25.01

Asm/x86/7zAsm.asm

@@ -1,7 +1,12 @@
 ; 7zAsm.asm -- ASM macros
-; 2021-12-25 : Igor Pavlov : Public domain
+; 2023-12-08 : Igor Pavlov : Public domain
 
 
+; UASM can require these changes
+; OPTION FRAMEPRESERVEFLAGS:ON
+; OPTION PROLOGUE:NONE
+; OPTION EPILOGUE:NONE
+
 ifdef @wordsize
 ; @wordsize is defined only in JWASM and ASMC and is not defined in MASM
 ; @wordsize eq 8 for 64-bit x64
@@ -116,10 +121,29 @@ endif
   x2_H equ DH
   x3_H equ BH
 
+;  r0_L equ AL
+;  r1_L equ CL
+;  r2_L equ DL
+;  r3_L equ BL
+
+;  r0_H equ AH
+;  r1_H equ CH
+;  r2_H equ DH
+;  r3_H equ BH
+
+
 ifdef x64
   x5_L equ BPL
   x6_L equ SIL
   x7_L equ DIL
+  x8_L equ r8b
+  x9_L equ r9b
+  x10_L equ r10b
+  x11_L equ r11b
+  x12_L equ r12b
+  x13_L equ r13b
+  x14_L equ r14b
+  x15_L equ r15b
 
   r0 equ RAX
   r1 equ RCX
@@ -148,6 +172,22 @@ else
   r7 equ x7
 endif
 
+  x0_R equ r0
+  x1_R equ r1
+  x2_R equ r2
+  x3_R equ r3
+  x4_R equ r4
+  x5_R equ r5
+  x6_R equ r6
+  x7_R equ r7
+  x8_R equ r8
+  x9_R equ r9
+  x10_R equ r10
+  x11_R equ r11
+  x12_R equ r12
+  x13_R equ r13
+  x14_R equ r14
+  x15_R equ r15
 
 ifdef x64
 ifdef ABI_LINUX
@@ -195,6 +235,14 @@ REG_ABI_PARAM_0   equ REG_PARAM_0
 REG_ABI_PARAM_1_x equ REG_PARAM_1_x
 REG_ABI_PARAM_1   equ REG_PARAM_1
 
+MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11 macro
+        MY_PUSH_4_REGS
+endm
+
+MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11 macro
+        MY_POP_4_REGS
+endm
+
 else
 ; x64
 
@@ -256,12 +304,25 @@ endm
 endif ; IS_LINUX
 
 
-MY_PUSH_PRESERVED_ABI_REGS macro
+MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11 macro
     if  (IS_LINUX gt 0)
         MY_PUSH_2_REGS
     else
         MY_PUSH_4_REGS
     endif
+endm
+
+MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11 macro
+    if  (IS_LINUX gt 0)
+        MY_POP_2_REGS
+    else
+        MY_POP_4_REGS
+    endif
+endm
+
+
+MY_PUSH_PRESERVED_ABI_REGS macro
+    MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
         push    r12
         push    r13
         push    r14
@@ -274,11 +335,7 @@ MY_POP_PRESERVED_ABI_REGS macro
         pop     r14
         pop     r13
         pop     r12
-    if  (IS_LINUX gt 0)
+    MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
-        MY_POP_2_REGS
-    else
-        MY_POP_4_REGS
-    endif
 endm
 
 endif ; x64

Asm/x86/7zCrcOpt.asm

@@ -1,180 +1,258 @@
 ; 7zCrcOpt.asm -- CRC32 calculation : optimized version
-; 2021-02-07 : Igor Pavlov : Public domain
+; 2023-12-08 : Igor Pavlov : Public domain
 
 include 7zAsm.asm
 
 MY_ASM_START
 
-rD   equ  r2
+NUM_WORDS       equ     3
-rN   equ  r7
+UNROLL_CNT      equ     2
-rT   equ  r5
 
-ifdef x64
+if (NUM_WORDS lt 1) or (NUM_WORDS gt 64)
-    num_VAR     equ r8
+.err <NUM_WORDS_IS_INCORRECT>
-    table_VAR   equ r9
+endif
-else
+if (UNROLL_CNT lt 1)
-  if (IS_CDECL gt 0)
+.err <UNROLL_CNT_IS_INCORRECT>
-    crc_OFFS    equ (REG_SIZE * 5)
-    data_OFFS   equ (REG_SIZE + crc_OFFS)
-    size_OFFS   equ (REG_SIZE + data_OFFS)
-  else
-    size_OFFS   equ (REG_SIZE * 5)
-  endif
-    table_OFFS  equ (REG_SIZE + size_OFFS)
-    num_VAR     equ [r4 + size_OFFS]
-    table_VAR   equ [r4 + table_OFFS]
 endif
 
-SRCDAT  equ  rD + rN * 1 + 4 *
+rD      equ  r2
+rD_x    equ  x2
+rN      equ  r7
+rT      equ  r5
+
+ifndef x64
+    if (IS_CDECL gt 0)
+        crc_OFFS    equ (REG_SIZE * 5)
+        data_OFFS   equ (REG_SIZE + crc_OFFS)
+        size_OFFS   equ (REG_SIZE + data_OFFS)
+    else
+        size_OFFS   equ (REG_SIZE * 5)
+    endif
+        table_OFFS  equ (REG_SIZE + size_OFFS)
+endif
+
+; rN + rD is same speed as rD, but we reduce one instruction in loop
+SRCDAT_1        equ     rN + rD * 1 + 1 *
+SRCDAT_4        equ     rN + rD * 1 + 4 *
 
 CRC macro op:req, dest:req, src:req, t:req
-    op      dest, DWORD PTR [rT + src * 4 + 0400h * t]
+        op      dest, dword ptr [rT + @CatStr(src, _R) * 4 + 0400h * (t)]
 endm
 
 CRC_XOR macro dest:req, src:req, t:req
-    CRC xor, dest, src, t
+        CRC     xor, dest, src, t
 endm
 
 CRC_MOV macro dest:req, src:req, t:req
-    CRC mov, dest, src, t
+        CRC     mov, dest, src, t
 endm
 
+MOVZXLO macro dest:req, src:req
+        movzx   dest, @CatStr(src, _L)
+endm
+
+MOVZXHI macro dest:req, src:req
+        movzx   dest, @CatStr(src, _H)
+endm
+
+; movzx x0, x0_L - is slow in some cpus (ivb), if same register for src and dest
+; movzx x3, x0_L sometimes is 0   cycles latency (not always)
+; movzx x3, x0_L sometimes is 0.5 cycles latency
+; movzx x3, x0_H is 2 cycles latency in some cpus
+
 CRC1b macro
-    movzx   x6, BYTE PTR [rD]
+        movzx   x6, byte ptr [rD]
-    inc     rD
+        MOVZXLO x3, x0
-    movzx   x3, x0_L
+        inc     rD
-    xor     x6, x3
+        shr     x0, 8
-    shr     x0, 8
+        xor     x6, x3
-    CRC     xor, x0, r6, 0
+        CRC_XOR x0, x6, 0
-    dec     rN
+        dec     rN
 endm
 
-MY_PROLOG macro crc_end:req
+LOAD_1 macro dest:req, t:req, iter:req, index:req
+        movzx   dest, byte ptr [SRCDAT_1 (4 * (NUM_WORDS - 1 - t + iter * NUM_WORDS) + index)]
+endm
 
+LOAD_2 macro dest:req, t:req, iter:req, index:req
+        movzx   dest, word ptr [SRCDAT_1 (4 * (NUM_WORDS - 1 - t + iter * NUM_WORDS) + index)]
+endm
+
+CRC_QUAD macro nn, t:req, iter:req
+ifdef x64
+        ; paired memory loads give 1-3% speed gain, but it uses more registers
+        LOAD_2  x3, t, iter, 0
+        LOAD_2  x9, t, iter, 2
+        MOVZXLO x6, x3
+        shr     x3, 8
+        CRC_XOR nn, x6, t * 4 + 3
+        MOVZXLO x6, x9
+        shr     x9, 8
+        CRC_XOR nn, x3, t * 4 + 2
+        CRC_XOR nn, x6, t * 4 + 1
+        CRC_XOR nn, x9, t * 4 + 0
+elseif 0
+        LOAD_2  x3, t, iter, 0
+        MOVZXLO x6, x3
+        shr     x3, 8
+        CRC_XOR nn, x6, t * 4 + 3
+        CRC_XOR nn, x3, t * 4 + 2
+        LOAD_2  x3, t, iter, 2
+        MOVZXLO x6, x3
+        shr     x3, 8
+        CRC_XOR nn, x6, t * 4 + 1
+        CRC_XOR nn, x3, t * 4 + 0
+elseif 0
+        LOAD_1  x3, t, iter, 0
+        LOAD_1  x6, t, iter, 1
+        CRC_XOR nn, x3, t * 4 + 3
+        CRC_XOR nn, x6, t * 4 + 2
+        LOAD_1  x3, t, iter, 2
+        LOAD_1  x6, t, iter, 3
+        CRC_XOR nn, x3, t * 4 + 1
+        CRC_XOR nn, x6, t * 4 + 0
+else
+        ; 32-bit load is better if there is only one read port (core2)
+        ; but that code can be slower if there are 2 read ports (snb)
+        mov     x3, dword ptr [SRCDAT_1 (4 * (NUM_WORDS - 1 - t + iter *  NUM_WORDS) + 0)]
+        MOVZXLO x6, x3
+        CRC_XOR nn, x6, t * 4 + 3
+        MOVZXHI x6, x3
+        shr     x3, 16
+        CRC_XOR nn, x6, t * 4 + 2
+        MOVZXLO x6, x3
+        shr     x3, 8
+        CRC_XOR nn, x6, t * 4 + 1
+        CRC_XOR nn, x3, t * 4 + 0
+endif
+endm
+
+
+LAST    equ     (4 * (NUM_WORDS - 1))
+
+CRC_ITER macro qq, nn, iter
+        mov     nn, [SRCDAT_4 (NUM_WORDS * (1 + iter))]
+
+    i = 0
+    rept NUM_WORDS - 1
+        CRC_QUAD nn, i, iter
+        i = i + 1
+    endm
+
+        MOVZXLO x6, qq
+        mov     x3, qq
+        shr     x3, 24
+        CRC_XOR nn, x6, LAST + 3
+        CRC_XOR nn, x3, LAST + 0
+        ror     qq, 16
+        MOVZXLO x6, qq
+        shr     qq, 24
+        CRC_XOR nn, x6, LAST + 1
+if ((UNROLL_CNT and 1) eq 1) and (iter eq (UNROLL_CNT - 1))
+        CRC_MOV qq, qq, LAST + 2
+        xor     qq, nn
+else
+        CRC_XOR nn, qq, LAST + 2
+endif
+endm
+
+
+; + 4 for prefetching next 4-bytes after current iteration
+NUM_BYTES_LIMIT equ     (NUM_WORDS * 4 * UNROLL_CNT + 4)
+ALIGN_MASK      equ     3
+
+
+; MY_PROC @CatStr(CrcUpdateT, 12), 4
+MY_PROC @CatStr(CrcUpdateT, %(NUM_WORDS * 4)), 4
+        MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
     ifdef x64
+        mov     x0, REG_ABI_PARAM_0_x   ; x0 = x1(win) / x7(linux)
+        mov     rT, REG_ABI_PARAM_3     ; r5 = r9(win) / x1(linux)
+        mov     rN, REG_ABI_PARAM_2     ; r7 = r8(win) / r2(linux)
+        ; mov     rD, REG_ABI_PARAM_1     ; r2 = r2(win)
       if  (IS_LINUX gt 0)
-        MY_PUSH_2_REGS
-        mov     x0, REG_ABI_PARAM_0_x   ; x0 = x7
-        mov     rT, REG_ABI_PARAM_3     ; r5 = r1
-        mov     rN, REG_ABI_PARAM_2     ; r7 = r2
         mov     rD, REG_ABI_PARAM_1     ; r2 = r6
-      else
-        MY_PUSH_4_REGS
-        mov     x0, REG_ABI_PARAM_0_x   ; x0 = x1
-        mov     rT, REG_ABI_PARAM_3     ; r5 = r9
-        mov     rN, REG_ABI_PARAM_2     ; r7 = r8
-        ; mov     rD, REG_ABI_PARAM_1     ; r2 = r2
       endif
     else
-        MY_PUSH_4_REGS
       if  (IS_CDECL gt 0)
         mov     x0, [r4 + crc_OFFS]
         mov     rD, [r4 + data_OFFS]
       else
         mov     x0, REG_ABI_PARAM_0_x
       endif
-        mov     rN, num_VAR
+        mov     rN, [r4 + size_OFFS]
-        mov     rT, table_VAR
+        mov     rT, [r4 + table_OFFS]
     endif
     
-    test    rN, rN
+        cmp     rN, NUM_BYTES_LIMIT + ALIGN_MASK
-    jz      crc_end
+        jb      crc_end
-  @@:
+@@:
-    test    rD, 7
+        test    rD_x, ALIGN_MASK    ; test    rD, ALIGN_MASK
-    jz      @F
+        jz      @F
-    CRC1b
+        CRC1b
-    jnz     @B
+        jmp     @B
-  @@:
+@@:
-    cmp     rN, 16
+        xor     x0, dword ptr [rD]
-    jb      crc_end
+        lea     rN, [rD + rN * 1 - (NUM_BYTES_LIMIT - 1)]
-    add     rN, rD
+        sub     rD, rN
-    mov     num_VAR, rN
+
-    sub     rN, 8
+align 16
-    and     rN, NOT 7
+@@:
-    sub     rD, rN
+unr_index = 0
-    xor     x0, [SRCDAT 0]
+while unr_index lt UNROLL_CNT
+    if (unr_index and 1) eq 0
+        CRC_ITER x0, x1, unr_index
+    else
+        CRC_ITER x1, x0, unr_index
+    endif
+        unr_index = unr_index + 1
 endm
 
-MY_EPILOG macro crc_end:req
+        add     rD, NUM_WORDS * 4 * UNROLL_CNT
-    xor     x0, [SRCDAT 0]
+        jnc     @B
-    mov     rD, rN
-    mov     rN, num_VAR
-    sub     rN, rD
-  crc_end:
-    test    rN, rN
-    jz      @F
-    CRC1b
-    jmp     crc_end
-  @@:
-      if (IS_X64 gt 0) and (IS_LINUX gt 0)
-        MY_POP_2_REGS
-      else
-        MY_POP_4_REGS
-      endif
-endm
 
