be held liable for data loss etc):

1. let the system boot (press <Shift>+<SPACE>, <ESC> and <Shift>+<F4> when requested)
2. call the module manager with <ctrl>+<F5>
3. load testio module
4. when loaded, switch to testio panel with <ctrl>+<F6>
5. run the "readMBR" test case. You should see something like the attached image ...

If testio hangs (when reading or on loading), try shifting to the log with <shift>+<F3> before pressing ENTER

--
http://clicker.sourceforge.net/pixs/c32 ... uccess.png

Played around in virtual pc a bit, loaded some blinkenlights applications (console 01 goes wild! ;).
The mbr test seemed to work. A sector-hex-viewer would be fun.
What I like best is the page-debug-viewer, it's just cool :)
But your apps seem to load slowly sometimes (e.g. I loaded the testio module, but it took about a second or two to show up).

Anyway, good work!

mpp.eox3 wrote: I loaded the testio module, but it took about a second or two to show up

Yup, that's normal: the 'testio' module need to load the whole I/O subsystem first, which in turns load the available bus drivers which scan your hardware ...

In particular, when the ATA busdriver is loaded, it scans ATA bus for devices, and when no device is present, it may require timeouts to discover no device is there, depending on your controller ...

Excellent Work!!

Tested on

Celeron 633MHz
Chipset Unknown (pre-built system)
224MB RAM

Pentium 4 3.0GHz
i875P Chipset
1GB RAM

Both cases seemed to go perfectly, detecting all installed hard disks, able to run through all info screens, no problems at all.

Cheers.

Hey pype,

I was looking through your source code (loader.asm) and noticed that the comments suggested something that wasn't the case in the code. Nothing major, in fact may never be a problem but thought I would just let you know anyway.

VIRTUALSTACKSIZE???equ 0x0100 ; v86 got a 512bytes stack
REALSTACKSIZE??????equ 0x0100 ; so does the real-mode
EXCEPTSTACKSIZE??????equ 0x0100 ; exception got a 1K stack
SYSTACK?????????equ 0x1000 ; so does kernel

EXCEPTSTACKSIZE is 512 bytes and your comments say that you intend it to be 1024 bytes.

Cheers.

Hi,

XStream wrote: EXCEPTSTACKSIZE is 512 bytes and your comments say that you intend it to be 1024 bytes.

You mean all of the sizes aren't consistant

VIRTUALSTACKSIZE and REALSTACKSIZE are 256 bytes, not 512. EXCEPTSTACKSIZE is 256 bytes not 1 Kb, and SYSTACK is 4 Kb instead of 1 Kb


Cheers,

Brendan

There you go, I noticed the difference of 0x100 and 0x1000 and took it at face value, never actually thought about the values. Well spotted.

mpp.eox3 wrote: Played around in virtual pc a bit, loaded some blinkenlights applications (console 01 goes wild! .
The mbr test seemed to work. A sector-hex-viewer would be fun.
What I like best is the page-debug-viewer, it's just cool

Note that if you head to the last page table and use the page-debug-viewer (pressing ENTER over a page), you get a chance to have a sector-hex-view of the MBR

Pype.Clicker wrote: Note that if you head to the last page table and use the page-debug-viewer (pressing ENTER over a page), you get a chance to have a sector-hex-view of the MBR :P

Ah undocumented cheats.. ;)
I will implement some in my kernel too, activated by some good old C=64 style "POKE 12345,242" commands ;P

mpp.eox3 wrote: Ah undocumented cheats..
I will implement some in my kernel too

Hmm ... I'll consider adding a cheat-sequence that brings you directly to the final boss arena