Mojo 0.1.2 released - please test and comment

[cyr1x]

Today I tested it on a IBM xSeries 235 and it panics .
Getting the same error as JackScott.

PANIC: kernel::memory_init::calculate_physical_memory_info() => Weird memory size

[clange]

Hi cyr1x

Did you test the 0.1.2 or 0.1.3-alpha image? If you tested the old could you please test with the newest image. You should use http://softwarewizard.dk/mojo/downloads ... mm-iso.zip since it will certainly display the debug output on the kernel console. If it still doesn't work could you please transcribe the memory map. That will allow me to tests how my code handles you memory mapping.

clange

[cyr1x]

I tried it with 0.1.3-alpha (at least this is what the filename states). If I have the time I'll try to get the memory map, but currently I can't just shut down the server as it is in use.

[cyr1x]

Ok, here we go:

Code: Select all

mem block (multiboot): 0x00000000 - 0x0009c000, 0x0009c000; type = 1
mem block (multiboot): 0x0009c000 - 0x000a0000, 0x00004000; type = 2
mem block (multiboot): 0x000e0000 - 0x00100000, 0x00020000; type = 2
mem block (multiboot): 0x00100000 - 0x7ffd8740, 0x7fed8740; type = 1
mem block (multiboot): 0x7ffe0000 - 0x80000000, 0x00020000; type = 2
mem block (multiboot): 0x7ffd8740 - 0x7ffe0000, 0x000078c0; type = 3
mem block (multiboot): 0xfec00000 - 0x00000000, 0x01400000; type = 2

[clange]

Hi cyr1x

Thanks for the info. I can easily see that my current code can't handle your memory map. The ACPI reclaimable memory area is not page aligned and I assume that all memory blocks are page aligned. So this is great test data. I will add it to my test suite and as soon as I can handle it I will let you now.

Merry Christmas

[clange]

Hi cyr1x

I have improved my memory initialization code to be able to handle your memory map. It was great test data since I was forced to unit test yet another part of the kernel code. I also had to improve the implementation. Hopefully everything will work on your machine now.

There is a small typo in the memory map you transcribed. A '0' has to be changed to a '9'. I have highlighted the change below. Just in case others wants to use the data.

mem block (multiboot): 0x00000000 - 0x0009c000, 0x0009c000; type = 1
mem block (multiboot): 0x0009c000 - 0x000a0000, 0x00004000; type = 2
mem block (multiboot): 0x000e0000 - 0x00100000, 0x00020000; type = 2
mem block (multiboot): 0x00100000 - 0x7ffd8740, 0x7fed8740; type = 1
mem block (multiboot): 0x7ffe0000 - 0x80000000, 0x00020000; type = 2
mem block (multiboot): 0x7ffd8740 - 0x7ffe0000, 0x000078c0; type = 3
mem block (multiboot): 0xfec00000 - 0x00000000, 0x01400000; type = 2

I have also added support for German.

Could you please test again when you find the time. Thanks.

clange

EDIT: fixed an (embarrassing) typo.

[giszo]

Hi!

I just tried to boot the 0.1.3 alpha version of your OS but it resulted in a KERNEL PANIC ASSERT right after boot.

Here's what I got:

Code: Select all

KERNEL PANIC ASSERT: src/kernel/interrupt/interrupt.c:161 failed expression: `g_exit_kernel_time < g_enter_kernel_time`

EDIT: 0.1.2 works fine

giszo

[cyr1x]

Ok now getting a Kernel panic at line 60 in page.c -> "0 == (base % PAGE_SIZE)"

mem block (multiboot): 0x000fc000 - 0x000a0000, 0x00004000; type = 2

This should rather be 0x0009c000 , but thanks for finding this typo anyway. I'll fix it in the table above.

I have also added support for German.

Cool

[clange]

@cyr1x:
Hehe, fixing a typo with a typo - I should have paid more attention

@gizso:
Does it fail every time or only occasional?

@cyr1x and gizso:
You are really exposing some weird bugs - I have looked at my code and can simply not explain what is happening. So clearly I have made a mistake I have added some more debug output could you please try again - thanks.

clange

[giszo]

@gizso:
Does it fail every time or only occasional?

0.1.3 fails on every boot.

@cyr1x and gizso:
You are really exposing some weird bugs - I have looked at my code and can simply not explain what is happening. So clearly I have made a mistake I have added some more debug output could you please try again - thanks.

I'll test it and do a new post with the results

Oh, and one more thing, my nick is giszo, not gizso

giszo

[giszo]

Tested the new 0.1.3 .iso and it works fine. Tried multiple times and I have no error

giszo

[clange]

Hi giszo

Sorry about misspelling your name - I even made an effort to do it correctly, but alas not enough

Thanks for testing again. I really glad to hear that the new image works. Could you please tell me which hardware you tested on. I will start to compile a list of machines that it runs successfully on.

clange

[giszo]

Hi!

I tested it on Sun xVM VirtualBox 2.0.4.

giszo

[clange]

Ok, I have added Virtual Box as a test target before releasing now. That should ensure that it will work in the future

clange

[cyr1x]

Now I'm getting the same error as above,but now on line 64.