-MY_PROC CrcUpdateT8, 4
+if 0
-    MY_PROLOG crc_end_8
+        ; byte verson
-    mov     x1, [SRCDAT 1]
+        add     rD, rN
-    align 16
+        xor     x0, dword ptr [rD]
-  main_loop_8:
+        add     rN, NUM_BYTES_LIMIT - 1
-    mov     x6, [SRCDAT 2]
+else
-    movzx   x3, x1_L
+        ; 4-byte version
-    CRC_XOR x6, r3, 3
+        add     rN, 4 * NUM_WORDS * UNROLL_CNT
-    movzx   x3, x1_H
+        sub     rD, 4 * NUM_WORDS * UNROLL_CNT
-    CRC_XOR x6, r3, 2
+@@:
-    shr     x1, 16
+        MOVZXLO x3, x0
-    movzx   x3, x1_L
+        MOVZXHI x1, x0
-    movzx   x1, x1_H
+        shr     x0, 16
-    CRC_XOR x6, r3, 1
+        MOVZXLO x6, x0
-    movzx   x3, x0_L
+        shr     x0, 8
-    CRC_XOR x6, r1, 0
+        CRC_MOV x0, x0, 0
+        CRC_XOR x0, x3, 3
+        CRC_XOR x0, x1, 2
+        CRC_XOR x0, x6, 1
 
-    mov     x1, [SRCDAT 3]
+        add     rD, 4
-    CRC_XOR x6, r3, 7
+if (NUM_WORDS * UNROLL_CNT) ne 1
-    movzx   x3, x0_H
+        jc      @F
-    shr     x0, 16
+        xor     x0, [SRCDAT_4 0]
-    CRC_XOR x6, r3, 6
+        jmp     @B
-    movzx   x3, x0_L
+@@:
-    CRC_XOR x6, r3, 5
+endif
-    movzx   x3, x0_H
+        add     rD, rN
-    CRC_MOV x0, r3, 4
+        add     rN, 4 - 1
-    xor     x0, x6
-    add     rD, 8
-    jnz     main_loop_8
         
-    MY_EPILOG crc_end_8
+endif
-MY_ENDP
         
-MY_PROC CrcUpdateT4, 4
+        sub     rN, rD
-    MY_PROLOG crc_end_4
+crc_end:
-    align 16
+        test    rN, rN
-  main_loop_4:
+        jz      func_end
-    movzx   x1, x0_L
+@@:
-    movzx   x3, x0_H
+        CRC1b
-    shr     x0, 16
+        jnz     @B
-    movzx   x6, x0_H
-    and     x0, 0FFh
-    CRC_MOV x1, r1, 3
-    xor     x1, [SRCDAT 1]
-    CRC_XOR x1, r3, 2
-    CRC_XOR x1, r6, 0
-    CRC_XOR x1, r0, 1
 
-    movzx   x0, x1_L
+func_end:
-    movzx   x3, x1_H
+        MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
-    shr     x1, 16
-    movzx   x6, x1_H
-    and     x1, 0FFh
-    CRC_MOV x0, r0, 3
-    xor     x0, [SRCDAT 2]
-    CRC_XOR x0, r3, 2
-    CRC_XOR x0, r6, 0
-    CRC_XOR x0, r1, 1
-    add     rD, 8
-    jnz     main_loop_4
-
-    MY_EPILOG crc_end_4
 MY_ENDP
 
 end

Asm/x86/LzFindOpt.asm

@@ -1,5 +1,5 @@
 ; LzFindOpt.asm -- ASM version of GetMatchesSpecN_2() function
-; 2021-07-21: Igor Pavlov : Public domain
+; 2024-06-18: Igor Pavlov : Public domain
 ;
 
 ifndef x64
@@ -11,10 +11,31 @@ include 7zAsm.asm
 
 MY_ASM_START
 
-_TEXT$LZFINDOPT SEGMENT ALIGN(64) 'CODE'
+ifndef Z7_LZ_FIND_OPT_ASM_USE_SEGMENT
+if (IS_LINUX gt 0)
+  Z7_LZ_FIND_OPT_ASM_USE_SEGMENT equ 1
+else
+  Z7_LZ_FIND_OPT_ASM_USE_SEGMENT equ 1
+endif
+endif
 
+ifdef Z7_LZ_FIND_OPT_ASM_USE_SEGMENT
+_TEXT$LZFINDOPT SEGMENT ALIGN(64) 'CODE'
 MY_ALIGN macro num:req
         align  num
+        ; align  16
+endm
+else
+MY_ALIGN macro num:req
+        ; We expect that ".text" is aligned for 16-bytes.
+        ; So we don't need large alignment inside our function.
+        align  16
+endm
+endif
+
+
+MY_ALIGN_16 macro
+        MY_ALIGN 16
 endm
 
 MY_ALIGN_32 macro
@@ -136,7 +157,11 @@ COPY_VAR_64 macro dest_var, src_var
 endm
 
 
+ifdef Z7_LZ_FIND_OPT_ASM_USE_SEGMENT
 ; MY_ALIGN_64
+else
+  MY_ALIGN_16
+endif
 MY_PROC GetMatchesSpecN_2, 13
 MY_PUSH_PRESERVED_ABI_REGS
         mov     r0, RSP
@@ -508,6 +533,8 @@ fin:
 MY_POP_PRESERVED_ABI_REGS
 MY_ENDP
 
+ifdef Z7_LZ_FIND_OPT_ASM_USE_SEGMENT
 _TEXT$LZFINDOPT ENDS
+endif
 
 end

Asm/x86/LzmaDecOpt.asm

@@ -1,5 +1,5 @@
 ; LzmaDecOpt.asm -- ASM version of LzmaDec_DecodeReal_3() function
-; 2021-02-23: Igor Pavlov : Public domain
+; 2024-06-18: Igor Pavlov : Public domain
 ;
 ; 3 - is the code compatibility version of LzmaDec_DecodeReal_*()
 ; function for check at link time.
@@ -17,11 +17,41 @@ include 7zAsm.asm
 
 MY_ASM_START
 
-_TEXT$LZMADECOPT SEGMENT ALIGN(64) 'CODE'
+; if Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT is     defined, we use additional SEGMENT with 64-byte alignment.
+; if Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT is not defined, we use default SEGMENT (where default 16-byte alignment of segment is expected).
+; The performance is almost identical in our tests.
+; But the performance can depend from position of lzmadec code inside instruction cache
+; or micro-op cache line (depending from low address bits in 32-byte/64-byte cache lines).
+; And 64-byte alignment provides a more consistent speed regardless
+; of the code's position in the executable.
+; But also it's possible that code without Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT can be
+; slightly faster than 64-bytes aligned code in some cases, if offset of lzmadec
+; code in 64-byte block after compilation provides better speed by some reason.
+; Note that Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT adds an extra section to the ELF file.
+; If you don't want to get that extra section, do not define Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT.
 
+ifndef Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT
+if (IS_LINUX gt 0)
+  Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT equ 1
+else
+  Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT equ 1
+endif
+endif
+
+ifdef Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT
+_TEXT$LZMADECOPT SEGMENT ALIGN(64) 'CODE'
 MY_ALIGN macro num:req
         align  num
+        ; align  16
 endm
+else
+MY_ALIGN macro num:req
+        ; We expect that ".text" is aligned for 16-bytes.
+        ; So we don't need large alignment inside out function.
+        align  16
+endm
+endif
+
 
 MY_ALIGN_16 macro
         MY_ALIGN 16
@@ -610,7 +640,11 @@ PARAM_lzma      equ REG_ABI_PARAM_0
 PARAM_limit     equ REG_ABI_PARAM_1
 PARAM_bufLimit  equ REG_ABI_PARAM_2
 
+ifdef Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT
 ; MY_ALIGN_64
+else
+  MY_ALIGN_16
+endif
 MY_PROC LzmaDec_DecodeReal_3, 3
 MY_PUSH_PRESERVED_ABI_REGS
 
@@ -1298,6 +1332,8 @@ fin:
 MY_POP_PRESERVED_ABI_REGS
 MY_ENDP
 
+ifdef Z7_LZMA_DEC_OPT_ASM_USE_SEGMENT
 _TEXT$LZMADECOPT ENDS
+endif
 
 end

Asm/x86/Sha1Opt.asm

@@ -1,5 +1,5 @@
 ; Sha1Opt.asm -- SHA-1 optimized code for SHA-1 x86 hardware instructions
-; 2021-03-10 : Igor Pavlov : Public domain
+; 2024-06-16 : Igor Pavlov : Public domain
 
 include 7zAsm.asm
 
@@ -20,7 +20,7 @@ MY_ASM_START
 
 
 
-CONST   SEGMENT
+CONST   SEGMENT READONLY
 
 align 16
 Reverse_Endian_Mask db 15,14,13,12, 11,10,9,8, 7,6,5,4, 3,2,1,0

Asm/x86/Sha256Opt.asm

@@ -1,5 +1,5 @@
 ; Sha256Opt.asm -- SHA-256 optimized code for SHA-256 x86 hardware instructions
-; 2021-03-10 : Igor Pavlov : Public domain
+; 2024-06-16 : Igor Pavlov : Public domain
 
 include 7zAsm.asm
 
@@ -20,7 +20,7 @@ endif
 EXTRN   K_CONST:xmmword
 @
 
-CONST   SEGMENT
+CONST   SEGMENT READONLY
 
 align 16
 Reverse_Endian_Mask db 3,2,1,0, 7,6,5,4, 11,10,9,8, 15,14,13,12
@@ -55,13 +55,19 @@ ifndef x64
     .xmm
 endif
         
+; jwasm-based assemblers for linux and linker from new versions of binutils
+; can generate incorrect code for load [ARRAY + offset] instructions.
+; 22.00: we load K_CONST offset to (rTable) register to avoid jwasm+binutils problem 
+        rTable  equ r0
+        ; rTable  equ K_CONST
+        
 ifdef x64
         rNum    equ REG_ABI_PARAM_2
     if (IS_LINUX eq 0)
         LOCAL_SIZE equ (16 * 2)
     endif
 else
-        rNum    equ r0
+        rNum    equ r3
         LOCAL_SIZE equ (16 * 1)
 endif
 
@@ -103,15 +109,18 @@ MY_PROLOG macro
         movdqa  [r4 + 16], xmm9
       endif
     else ; x86
-      if (IS_CDECL gt 0)
+        push    r3
-        mov     rState, [r4 + REG_SIZE * 1]
-        mov     rData,  [r4 + REG_SIZE * 2]
-        mov     rNum,   [r4 + REG_SIZE * 3]
-      else ; fastcall
-        mov     rNum,   [r4 + REG_SIZE * 1]
-      endif
         push    r5
         mov     r5, r4
+        NUM_PUSH_REGS   equ 2
+        PARAM_OFFSET    equ (REG_SIZE * (1 + NUM_PUSH_REGS))
+      if (IS_CDECL gt 0)
+        mov     rState, [r4 + PARAM_OFFSET]
+        mov     rData,  [r4 + PARAM_OFFSET + REG_SIZE * 1]
+        mov     rNum,   [r4 + PARAM_OFFSET + REG_SIZE * 2]
+      else ; fastcall
+        mov     rNum,   [r4 + PARAM_OFFSET]
+      endif
         and     r4, -16
         sub     r4, LOCAL_SIZE
     endif
@@ -129,6 +138,7 @@ MY_EPILOG macro
     else ; x86
         mov     r4, r5
         pop     r5
+        pop     r3
     endif
     MY_ENDP
 endm
@@ -171,7 +181,7 @@ pre2 equ 2
 
 
 RND4 macro k
-        movdqa  msg, xmmword ptr [K_CONST + (k) * 16]
+        movdqa  msg, xmmword ptr [rTable + (k) * 16]
         paddd   msg, @CatStr(xmm, %(w_regs + ((k + 0) mod 4)))
         MY_sha256rnds2 state0_N, state1_N
         pshufd   msg, msg, 0eH
@@ -210,6 +220,8 @@ endm
 MY_PROC Sha256_UpdateBlocks_HW, 3
     MY_PROLOG
 
+        lea     rTable, [K_CONST]
+
         cmp     rNum, 0
         je      end_c
 

