From fd61f403ad751b4dc70876f43dbd313298530fa3 Mon Sep 17 00:00:00 2001 From: introspec <31136975+specke@users.noreply.github.com> Date: Wed, 24 Jul 2019 17:30:37 +0100 Subject: [PATCH] LZSA1 decompressors with added support for -b. --- asm/z80/unlzsa1_fast_v1.asm | 145 +++++++++++++++++++++++++++++++++++ asm/z80/unlzsa1_small_v1.asm | 128 +++++++++++++++++++++++++++++++ 2 files changed, 273 insertions(+) create mode 100644 asm/z80/unlzsa1_fast_v1.asm create mode 100644 asm/z80/unlzsa1_small_v1.asm diff --git a/asm/z80/unlzsa1_fast_v1.asm b/asm/z80/unlzsa1_fast_v1.asm new file mode 100644 index 0000000..b8443cd --- /dev/null +++ b/asm/z80/unlzsa1_fast_v1.asm @@ -0,0 +1,145 @@ +; +; Speed-optimized LZSA decompressor by spke (v.1 03-25/04/2019, 110 bytes) +; +; The data must be compressed using the command line compressor by Emmanuel Marty +; The compression is done as follows: +; +; lzsa.exe -f1 -r +; +; where option -r asks for the generation of raw (frame-less) data. +; +; The decompression is done in the standard way: +; +; ld hl,FirstByteOfCompressedData +; ld de,FirstByteOfMemoryForDecompressedData +; call DecompressLZSA +; +; Backward compression is also supported; you can compress files backward using: +; +; lzsa.exe -f1 -r -b +; +; and decompress the resulting files using: +; +; ld hl,LastByteOfCompressedData +; ld de,LastByteOfMemoryForDecompressedData +; call DecompressLZSA +; +; (do not forget to uncomment the BACKWARD_DECOMPRESS option in the decompressor). +; +; Of course, LZSA compression algorithm is (c) 2019 Emmanuel Marty, +; see https://github.com/emmanuel-marty/lzsa for more information +; +; Drop me an email if you have any comments/ideas/suggestions: zxintrospec@gmail.com +; +; This software is provided 'as-is', without any express or implied +; warranty. In no event will the authors be held liable for any damages +; arising from the use of this software. +; +; Permission is granted to anyone to use this software for any purpose, +; including commercial applications, and to alter it and redistribute it +; freely, subject to the following restrictions: +; +; 1. The origin of this software must not be misrepresented; you must not +; claim that you wrote the original software. If you use this software +; in a product, an acknowledgment in the product documentation would be +; appreciated but is not required. +; 2. Altered source versions must be plainly marked as such, and must not be +; misrepresented as being the original software. +; 3. This notice may not be removed or altered from any source distribution. + +; DEFINE BACKWARD_DECOMPRESS + + IFDEF BACKWARD_DECOMPRESS + + MACRO NEXT_HL + dec hl + ENDM + + MACRO ADD_OFFSET + or a : sbc hl,de + ENDM + + MACRO BLOCKCOPY + lddr + ENDM + + ELSE + + MACRO NEXT_HL + inc hl + ENDM + + MACRO ADD_OFFSET + add hl,de + ENDM + + MACRO BLOCKCOPY + ldir + ENDM + + ENDIF + +@DecompressLZSA: + ld b,0 : jr ReadToken + +NoLiterals: xor (hl) : NEXT_HL + push de : ld e,(hl) : NEXT_HL : jp m,LongOffset + + ; short matches have length 0+3..14+3 +ShortOffset: ld d,#FF : add 3 : cp 15+3 : jr nc,LongerMatch + + ; placed here this saves a JP per iteration +CopyMatch: ld c,a +.UseC ex (sp),hl : push hl ; BC = len, DE = offset, HL = dest, SP ->[dest,src] + ADD_OFFSET : pop de ; BC = len, DE = dest, HL = dest-offset, SP->[src] + BLOCKCOPY : pop hl ; BC = 0, DE = dest, HL = src + +ReadToken: ; first a byte token "O|LLL|MMMM" is read from the stream, + ; where LLL is the number of literals and MMMM is + ; a length of the match that follows after the literals + ld a,(hl) : and #70 : jr z,NoLiterals + + cp #70 : jr z,MoreLiterals ; LLL=7 means 7+ literals... + rrca : rrca : rrca : rrca ; LLL<7 means 0..6 literals... + + ld c,a : ld a,(hl) : NEXT_HL + BLOCKCOPY + + ; next we read the first byte of the offset + push de : ld e,(hl) : NEXT_HL + ; the top bit of token is set if the offset contains two bytes + and #8F : jp p,ShortOffset + +LongOffset: ; read second byte of the offset + ld d,(hl) : NEXT_HL + add -128+3 : cp 15+3 : jp c,CopyMatch + + ; MMMM=15 indicates a multi-byte number of literals +LongerMatch: add (hl) : NEXT_HL : jr nc,CopyMatch + + ; the codes are designed to overflow; + ; the overflow value 1 means read 1 extra byte + ; and overflow value 0 means read 2 extra bytes +.code1 ld b,a : ld c,(hl) : NEXT_HL : jr nz,CopyMatch.UseC +.code0 ld b,(hl) : NEXT_HL + + ; the two-byte match length equal to zero + ; designates the end-of-data marker + ld a,b : or c : jr nz,CopyMatch.UseC + pop de : ret + +MoreLiterals: ; there are three possible situations here + xor (hl) : NEXT_HL : exa + ld a,7 : add (hl) : NEXT_HL : jr c,ManyLiterals + +CopyLiterals: ld c,a +.UseC BLOCKCOPY + + push de : ld e,(hl) : NEXT_HL + exa : jp p,ShortOffset : jr LongOffset + +ManyLiterals: +.code1 ld b,a : ld c,(hl) : NEXT_HL : jr nz,CopyLiterals.UseC +.code0 ld b,(hl) : NEXT_HL : jr CopyLiterals.UseC + + diff --git a/asm/z80/unlzsa1_small_v1.asm b/asm/z80/unlzsa1_small_v1.asm new file mode 100644 index 0000000..df4dfc5 --- /dev/null +++ b/asm/z80/unlzsa1_small_v1.asm @@ -0,0 +1,128 @@ +; +; Size-optimized LZSA decompressor by spke (v.1 23/04/2019, 69 bytes) +; +; The data must be compressed using the command line compressor by Emmanuel Marty +; The compression is done as follows: +; +; lzsa.exe -f1 -r +; +; where option -r asks for the generation of raw (frame-less) data. +; +; The decompression is done in the standard way: +; +; ld hl,FirstByteOfCompressedData +; ld de,FirstByteOfMemoryForDecompressedData +; call DecompressLZSA +; +; Backward compression is also supported; you can compress files backward using: +; +; lzsa.exe -f1 -r -b +; +; and decompress the resulting files using: +; +; ld hl,LastByteOfCompressedData +; ld de,LastByteOfMemoryForDecompressedData +; call DecompressLZSA +; +; (do not forget to uncomment the BACKWARD_DECOMPRESS option in the decompressor). +; +; Of course, LZSA compression algorithm is (c) 2019 Emmanuel Marty, +; see https://github.com/emmanuel-marty/lzsa for more information +; +; Drop me an email if you have any comments/ideas/suggestions: zxintrospec@gmail.com +; +; This software is provided 'as-is', without any express or implied +; warranty. In no event will the authors be held liable for any damages +; arising from the use of this software. +; +; Permission is granted to anyone to use this software for any purpose, +; including commercial applications, and to alter it and redistribute it +; freely, subject to the following restrictions: +; +; 1. The origin of this software must not be misrepresented; you must not +; claim that you wrote the original software. If you use this software +; in a product, an acknowledgment in the product documentation would be +; appreciated but is not required. +; 2. Altered source versions must be plainly marked as such, and must not be +; misrepresented as being the original software. +; 3. This notice may not be removed or altered from any source distribution. + +; DEFINE BACKWARD_DECOMPRESS + + IFDEF BACKWARD_DECOMPRESS + + MACRO NEXT_HL + dec hl + ENDM + + MACRO ADD_OFFSET + or a : sbc hl,de + ENDM + + MACRO BLOCKCOPY + lddr + ENDM + + ELSE + + MACRO NEXT_HL + inc hl + ENDM + + MACRO ADD_OFFSET + add hl,de + ENDM + + MACRO BLOCKCOPY + ldir + ENDM + + ENDIF + +@DecompressLZSA: + ld b,0 + + ; first a byte token "O|LLL|MMMM" is read from the stream, + ; where LLL is the number of literals and MMMM is + ; a length of the match that follows after the literals +ReadToken: ld a,(hl) : exa : ld a,(hl) : NEXT_HL + and #70 : jr z,NoLiterals + + rrca : rrca : rrca : rrca ; LLL<7 means 0..6 literals... + cp #07 : call z,ReadLongBA ; LLL=7 means 7+ literals... + + ld c,a : BLOCKCOPY + + ; next we read the low byte of the -offset +NoLiterals: push de : ld e,(hl) : NEXT_HL : ld d,#FF + ; the top bit of token is set if + ; the offset contains the high byte as well + exa : or a : jp p,ShortOffset + +LongOffset: ld d,(hl) : NEXT_HL + + ; last but not least, the match length is read +ShortOffset: and #0F : add 3 ; MMMM<15 means match lengths 0+3..14+3 + cp 15+3 : call z,ReadLongBA ; MMMM=15 means lengths 14+3+ + ld c,a + + ex (sp),hl : push hl ; BC = len, DE = -offset, HL = dest, SP ->[dest,src] + ADD_OFFSET : pop de ; BC = len, DE = dest, HL = dest+(-offset), SP->[src] + BLOCKCOPY : pop hl ; BC = 0, DE = dest, HL = src + jr ReadToken + + ; a standard routine to read extended codes + ; into registers B (higher byte) and A (lower byte). +ReadLongBA: add (hl) : NEXT_HL : ret nc + + ; the codes are designed to overflow; + ; the overflow value 1 means read 1 extra byte + ; and overflow value 0 means read 2 extra bytes +.code1: ld b,a : ld a,(hl) : NEXT_HL : ret nz +.code0: ld c,a : ld b,(hl) : NEXT_HL + + ; the two-byte match length equal to zero + ; designates the end-of-data marker + or b : ld a,c : ret nz + pop de : pop de : ret +