Option ROM loaded but no boot device

[TheWiz]

Well wouldn't the drive number be determined by what SATA port is plugged into?

[gabiz_ro]

yes but in my case was 00 ?
BIOS is not making a call to BCV as result option ROM doesn't know what number was assigned to respective disk.
Now I try to make this call to BCV using other option ROM (for network card,and set boot from network first in BIOS)
Calling BCV increase number of disks in BIOS data area.But after that computer freeze.
And still don't know how to put in ESI pointer to PnP installation check.
PnP installation check is at 000FE2D0h
How can I make ESI Pointer to System BIOS PnP Installation Check Structure?
All my tries resulted in wrong data in ESI

[TheWiz]

Yess 00 is right because the First SATA drive is numbered 0, not 1 (applies to a lot of things in computer programming, this is one of them). As for your second question, I'm still catching up with all you've posted in the last couple days

[gabiz_ro]

is a little confusion I think.
On Intel board ,where BIOS detect disk,inside initialized option ROM I have BIOS disk number
80 if is first drive
81 if is second drive

On Dell board where BIOS doesn't detect disk this field is empty , 00
I use another option ROM (for network card and set boot from network so BIOS is loading and exectute ) which do next things
-display int13 vector
-display number of disks from BIOS data area
-set ESI to point to PnP installation check
-display ESI
-make a far call to BCV
-display int13 vector
-display number of disks from BIOS data area
-exit by choice with retf or int18

int13 vector is changed and in BDA number of disks is increased after calling BCV but ESI is not right.
exiting from that option ROM with retf is freezing computer and exiting with int18 is restarting at any keypress.

[TheWiz]

Man you are really great at ASM, something I still haven't quite gripped yet as Im a C# programmer etc. How's it coming along?

[gabiz_ro]

I'm not great at ASM, I just use info collected and get help on other forums.
Now I don't know where thing go wrong.
BIOS load Option ROM > Option ROM initialization > BIOS write protect shadow RAM area and continue POST > boot
In this time if BIOS was follow PnP BIOS specification then it must see that is a disk device (Device type code in expansion PnP header,which is OK inside and is working with other manufacturer BIOS's) then after Option ROM initialization make a call to its BCV.

I don't know when Option ROM decide what BIOS disk number have their attached drive.
(80 is first disk;81 is second disk ... in my case on Dell is 00 and on Intel board 80 if is single disk or 81 if I have another one attached to onboard ICH chip.)
But since in BDA number of disk is not touched until BCV is called and in any case Option ROM read offset 75h of BDA (or 0475h of RAM) then decide what BIOS disk number get and I get 00 always I think sequence is
BIOS load Option ROM > Option ROM initialization > BIOS call BCV > Option ROM increase number of disks detected in BDA > depending on what value was readed from BDA assign a number (80 81 82 ... ) to their attached disk > BIOS write protect shadow RAM area and continue POST > boot

The problem now I think is when I force a call to BCV using that trick from network boot RAM region where Option rom is shadowed is read-only.
Since number of disks in BDA is increased,INT13 vector is changed but option rom can't update itself with assigned BIOS disk number.

Maybe if I can re-enable write in shadow RAM region I can get more results.

[gabiz_ro]

Nobody ,any idea if and how can shadow ram region be write possible?

I have patched Sil option ROM so it keep BIOS disk number assigned in initialization stage.But calling BCV and trying to boot still not success.

I have found this in RAM,is coming from BIOS

Code:

Offset 0 1 2 3 4 5 6 7 8 9 A B C D E F

00000000 F0 01 F6 03 E0 00 0E 08 00 04 95 08 00 00 11 6E ð.ö.à.....•....n
00000010 78 00 37 11 30 22 A5 0B 00 00 A0 BF FA 00 67 81 x.7.0"¥... ¿ú.g
00000020 F0 01 F6 03 B0 00 0E 00 00 00 00 00 00 00 11 00 ð.ö.°...........
00000030 58 02 00 00 00 00 00 00 00 00 00 00 FA 01 50 81 X...........ú.P
00000040 70 01 76 03 A0 00 0F 00 00 00 00 00 00 00 11 00 p.v. ...........
00000050 58 02 00 00 00 00 00 00 00 00 00 00 FA 02 75 81 X...........ú.u
00000060 70 01 76 03 B0 00 0F 00 00 00 00 00 00 00 11 00 p.v.°...........
00000070 58 02 00 00 00 00 00 00 00 00 00 00 FA 03 50 81 X...........ú.P
00000080 F0 FF F0 FF B0 00 FF 00 00 00 00 00 00 00 11 AC ðÿðÿ°.ÿ........¬
00000090 58 02 35 01 00 00 00 00 00 00 00 00 00 40 57 81 X.5..........@W

It looks like an IPL table with entries for primary -master slave secondary -master slave and one that could be for USB boot device.

As far as I can interpret this data for first entry:
F0 01 1F0h controller I/O base address of primary ide channel
F6 03 3F6h controller port address of primary ide channel
E0 00 E0 ???
0E 08 0Eh IRQ14 08h ??
00 04 04h PIO mode?
95 08 0895h DPTE flag
00 00 ???
11 6E ???
78 00 device present? (58 no dev,78 dev exist?)
37 11 ???
30 22 - 0BA52230h 195371568 LBA sectors of attached HDD
A5 0B /
00 00 ???
A0 BF BFA0h I/O range of ICH7-M D31:F2
FA 00 ???
67 81 ???

I'm thinking to try to mod one of these entries to see what is happening but I don't know some fields.

[TheWiz]

Basically PIO is a method of data transfer on the harddisk. Today though DMA is highly perferred and faster and PIO has become a last resort really. Can you post the ASM output?

[gabiz_ro]

Early I was thinking to mod IPL table builded by BIOS.
I looked into BIOS and find this field blank and in RAM was populated with some info then looking at some info gathered with various soft I see some matching data

This is with HDD connected to onboard sata.

Code:

EDD version: 2.1
EDD features: Fixed disk access, EDD support
EDD flags: 0001h
Physical drive #0 geometry LBA sectors: 195371568 (100 GB), sector=512 B
Device Path Information: absent
Device Parameter Table Extension pointer: F000:81A9h
DPTE revision: 1.1 [OK]
DPTE flags: 0895h
Controller I/O base address: 1F0h, control port: 3F6h
IRQ: 14, DMA channel: 0
PIO mode: 4, Ultra DMA: enabled
32-bit transfer mode: enabled, LBA: enabled, LBA48: disabled
ATA read/write multiple command max sectors: 8, enabled
ATAPI: no, uses command packet interrupt: no, nonremovable media
Host Protected Area: inactive

But any attempt to mod one of these five blocks ended with a bricked BIOS.

[gabiz_ro]

Almost done.
Still BIOS can't boot native from disks.
Main problem was:
-BIOS doesn't make a call to BCV (after initialization OROM resize itself but also clear any configuration,that is make so when calling BCV, OROM get his disk BIOS assigned number.
-trying to make later a far call to this BCV shadow RAM area was RO so OROM can't configure well.

After writing few chipset registers of DRAM controller,first,to enabe RW on shadow RAM and then next making a call to BCV and booting from USB I can see and get access to HDD,now using GRUB to remap disks and then boot from one attached to sata card is working fine.