| 1 |
|
|---|
| 2 | Notes on BIOS usage
|
|---|
| 3 | -------------------
|
|---|
| 4 |
|
|---|
| 5 | - DOS (including 6.22/7.1) does not need INT 15h or INT 1Ah. Most other
|
|---|
| 6 | operating systems require INT 15h to detect installed memory.
|
|---|
| 7 |
|
|---|
| 8 | - OS/2 (WSeB/MCP/ACP) appears to be one of the very few operating systems
|
|---|
| 9 | which use the El Torito floppy emulation.
|
|---|
| 10 |
|
|---|
| 11 | - MS-DOS 6.2/V is a rare user of the INT 15h keyboard intercept routines.
|
|---|
| 12 |
|
|---|
| 13 | - Some software uses the model byte at F000:FFFE to determine the system
|
|---|
| 14 | type (PC-DOS 3.0, Norton Utilities 8). Other software first tries INT 15h,
|
|---|
| 15 | fn C0h instead (PC-DOS 3.1, MSD).
|
|---|
| 16 |
|
|---|
| 17 | - DOS 4.01 (both IBM and Microsoft) calls INT 13h to read from disk with less
|
|---|
| 18 | than 100 bytes of stack space early in the boot sequence.
|
|---|
| 19 |
|
|---|
| 20 | - Very few guests use the 32-bit BIOS interface. One is OS/2 (but falls back),
|
|---|
| 21 | another is Etherboot.
|
|---|
| 22 |
|
|---|
| 23 | - OS/2 is the only known guest which can run the 16-bit PCI BIOS in protected
|
|---|
| 24 | mode (but only if the 32-bit PCI BIOS is unavailable).
|
|---|
| 25 |
|
|---|
| 26 | - Any disk reads which use bus-master DMA (AHCI, IDE BM) must use VDS
|
|---|
| 27 | (Virtual DMA Services) when present. Otherwise any reads/writes when the
|
|---|
| 28 | real mode addresses don't map directly to physical addresses will fail
|
|---|
| 29 | horribly. DOS 6.x with EMM386 is a good testcase (esp. loading drivers
|
|---|
| 30 | into UMBs).
|
|---|
| 31 |
|
|---|
| 32 | - Many older OSes (especially UNIX based) require the FDPT to contain
|
|---|
| 33 | physical ATA disk geometry; for that reason, disks smaller than ~500MB are
|
|---|
| 34 | easiest to use. Otherwise a "large" BIOS disk option would be required.
|
|---|
| 35 |
|
|---|
| 36 | - Some really old OSes (Xenix circa 1986-7) do not understand the EBDA idea
|
|---|
| 37 | and clear the memory. For those, the FDPT must be in the BIOS ROM area, or
|
|---|
| 38 | the OS will destroy it (even when it's at 0:300 in the IVT).
|
|---|
| 39 |
|
|---|
| 40 |
|
|---|
| 41 | Notes on BIOS implementation
|
|---|
| 42 | ----------------------------
|
|---|
| 43 |
|
|---|
| 44 | - To return values from functions not declared as __interrupt, the arguments
|
|---|
| 45 | may need to be declared volatile (not ideal, but does the job).
|
|---|
| 46 |
|
|---|
| 47 | - The way the POST code selectively clears or doesn't clear memory
|
|---|
| 48 | is extremely suspect and will need reworking.
|
|---|
| 49 |
|
|---|
| 50 | - Need to review string routines wrt direction flag (should be OK now).
|
|---|
| 51 |
|
|---|
| 52 | - Need to review CMOS access wrt interrupts (possible index reg change by
|
|---|
| 53 | an interrupt handler).
|
|---|
| 54 |
|
|---|
| 55 | - The POST code zeroes the entire BDA, and then various bits zero specific
|
|---|
| 56 | parts of the BDA again. That's a waste of time.
|
|---|
| 57 |
|
|---|
| 58 | - After a reset, all interrupts are unmasked. Not sure if that's OK.
|
|---|
| 59 |
|
|---|
| 60 | - BCC mishandles the following (where buf is an uint8_t array):
|
|---|
| 61 | lba=buf[0x2B]*0x1000000+buf[0x2A]*0x10000+buf[0x29]*0x100+buf[0x28];
|
|---|
| 62 | The buf[x]*100 expression should end up being of type signed int, which
|
|---|
| 63 | causes the sign to be incorrectly propagated. BCC incorrectly keeps
|
|---|
| 64 | the type unsigned.
|
|---|
| 65 |
|
|---|
| 66 |
|
|---|
| 67 |
|
|---|
| 68 | Code size notes (code as of 7/6/2011):
|
|---|
| 69 |
|
|---|
| 70 | The following values are the size of the _TEXT segment, i.e. only C code;
|
|---|
| 71 | data defined in C is not included, neither are assembly modules.
|
|---|
| 72 |
|
|---|
| 73 | Options: Size (hex):
|
|---|
| 74 | -------- -----------
|
|---|
| 75 | -0 -zu -s -oas -ecc 631A
|
|---|
| 76 | -3 -zu -s -oas -ecc 5C1E
|
|---|
| 77 | -0 -zu -s -oas 578A
|
|---|
| 78 | -3 -zu -s -oas 5452
|
|---|
| 79 |
|
|---|
| 80 | Both generating 386 code and using register-based calling convention for
|
|---|
| 81 | internal functions brings significant size savings (15% when combined).
|
|---|