Asm/x86/Sort.asm

@@ -0,0 +1,860 @@
+; SortTest.asm -- ASM version of HeapSort() function
+; Igor Pavlov : Public domain
+
+include ../../../../Asm/x86/7zAsm.asm
+
+MY_ASM_START
+
+ifndef Z7_SORT_ASM_USE_SEGMENT
+if (IS_LINUX gt 0)
+  ; Z7_SORT_ASM_USE_SEGMENT equ 1
+else
+  ; Z7_SORT_ASM_USE_SEGMENT equ 1
+endif
+endif
+
+ifdef Z7_SORT_ASM_USE_SEGMENT
+_TEXT$Z7_SORT SEGMENT ALIGN(64) 'CODE'
+MY_ALIGN macro num:req
+        align  num
+endm
+else
+MY_ALIGN macro num:req
+        ; We expect that ".text" is aligned for 16-bytes.
+        ; So we don't need large alignment inside our function.
+        align  16
+endm
+endif
+
+
+MY_ALIGN_16 macro
+        MY_ALIGN 16
+endm
+
+MY_ALIGN_32 macro
+        MY_ALIGN 32
+endm
+
+MY_ALIGN_64 macro
+        MY_ALIGN 64
+endm
+
+ifdef x64
+
+NUM_PREFETCH_LEVELS     equ 3 ; to prefetch 1x 64-bytes line (is good for most cases)
+; NUM_PREFETCH_LEVELS     equ 4 ; to prefetch 2x 64-bytes lines (better for big arrays)
+
+acc           equ x0
+k             equ r0
+k_x           equ x0
+
+p             equ r1
+
+s             equ r2
+s_x           equ x2
+
+a0            equ x3
+t0            equ a0
+
+a3            equ x5
+qq            equ a3
+
+a1            equ x6
+t1            equ a1
+t1_r          equ r6
+
+a2            equ x7
+t2            equ a2
+
+i             equ r8
+e0            equ x8
+
+e1            equ x9
+
+num_last      equ r10
+num_last_x    equ x10
+
+next4_lim     equ r11
+pref_lim      equ r12
+
+
+
+SORT_2_WITH_TEMP_REG macro b0, b1, temp_reg
+        mov     temp_reg, b0
+        cmp     b0, b1
+        cmovae  b0, b1 ; min
+        cmovae  b1, temp_reg ; max
+endm
+
+SORT macro b0, b1
+        SORT_2_WITH_TEMP_REG b0, b1, acc
+endm
+
+LOAD macro dest:req, index:req
+        mov     dest, [p + 4 * index]
+endm
+
+STORE macro reg:req, index:req
+        mov     [p + 4 * index], reg
+endm
+
+
+if (NUM_PREFETCH_LEVELS gt 3)
+   num_prefetches equ (1 SHL (NUM_PREFETCH_LEVELS - 3))
+else
+   num_prefetches equ 1
+endif
+
+PREFETCH_OP macro offs
+  cur_offset = 7 * 4    ; it's average offset in 64-bytes cache line.
+  ; cur_offset = 0      ; we can use zero offset, if we are sure that array is aligned for 64-bytes.
+  rept num_prefetches
+    if 1
+        prefetcht0 byte ptr [p + offs + cur_offset]
+    else
+        mov     pref_x, dword ptr [p + offs + cur_offset]
+    endif
+        cur_offset = cur_offset + 64
+  endm
+endm
+
+PREFETCH_MY macro
+if 1
+    if 1
+        shl     k, NUM_PREFETCH_LEVELS + 3
+    else
+        ; we delay prefetch instruction to improve main loads
+        shl     k, NUM_PREFETCH_LEVELS
+        shl     k, 3
+        ; shl     k, 0
+    endif
+        PREFETCH_OP k
+elseif 1
+        shl     k, 3
+        PREFETCH_OP k * (1 SHL NUM_PREFETCH_LEVELS) ; change it
+endif
+endm
+
+
+STEP_1 macro exit_label, prefetch_macro
+use_cmov_1 equ 1  ; set 1 for cmov, but it's slower in some cases
+                  ; set 0 for LOAD after adc s, 0
+        cmp     t0, t1
+    if use_cmov_1
+        cmovb   t0, t1
+        ; STORE   t0, k
+    endif
+        adc     s, 0
+    if use_cmov_1 eq 0
+        LOAD    t0, s
+    endif
+        cmp     qq, t0
+        jae     exit_label
+    if 1 ; use_cmov_1 eq 0
+        STORE   t0, k
+    endif
+        prefetch_macro
+        mov     t0, [p + s * 8]
+        mov     t1, [p + s * 8 + 4]
+        mov     k, s
+        add     s, s                 ; slower  for some cpus
+        ; lea     s, dword ptr [s + s] ; slower for some cpus
+        ; shl     s, 1               ; faster for some cpus
+        ; lea     s, dword ptr [s * 2] ; faster for some cpus
+    rept 0 ; 1000 for debug : 0 for normal
+        ; number of calls in generate_stage : ~0.6 of number of items
+        shl     k, 0
+    endm
+endm
+
+
+STEP_2 macro exit_label, prefetch_macro
+use_cmov_2 equ 0  ; set 1 for cmov, but it's slower in some cases
+                  ; set 0 for LOAD after adc s, 0
+        cmp     t0, t1
+    if use_cmov_2
+        mov     t2, t0
+        cmovb   t2, t1
+        ; STORE   t2, k
+    endif
+        mov     t0, [p + s * 8]
+        mov     t1, [p + s * 8 + 4]
+        cmovb   t0, [p + s * 8 + 8]
+        cmovb   t1, [p + s * 8 + 12]
+        adc     s, 0
+    if use_cmov_2 eq 0
+        LOAD    t2, s
+    endif
+        cmp     qq, t2
+        jae     exit_label
+    if 1 ; use_cmov_2 eq 0
+        STORE   t2, k
+    endif
+        prefetch_macro
+        mov     k, s
+        ; add     s, s
+        ; lea     s, [s + s]
+        shl     s, 1
+        ; lea     s, [s * 2]
+endm
+
+
+MOVE_SMALLEST_UP macro STEP, use_prefetch, num_unrolls
+        LOCAL exit_1, exit_2, leaves, opt_loop, last_nodes
+
+        ; s == k * 2
+        ; t0 == (p)[s]
+        ; t1 == (p)[s + 1]
+        cmp     k, next4_lim
+        jae     leaves
+
+    rept num_unrolls
+        STEP    exit_2
+        cmp     k, next4_lim
+        jae     leaves
+    endm
+  
+    if use_prefetch
+        prefetch_macro  equ PREFETCH_MY
+        pref_lim_2      equ pref_lim
+        ; lea     pref_lim, dword ptr [num_last + 1]
+        ; shr     pref_lim, NUM_PREFETCH_LEVELS + 1
+        cmp     k, pref_lim_2
+        jae     last_nodes
+    else
+        prefetch_macro  equ
+        pref_lim_2      equ next4_lim
+    endif
+  
+MY_ALIGN_16
+opt_loop:
+        STEP    exit_2, prefetch_macro
+        cmp     k, pref_lim_2
+        jb      opt_loop
+
+last_nodes:
+        ; k >= pref_lim_2
+        ; 2 cases are possible:
+        ;   case-1: num_after_prefetch_levels == 0 && next4_lim = pref_lim_2
+        ;   case-2: num_after_prefetch_levels == NUM_PREFETCH_LEVELS - 1 &&
+        ;        next4_lim = pref_lim_2 / (NUM_PREFETCH_LEVELS - 1)
+  if use_prefetch
+    yyy = NUM_PREFETCH_LEVELS - 1
+    while yyy
+        yyy = yyy - 1
+        STEP    exit_2
+      if yyy
+        cmp     k, next4_lim
+        jae     leaves
+      endif
+    endm
+  endif
+
+leaves:
+        ; k >= next4_lim == (num_last + 1) / 4 must be provided by previous code.
+        ;   we     have    2 nodes in (s)     level :  always
+        ;   we can have some nodes in (s * 2) level :  low probability case
+        ;   we     have   no nodes in (s * 4) level
+        ; s == k * 2
+        ; t0 == (p)[s]
+        ; t1 == (p)[s + 1]
+        cmp     t0, t1
+        cmovb   t0, t1
+        adc     s, 0
+        STORE   t0, k
+
+        ; t0 == (p)[s]
+        ; s / 2 == k  : (s) is index of max item from (p)[k * 2], (p)[k * 2 + 1]
+        ; we have 3 possible cases here:
+        ;   s * 2 >  num_last : (s) node has no childs
+        ;   s * 2 == num_last : (s) node has 1 leaf child that is last item of array
+        ;   s * 2 <  num_last : (s) node has 2 leaf childs. We provide (s * 4 > num_last)
+        ; we check for (s * 2 > num_last) before "cmp qq, t0" check, because
+        ; we will replace conditional jump with cmov instruction later.
+        lea     t1_r, dword ptr [s + s]
+        cmp     t1_r, num_last
+        ja      exit_1 ; if (s * 2 > num_last), we have no childs : it's high probability branch
+        
+        ; it's low probability branch
+        ; s * 2 <= num_last
+        cmp     qq, t0
+        jae     exit_2
+
+        ; qq < t0, so we go to next level
+        ; we check 1 or 2 childs in next level
+        mov     t0, [p + s * 8]
+        mov     k, s
+        mov     s, t1_r
+        cmp     t1_r, num_last
+        je      @F ; (s == num_last) means that we have single child in tree
+
+        ; (s < num_last) : so we must read both childs and select max of them.
+        mov     t1, [p + k * 8 + 4]
+        cmp     t0, t1
+        cmovb   t0, t1
+        adc     s, 0
+@@:
+        STORE   t0, k
+exit_1:
+        ; t0 == (p)[s],  s / 2 == k  : (s) is index of max item from (p)[k * 2], (p)[k * 2 + 1]
+        cmp     qq, t0
+        cmovb   k, s
+exit_2:
+        STORE   qq, k
+endm
+
+
+
+
+ifdef Z7_SORT_ASM_USE_SEGMENT
+; MY_ALIGN_64
+else
+  MY_ALIGN_16
+endif
+
+MY_PROC HeapSort, 2
+
+if (IS_LINUX gt 0)
+        mov     p, REG_ABI_PARAM_0    ; r1 <- r7 : linux
+endif
+        mov     num_last, REG_ABI_PARAM_1  ; r10 <- r6 : linux
+                                      ; r10 <- r2 : win64
+        cmp     num_last, 2
+        jb      end_1
+        
+        ; MY_PUSH_PRESERVED_ABI_REGS
+        MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
+        push    r12
+        
+        cmp     num_last, 4
+        ja      sort_5
+        
+        LOAD    a0, 0
+        LOAD    a1, 1
+        SORT    a0, a1
+        cmp     num_last, 3
+        jb      end_2
+        
+        LOAD    a2, 2
+        je      sort_3
+        
+        LOAD    a3, 3
+        SORT    a2, a3
+        SORT    a1, a3
+        STORE   a3, 3
+sort_3:
+        SORT    a0, a2
+        SORT    a1, a2
+        STORE   a2, 2
+        jmp     end_2
+        
+sort_5:
+        ; (num_last > 4) is required here
+        ; if (num_last >= 6) : we will use optimized loop for leaf nodes loop_down_1
+        mov     next4_lim, num_last
+        shr     next4_lim, 2
+        
+        dec     num_last
+        mov     k, num_last
+        shr     k, 1
+        mov     i, num_last
+        shr     i, 2
+        test    num_last, 1
+        jnz     size_even
+
+        ; ODD number of items. So we compare parent with single child
+        LOAD    t1, num_last
+        LOAD    t0, k
+        SORT_2_WITH_TEMP_REG  t1, t0, t2
+        STORE   t1, num_last
+        STORE   t0, k
+        dec     k
+
+size_even:
+        cmp     k, i
+        jbe     loop_down ; jump for num_last == 4 case
+
+if 0 ; 1 for debug
+        mov     r15, k
+        mov     r14d, 1 ; 100
+loop_benchmark:
+endif
+        ; optimized loop for leaf nodes:
+        mov     t0, [p + k * 8]
+        mov     t1, [p + k * 8 + 4]
+
+MY_ALIGN_16
+loop_down_1:
+        ; we compare parent with max of childs:
+        ; lea     s, dword ptr [2 * k]
+        mov     s, k
+        cmp     t0, t1
+        cmovb   t0, t1
+        adc     s, s
+        LOAD    t2, k
+        STORE   t0, k
+        cmp     t2, t0
+        cmovae  s, k
+        dec     k
+        ; we preload next items before STORE operation for calculated address
+        mov     t0, [p + k * 8]
+        mov     t1, [p + k * 8 + 4]
+        STORE   t2, s
+        cmp     k, i
+        jne     loop_down_1
+
+if 0 ; 1 for debug
+        mov     k, r15
+        dec     r14d
+        jnz     loop_benchmark
+        ; jmp end_debug
+endif
+       
+MY_ALIGN_16
+loop_down:
+        mov     t0, [p + i * 8]
+        mov     t1, [p + i * 8 + 4]
+        LOAD    qq, i
+        mov     k, i
+        lea     s, dword ptr [i + i]
+        ; jmp end_debug
+    DOWN_use_prefetch  equ 0
+    DOWN_num_unrolls   equ 0
+        MOVE_SMALLEST_UP  STEP_1, DOWN_use_prefetch, DOWN_num_unrolls
+        sub     i, 1
+        jnb     loop_down
+
+        ; jmp end_debug
+        LOAD    e0, 0
+        LOAD    e1, 1
+
+   LEVEL_3_LIMIT equ 8   ; 8 is default, but 7 also can work
+
+        cmp     num_last, LEVEL_3_LIMIT + 1
+        jb      main_loop_sort_5
+
+MY_ALIGN_16
+main_loop_sort:
+        ; num_last > LEVEL_3_LIMIT
+        ; p[size--] = p[0];
+        LOAD    qq, num_last
+        STORE   e0, num_last
+        mov     e0, e1
+        
+        mov     next4_lim, num_last
+        shr     next4_lim, 2
+        mov     pref_lim, num_last
+        shr     pref_lim, NUM_PREFETCH_LEVELS + 1
+        
+        dec     num_last
+if 0    ; 1 for debug
+        ; that optional optimization can improve the performance, if there are identical items in array
+        ;    3 times improvement : if all items in array are identical
+        ;   20%  improvement : if items are different for 1 bit only
+        ; 1-10%  improvement : if items are different for (2+) bits
+        ; no gain : if items are different
+        cmp     qq, e1
+        jae     next_iter_main
+endif
+        LOAD    e1, 2
+        LOAD    t0, 3
+        mov     k_x, 2
+        cmp     e1, t0
+        cmovb   e1, t0
+        mov     t0, [p + 4 * (4 + 0)]
+        mov     t1, [p + 4 * (4 + 1)]
+        cmovb   t0, [p + 4 * (4 + 2)]
+        cmovb   t1, [p + 4 * (4 + 3)]
+        adc     k_x, 0
+        ; (qq <= e1), because the tree is correctly sorted
+        ; also here we could check (qq >= e1) or (qq == e1) for faster exit
+        lea     s, dword ptr [k + k]
+    MAIN_use_prefetch  equ 1
+    MAIN_num_unrolls   equ 0
+        MOVE_SMALLEST_UP  STEP_2, MAIN_use_prefetch, MAIN_num_unrolls
+
+next_iter_main:
+        cmp     num_last, LEVEL_3_LIMIT
+        jne     main_loop_sort
+
+        ; num_last == LEVEL_3_LIMIT
+main_loop_sort_5:
+        ; 4 <= num_last <= LEVEL_3_LIMIT
+        ; p[size--] = p[0];
+        LOAD    qq, num_last
+        STORE   e0, num_last
+        mov     e0, e1
+        dec     num_last_x
+        
+        LOAD    e1, 2
+        LOAD    t0, 3
+        mov     k_x, 2
+        cmp     e1, t0
+        cmovb   e1, t0
+        adc     k_x, 0
+
+        lea     s_x, dword ptr [k * 2]
+        cmp     s_x, num_last_x
+        ja      exit_2
+
+        mov     t0, [p + k * 8]
+        je      exit_1
+
+        ; s < num_last
+        mov     t1, [p + k * 8 + 4]
+        cmp     t0, t1
+        cmovb   t0, t1
+        adc     s_x, 0
+exit_1:
+        STORE   t0, k
+        cmp     qq, t0
+        cmovb   k_x, s_x
+exit_2:
+        STORE   qq, k
+        cmp     num_last_x, 3
+        jne     main_loop_sort_5
+
+        ; num_last == 3 (real_size == 4)
+        LOAD    a0, 2
+        LOAD    a1, 3
+        STORE   e1, 2
+        STORE   e0, 3
+        SORT    a0, a1
+end_2:
+        STORE   a0, 0
+        STORE   a1, 1
+; end_debug:
+        ; MY_POP_PRESERVED_ABI_REGS
+        pop     r12
+        MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
+end_1:
+MY_ENDP
+
+
+
+else
+; ------------ x86 32-bit ------------
+
+ifdef x64
+IS_CDECL = 0
+endif
+
+acc           equ x0
+k             equ r0
+k_x           equ acc
+
+p             equ r1
+
+num_last      equ r2
+num_last_x    equ x2
+
+a0            equ x3
+t0            equ a0
+
+a3            equ x5
+i             equ r5
+e0            equ a3
+
+a1            equ x6
+qq            equ a1
+ 
+a2            equ x7
+s             equ r7
+s_x           equ a2
+
+
+SORT macro b0, b1
+        cmp     b1, b0
+        jae     @F
+    if 1
+        xchg    b0, b1
+    else
+        mov     acc, b0
+        mov     b0, b1 ; min
+        mov     b1, acc ; max
+    endif
+@@:
+endm
+
+LOAD macro dest:req, index:req
+        mov     dest, [p + 4 * index]
+endm
+
+STORE macro reg:req, index:req
+        mov     [p + 4 * index], reg
+endm
+
+
+STEP_1 macro exit_label
+        mov     t0, [p + k * 8]
+        cmp     t0, [p + k * 8 + 4]
+        adc     s, 0
+        LOAD    t0, s
+        STORE   t0, k ; we lookahed stooring for most expected branch
+        cmp     qq, t0
+        jae     exit_label
+        ; STORE   t0, k  ; use if
+        mov     k, s
+        add     s, s
+        ; lea     s, dword ptr [s + s]
+        ; shl     s, 1
+        ; lea     s, dword ptr [s * 2]
+endm
+
+STEP_BRANCH macro exit_label
+        mov     t0, [p + k * 8]
+        cmp     t0, [p + k * 8 + 4]
+        jae     @F
+        inc     s
+        mov     t0, [p + k * 8 + 4]
+@@:
+        cmp     qq, t0
+        jae     exit_label
+        STORE   t0, k
+        mov     k, s
+        add     s, s
+endm
+
+
+
+MOVE_SMALLEST_UP macro STEP, num_unrolls, exit_2
+        LOCAL leaves, opt_loop, single
+
+        ; s == k * 2
+    rept num_unrolls
+        cmp     s, num_last
+        jae     leaves
+        STEP_1  exit_2
+    endm
+        cmp     s, num_last
+        jb      opt_loop
+
+leaves:
+        ; (s >= num_last)
+        jne     exit_2
+single:
+        ; (s == num_last)
+        mov     t0, [p + k * 8]
+        cmp     qq, t0
+        jae     exit_2
+        STORE   t0, k
+        mov     k, s
+        jmp     exit_2
+ 
+MY_ALIGN_16
+opt_loop:
+        STEP    exit_2
+        cmp     s, num_last
+        jb      opt_loop
+        je      single
+exit_2:
+        STORE   qq, k
+endm
+
+
+
+
+ifdef Z7_SORT_ASM_USE_SEGMENT
+; MY_ALIGN_64
+else
+  MY_ALIGN_16
+endif
+
+MY_PROC HeapSort, 2
+  ifdef x64
+    if (IS_LINUX gt 0)
+        mov     num_last, REG_ABI_PARAM_1  ; r2 <- r6 : linux
+        mov     p,        REG_ABI_PARAM_0  ; r1 <- r7 : linux
+    endif
+  elseif (IS_CDECL gt 0)
+        mov     num_last, [r4 + REG_SIZE * 2]
+        mov     p,        [r4 + REG_SIZE * 1]
+  endif
+        cmp     num_last, 2
+        jb      end_1
+        MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
+        
+        cmp     num_last, 4
+        ja      sort_5
+        
+        LOAD    a0, 0
+        LOAD    a1, 1
+        SORT    a0, a1
+        cmp     num_last, 3
+        jb      end_2
+        
+        LOAD    a2, 2
+        je      sort_3
+        
+        LOAD    a3, 3
+        SORT    a2, a3
+        SORT    a1, a3
+        STORE   a3, 3
+sort_3:
+        SORT    a0, a2
+        SORT    a1, a2
+        STORE   a2, 2
+        jmp     end_2
+        
+sort_5:
+        ; num_last > 4
+        lea     i, dword ptr [num_last - 2]
+        dec     num_last
+        test    i, 1
+        jz      loop_down
+
+        ; single child
+        mov     t0, [p + num_last * 4]
+        mov     qq, [p + num_last * 2]
+        dec     i
+        cmp     qq, t0
+        jae     loop_down
+
+        mov     [p + num_last * 2], t0
+        mov     [p + num_last * 4], qq
+       
+MY_ALIGN_16
+loop_down:
+        mov     t0, [p + i * 4]
+        cmp     t0, [p + i * 4 + 4]
+        mov     k, i
+        mov     qq, [p + i * 2]
+        adc     k, 0
+        LOAD    t0, k
+        cmp     qq, t0
+        jae     down_next
+        mov     [p + i * 2], t0
+        lea     s, dword ptr [k + k]
+        
+        DOWN_num_unrolls   equ 0
+        MOVE_SMALLEST_UP  STEP_1, DOWN_num_unrolls, down_exit_label
+down_next:
+        sub     i, 2
+        jnb     loop_down
+        ; jmp end_debug
+
+        LOAD    e0, 0
+
+MY_ALIGN_16
+main_loop_sort:
+        ; num_last > 3
+        mov     t0, [p + 2 * 4]
+        cmp     t0, [p + 3 * 4]
+        LOAD    qq, num_last
+        STORE   e0, num_last
+        LOAD    e0, 1
+        mov     s_x, 2
+        mov     k_x, 1
+        adc     s, 0
+        LOAD    t0, s
+        dec     num_last
+        cmp     qq, t0
+        jae     main_exit_label
+        STORE   t0, 1
+        mov     k, s
+        add     s, s
+    if 1
+        ; for branch data prefetch mode :
+        ; it's faster for large arrays : larger than (1 << 13) items.
+        MAIN_num_unrolls   equ 10
+        STEP_LOOP          equ STEP_BRANCH
+    else
+        MAIN_num_unrolls   equ 0
+        STEP_LOOP          equ STEP_1
+    endif
+        
+        MOVE_SMALLEST_UP  STEP_LOOP, MAIN_num_unrolls, main_exit_label
+        
+        ; jmp end_debug
+        cmp     num_last, 3
+        jne     main_loop_sort
+
+        ; num_last == 3 (real_size == 4)
+        LOAD    a0, 2
+        LOAD    a1, 3
+        LOAD    a2, 1
+        STORE   e0, 3  ; e0 is alias for a3
+        STORE   a2, 2
+        SORT    a0, a1
+end_2:
+        STORE   a0, 0
+        STORE   a1, 1
+; end_debug:
+        MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
+end_1:
+MY_ENDP
+
+endif
+
+ifdef Z7_SORT_ASM_USE_SEGMENT
+_TEXT$Z7_SORT ENDS
+endif
+
+if 0
+LEA_IS_D8 (R64) [R2 * 4 + 16]
+ Lat : TP
+   2 :  1 :      adl-e
+   2 :  3   p056 adl-p
+   1 :  2 : p15  hsw-rocket
+   1 :  2 : p01  snb-ivb
+   1 :  1 : p1   conroe-wsm
+   1 :  4 : zen3,zen4
+   2 :  4 : zen1,zen2
+
+LEA_B_IS (R64) [R2 + R3 * 4]
+ Lat : TP
+   1 :  1 :      adl-e
+   2 :  3   p056 adl-p
+   1 :  2 : p15  hsw-rocket
+   1 :  2 : p01  snb-ivb
+   1 :  1 : p1   nhm-wsm
+   1 :  1 : p0   conroe-wsm
+   1 :  4 : zen3,zen4
+   2 :2,4 : zen1,zen2
+   
+LEA_B_IS_D8 (R64) [R2 + R3 * 4 + 16]
+ Lat : TP
+   2 :  1 :      adl-e
+   2 :  3   p056 adl-p
+   1 :  2 : p15  ice-rocket
+   3 :  1 : p1/p15 hsw-rocket
+   3 :  1 : p01  snb-ivb
+   1 :  1 : p1   nhm-wsm
+   1 :  1 : p0   conroe-wsm
+ 2,1 :  2 : zen3,zen4
+   2 :  2 : zen1,zen2
+   
+CMOVB (R64, R64)
+ Lat : TP
+ 1,2 :  2 :      adl-e
+   1 :  2   p06  adl-p
+   1 :  2 : p06  bwd-rocket
+ 1,2 :  2 : p0156+p06 hsw
+ 1,2 :1.5 : p015+p05  snb-ivb
+ 1,2 :  1 : p015+p05  nhm
+   1 :  1 : 2*p015  conroe
+   1 :  2 : zen3,zen4
+   1 :  4 : zen1,zen2
+
+ADC (R64, 0)
+ Lat : TP
+ 1,2 :  2 :      adl-e
+   1 :  2   p06  adl-p
+   1 :  2 : p06  bwd-rocket
+   1 :1.5 : p0156+p06 hsw
+   1 :1.5 : p015+p05  snb-ivb
+   2 :  1 : 2*p015    conroe-wstm
+   1 :  2 : zen1,zen2,zen3,zen4
+   
+PREFETCHNTA : fetch data into non-temporal cache close to the processor, minimizing cache pollution.
+  L1 : Pentium3
+  L2 : NetBurst
+  L1, not L2: Core duo, Core 2, Atom processors
+  L1, not L2, may fetch into L3 with fast replacement: Nehalem, Westmere, Sandy Bridge, ...
+      NEHALEM: Fills L1/L3, L1 LRU is not updated
+  L3 with fast replacement: Xeon Processors based on Nehalem, Westmere, Sandy Bridge, ...
+PREFETCHT0 : fetch data into all cache levels.
+PREFETCHT1 : fetch data into L2 and L3
+endif
+
+end

