Re: [Bochs-developers] bx_debug and print-stack

[Stanislav S. <stl...@gm...>]

> The printed offset is wrong, but i expected it because i have modified
linear_sp.

The printed offset is SS.BASE + offset, which is correct linear SP ...

Stanislav

-----Original Message-----
From: Sebastian Herbszt [mailto:he...@gm...] 
Sent: Sunday, January 10, 2010 2:58 PM
To: Stanislav Shwartsman; boc...@li...
Subject: Re: [Bochs-developers] bx_debug and print-stack

Stanislav Shwartsman wrote:
>I should see your test program and look what are you actually doing.

The test program (16-bit real mode) gets loaded at 0x33100 and is

<bochs:11> disasm /7 0x33100
00033100: (                    ): mov ax, 0x1234            ; b83412
00033103: (                    ): push ax                   ; 50
00033104: (                    ): mov ax, 0x5678            ; b87856
00033107: (                    ): push ax                   ; 50
00033108: (                    ): pop ax                    ; 58
00033109: (                    ): pop ax                    ; 58
0003310a: (                    ): int 0x20                  ; cd20

on entry the registers are as follows

<bochs:13> r
rax: 0x00000000:00003300 rcx: 0x00000000:0000000c
rdx: 0x00000000:00000580 rbx: 0x00000000:c5390100
rsp: 0x00000000:0000fffe rbp: 0x00000000:00032ca2
rsi: 0x00000000:1fea0000 rdi: 0x00000000:00000083
r8 : 0x00000000:00000000 r9 : 0x00000000:00000000
r10: 0x00000000:00000000 r11: 0x00000000:00000000
r12: 0x00000000:00000000 r13: 0x00000000:00000000
r14: 0x00000000:00000000 r15: 0x00000000:00000000
rip: 0x00000000:00000100
eflags 0x00000246: id vip vif ac vm rf nt IOPL=0 of df IF tf sf ZF af PF cf

<bochs:14> sreg
es:0x3300, dh=0x00009303, dl=0x3000ffff, valid=1
        Data segment, base=0x00033000, limit=0x0000ffff, Read/Write,
Accessed
cs:0x3300, dh=0x00009303, dl=0x3000ffff, valid=1
        Data segment, base=0x00033000, limit=0x0000ffff, Read/Write,
Accessed
ss:0x3300, dh=0x00009303, dl=0x3000ffff, valid=7
        Data segment, base=0x00033000, limit=0x0000ffff, Read/Write,
Accessed
ds:0x3300, dh=0x00009303, dl=0x3000ffff, valid=1
        Data segment, base=0x00033000, limit=0x0000ffff, Read/Write,
Accessed
fs:0x9c72, dh=0x00009309, dl=0xc720ffff, valid=1
        Data segment, base=0x0009c720, limit=0x0000ffff, Read/Write,
Accessed
gs:0x9c72, dh=0x00009309, dl=0xc720ffff, valid=1
        Data segment, base=0x0009c720, limit=0x0000ffff, Read/Write,
Accessed
ldtr:0x0000, dh=0x00008200, dl=0x0000ffff, valid=0
tr:0x0008, dh=0x00008b00, dl=0x05800067, valid=1
gdtr:base=0x000000000000b160, limit=0x2f
idtr:base=0x0000000000000000, limit=0xffff

so the stack is at 0x3300:0xfffe

After the first "push ax" i have
<bochs:17> print-stack 4
Stack address size 2
 | STACK 0xfffc [0x0000]
 | STACK 0xfffe [0x0000]
 | STACK 0x10000 [0x0000]
 | STACK 0x10002 [0x0000]

ss:sp is 0x3300:0xfffc

<bochs:24> x/1h 0x3300:0xfffc
[bochs]:
0x0000000000042ffc <bogus+       0>:    0x1234

> bx_dbg_read_linear is too generic, I don't believe it doesn't work. You
> won't boot anything if it is the case.
> Are you on big endian host (this might be real issue) ?

Vista 32-bit host with Intel cpu.

If i change the code in dbg_main.c like the following i get the correct
values

    else {
//      linear_sp = BX_CPU(dbg_cpu)->get_reg16(BX_16BIT_REG_SP);
      linear_sp = BX_CPU(dbg_cpu)->sregs[BX_SEG_REG_SS].cache.u.segment.base
+ 
BX_CPU(dbg_cpu)->get_reg16(BX_16BIT_REG_SP);
      len = 2;
    }
  }

<bochs:5> print-stack 4
Stack address size 2
 | STACK 0x42ffc [0x1234]
 | STACK 0x42ffe [0x0000]
 | STACK 0x43000 [0x0000]
 | STACK 0x43002 [0x0000]

The printed offset is wrong, but i expected it because i have modified
linear_sp.

- Sebastian