OSDev.org

disk image, kernel&modules sources
compilation tools

Any feedback is welcome. 0.8.14 mainly brings back address space creation to the new-and-improved memory management system. you may wish to test "procCreate" in "testp2" module ...
use [SHIFT+SPACE] to launch the boot sequence and [ESC] when boot completes to dismiss the kinit thread.
[SHIFT+F4] [CTRL+F5] will bring you to the module manager panel.

http://clicker.sourceforge.net/docs/cake/debug.html and http://clicker.sourceforge.net/pixs/snapshots.html may help if you feel lost in the E.S.S.U.I (ExtremelySimpleSuperUserInterface)

Have given it a look in VMware. Works like a charm, loads and sits there - and is happy.

what exactly should I give a test? special behaviour? Unexpected behaviour? (hmm ... sometimes I encounter a bit of it in my own kernel - mostly due to my own faults).

It works fine ;D
But the screen should scroll, it's better to read

[SHIFT+SPACE] to launch boot sequence.
[ESC] to close initialization thread.
[SHIFT+F4] to show Info shower console.
[CTRL+F5] to show module manager panel.
Install testp2 module from available list.
[CTRL+F6] to show panel added by testp2.
Run procCreate - displays "test successful"
[SHIFT+F1] to check output on main console

ASCII <=> <=> )-] <=> <=> scrolls continuously across 2nd line of console.

Is this the output of the thread created by procCreate, and if so, should it be killed (as stated on panel 6), and stop scrolling the ASCII on line 2 ?

Tested disk image on Bochs, under winXP.

[edit]For each thread created with procCreate, a new line is added on the main console for the ASCII output[/edit]

i failed to boot with floppy image with bochs

hmm? can you detail your BOCHS version and config file ?





thanks all for testing

I use bochs 2.1.1, with the same config as you, except hdd, i have no ata disk atached(but i don't think this is a problem).

> the kenrel hangs after pressing shift+space ??
> i can switch consoles, but, i can't enter commands.
> the os indicate a kernel panic

there are no way to 'invoke commands" in the microkernel. All you can do is switching to the "panel system" (SHIFT+F4) and use CTRL+F5 to display the module manager panel.

Can you tell the registers value and the stack trace of the kernel panic you got ?

How many megs did you gave to the system ?

I have downloaded the image and tested it on real hardware. It runs perfectly but it is very hard to figure out what the kernel is doing/not doing, if it is doing good or bad, because of the strange and too cryptic "UI", IMHO.

I'd like to see a better organized interface in the next versions, maybe I will be more useful. ;D

Hi,

I tested the image "c32-img-0.8.14.gz" on 6 of my computers and a couple different Bochs 2.1.1 configurations.

On Bochs I (almost - see below) had no problems. Bochs was configured for single CPU 80586, a single CPU 80686 and 80686 with 8 logical CPUs.

On 3 of the computers it crashed during boot. The error message is the same each time:



The computers are:



I don't know if your OS supports pentium compatibles or not. If it doesn't perhaps the OS should check what it's running on and display some sort of "CPU too old" error message?

On one computer I hit enter when cowtest was selected [shiftF4][controlF5] and it displayed "loading cowtest" at the bottom of the screen permanently - as if it was trying to load it forever. Then (in the same window) I held my finger down on the enter key (key repeat) while procCreate was selected and it crashed. Error message was:



This error makes me wonder if your IRQ handling is re-entrant. I also notice that syscow::init is listed so I guess it might have locked up during init

The second time I tested it on this computer I started procCreate at [shiftF4][controlF6] many times (trying to re-create the previous crash), and the screen at [shiftF4][controlF7] reported 3 instances of procCreate with a thread running each (but no crash).

