protected mode, working..almost

[ggodw000]

made a lot of progress since I posted this issue earlier:
http://forum.osdev.org/viewtopic.php?f=1&t=29900

now my code enters and exits protected mode just fine after following fixes:
-corrected the way GDT PTR is calculated (use flat address instead of SEG:OFF). that means before correction. lgdt was loading some arbitrary address and system was still stable before loading segment selector!
-did spend lot of time on descriptor format and make sure every bits are set as intended.
-I did parseDesc to printout the every segment descriptors in a GDT in very verbose manner (pasted below)

Now I enter pmode grab few bytes (dword) from memory and display back after exiting using segment selector+offset.
I did two different known addresses:
BDA at 0:0400h - this is known memory location as first dword has serial port address
RSD PTR at 0f45f0h - another known memory location that reside somewhere between e000-ffffh segment and has "RSD "

Once pass these address to function that enters pmode and exit, I see dword at BDA returned right bytes: serial port addresses "03f802f8"
But RSD PTR is not returning right bytes: it should return part of signature "RSD " in 1st dowrd.

If I read the same location from real-mode at the same time in same program, it DOES return the correct signature.

I am wondering what could be the culprit for the latter. Both location are read exactly the same way.
I will be trying few more known memory locations to see if I can find out any pattern.

Code: Select all

==============
Descriptor No: 01
===============:
Descriptor Info:
---------------:
0000FFFF:00CF9A00
Segment base:       00000000
Segment size(lim):  000FFFFF
Rx06[7]:     01 - Granularity (0: multiplier 1, 1: multiplier 4K):
Rx06[6]:     01 - Default Rx size: (0: 16-bit, 1: 32-bit):
Rx06[5:4]:   00- AVL (for O/S use any way):
Rx05[7]:     01 - Present bit: (0: not present, 1: present):
Rx05[6:5]:   00 - DPL(desc.priv.lvl bit:
  Rx05[4]:   01 - descriptor type (0: system segment, 1: data/code segment):
Rx05[4:1]:   0A - segment descriptor type:
  Rx05[3]:   01 - code/data bit (0: data/stack/non-exec, 1: code segment):
    Rx05[2]: 00 - expansion direction(data)/conform(code) bit:
    code:       0=code can only be exec-d from priv.lvl set in ring
                1=code can be exec-d from same or lower priv.level
    Rx05[1]: 01 - W(data)/R(code)
    code:       0=read access not allowed
                1=read access allowed (write access never allowed)
Rx05[0]:     00 - A bit(accessed, each time CPU accesses it)
==============
Descriptor No: 02
===============:
Descriptor Info:
---------------:
0000FFFF:00CF9200
Segment base:       00000000
Segment size(lim):  000FFFFF
Rx06[7]:     01 - Granularity (0: multiplier 1, 1: multiplier 4K):
Rx06[6]:     01 - Default Rx size: (0: 16-bit, 1: 32-bit):
Rx06[5:4]:   00- AVL (for O/S use any way):
Rx05[7]:     01 - Present bit: (0: not present, 1: present):
Rx05[6:5]:   00 - DPL(desc.priv.lvl bit:
  Rx05[4]:   01 - descriptor type (0: system segment, 1: data/code segment):
Rx05[4:1]:   02 - segment descriptor type:
  Rx05[3]:   00 - code/data bit (0: data/stack/non-exec, 1: code segment):
    Rx05[2]: 00 - expansion direction(data)/conform(code) bit:
    data:       0=segment grows up
                1=segment growns down(SS)
    Rx05[1]: 01 - W(data)/R(code)
    data:       0=write access not allowed
                1=write access allowed (read always allowed)
Rx05[0]:     00 - A bit(accessed, each time CPU accesses it)

[Combuster]

RSD PTR at 0f45f0h - another known memory location

That table can be anywhere, so you might better not bother trying single addresses in the first place - just find the 4 bytes you're looking for in the entire space the BIOS occupies.

[ggodw000]

RSD PTR at 0f45f0h - another known memory location

That table can be anywhere, so you might better not bother trying single addresses in the first place - just find the 4 bytes you're looking for in the entire space the BIOS occupies.

yes i am aware, to clarify: i search for signature in e000 and f000 segment before calling pmode function and save the RSD PTR address.
also note that i mentioned: before calling pmode, i printout the RSD PTR bytes at the address found using the SEG:OFF address (which prints it correctly) and after that I do it using pmode SEG.SEL:32-BIT FLAT address (in which case it does not print correct values). In other words, both times, used same address but pmode prints out incorrectly.

[Combuster]

That just sounds like you're not dealing with non-zero segment bases properly.

Debugger homework (If you haven't used one yet, try bochs with built-in debugger enabled):
1: What segment register and general purpose register do you use to find the structure in real mode?
2: What are the register values when the structure is found? What is the segment base?
3: What is the physical address of the structure?
4: Can you use that physical address to print the structure from the debugger?
5: What segment register and general purpose register do you use to address the structure in protected mode?
6: What are the register values at the point of lookup? What is the segment's base address in the GDT? What is the actual base stored in the segment's descriptor cache?
7: Is the structure still there using it's physical address? Is it also still there in linear address space using the same address?

[ggodw000]

lot of code fixing and pmode is working perfectly now. i am working on IDT now. Thanks.,