Asm/x86/XzCrc64Opt.asm

@@ -1,113 +1,231 @@
 ; XzCrc64Opt.asm -- CRC64 calculation : optimized version
-; 2021-02-06 : Igor Pavlov : Public domain
+; 2023-12-08 : Igor Pavlov : Public domain
 
 include 7zAsm.asm
 
 MY_ASM_START
 
+NUM_WORDS       equ     3
+
+if (NUM_WORDS lt 1) or (NUM_WORDS gt 64)
+.err <num_words_IS_INCORRECT>
+endif
+
+NUM_SKIP_BYTES  equ     ((NUM_WORDS - 2) * 4)
+
+
+MOVZXLO macro dest:req, src:req
+        movzx   dest, @CatStr(src, _L)
+endm
+
+MOVZXHI macro dest:req, src:req
+        movzx   dest, @CatStr(src, _H)
+endm
+
+
 ifdef x64
 
-rD      equ  r9
+rD      equ  r11
 rN      equ  r10
-rT      equ  r5
+rT      equ  r9
-num_VAR equ  r8
   
-SRCDAT4 equ  dword ptr [rD + rN * 1]
+CRC_OP macro op:req, dest:req, src:req, t:req
+        op      dest, QWORD PTR [rT + @CatStr(src, _R) * 8 + 0800h * (t)]
+endm
     
 CRC_XOR macro dest:req, src:req, t:req
-    xor     dest, QWORD PTR [rT + src * 8 + 0800h * t]
+        CRC_OP  xor, dest, src, t
+endm
+
+CRC_MOV macro dest:req, src:req, t:req
+        CRC_OP  mov, dest, src, t
 endm
 
 CRC1b macro
-    movzx   x6, BYTE PTR [rD]
+        movzx   x6, BYTE PTR [rD]
-    inc     rD
+        inc     rD
-    movzx   x3, x0_L
+        MOVZXLO x3, x0
-    xor     x6, x3
+        xor     x6, x3
-    shr     r0, 8
+        shr     r0, 8
-    CRC_XOR r0, r6, 0
+        CRC_XOR r0, x6, 0
-    dec     rN
+        dec     rN
 endm
 
-MY_PROLOG macro crc_end:req
+
-  ifdef ABI_LINUX
+; ALIGN_MASK is 3 or 7 bytes alignment:
-    MY_PUSH_2_REGS
+ALIGN_MASK      equ  (7 - (NUM_WORDS and 1) * 4)
-  else
+
-    MY_PUSH_4_REGS
+if NUM_WORDS eq 1
-  endif
+
-    mov     r0, REG_ABI_PARAM_0
+src_rN_offset   equ  4
-    mov     rN, REG_ABI_PARAM_2
+; + 4 for prefetching next 4-bytes after current iteration
-    mov     rT, REG_ABI_PARAM_3
+NUM_BYTES_LIMIT equ  (NUM_WORDS * 4 + 4)
-    mov     rD, REG_ABI_PARAM_1
+SRCDAT4         equ  DWORD PTR [rN + rD * 1]
-    test    rN, rN
+
-    jz      crc_end
+XOR_NEXT macro
-  @@:
+        mov     x1, [rD]
-    test    rD, 3
+        xor     r0, r1
-    jz      @F
-    CRC1b
-    jnz     @B
-  @@:
-    cmp     rN, 8
-    jb      crc_end
-    add     rN, rD
-    mov     num_VAR, rN
-    sub     rN, 4
-    and     rN, NOT 3
-    sub     rD, rN
-    mov     x1, SRCDAT4
-    xor     r0, r1
-    add     rN, 4
 endm
 
