Question about which tools to use, bugs, the best way to implement a function, etc should go here. Don't forget to see if your question is answered in the wiki first! When in doubt post here.
I'm new to OSDev.org and I'm also new to OS development. I got some practice with BOCHS and x86 assembly but I still got a lot to learn. My question is:
I got trouble with my bootloader code. I try to use the INT 13h AH=42h: "Extended Read Sectors From Drive" BIOS function of version 3.0 (EDD-3.0) with BOCHS to copy my kernel code from a hard disk image to a memory location above 0x100000 in real mode.
When I use the INT13 AH=41h BX=55AAh function to check for support of EDD-3.0 I get a positive result. I'm using BOCHS 2.6.8.
Disk Address Packet:
db 0x18 ; size of packet
db 0x00 ; reserved
db 0x04 ; # blocks to transfer
db 0x00 ; reserved
dw 0xFFFF ; offset
dw 0xFFFF ; segment
dq 0x1 ; starting absolute block number (LBA)
dq 0x100000; 64-bit flat target address
I use this DAP. What BOCHS does is copy 4 blocks from disk to FFFF:FFFF mod 100000h = FFEFh. What am I doing wrong? I use Ralf Brown's Interrupt List (http://www.ctyme.com/intr/rb-0708.htm) as reference.
Bochs (or rather, the BIOS you're using) does not support that extension.
RBIL says it should be supported, but RBIL isn't always the most accurate source - it could very well be based on a draft that was never officially accepted as standard. At least one draft says function 41h returns a bit in CX that indicates whether the 64-bit flat addresses are supported or not. Going by that draft, Bochs is correct.
You are right. I searched the web for the specification of EDD-3.0 and found a draft but I didnt check the exact definition of INT13h AH=41h BX=55AAh. The draft differs from what RBIL says I didnt expect that.
RBIL isn't the bible, good to know
I didnt expect that.