I decided to try out key repeat on pageFrameViewer at [shiftF4][controlF7] too and noticed that the screen at [shiftF3] was full of messages saying "04BE (0070, 2744):> Warning: kdsSend error 0000000B processing showers.p2-check:" followed by a line containing text that doesn't give any indication that there's something to give a warning about (e.g. "virtualCopyTable(00000000) job completed. pgSpace ready." or "virtualCopyTable(00000000)>thread 0007F2D8 calls hyde in xpSYNC mode (notify=000") I don't know if this is normal or not...

This computer is:




AAARGHHH! This message was too long so I've split it into 2 seperate posts

Hi again!

The next computer I think worked fine, but on the screen at [shiftF3] it was repeatedly (non-stop) displaying the line:



And at the bottom of the screen:


I'm not sure what I did to trigger this, and after pressing more keys it finally disappeared. Not sure if it's normal or not...

The second time I booted on this computer I hit enter when cowtest was selected [shiftF4][controlF5] and it displayed "loading cowtest" at the bottom of the screen permanently. I've tested this more under Bochs and it seems to always happen if cowtest is the first thing you try on this screen (if you try something else first, then cowtest will work as it should). Anyway, the second time I booted on this computer I couldn't get the repeating "suspending self" messages to happen.

The computer was:




On the next computer it did work mostly fine, although I only booted once:



There are a few problems that occur on everything that boots. Pressing enter when spaceDTOR or virtualCopyTable is selected [shiftF4][controlF6] gives a "error 11 :" down the bottom. The floppy disk drive motor is never turned off. Also the keyboard driver seems a little funny - on the screen at [shiftF4][controlF5] I've been hitting enter on everything in the left hand list. I hit enter, down cursor, enter, down cursor, etc and half the time down cursor moves twice. Then there's the Info shower console screens, where all of them except [controlF3] don't clear the top lines - you can check this by pressing [shiftF4][controlF8][controlF1][controlF8][controlF2], etc.

Then there's various spelling and/or text positioning errors. E.g. "Info shower console -use CTRL+Fx to switch between infos pannels a l0t of Fun --" where the end part ("a l0t of Fun --") looks like it should've been cleared, or the screen at [shiftF4][controlF2] where "root@003E6248" is messing up the borders on the left, or the screen at [shiftF4][controlF1] where the right hand list seems an item too long. Sorry if I sound too much like a critic

Ok, the last problem I've found is with KVM switches (Keyboard, Video, Mouse). The KVM's allow me to connect multiple computers to the same Keyboard, Monitor and Mouse, and select the computer via an old fashioned switch or by pressing a magic key sequence. I'm actually using 3 of these things. Your OS works fine on the older (manual switch) versions, but on the newer electronic (magic key sequence) KVM's your OS stops accepting keypresses when I switch back to it. I have no idea why (the same KVM's work fine under DOS, windows, linux and my OS).

That's all I could find for now ;D


Cheers,

Brendan

Thanks for reminding me of this. Clicker reserves 16MB of physical memory for I/O. I'll try to make it check the avl. size and report a complaint if it cannot work with what's there. Also, considering the amount of 16MB systems you have, i'll make it reserve less IO memory if it appears the system has not much RAM (like, only 4MB on >32MB systems).

Actually, there's two things (which makes me believe i should really add a KEXT telling what the module is expected to do when the user points at it
- the 'syscow' test is an attempt to check that System-Level pages may be write-protected ... If the CPU supports the extension, the module execution results in a page fault. I should warn the user about this...

- there's an issue with the post-a-job feature in 0.8.14 that i discovered recently and (at least hope) fixed ... as soon as the KEXT-thing is added aswell, i'll put a new release, which should also provide hints about what key could be used in which panel (hoping these two things will make the ESSUI a bit less cryptic)

a **huge** thanks for you to have carried all these tests.

I booted your image in Bochs and got "ERROR cowardly refusing..." (see image). I don't know how bad this effects things but I can still change from consoles 1-4, but can't get to the 5th.

http://www.geocities.com/idontthinkalot/clicker.png

I get this each time I boot, so if you need any more information I could probably get it...


PS, was really looking forward to trying the new Clicker since I hadn't seen it in a while