-MY_EPILOG macro crc_end:req
+else ; NUM_WORDS > 1
-    sub     rN, 4
+
-    mov     x1, SRCDAT4
+src_rN_offset   equ 8
-    xor     r0, r1
+; + 8 for prefetching next 8-bytes after current iteration
-    mov     rD, rN
+NUM_BYTES_LIMIT equ (NUM_WORDS * 4 + 8)
-    mov     rN, num_VAR
+
-    sub     rN, rD
+XOR_NEXT macro
-  crc_end:
+        xor     r0, QWORD PTR [rD] ; 64-bit read, can be unaligned
-    test    rN, rN
-    jz      @F
-    CRC1b
-    jmp     crc_end
-  @@:
-  ifdef ABI_LINUX
-    MY_POP_2_REGS
-  else
-    MY_POP_4_REGS
-  endif
 endm
 
-MY_PROC XzCrc64UpdateT4, 4
+; 32-bit or 64-bit
-    MY_PROLOG crc_end_4
+LOAD_SRC_MULT4 macro dest:req, word_index:req
-    align 16
+        mov     dest, [rN + rD * 1 + 4 * (word_index) - src_rN_offset];
-  main_loop_4:
+endm
-    mov     x1, SRCDAT4
-    movzx   x2, x0_L
-    movzx   x3, x0_H
-    shr     r0, 16
-    movzx   x6, x0_L
-    movzx   x7, x0_H
-    shr     r0, 16
-    CRC_XOR r1, r2, 3
-    CRC_XOR r0, r3, 2
-    CRC_XOR r1, r6, 1
-    CRC_XOR r0, r7, 0
-    xor     r0, r1
 
-    add     rD, 4
+endif
-    jnz     main_loop_4
 
-    MY_EPILOG crc_end_4
+
+
+MY_PROC @CatStr(XzCrc64UpdateT, %(NUM_WORDS * 4)), 4
+        MY_PUSH_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
+
+        mov     r0, REG_ABI_PARAM_0   ; r0  <- r1 / r7
+        mov     rD, REG_ABI_PARAM_1   ; r11 <- r2 / r6
+        mov     rN, REG_ABI_PARAM_2   ; r10 <- r8 / r2
+if  (IS_LINUX gt 0)
+        mov     rT, REG_ABI_PARAM_3   ; r9  <- r9 / r1
+endif
+
+        cmp     rN, NUM_BYTES_LIMIT + ALIGN_MASK
+        jb      crc_end
+@@:
+        test    rD, ALIGN_MASK
+        jz      @F
+        CRC1b
+        jmp     @B
+@@:
+        XOR_NEXT
+        lea     rN, [rD + rN * 1 - (NUM_BYTES_LIMIT - 1)]
+        sub     rD, rN
+        add     rN, src_rN_offset
+
+align 16
+@@:
+
+if NUM_WORDS eq 1
+  
+        mov     x1, x0
+        shr     x1, 8
+        MOVZXLO x3, x1
+        MOVZXLO x2, x0
+        shr     x1, 8
+        shr     r0, 32
+        xor     x0, SRCDAT4
+        CRC_XOR r0, x2, 3
+        CRC_XOR r0, x3, 2
+        MOVZXLO x2, x1
+        shr     x1, 8
+        CRC_XOR r0, x2, 1
+        CRC_XOR r0, x1, 0
+
+else ; NUM_WORDS > 1
+    
+if NUM_WORDS ne 2
+  k = 2
+  while k lt NUM_WORDS
+
+        LOAD_SRC_MULT4  x1, k
+    crc_op1  textequ <xor>
+
+    if k eq 2
+      if (NUM_WORDS and 1)
+        LOAD_SRC_MULT4  x7, NUM_WORDS       ; aligned 32-bit
+        LOAD_SRC_MULT4  x6, NUM_WORDS + 1   ; aligned 32-bit
+        shl     r6, 32
+      else
+        LOAD_SRC_MULT4  r6, NUM_WORDS       ; aligned 64-bit
+        crc_op1  textequ <mov>
+      endif
+    endif
+        table = 4 * (NUM_WORDS - 1 - k)
+        MOVZXLO x3, x1
+        CRC_OP crc_op1, r7, x3, 3 + table
+        MOVZXHI x3, x1
+        shr     x1, 16
+        CRC_XOR r6, x3, 2 + table
+        MOVZXLO x3, x1
+        shr     x1, 8
+        CRC_XOR r7, x3, 1 + table
+        CRC_XOR r6, x1, 0 + table
+        k = k + 1
+  endm
+        crc_op2  textequ <xor>
+
+else ; NUM_WORDS == 2
+        LOAD_SRC_MULT4  r6, NUM_WORDS       ; aligned 64-bit
+        crc_op2  textequ <mov>
+endif ; NUM_WORDS == 2
+
+        MOVZXHI x3, x0
+        MOVZXLO x2, x0
+        mov     r1, r0
+        shr     r1, 32
+        shr     x0, 16
+        CRC_XOR r6, x2, NUM_SKIP_BYTES + 7
+        CRC_OP  crc_op2, r7, x3, NUM_SKIP_BYTES + 6
+        MOVZXLO x2, x0
+        MOVZXHI x5, x1
+        MOVZXLO x3, x1
+        shr     x0, 8
+        shr     x1, 16
+        CRC_XOR r7, x2, NUM_SKIP_BYTES + 5
+        CRC_XOR r6, x3, NUM_SKIP_BYTES + 3
+        CRC_XOR r7, x0, NUM_SKIP_BYTES + 4
+        CRC_XOR r6, x5, NUM_SKIP_BYTES + 2
+        MOVZXLO x2, x1
+        shr     x1, 8
+        CRC_XOR r7, x2, NUM_SKIP_BYTES + 1
+        CRC_MOV r0, x1, NUM_SKIP_BYTES + 0
+        xor     r0, r6
+        xor     r0, r7
+
+endif ; NUM_WORDS > 1
+        add     rD, NUM_WORDS * 4
+        jnc     @B
+
+        sub     rN, src_rN_offset
+        add     rD, rN
+        XOR_NEXT
+        add     rN, NUM_BYTES_LIMIT - 1
+        sub     rN, rD
+
+crc_end:
+        test    rN, rN
+        jz      func_end
+@@:
+        CRC1b
+        jnz      @B
+func_end:
+        MY_POP_PRESERVED_ABI_REGS_UP_TO_INCLUDING_R11
 MY_ENDP
 
+
+
 else
+; ==================================================================
 ; x86 (32-bit)
 
-rD      equ  r1
+rD      equ  r7
-rN      equ  r7
+rN      equ  r1
 rT      equ  r5
 
+xA      equ  x6
+xA_R    equ  r6
+
+ifdef x64
+    num_VAR     equ  r8
+else
+
 crc_OFFS  equ  (REG_SIZE * 5)
 
 if (IS_CDECL gt 0) or (IS_LINUX gt 0)
@@ -133,107 +251,273 @@ else
     table_VAR   equ  [r4 + table_OFFS]
     num_VAR     equ  table_VAR
 endif
+endif ; x64
 
-SRCDAT4 equ  dword ptr [rD + rN * 1]
+SRCDAT4         equ     DWORD PTR [rN + rD * 1]
+
+CRC_1 macro op:req, dest:req, src:req, t:req, word_index:req
+        op      dest, DWORD PTR [rT + @CatStr(src, _R) * 8 + 0800h * (t) + (word_index) * 4]
+endm
 
 CRC macro op0:req, op1:req, dest0:req, dest1:req, src:req, t:req
-    op0     dest0, DWORD PTR [rT + src * 8 + 0800h * t]
+        CRC_1   op0, dest0, src, t, 0
-    op1     dest1, DWORD PTR [rT + src * 8 + 0800h * t + 4]
+        CRC_1   op1, dest1, src, t, 1
 endm
 
 CRC_XOR macro dest0:req, dest1:req, src:req, t:req
-    CRC xor, xor, dest0, dest1, src, t
+        CRC xor, xor, dest0, dest1, src, t
 endm
 
 
 CRC1b macro
-    movzx   x6, BYTE PTR [rD]
+        movzx   xA, BYTE PTR [rD]
-    inc     rD
+        inc     rD
-    movzx   x3, x0_L
+        MOVZXLO x3, x0
-    xor     x6, x3
+        xor     xA, x3
-    shrd    r0, r2, 8
+        shrd    x0, x2, 8
-    shr     r2, 8
+        shr     x2, 8
-    CRC_XOR r0, r2, r6, 0
+        CRC_XOR x0, x2, xA, 0
-    dec     rN
+        dec     rN
 endm
 
-MY_PROLOG macro crc_end:req
-    MY_PUSH_4_REGS
 
-  if (IS_CDECL gt 0) or (IS_LINUX gt 0)
+MY_PROLOG_BASE macro
-    proc_numParams = proc_numParams + 2 ; for ABI_LINUX
+        MY_PUSH_4_REGS
-    mov     rN, [r4 + size_OFFS]
+ifdef x64
-    mov     rD, [r4 + data_OFFS]
+        mov     r0, REG_ABI_PARAM_0     ; r0 <- r1 / r7
-  else
+        mov     rT, REG_ABI_PARAM_3     ; r5 <- r9 / r1
-    mov     rN, r2
+        mov     rN, REG_ABI_PARAM_2     ; r1 <- r8 / r2
-  endif
+        mov     rD, REG_ABI_PARAM_1     ; r7 <- r2 / r6
-
+        mov     r2, r0
-    mov     x0, [r4 + crc_OFFS]
+        shr     r2, 32
-    mov     x2, [r4 + crc_OFFS + 4]
+        mov     x0, x0
-    mov     rT, table_VAR
+else
-    test    rN, rN
+    if (IS_CDECL gt 0) or (IS_LINUX gt 0)
-    jz      crc_end
+        proc_numParams = proc_numParams + 2 ; for ABI_LINUX
-  @@:
+        mov     rN, [r4 + size_OFFS]
-    test    rD, 3
+        mov     rD, [r4 + data_OFFS]
-    jz      @F
+    else
-    CRC1b
+        mov     rD, REG_ABI_PARAM_0     ; r7 <- r1 : (data)
-    jnz     @B
+        mov     rN, REG_ABI_PARAM_1     ; r1 <- r2 : (size)
-  @@:
+    endif
-    cmp     rN, 8
+        mov     x0, [r4 + crc_OFFS]
-    jb      crc_end
+        mov     x2, [r4 + crc_OFFS + 4]
-    add     rN, rD
+        mov     rT, table_VAR
-
+endif
-    mov     num_VAR, rN
-
-    sub     rN, 4
-    and     rN, NOT 3
-    sub     rD, rN
-    xor     r0, SRCDAT4
-    add     rN, 4
 endm
 
-MY_EPILOG macro crc_end:req
-    sub     rN, 4
-    xor     r0, SRCDAT4
 
-    mov     rD, rN
+MY_EPILOG_BASE macro crc_end:req, func_end:req
-    mov     rN, num_VAR
+crc_end:
-    sub     rN, rD
+        test    rN, rN
-  crc_end:
+        jz      func_end
-    test    rN, rN
+@@:
-    jz      @F
+        CRC1b
-    CRC1b
+        jnz      @B
-    jmp     crc_end
+func_end:
-  @@:
+ifdef x64
-    MY_POP_4_REGS
+        shl     r2, 32
+        xor     r0, r2
+endif
+        MY_POP_4_REGS
 endm
 
-MY_PROC XzCrc64UpdateT4, 5
-    MY_PROLOG crc_end_4
-    movzx   x6, x0_L
-    align 16
-  main_loop_4:
-    mov     r3, SRCDAT4
-    xor     r3, r2
 
-    CRC xor, mov, r3, r2, r6, 3
+; ALIGN_MASK is 3 or 7 bytes alignment:
-    movzx   x6, x0_H
+ALIGN_MASK  equ     (7 - (NUM_WORDS and 1) * 4)
-    shr     r0, 16
-    CRC_XOR r3, r2, r6, 2
 
-    movzx   x6, x0_L
+if (NUM_WORDS eq 1)
-    movzx   x0, x0_H
-    CRC_XOR r3, r2, r6, 1
-    CRC_XOR r3, r2, r0, 0
-    movzx   x6, x3_L
-    mov     r0, r3
 
-    add     rD, 4
+NUM_BYTES_LIMIT_T4 equ (NUM_WORDS * 4 + 4)
-    jnz     main_loop_4
 
-    MY_EPILOG crc_end_4
+MY_PROC @CatStr(XzCrc64UpdateT, %(NUM_WORDS * 4)), 5
+        MY_PROLOG_BASE
+
+        cmp     rN, NUM_BYTES_LIMIT_T4 + ALIGN_MASK
+        jb      crc_end_4
+@@:
+        test    rD, ALIGN_MASK
+        jz      @F
+        CRC1b
+        jmp     @B
+@@:
+        xor     x0, [rD]
+        lea     rN, [rD + rN * 1 - (NUM_BYTES_LIMIT_T4 - 1)]
+        sub     rD, rN
+        add     rN, 4
+
+        MOVZXLO xA, x0
+align 16
+@@:
+        mov     x3, SRCDAT4
+        xor     x3, x2
+        shr     x0, 8
+        CRC xor, mov, x3, x2, xA, 3
+        MOVZXLO xA, x0
+        shr     x0, 8
+        ; MOVZXHI  xA, x0
+        ; shr     x0, 16
+        CRC_XOR x3, x2, xA, 2
+
+        MOVZXLO xA, x0
+        shr     x0, 8
+        CRC_XOR x3, x2, xA, 1
+        CRC_XOR x3, x2, x0, 0
+        MOVZXLO xA, x3
+        mov     x0, x3
+
+        add     rD, 4
+        jnc     @B
+
+        sub     rN, 4
+        add     rD, rN
+        xor     x0, [rD]
+        add     rN, NUM_BYTES_LIMIT_T4 - 1
+        sub     rN, rD
+        MY_EPILOG_BASE crc_end_4, func_end_4
 MY_ENDP
 
