OSDev.org

Hello! I am implementing some COM port stuff and want to test it on real hardware, the only problem is I don't have 2 serial ports on my PC(I use virtual serial ports for testing) so can someone test this
what they need on the pc's is at least 4mb of ram, a floppy drive, and a com port that connects them together; This will not access your harddrive or anything so no worries

now then what you do is:
1. download this floppy image: http://JouleOS.galekus.com/temp/tmp1.zip and copy it to floppies for booting

2.start the pc's with the floppies in it and wait for them both to get to a command line prompt thingy

3.then you type "test;t36" and press enter on both of them (without quotes though)

4. once you did that to both of them push a key on 1 of them and wait a few seconds then push a key on the other one

5. then once they both say "done" press a key on both of them and it should say "test" on both of them


btw another thing to check on is if when you booted it said Com port hardware: [failed] or com port line: [failed]

thanks in advance

Both computers report the same thing
I could not get a change in this message by connecting to nothing, the wrong com port, or by leaving out the null modem (tested in all combinations, main reason was that Mercury has two COMports and the casing doesn't read '1' or '2').

I ran the test between Io (2xPII) and Mercury (486). Uranus refused to boot it (Celeron,466mhz). Other pc's left untested.

Both reported all ok's during startup
However, Mercury reported a ps2 mouse which was not connected (the port was present however).
Some other quirks include: the mouse on Io was way off and for some reason i managed to leave a spare copy of the mouse pointer somewhere on the screen. (race condition/scrolling screen?) Secondly, had to guess keys for an us keyboard when an dutch keyboard was used (semicolon is located one row down on these)

Specs this page - just find my post

Maybe it helps sharing source code so I or some other people can figure where the error goes.

If you want to try again, just let me know.

Ok thanks for testing it!

Uranus refused to boot it (Celeron,466mhz)

celeron is 486+ isn't it(OS requires a 486)
what exactly are null modems? are they just like com ports with a little attachment on them so they will go across phone lines?


as far as the mouse, that is very wierd on the mosue failing but the mouse screen stuff is way off(trying to fix that)

I don't have anything but a US scancode-to-ascii map so...



now I want to change the test a bit
1. boot it and such on both pc's

2.type terminal and press enter for both pc's

3. just try typing stuff and see if it matchs on both pc's(esc to exit)


known bugs:
backspace don't fully behave like it should
arrow keys to set cursor doesn't work
tab don't work
a problem may occur if you escape from terminal on 1 pc and then start it back or something(just found that out..)
newline stuff isn't working(I know the problem and a solution just haven't got it done yet


code:(condensed)


a new image file is uploaded at
http://jouleos.galekus.com/temp/tmp2.zip


also another thing you can try is use window's hyperterminal @9600kbit/s, data bits:8bits, parity: none, stop bits: 1, flow control: none
and interface with it to the JouleOS terminal(this works with virtual com ports completely[excluding files of course though])


lol my message is too big!
excluding source...

source is now attached

Jordan3 wrote: what exactly are null modems? are they just like com ports with a little attachment on them so they will go across phone lines?

A null-modem cable is a serial cable used to connect directly two pc via serial ports.

Celeron should be a 686, IIRC. I don't think, however, that a code that can run on a 486 is not able to run on a superior class pc... except if you put some race conditions hardcoded into your source ( example: if ARCH > 486 do not start )

Axel