Problem with code running in Bochs

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 11:47 am

I am having some trouble with getting my code to work and when I ran it in Bochs, I got a weird error. Does anybody know what this error means and how I would go about solving it.

Code: Select all

00000542541p[CPU  ] >>PANIC<< can_pop(): found SP=ffff
00000542541i[SYS  ] Last time is 1042997919
00000542541i[CPU  ] protected mode
00000542541i[CPU  ] CS.d_b = 32 bit
00000542541i[CPU  ] SS.d_b = 32 bit
00000542541i[CPU  ] | EAX=60000010  EBX=00001000  ECX=000a0002  EDX=00000000
00000542541i[CPU  ] | ESP=ffffffff  EBP=ffffffff  ESI=00000000  EDI=0000ffe4
00000542541i[CPU  ] | IOPL=0 NV UP DI NG NZ AC PE NC
00000542541i[CPU  ] | SEG selector     base    limit G D
00000542541i[CPU  ] | SEG sltr(index|ti|rpl)     base    limit G D
00000542541i[CPU  ] |  DS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000542541i[CPU  ] |  ES:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000542541i[CPU  ] |  FS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000542541i[CPU  ] |  GS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000542541i[CPU  ] |  SS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000542541i[CPU  ] |  CS:0008( 0001| 0|  0) 00000000 000fffff 1 1
00000542541i[CPU  ] | EIP=00001002 (00001001)
00000542541i[CPU  ] | CR0=0x60000011 CR1=0x00000000 CR2=0x00000000
00000542541i[CPU  ] | CR3=0x00000000 CR4=0x00000000

Slasher · Post by **Slasher** » Sun Jan 19, 2003 11:54 am

Your Os is trying to return from an interrupt with IRET but the values on the stack are not accurate or appropriate. Its a problem with the interrupt and stack working together. If you are trying tp set up a stack frame for stack based task switching then you frame setup code is not working properly!
if you paste code, it will be easier to know exactly what is the problem.

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 11:59 am

[attachment deleted by admin]

Slasher · Post by **Slasher** » Sun Jan 19, 2003 12:12 pm

HI,
you are setting ESP to 4GB without paging on in your boot code!!! You should put it an adress that EXISTS such as 1mb,2mb or 4mb!
Try that and see.

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 12:17 pm

When I do that, I get:

Code: Select all

>>PANIC<< exception(): 3rd (13) exception with no resolution

Slasher · Post by **Slasher** » Sun Jan 19, 2003 12:20 pm

Disable interrupts cause you do not have an IDT in place!

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 12:23 pm

I never enabled interrupts in my pmode bootsector.

Slasher · Post by **Slasher** » Sun Jan 19, 2003 12:30 pm

then the problem occurs when it wants to execute the c kernel try this, put
hlt
before
jmp LINEAR_CODE_SEL:0x1000
this will halt the cpu, if there isn't any problem then the boot code is fine and the problem is with the kernel part.
how did you write the kernel binary file to the disk?
all so add
-fwritable-string --nostdlib --freestanding (check the syntac in gcc info files) when you are compiling the kernel and remove the del *.bin and *.o commands from you compile.bat, you might be deleting the kernel before you even write it to disk!

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 12:57 pm

I changed a few things and I now I get a: running in bogus memory.

Slasher · Post by **Slasher** » Sun Jan 19, 2003 1:01 pm

did you first try halting the cpu as i suggested above?
what happened when you did?(if everything is fine, you should not get any messages from BOCHS, it will just stop)
Then tell me how you compiled the kernel(you can paste the lines in you next post)
ans how you wrote it to the disk?

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 1:08 pm

Code: Select all

Did you first try halting the cpu as i suggested above?

YES

Code: Select all

what happened when you did?(if everything is fine, you should not get any messages from BOCHS, it will just stop)

It stopped.