-endif ; ! x64
+else ; NUM_WORDS > 1
 
+SHR_X macro x, imm
+        shr x, imm
+endm
+
+
+ITER_1 macro v0, v1, a, off
+        MOVZXLO xA, a
+        SHR_X   a, 8
+        CRC_XOR v0, v1, xA, off
+endm
+
+
+ITER_4 macro v0, v1, a, off
+if 0 eq 0
+        ITER_1  v0, v1, a, off + 3
+        ITER_1  v0, v1, a, off + 2
+        ITER_1  v0, v1, a, off + 1
+        CRC_XOR v0, v1, a, off
+elseif 0 eq 0
+        MOVZXLO xA, a
+        CRC_XOR v0, v1, xA, off + 3
+        mov     xA, a
+        ror     a, 16   ; 32-bit ror
+        shr     xA, 24
+        CRC_XOR v0, v1, xA, off
+        MOVZXLO xA, a
+        SHR_X   a, 24
+        CRC_XOR v0, v1, xA, off + 1
+        CRC_XOR v0, v1, a, off + 2
+else
+        ; MOVZXHI provides smaller code, but MOVZX_HI_BYTE is not fast instruction
+        MOVZXLO xA, a
+        CRC_XOR v0, v1, xA, off + 3
+        MOVZXHI xA, a
+        SHR_X   a, 16
+        CRC_XOR v0, v1, xA, off + 2
+        MOVZXLO xA, a
+        SHR_X   a, 8
+        CRC_XOR v0, v1, xA, off + 1
+        CRC_XOR v0, v1, a, off
+endif
+endm
+
+
+
+ITER_1_PAIR macro v0, v1, a0, a1, off
+        ITER_1 v0, v1, a0, off + 4
+        ITER_1 v0, v1, a1, off
+endm
+
+src_rD_offset equ 8
+STEP_SIZE       equ     (NUM_WORDS * 4)
+
+ITER_12_NEXT macro op, index, v0, v1
+        op     v0, DWORD PTR [rD + (index + 1) * STEP_SIZE     - src_rD_offset]
+        op     v1, DWORD PTR [rD + (index + 1) * STEP_SIZE + 4 - src_rD_offset]
+endm
+
+ITER_12 macro index, a0, a1, v0, v1
+  
+  if NUM_SKIP_BYTES  eq 0
+        ITER_12_NEXT mov, index, v0, v1
+  else
+    k = 0
+    while k lt NUM_SKIP_BYTES
+        movzx   xA, BYTE PTR [rD + (index) * STEP_SIZE + k + 8 - src_rD_offset]
+      if k eq 0
+        CRC mov, mov,   v0, v1, xA, NUM_SKIP_BYTES - 1 - k
+      else
+        CRC_XOR         v0, v1, xA, NUM_SKIP_BYTES - 1 - k
+      endif
+      k = k + 1
+    endm
+        ITER_12_NEXT xor, index, v0, v1
+  endif
+
+if 0 eq 0
+        ITER_4  v0, v1, a0, NUM_SKIP_BYTES + 4
+        ITER_4  v0, v1, a1, NUM_SKIP_BYTES
+else ; interleave version is faster/slower for different processors
+        ITER_1_PAIR v0, v1, a0, a1, NUM_SKIP_BYTES + 3
+        ITER_1_PAIR v0, v1, a0, a1, NUM_SKIP_BYTES + 2
+        ITER_1_PAIR v0, v1, a0, a1, NUM_SKIP_BYTES + 1
+        CRC_XOR     v0, v1, a0,     NUM_SKIP_BYTES + 4
+        CRC_XOR     v0, v1, a1,     NUM_SKIP_BYTES
+endif
+endm
+
+; we use (UNROLL_CNT > 1) to reduce read ports pressure (num_VAR reads)
+UNROLL_CNT      equ     (2 * 1)
+NUM_BYTES_LIMIT equ     (STEP_SIZE * UNROLL_CNT + 8)
+
+MY_PROC @CatStr(XzCrc64UpdateT, %(NUM_WORDS * 4)), 5
+        MY_PROLOG_BASE
+
+        cmp     rN, NUM_BYTES_LIMIT + ALIGN_MASK
+        jb      crc_end_12
+@@:
+        test    rD, ALIGN_MASK
+        jz      @F
+        CRC1b
+        jmp     @B
+@@:
+        xor     x0, [rD]
+        xor     x2, [rD + 4]
+        add     rD, src_rD_offset
+        lea     rN, [rD + rN * 1 - (NUM_BYTES_LIMIT - 1)]
+        mov     num_VAR, rN
+
+align 16
+@@:
+    i = 0
+    rept UNROLL_CNT
+      if (i and 1) eq 0
+        ITER_12     i, x0, x2,  x1, x3
+      else
+        ITER_12     i, x1, x3,  x0, x2
+      endif
+      i = i + 1
+    endm
+
+    if (UNROLL_CNT and 1)
+        mov     x0, x1
+        mov     x2, x3
+    endif
+        add     rD, STEP_SIZE * UNROLL_CNT
+        cmp     rD, num_VAR
+        jb      @B
+
+        mov     rN, num_VAR
+        add     rN, NUM_BYTES_LIMIT - 1
+        sub     rN, rD
+        sub     rD, src_rD_offset
+        xor     x0, [rD]
+        xor     x2, [rD + 4]
+
+        MY_EPILOG_BASE crc_end_12, func_end_12
+MY_ENDP
+
+endif ; (NUM_WORDS > 1)
+endif ; ! x64
 end

C/7z.h

@@ -1,8 +1,8 @@
 /* 7z.h -- 7z interface
-2018-07-02 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_H
+#ifndef ZIP7_INC_7Z_H
-#define __7Z_H
+#define ZIP7_INC_7Z_H
 
 #include "7zTypes.h"
 
@@ -98,7 +98,7 @@ typedef struct
 UInt64 SzAr_GetFolderUnpackSize(const CSzAr *p, UInt32 folderIndex);
 
 SRes SzAr_DecodeFolder(const CSzAr *p, UInt32 folderIndex,
-    ILookInStream *stream, UInt64 startPos,
+    ILookInStreamPtr stream, UInt64 startPos,
     Byte *outBuffer, size_t outSize,
     ISzAllocPtr allocMain);
 
@@ -174,7 +174,7 @@ UInt16 *SzArEx_GetFullNameUtf16_Back(const CSzArEx *p, size_t fileIndex, UInt16
 
 SRes SzArEx_Extract(
     const CSzArEx *db,
-    ILookInStream *inStream,
+    ILookInStreamPtr inStream,
     UInt32 fileIndex,         /* index of file */
     UInt32 *blockIndex,       /* index of solid block */
     Byte **outBuffer,         /* pointer to pointer to output buffer (allocated with allocMain) */
@@ -196,7 +196,7 @@ SZ_ERROR_INPUT_EOF
 SZ_ERROR_FAIL
 */
 
-SRes SzArEx_Open(CSzArEx *p, ILookInStream *inStream,
+SRes SzArEx_Open(CSzArEx *p, ILookInStreamPtr inStream,
     ISzAllocPtr allocMain, ISzAllocPtr allocTemp);
 
 EXTERN_C_END

C/7zAlloc.c

@@ -1,5 +1,5 @@
-/* 7zAlloc.c -- Allocation functions
+/* 7zAlloc.c -- Allocation functions for 7z processing
-2017-04-03 : Igor Pavlov : Public domain */
+2023-03-04 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -7,74 +7,83 @@
 
 #include "7zAlloc.h"
 
-/* #define _SZ_ALLOC_DEBUG */
+/* #define SZ_ALLOC_DEBUG */
-/* use _SZ_ALLOC_DEBUG to debug alloc/free operations */
+/* use SZ_ALLOC_DEBUG to debug alloc/free operations */
 
-#ifdef _SZ_ALLOC_DEBUG
+#ifdef SZ_ALLOC_DEBUG
 
+/*
 #ifdef _WIN32
-#include <windows.h>
+#include "7zWindows.h"
 #endif
+*/
 
 #include <stdio.h>
-int g_allocCount = 0;
+static int g_allocCount = 0;
-int g_allocCountTemp = 0;
+static int g_allocCountTemp = 0;
 
+static void Print_Alloc(const char *s, size_t size, int *counter)
+{
+  const unsigned size2 = (unsigned)size;
+  fprintf(stderr, "\n%s count = %10d : %10u bytes; ", s, *counter, size2);
+  (*counter)++;
+}
+static void Print_Free(const char *s, int *counter)
+{
+  (*counter)--;
+  fprintf(stderr, "\n%s count = %10d", s, *counter);
+}
 #endif
 
 void *SzAlloc(ISzAllocPtr p, size_t size)
 {
-  UNUSED_VAR(p);
+  UNUSED_VAR(p)
   if (size == 0)
     return 0;
-  #ifdef _SZ_ALLOC_DEBUG
+  #ifdef SZ_ALLOC_DEBUG
-  fprintf(stderr, "\nAlloc %10u bytes; count = %10d", (unsigned)size, g_allocCount);
+  Print_Alloc("Alloc", size, &g_allocCount);
-  g_allocCount++;
   #endif
   return malloc(size);
 }
 
 void SzFree(ISzAllocPtr p, void *address)
 {
-  UNUSED_VAR(p);
+  UNUSED_VAR(p)
-  #ifdef _SZ_ALLOC_DEBUG
+  #ifdef SZ_ALLOC_DEBUG
-  if (address != 0)
+  if (address)
-  {
+    Print_Free("Free ", &g_allocCount);
-    g_allocCount--;
-    fprintf(stderr, "\nFree; count = %10d", g_allocCount);
-  }
   #endif
   free(address);
 }
 
 void *SzAllocTemp(ISzAllocPtr p, size_t size)
 {
-  UNUSED_VAR(p);
+  UNUSED_VAR(p)
   if (size == 0)
     return 0;
-  #ifdef _SZ_ALLOC_DEBUG
+  #ifdef SZ_ALLOC_DEBUG
-  fprintf(stderr, "\nAlloc_temp %10u bytes;  count = %10d", (unsigned)size, g_allocCountTemp);
+  Print_Alloc("Alloc_temp", size, &g_allocCountTemp);
-  g_allocCountTemp++;
+  /*
   #ifdef _WIN32
   return HeapAlloc(GetProcessHeap(), 0, size);
   #endif
+  */
   #endif
   return malloc(size);
 }
 
 void SzFreeTemp(ISzAllocPtr p, void *address)
 {
-  UNUSED_VAR(p);
+  UNUSED_VAR(p)
-  #ifdef _SZ_ALLOC_DEBUG
+  #ifdef SZ_ALLOC_DEBUG
-  if (address != 0)
+  if (address)
-  {
+    Print_Free("Free_temp ", &g_allocCountTemp);
-    g_allocCountTemp--;
+  /*
-    fprintf(stderr, "\nFree_temp; count = %10d", g_allocCountTemp);
-  }
   #ifdef _WIN32
   HeapFree(GetProcessHeap(), 0, address);
   return;
   #endif
+  */
   #endif
   free(address);
 }

C/7zAlloc.h

@@ -1,8 +1,8 @@
 /* 7zAlloc.h -- Allocation functions
-2017-04-03 : Igor Pavlov : Public domain */
+2023-03-04 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_ALLOC_H
+#ifndef ZIP7_INC_7Z_ALLOC_H
-#define __7Z_ALLOC_H
+#define ZIP7_INC_7Z_ALLOC_H
 
 #include "7zTypes.h"
 

C/7zBuf.h

@@ -1,8 +1,8 @@
 /* 7zBuf.h -- Byte Buffer
-2017-04-03 : Igor Pavlov : Public domain */
+2023-03-04 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_BUF_H
+#ifndef ZIP7_INC_7Z_BUF_H
-#define __7Z_BUF_H
+#define ZIP7_INC_7Z_BUF_H
 
 #include "7zTypes.h"
 

C/7zCrc.c

@@ -1,182 +1,218 @@
-/* 7zCrc.c -- CRC32 init
+/* 7zCrc.c -- CRC32 calculation and init
-2021-04-01 : Igor Pavlov : Public domain */
+2024-03-01 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include "7zCrc.h"
 #include "CpuArch.h"
 
-#define kCrcPoly 0xEDB88320
+// for debug:
+// #define __ARM_FEATURE_CRC32 1
 
-#ifdef MY_CPU_LE
+#ifdef __ARM_FEATURE_CRC32
-  #define CRC_NUM_TABLES 8
+// #pragma message("__ARM_FEATURE_CRC32")
+#define Z7_CRC_HW_FORCE
+#endif
+
+// #define Z7_CRC_DEBUG_BE
+#ifdef Z7_CRC_DEBUG_BE
+#undef MY_CPU_LE
+#define MY_CPU_BE
+#endif
+
+#ifdef Z7_CRC_HW_FORCE
+  #define Z7_CRC_NUM_TABLES_USE  1
 #else
-  #define CRC_NUM_TABLES 9
+#ifdef Z7_CRC_NUM_TABLES
-
+  #define Z7_CRC_NUM_TABLES_USE  Z7_CRC_NUM_TABLES
-  #define CRC_UINT32_SWAP(v) ((v >> 24) | ((v >> 8) & 0xFF00) | ((v << 8) & 0xFF0000) | (v << 24))
+#else
-
+  #define Z7_CRC_NUM_TABLES_USE  12
-  UInt32 MY_FAST_CALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, const UInt32 *table);
+#endif
-  UInt32 MY_FAST_CALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, const UInt32 *table);
 #endif
 
-#ifndef MY_CPU_BE
+#if Z7_CRC_NUM_TABLES_USE < 1
-  UInt32 MY_FAST_CALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table);
+  #error Stop_Compiling_Bad_Z7_CRC_NUM_TABLES
-  UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table);
 #endif
 
-typedef UInt32 (MY_FAST_CALL *CRC_FUNC)(UInt32 v, const void *data, size_t size, const UInt32 *table);
+#if defined(MY_CPU_LE) || (Z7_CRC_NUM_TABLES_USE == 1)
+  #define Z7_CRC_NUM_TABLES_TOTAL  Z7_CRC_NUM_TABLES_USE
+#else
+  #define Z7_CRC_NUM_TABLES_TOTAL  (Z7_CRC_NUM_TABLES_USE + 1)
+#endif
 
-extern
+#ifndef Z7_CRC_HW_FORCE
-CRC_FUNC g_CrcUpdateT4;
-CRC_FUNC g_CrcUpdateT4;
-extern
-CRC_FUNC g_CrcUpdateT8;
-CRC_FUNC g_CrcUpdateT8;
-extern
-CRC_FUNC g_CrcUpdateT0_32;
-CRC_FUNC g_CrcUpdateT0_32;
-extern
-CRC_FUNC g_CrcUpdateT0_64;
-CRC_FUNC g_CrcUpdateT0_64;
-extern
-CRC_FUNC g_CrcUpdate;
-CRC_FUNC g_CrcUpdate;
 
