Retro/pcjs
Watch
1
0
Fork
You've already forked pcjs
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Some undocumented instruction updates

Browse source
This commit is contained in:
Jeff Parsons 2016-02-16 12:21:58 -08:00
commit 025b1abb93
19 changed files with 2531 additions and 1582 deletions
Download patch file Download diff file Expand all files Collapse all files

204
docs/x86/ops/AAD/AAD.ASM Normal file
View file

@ -0,0 +1,204 @@
;
; Saved on February 16, 2015 from http://www.rcollins.org/ftp/source/aad/aad.asm
;
.386p
;-----------------------------------------------------------------------------
;
; AAD.ASM Copyright (c) 1991, 1995-Present, Robert Collins
;
; You have my permission to copy and distribute this software for
; non-commercial purposes. Any commercial use of this software or
; source code is allowed, so long as the appropriate copyright
; attributions (to me) are intact, *AND* my email address is properly
; displayed.
;
; Basically, give me credit, where credit is due, and show my email
; address.
;
;-----------------------------------------------------------------------------
;
; Robert R. Collins email: rcollins@x86.org
;
;-----------------------------------------------------------------------------
.model small
.code
.286
;-----------------------------------------------------------------------------
; Interrupt vector segment
;-----------------------------------------------------------------------------
ABS0 segment at 0
org 0*4
Orig_INT0 label word
ABS0 ends
;-----------------------------------------------------------------------------
; Local stack frame variable(s)
;-----------------------------------------------------------------------------
INT0 equ [bp-4]
;-----------------------------------------------------------------------------
; Instruction macro definition
;-----------------------------------------------------------------------------
AADI MACRO VALUE
db 0d5h,VALUE
ENDM
;-----------------------------------------------------------------------------
TEST_AAD proc near ; Test AAD IMMED08 instruction functionality.
;-----------------------------------------------------------------------------
; AAD:
; {
; AL = AH*IMMED08 + AL
; AH = 0
; CF = AL[b7] Overflow?
; OF = AL[b7] set, but not overflow? or AL[b7] overflow, but not set?
; AF = AL[b3] overflow, or AL[b3] borrow?
; SF = AL[b7]=1?
; ZF = AL==0?
; PF = Even/Odd parity
; }
; Input: None
; Output: BX = Bit mask of results (3FFF if all tests passed)
; [b13] = 1, NS flag test passed
; [b12] = 1, SF flag test passed
; [b11] = 1, NZ flag test passed
; [b10] = 1, ZF flag test passed
; [b09] = 1, PO flag test passed
; [b08] = 1, PE flag test passed
; [b07] = 1, AF (test 2) flag test passed
; [b06] = 1, AF (test 1) flag test passed
; [b05] = 1, NA flag test passed
; [b04] = 1, OF (test 2) flag test passed
; [b03] = 1, NO flag test passed
; [b02] = 1, OF (test 1) flag test passed
; [b01] = 1, CY flag test passed
; [b00] = 1, NC flag test passed
; Register(s) modified: AX, BX, CX
;-----------------------------------------------------------------------------
xor bx,bx ; clear result flags
xor cx,cx
;-----------------------------------------------------------------------------
; Test Carry Flag set. According to Intel, CF is undefined after AAD, but
; should be set according to the results. Since this is an arithmatic
; operation, CF should be set according to the results. Let's find out!
;-----------------------------------------------------------------------------
mov ax,3300h ; set AH=51, AL=0 and perform 51*5.
AADI 5 ; result should not produce a CF
jc @F ; oops
or bl,1 ; set NC passed
@@: mov ax,3301h ; set AH=51, AL=1 and perform
AADI 5 ; (55*5)+1 should set CF
jnc @F ; oops, didn't work
or bl,2 ; set CF passed
;-----------------------------------------------------------------------------
; Test Overflow Flag set. There are two ways the OF can be set:
; 1) If AL[b7] is set, but doesn't overflow into CF;
; 2) If AL[b7] overflows, but doesn't get set.
;-----------------------------------------------------------------------------
; 1) If AL[b7] is set, but doesn't overflow into CF;
;-----------------------------------------------------------------------------
mov ax,2a02h ; perform (42*3)+2 = 128, should set
AADI 3 ; OF.
jno @F ; oops, didn't work
or bl,4 ; set OF passed
@@: mov ax,8080h ; perform (128*2)+128, should not set
AADI 2 ; OF because it sets CF, & AL[b7]
jo @F ; oops, didn't work
or bl,8 ; set NO passed
;-----------------------------------------------------------------------------
; 2) If AL[b7] overflows, but doesn't get set.
;-----------------------------------------------------------------------------
@@: mov ax,8080h ; perform (128*2)+128, should not set
AADI 1 ; OF because it sets CF, & AL[b7]
jno @F ; oops, didn't work
or bl,10h
;-----------------------------------------------------------------------------
; Test Auxiliary carry Flag (AF) set. AF is set in two ways:
; 1) If there is a carry out of bit3;
; 2) If there is a borrow out of bit3.
;-----------------------------------------------------------------------------
@@: mov ax,2200h ; perform (34*5)+0, should not set
AADI 5 ; AM because no carry from bit3
lahf ; get flags
test ah,10h ; AF set?
jnz @F ; yes, must have been a mistake
or bl,20h ; set NA flag passed
;-----------------------------------------------------------------------------
; 1) If there is a carry out of bit3;
;-----------------------------------------------------------------------------
@@: mov ax,2208h ; perform (34*5)+8, should set AF
AADI 5 ; because a bit3 will carry
lahf ; get flags
test ah,10h ; AF set?
jz @F ; nope, oops
or bl,40h ; set AF flag passed
;-----------------------------------------------------------------------------
; 2) If there is a borrow out of bit3.
; (This may not be an accurate test of borrowing out of bit 3, because
; this test adds a -8 to the result of the multplication. In other
; words, the addition algorithm is still used, not the subtraction
; algorithm -- if they are even different in the first place.)
;-----------------------------------------------------------------------------
@@: mov ax,22f8h ; perform (34*5)-8, should set AF
AADI 5 ; because a bit3 will borrow
lahf ; get flags
test ah,10h ; AF set?
jz @F ; nope, oops
or bl,80h ; set AF flag passed
;-----------------------------------------------------------------------------
; Test EVEN and ODD parity by generating results in the low byte that
; contain even and odd parity respectively.
;-----------------------------------------------------------------------------
@@: mov ax,0a00h ; 10*17+0 which is EVEN parity
AADI 11h
jpo @F ; didn't generate even parity
or bh,1 ; set even parity flag passed
@@: mov ax,0a01h ; 10*17+1 which is ODD parity
AADI 11h
jpe @F ; didn't generate odd parity
or bh,2 ; set odd parity flag passed
;-----------------------------------------------------------------------------
; Test ZERO FLAG by generating results that produce a zero, and non-zero.
;-----------------------------------------------------------------------------
@@: mov ax,0880h ; 8*16+128 will generate a zero result
AADI 10h ; This should force ZF=1
jnz @F ; didn't work as expected
or bh,4
@@: mov ax,0881h ; 8*16+129 will be non-zero
AADI 10h ; This should force ZF=0,
jz @F ; didn't work as expected
or bh,8
;-----------------------------------------------------------------------------
; Test Sign Flag by generating results whose highest bit is on.
; I'll try and do this in a manner that doesn't set OF.
;-----------------------------------------------------------------------------
mov ax,80c0h ; (128*128)+192 will generate SF
AADI 80h ; without generating OF.
jns @F ; oops
or bh,10h ; set SF flag passed
@@: mov ax,8040h ;
AADI 80h
js @F
or bh,20h
@@: ret
Test_AAD endp
end