This is how I compiled it

Code: Select all

nasm -f bin bootpm.asm -o bootpm.bin
nasm -f aout km.asm -o km.o

gcc -c kernel.c -o kernel.o 
ld -T link.ld -o kernel.bin km.o kernel.o

bootcopy bootpm.bin 0
bootcopy kernel.bin 1

makeboot test.img bootpm.bin kernel.bin

km.asm

Code: Select all

[BITS 32]
[global start]
[extern _k_main] ; this is in the c file

start:
  call _k_main

  cli  ; stop interrupts
  hlt  ; halt the CPU

link.ld

Code: Select all

OUTPUT_FORMAT("binary")
ENTRY(start)
SECTIONS
{
  .text  0x1000 : {
    code = .; _code = .; __code = .;
    *(.text)
    . = ALIGN(4096);
  }
  .data  : {
    data = .; _data = .; __data = .;
    *(.data)
    . = ALIGN(4096);
  }
  .bss  :
  {
    bss = .; _bss = .; __bss = .;
    *(.bss)
    . = ALIGN(4096);
  }
  end = .; _end = .; __end = .;
}

I just took out the stub function of my kernel.
I set my pmode stack to 0x100000.

Code: Select all

Slasher · Post by **Slasher** » Sun Jan 19, 2003 1:14 pm

hi,
i once had a problem when i started my kernel and the only way i was able to solve it was by writing out my adresses in full in the linker script! so try
0x00001000 instead of 0x1000!

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 1:24 pm

I changed it.

I noticed something. When I set my pmode stack which is esp = 0xffff, it says running in bogus memory and when I look at the bochs dump, it says 00000fff.

If I set esp = 0x100000 (1 MB) it loads, displays a whole screen of blue symbols and then halts = running in bogus memory.

If I set esp = 0xFFFFFFFF it loads and halts like a: hlt.

What could be the problem?

Slasher · Post by **Slasher** » Sun Jan 19, 2003 1:42 pm

its the kernel code!
what is makeboot?
I think you should use a program like partcopy to copy the bootcode to the first sector and then the kernel to the 2nd sector of a real disk and then use partcopy to make an image of the disk
eg
partcopy -f0 0 1680000 test.img will create an image of the entire floppy called test.img
partcopy kernel.bin 200 <size of the kernel> -f0 will copy the kernel.bin file to sector 2 of the floppy(sectors numbers start from 1 not 0)

beyondsociety · Post by **beyondsociety** » Sun Jan 19, 2003 1:56 pm

I tried what you suggested and I get the same results. Look at the bochs memory dump and see if you see anything weird. In my pmode bootsector, ESP = 0x100000.

Code: Select all

 >>PANIC<< prefetch: running in bogus memory
00000803660i[SYS  ] Last time is 1043006043
00000803660i[CPU  ] protected mode
00000803660i[CPU  ] CS.d_b = 32 bit
00000803660i[CPU  ] SS.d_b = 32 bit
00000803660i[CPU  ] | EAX=60000010  EBX=00001000  ECX=000c0002  EDX=00000000
00000803660i[CPU  ] | ESP=00001000  EBP=00000000  ESI=00000000  EDI=0000ffe4
00000803660i[CPU  ] | IOPL=0 NV UP DI PL NZ NA PE NC
00000803660i[CPU  ] | SEG selector     base    limit G D
00000803660i[CPU  ] | SEG sltr(index|ti|rpl)     base    limit G D
00000803660i[CPU  ] |  DS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000803660i[CPU  ] |  ES:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000803660i[CPU  ] |  FS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000803660i[CPU  ] |  GS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000803660i[CPU  ] |  SS:0010( 0002| 0|  0) 00000000 000fffff 1 1
00000803660i[CPU  ] |  CS:0008( 0001| 0|  0) 00000000 000fffff 1 1
00000803660i[CPU  ] | EIP=06001000 (06001000)
00000803660i[CPU  ] | CR0=0x60000011 CR1=0x00000000 CR2=0x00000000
00000803660i[CPU  ] | CR3=0x00000000 CR4=0x00000000

OSDev.org

Problem with code running in Bochs

Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs

Re:Problem with code running in Bochs