-UInt32 g_CrcTable[256 * CRC_NUM_TABLES];
+#if Z7_CRC_NUM_TABLES_USE == 1 \
-
+   || (!defined(MY_CPU_LE) && !defined(MY_CPU_BE))
-UInt32 MY_FAST_CALL CrcUpdate(UInt32 v, const void *data, size_t size)
+#define CRC_UPDATE_BYTE_2(crc, b)   (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8))
-{
+#define Z7_CRC_UPDATE_T1_FUNC_NAME  CrcUpdateGT1
-  return g_CrcUpdate(v, data, size, g_CrcTable);
+static UInt32 Z7_FASTCALL Z7_CRC_UPDATE_T1_FUNC_NAME(UInt32 v, const void *data, size_t size)
-}
-
-UInt32 MY_FAST_CALL CrcCalc(const void *data, size_t size)
-{
-  return g_CrcUpdate(CRC_INIT_VAL, data, size, g_CrcTable) ^ CRC_INIT_VAL;
-}
-
-#define CRC_UPDATE_BYTE_2(crc, b) (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8))
-
-UInt32 MY_FAST_CALL CrcUpdateT1(UInt32 v, const void *data, size_t size, const UInt32 *table);
-UInt32 MY_FAST_CALL CrcUpdateT1(UInt32 v, const void *data, size_t size, const UInt32 *table)
 {
+  const UInt32 *table = g_CrcTable;
   const Byte *p = (const Byte *)data;
-  const Byte *pEnd = p + size;
+  const Byte *lim = p + size;
-  for (; p != pEnd; p++)
+  for (; p != lim; p++)
     v = CRC_UPDATE_BYTE_2(v, *p);
   return v;
 }
+#endif
 
 
+#if Z7_CRC_NUM_TABLES_USE != 1
+#ifndef MY_CPU_BE
+  #define FUNC_NAME_LE_2(s)   CrcUpdateT ## s
+  #define FUNC_NAME_LE_1(s)   FUNC_NAME_LE_2(s)
+  #define FUNC_NAME_LE        FUNC_NAME_LE_1(Z7_CRC_NUM_TABLES_USE)
+  UInt32 Z7_FASTCALL FUNC_NAME_LE (UInt32 v, const void *data, size_t size, const UInt32 *table);
+#endif
+#ifndef MY_CPU_LE
+  #define FUNC_NAME_BE_2(s)   CrcUpdateT1_BeT ## s
+  #define FUNC_NAME_BE_1(s)   FUNC_NAME_BE_2(s)
+  #define FUNC_NAME_BE        FUNC_NAME_BE_1(Z7_CRC_NUM_TABLES_USE)
+  UInt32 Z7_FASTCALL FUNC_NAME_BE (UInt32 v, const void *data, size_t size, const UInt32 *table);
+#endif
+#endif
+
+#endif // Z7_CRC_HW_FORCE
+
 /* ---------- hardware CRC ---------- */
 
 #ifdef MY_CPU_LE
 
 #if defined(MY_CPU_ARM_OR_ARM64)
-
 // #pragma message("ARM*")
 
-  #if defined(_MSC_VER)
+  #if (defined(__clang__) && (__clang_major__ >= 3)) \
-    #if defined(MY_CPU_ARM64)
+     || defined(__GNUC__) && (__GNUC__ >= 6) && defined(MY_CPU_ARM64) \
-    #if (_MSC_VER >= 1910)
+     || defined(__GNUC__) && (__GNUC__ >= 8)
-        #define USE_ARM64_CRC
-    #endif
-    #endif
-  #elif (defined(__clang__) && (__clang_major__ >= 3)) \
-     || (defined(__GNUC__) && (__GNUC__ > 4))
       #if !defined(__ARM_FEATURE_CRC32)
+//        #pragma message("!defined(__ARM_FEATURE_CRC32)")
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
         #define __ARM_FEATURE_CRC32 1
-          #if (!defined(__clang__) || (__clang_major__ > 3)) // fix these numbers
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
-            #define ATTRIB_CRC __attribute__((__target__("arch=armv8-a+crc")))
+        #define Z7_ARM_FEATURE_CRC32_WAS_SET
+        #if defined(__clang__)
+          #if defined(MY_CPU_ARM64)
+            #define ATTRIB_CRC __attribute__((__target__("crc")))
+          #else
+            #define ATTRIB_CRC __attribute__((__target__("armv8-a,crc")))
           #endif
+        #else
+          #if defined(MY_CPU_ARM64)
+#if !defined(Z7_GCC_VERSION) || (Z7_GCC_VERSION >= 60000)
+            #define ATTRIB_CRC __attribute__((__target__("+crc")))
+#endif
+          #else
+#if !defined(Z7_GCC_VERSION) || (__GNUC__  >= 8)
+#if defined(__ARM_FP) && __GNUC__ >= 8
+// for -mfloat-abi=hard: similar to <arm_acle.h>
+            #define ATTRIB_CRC __attribute__((__target__("arch=armv8-a+crc+simd")))
+#else
+            #define ATTRIB_CRC __attribute__((__target__("arch=armv8-a+crc")))
+#endif
+#endif
+          #endif
+        #endif
       #endif
       #if defined(__ARM_FEATURE_CRC32)
-        #define USE_ARM64_CRC
+      // #pragma message("<arm_acle.h>")
+/*
+arm_acle.h (GGC):
+    before Nov 17, 2017:
+#ifdef __ARM_FEATURE_CRC32
+
+    Nov 17, 2017: gcc10.0  (gcc 9.2.0) checked"
+#if __ARM_ARCH >= 8
+#pragma GCC target ("arch=armv8-a+crc")
+
+    Aug 22, 2019: GCC 8.4?, 9.2.1, 10.1:
+#ifdef __ARM_FEATURE_CRC32
+#ifdef __ARM_FP
+#pragma GCC target ("arch=armv8-a+crc+simd")
+#else
+#pragma GCC target ("arch=armv8-a+crc")
+#endif
+*/
+#if defined(__ARM_ARCH) && __ARM_ARCH < 8
+#if defined(Z7_GCC_VERSION) && (__GNUC__ ==   8) && (Z7_GCC_VERSION <  80400) \
+ || defined(Z7_GCC_VERSION) && (__GNUC__ ==   9) && (Z7_GCC_VERSION <  90201) \
+ || defined(Z7_GCC_VERSION) && (__GNUC__ ==  10) && (Z7_GCC_VERSION < 100100)
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+// #pragma message("#define __ARM_ARCH 8")
+#undef  __ARM_ARCH
+#define __ARM_ARCH 8
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+#endif
+#endif
+        #define Z7_CRC_HW_USE
         #include <arm_acle.h>
       #endif
+  #elif defined(_MSC_VER)
+    #if defined(MY_CPU_ARM64)
+    #if (_MSC_VER >= 1910)
+    #ifdef __clang__
+       // #define Z7_CRC_HW_USE
+       // #include <arm_acle.h>
+    #else
+       #define Z7_CRC_HW_USE
+       #include <intrin.h>
+    #endif
+    #endif
+    #endif
   #endif
 
-#else
+#else // non-ARM*
 
-// no hardware CRC
+// #define Z7_CRC_HW_USE // for debug : we can test HW-branch of code
-
+#ifdef Z7_CRC_HW_USE
-// #define USE_CRC_EMU
+#include "7zCrcEmu.h"
-
-#ifdef USE_CRC_EMU
-
-#pragma message("ARM64 CRC emulation")
-
-MY_FORCE_INLINE
-UInt32 __crc32b(UInt32 v, UInt32 data)
-{
-  const UInt32 *table = g_CrcTable;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data);
-  return v;
-}
-
-MY_FORCE_INLINE
-UInt32 __crc32w(UInt32 v, UInt32 data)
-{
-  const UInt32 *table = g_CrcTable;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  return v;
-}
-
-MY_FORCE_INLINE
-UInt32 __crc32d(UInt32 v, UInt64 data)
-{
-  const UInt32 *table = g_CrcTable;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  v = CRC_UPDATE_BYTE_2(v, (Byte)data); data >>= 8;
-  return v;
-}
-
-#endif // USE_CRC_EMU
-
-#endif // defined(MY_CPU_ARM64) && defined(MY_CPU_LE)
-
-
-
-#if defined(USE_ARM64_CRC) || defined(USE_CRC_EMU)
-
-#define T0_32_UNROLL_BYTES (4 * 4)
-#define T0_64_UNROLL_BYTES (4 * 8)
-
-#ifndef ATTRIB_CRC
-#define ATTRIB_CRC
 #endif
+
+#endif // non-ARM*
+
+
+
+#if defined(Z7_CRC_HW_USE)
+
 // #pragma message("USE ARM HW CRC")
 
-ATTRIB_CRC
+#ifdef MY_CPU_64BIT
-UInt32 MY_FAST_CALL CrcUpdateT0_32(UInt32 v, const void *data, size_t size, const UInt32 *table);
+  #define CRC_HW_WORD_TYPE  UInt64
-ATTRIB_CRC
+  #define CRC_HW_WORD_FUNC  __crc32d
-UInt32 MY_FAST_CALL CrcUpdateT0_32(UInt32 v, const void *data, size_t size, const UInt32 *table)
+#else
+  #define CRC_HW_WORD_TYPE  UInt32
+  #define CRC_HW_WORD_FUNC  __crc32w
+#endif
+
+#define CRC_HW_UNROLL_BYTES (sizeof(CRC_HW_WORD_TYPE) * 4)
+
+#ifdef ATTRIB_CRC
+  ATTRIB_CRC
+#endif
+Z7_NO_INLINE
+#ifdef Z7_CRC_HW_FORCE
+         UInt32 Z7_FASTCALL CrcUpdate
+#else
+  static UInt32 Z7_FASTCALL CrcUpdate_HW
+#endif
+    (UInt32 v, const void *data, size_t size)
 {
   const Byte *p = (const Byte *)data;
-  UNUSED_VAR(table);
+  for (; size != 0 && ((unsigned)(ptrdiff_t)p & (CRC_HW_UNROLL_BYTES - 1)) != 0; size--)
-
-  for (; size != 0 && ((unsigned)(ptrdiff_t)p & (T0_32_UNROLL_BYTES - 1)) != 0; size--)
     v = __crc32b(v, *p++);
-
+  if (size >= CRC_HW_UNROLL_BYTES)
-  if (size >= T0_32_UNROLL_BYTES)
   {
     const Byte *lim = p + size;
-    size &= (T0_32_UNROLL_BYTES - 1);
+    size &= CRC_HW_UNROLL_BYTES - 1;
     lim -= size;
     do
     {
-      v = __crc32w(v, *(const UInt32 *)(const void *)(p));
+      v = CRC_HW_WORD_FUNC(v, *(const CRC_HW_WORD_TYPE *)(const void *)(p));
-      v = __crc32w(v, *(const UInt32 *)(const void *)(p + 4)); p += 2 * 4;
+      v = CRC_HW_WORD_FUNC(v, *(const CRC_HW_WORD_TYPE *)(const void *)(p + sizeof(CRC_HW_WORD_TYPE)));
-      v = __crc32w(v, *(const UInt32 *)(const void *)(p));
+      p += 2 * sizeof(CRC_HW_WORD_TYPE);
-      v = __crc32w(v, *(const UInt32 *)(const void *)(p + 4)); p += 2 * 4;
+      v = CRC_HW_WORD_FUNC(v, *(const CRC_HW_WORD_TYPE *)(const void *)(p));
+      v = CRC_HW_WORD_FUNC(v, *(const CRC_HW_WORD_TYPE *)(const void *)(p + sizeof(CRC_HW_WORD_TYPE)));
+      p += 2 * sizeof(CRC_HW_WORD_TYPE);
     }
     while (p != lim);
   }
@@ -187,136 +223,198 @@ UInt32 MY_FAST_CALL CrcUpdateT0_32(UInt32 v, const void *data, size_t size, cons
   return v;
 }
 
-ATTRIB_CRC
+#ifdef Z7_ARM_FEATURE_CRC32_WAS_SET
-UInt32 MY_FAST_CALL CrcUpdateT0_64(UInt32 v, const void *data, size_t size, const UInt32 *table);
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
-ATTRIB_CRC
+#undef __ARM_FEATURE_CRC32
-UInt32 MY_FAST_CALL CrcUpdateT0_64(UInt32 v, const void *data, size_t size, const UInt32 *table)
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
-{
+#undef Z7_ARM_FEATURE_CRC32_WAS_SET
-  const Byte *p = (const Byte *)data;
+#endif
-  UNUSED_VAR(table);
-
-  for (; size != 0 && ((unsigned)(ptrdiff_t)p & (T0_64_UNROLL_BYTES - 1)) != 0; size--)
-    v = __crc32b(v, *p++);
-
-  if (size >= T0_64_UNROLL_BYTES)
-  {
-    const Byte *lim = p + size;
-    size &= (T0_64_UNROLL_BYTES - 1);
-    lim -= size;
-    do
-    {
-      v = __crc32d(v, *(const UInt64 *)(const void *)(p));
-      v = __crc32d(v, *(const UInt64 *)(const void *)(p + 8)); p += 2 * 8;
-      v = __crc32d(v, *(const UInt64 *)(const void *)(p));
-      v = __crc32d(v, *(const UInt64 *)(const void *)(p + 8)); p += 2 * 8;
-    }
-    while (p != lim);
-  }
-  
-  for (; size != 0; size--)
-    v = __crc32b(v, *p++);
-
-  return v;
-}
-
-#endif // defined(USE_ARM64_CRC) || defined(USE_CRC_EMU)
 
+#endif // defined(Z7_CRC_HW_USE)
 #endif // MY_CPU_LE
 
 
 