39
docs/x86/ops/AAD/README.md Normal file
View file

@ -0,0 +1,39 @@
---
layout: page
title: "x86 Instructions: AAD"
permalink: /docs/x86/ops/AAD/
---
AAD (0xD5)
---
### Description
From [http://www.rcollins.org/secrets/opcodes/AAD.html](http://www.rcollins.org/secrets/opcodes/AAD.html):
Undocumented: Available to all Intel x86 processors
Useful in production source code.
AAD
Flags: ASCII Adjust before Division
+-+-+-+-+-+-+-+-+-+ +----------+----------+
|O|D|I|T|S|Z|A|P|C| | 11010101 | DATA |
+-+-+-+-+-+-+-+-+-+ +----------+----------+
|+| | | |+|+|+|+|+| | D5 | IMM8 |
+-+-+-+-+-+-+-+-+-+ +----------+----------+
This instruction is the multiplication counterpart to AAM. As is the case with AAM, AAD uses the second byte
as an operand. This operand is the multiplicand for AAD. Like AAM, AAD provides a way to execute a MUL IMM8 that
is unavailable through any other means in the CPU. Unlike MUL, or IMUL, AAD sets all of the CPU status flags
according to the result. Intel states that the Overflow Flag (OF), Auxiliary carry Flag (AF), and Carry Flag (CF)
are undefined. This assertion is incorrect. These flags are fully defined, and are set consistently with respect
to any other integer operations. And again, like AMM, beginning with the Pentium, Intel has finally acknowledged
the existence of the second byte of this instruction as its operand. Intel says:
Note: imm8 has the value of the instruction's second byte. The second byte under normally assembly [sic] of
this instruction will be 0A, however, explicit modification of this byte will result in the operation described
above and may alter results.
This instruction exists in this form on all Intel x86 processors. See the file AAD.ASM for diagnostics source code
for this instruction.
See [AAD.ASM](AAD.ASM) for the source code mentioned above.