+#ifndef Z7_CRC_HW_FORCE
 
-void MY_FAST_CALL CrcGenerateTable()
+#if defined(Z7_CRC_HW_USE) || defined(Z7_CRC_UPDATE_T1_FUNC_NAME)
+/*
+typedef UInt32 (Z7_FASTCALL *Z7_CRC_UPDATE_WITH_TABLE_FUNC)
+    (UInt32 v, const void *data, size_t size, const UInt32 *table);
+Z7_CRC_UPDATE_WITH_TABLE_FUNC g_CrcUpdate;
+*/
+static unsigned g_Crc_Algo;
+#if (!defined(MY_CPU_LE) && !defined(MY_CPU_BE))
+static unsigned g_Crc_Be;
+#endif
+#endif // defined(Z7_CRC_HW_USE) || defined(Z7_CRC_UPDATE_T1_FUNC_NAME)
+
+
+
+Z7_NO_INLINE
+#ifdef Z7_CRC_HW_USE
+  static UInt32 Z7_FASTCALL CrcUpdate_Base
+#else
+         UInt32 Z7_FASTCALL CrcUpdate
+#endif
+    (UInt32 crc, const void *data, size_t size)
+{
+#if Z7_CRC_NUM_TABLES_USE == 1
+    return Z7_CRC_UPDATE_T1_FUNC_NAME(crc, data, size);
+#else // Z7_CRC_NUM_TABLES_USE != 1
+#ifdef Z7_CRC_UPDATE_T1_FUNC_NAME
+  if (g_Crc_Algo == 1)
+    return Z7_CRC_UPDATE_T1_FUNC_NAME(crc, data, size);
+#endif
+
+#ifdef MY_CPU_LE
+    return FUNC_NAME_LE(crc, data, size, g_CrcTable);
+#elif defined(MY_CPU_BE)
+    return FUNC_NAME_BE(crc, data, size, g_CrcTable);
+#else
+  if (g_Crc_Be)
+    return FUNC_NAME_BE(crc, data, size, g_CrcTable);
+  else
+    return FUNC_NAME_LE(crc, data, size, g_CrcTable);
+#endif
+#endif // Z7_CRC_NUM_TABLES_USE != 1
+}
+
+
+#ifdef Z7_CRC_HW_USE
+Z7_NO_INLINE
+UInt32 Z7_FASTCALL CrcUpdate(UInt32 crc, const void *data, size_t size)
+{
+  if (g_Crc_Algo == 0)
+    return CrcUpdate_HW(crc, data, size);
+  return CrcUpdate_Base(crc, data, size);
+}
+#endif
+
+#endif // !defined(Z7_CRC_HW_FORCE)
+
+
+
+UInt32 Z7_FASTCALL CrcCalc(const void *data, size_t size)
+{
+  return CrcUpdate(CRC_INIT_VAL, data, size) ^ CRC_INIT_VAL;
+}
+
+
+MY_ALIGN(64)
+UInt32 g_CrcTable[256 * Z7_CRC_NUM_TABLES_TOTAL];
+
+
+void Z7_FASTCALL CrcGenerateTable(void)
 {
   UInt32 i;
   for (i = 0; i < 256; i++)
   {
+#if defined(Z7_CRC_HW_FORCE)
+    g_CrcTable[i] = __crc32b(i, 0);
+#else
+    #define kCrcPoly 0xEDB88320
     UInt32 r = i;
     unsigned j;
     for (j = 0; j < 8; j++)
       r = (r >> 1) ^ (kCrcPoly & ((UInt32)0 - (r & 1)));
     g_CrcTable[i] = r;
+#endif
   }
-  for (i = 256; i < 256 * CRC_NUM_TABLES; i++)
+  for (i = 256; i < 256 * Z7_CRC_NUM_TABLES_USE; i++)
   {
-    UInt32 r = g_CrcTable[(size_t)i - 256];
+    const UInt32 r = g_CrcTable[(size_t)i - 256];
     g_CrcTable[i] = g_CrcTable[r & 0xFF] ^ (r >> 8);
   }
 
-  #if CRC_NUM_TABLES < 4
+#if !defined(Z7_CRC_HW_FORCE) && \
+    (defined(Z7_CRC_HW_USE) || defined(Z7_CRC_UPDATE_T1_FUNC_NAME) || defined(MY_CPU_BE))
 
-  g_CrcUpdate = CrcUpdateT1;
+#if Z7_CRC_NUM_TABLES_USE <= 1
+    g_Crc_Algo = 1;
+#else // Z7_CRC_NUM_TABLES_USE <= 1
 
-  #else
+#if defined(MY_CPU_LE)
- 
+    g_Crc_Algo = Z7_CRC_NUM_TABLES_USE;
-  #ifdef MY_CPU_LE
+#else // !defined(MY_CPU_LE)
-
-    g_CrcUpdateT4 = CrcUpdateT4;
-    g_CrcUpdate = CrcUpdateT4;
-
-    #if CRC_NUM_TABLES >= 8
-      g_CrcUpdateT8 = CrcUpdateT8;
-  
-      #ifdef MY_CPU_X86_OR_AMD64
-      if (!CPU_Is_InOrder())
-      #endif
-        g_CrcUpdate = CrcUpdateT8;
-    #endif
-
-  #else
   {
-    #ifndef MY_CPU_BE
+#ifndef MY_CPU_BE
     UInt32 k = 0x01020304;
     const Byte *p = (const Byte *)&k;
     if (p[0] == 4 && p[1] == 3)
-    {
+      g_Crc_Algo = Z7_CRC_NUM_TABLES_USE;
-      g_CrcUpdateT4 = CrcUpdateT4;
-      g_CrcUpdate = CrcUpdateT4;
-      #if CRC_NUM_TABLES >= 8
-      g_CrcUpdateT8 = CrcUpdateT8;
-      g_CrcUpdate = CrcUpdateT8;
-      #endif
-    }
     else if (p[0] != 1 || p[1] != 2)
-      g_CrcUpdate = CrcUpdateT1;
+      g_Crc_Algo = 1;
     else
-    #endif
+#endif // MY_CPU_BE
     {
-      for (i = 256 * CRC_NUM_TABLES - 1; i >= 256; i--)
+      for (i = 256 * Z7_CRC_NUM_TABLES_TOTAL - 1; i >= 256; i--)
       {
-        UInt32 x = g_CrcTable[(size_t)i - 256];
+        const UInt32 x = g_CrcTable[(size_t)i - 256];
-        g_CrcTable[i] = CRC_UINT32_SWAP(x);
+        g_CrcTable[i] = Z7_BSWAP32(x);
       }
-      g_CrcUpdateT4 = CrcUpdateT1_BeT4;
+#if defined(Z7_CRC_UPDATE_T1_FUNC_NAME)
-      g_CrcUpdate = CrcUpdateT1_BeT4;
+      g_Crc_Algo = Z7_CRC_NUM_TABLES_USE;
-      #if CRC_NUM_TABLES >= 8
+#endif
-      g_CrcUpdateT8 = CrcUpdateT1_BeT8;
+#if (!defined(MY_CPU_LE) && !defined(MY_CPU_BE))
-      g_CrcUpdate = CrcUpdateT1_BeT8;
+      g_Crc_Be = 1;
-      #endif
+#endif
     }
   }
-  #endif
+#endif  // !defined(MY_CPU_LE)
-  #endif
 
-  #ifdef MY_CPU_LE
+#ifdef MY_CPU_LE
-    #ifdef USE_ARM64_CRC
+#ifdef Z7_CRC_HW_USE
-      if (CPU_IsSupported_CRC32())
+  if (CPU_IsSupported_CRC32())
-      {
+    g_Crc_Algo = 0;
-        g_CrcUpdateT0_32 = CrcUpdateT0_32;
+#endif // Z7_CRC_HW_USE
-        g_CrcUpdateT0_64 = CrcUpdateT0_64;
+#endif // MY_CPU_LE
-        g_CrcUpdate =
-          #if defined(MY_CPU_ARM)
-            CrcUpdateT0_32;
-          #else
-            CrcUpdateT0_64;
-          #endif
-      }
-    #endif
 
-    #ifdef USE_CRC_EMU
+#endif // Z7_CRC_NUM_TABLES_USE <= 1
-      g_CrcUpdateT0_32 = CrcUpdateT0_32;
+#endif // g_Crc_Algo was declared
-      g_CrcUpdateT0_64 = CrcUpdateT0_64;
-      g_CrcUpdate = CrcUpdateT0_64;
-    #endif
-  #endif
 }
+
+Z7_CRC_UPDATE_FUNC z7_GetFunc_CrcUpdate(unsigned algo)
+{
+  if (algo == 0)
+    return &CrcUpdate;
+
+#if defined(Z7_CRC_HW_USE)
+  if (algo == sizeof(CRC_HW_WORD_TYPE) * 8)
+  {
+#ifdef Z7_CRC_HW_FORCE
+    return &CrcUpdate;
+#else
+    if (g_Crc_Algo == 0)
+      return &CrcUpdate_HW;
+#endif
+  }
+#endif
+
+#ifndef Z7_CRC_HW_FORCE
+  if (algo == Z7_CRC_NUM_TABLES_USE)
+    return
+  #ifdef Z7_CRC_HW_USE
+      &CrcUpdate_Base;
+  #else
+      &CrcUpdate;
+  #endif
+#endif
+
+  return NULL;
+}
+
+#undef kCrcPoly
+#undef Z7_CRC_NUM_TABLES_USE
+#undef Z7_CRC_NUM_TABLES_TOTAL
+#undef CRC_UPDATE_BYTE_2
+#undef FUNC_NAME_LE_2
+#undef FUNC_NAME_LE_1
+#undef FUNC_NAME_LE
+#undef FUNC_NAME_BE_2
+#undef FUNC_NAME_BE_1
+#undef FUNC_NAME_BE
+
+#undef CRC_HW_UNROLL_BYTES
+#undef CRC_HW_WORD_FUNC
+#undef CRC_HW_WORD_TYPE

C/7zCrc.h

@@ -1,8 +1,8 @@
 /* 7zCrc.h -- CRC32 calculation
-2013-01-18 : Igor Pavlov : Public domain */
+2024-01-22 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_CRC_H
+#ifndef ZIP7_INC_7Z_CRC_H
-#define __7Z_CRC_H
+#define ZIP7_INC_7Z_CRC_H
 
 #include "7zTypes.h"
 
@@ -11,14 +11,17 @@ EXTERN_C_BEGIN
 extern UInt32 g_CrcTable[];
 
 /* Call CrcGenerateTable one time before other CRC functions */
-void MY_FAST_CALL CrcGenerateTable(void);
+void Z7_FASTCALL CrcGenerateTable(void);
 
 #define CRC_INIT_VAL 0xFFFFFFFF
 #define CRC_GET_DIGEST(crc) ((crc) ^ CRC_INIT_VAL)
 #define CRC_UPDATE_BYTE(crc, b) (g_CrcTable[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8))
 
-UInt32 MY_FAST_CALL CrcUpdate(UInt32 crc, const void *data, size_t size);
+UInt32 Z7_FASTCALL CrcUpdate(UInt32 crc, const void *data, size_t size);
-UInt32 MY_FAST_CALL CrcCalc(const void *data, size_t size);
+UInt32 Z7_FASTCALL CrcCalc(const void *data, size_t size);
+
+typedef UInt32 (Z7_FASTCALL *Z7_CRC_UPDATE_FUNC)(UInt32 v, const void *data, size_t size);
+Z7_CRC_UPDATE_FUNC z7_GetFunc_CrcUpdate(unsigned algo);
 
 EXTERN_C_END
 

C/7zCrcOpt.c

@@ -1,117 +1,199 @@
-/* 7zCrcOpt.c -- CRC32 calculation
+/* 7zCrcOpt.c -- CRC32 calculation (optimized functions)
-2021-02-09 : Igor Pavlov : Public domain */
+2023-12-07 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include "CpuArch.h"
 
+#if !defined(Z7_CRC_NUM_TABLES) || Z7_CRC_NUM_TABLES > 1
+
+// for debug only : define Z7_CRC_DEBUG_BE to test big-endian code in little-endian cpu
+// #define Z7_CRC_DEBUG_BE
+#ifdef Z7_CRC_DEBUG_BE
+#undef MY_CPU_LE
+#define MY_CPU_BE
+#endif
+
+// the value Z7_CRC_NUM_TABLES_USE must be defined to same value as in 7zCrc.c
+#ifdef Z7_CRC_NUM_TABLES
+#define Z7_CRC_NUM_TABLES_USE  Z7_CRC_NUM_TABLES
+#else
+#define Z7_CRC_NUM_TABLES_USE  12
+#endif
+
+#if Z7_CRC_NUM_TABLES_USE % 4     || \
+    Z7_CRC_NUM_TABLES_USE < 4 * 1 || \
+    Z7_CRC_NUM_TABLES_USE > 4 * 6
+  #error Stop_Compiling_Bad_Z7_CRC_NUM_TABLES
+#endif
+
+
 #ifndef MY_CPU_BE
 
-#define CRC_UPDATE_BYTE_2(crc, b) (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8))
+#define CRC_UPDATE_BYTE_2(crc, b)  (table[((crc) ^ (b)) & 0xFF] ^ ((crc) >> 8))
 
-UInt32 MY_FAST_CALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table);
+#define Q(n, d) \
-UInt32 MY_FAST_CALL CrcUpdateT4(UInt32 v, const void *data, size_t size, const UInt32 *table)
+    ( (table + ((n) * 4 + 3) * 0x100)[(Byte)(d)] \
+    ^ (table + ((n) * 4 + 2) * 0x100)[((d) >> 1 * 8) & 0xFF] \
+    ^ (table + ((n) * 4 + 1) * 0x100)[((d) >> 2 * 8) & 0xFF] \
+    ^ (table + ((n) * 4 + 0) * 0x100)[((d) >> 3 * 8)] )
+
+#define R(a)  *((const UInt32 *)(const void *)p + (a))
+
+#define CRC_FUNC_PRE_LE2(step) \
+UInt32 Z7_FASTCALL CrcUpdateT ## step (UInt32 v, const void *data, size_t size, const UInt32 *table)
+
+#define CRC_FUNC_PRE_LE(step)   \
+        CRC_FUNC_PRE_LE2(step); \
+        CRC_FUNC_PRE_LE2(step)
+
+CRC_FUNC_PRE_LE(Z7_CRC_NUM_TABLES_USE)
 {
   const Byte *p = (const Byte *)data;
-  for (; size > 0 && ((unsigned)(ptrdiff_t)p & 3) != 0; size--, p++)
+  const Byte *lim;
+  for (; size && ((unsigned)(ptrdiff_t)p & (7 - (Z7_CRC_NUM_TABLES_USE & 4))) != 0; size--, p++)
     v = CRC_UPDATE_BYTE_2(v, *p);
-  for (; size >= 4; size -= 4, p += 4)
+  lim = p + size;
+  if (size >= Z7_CRC_NUM_TABLES_USE)
   {
-    v ^= *(const UInt32 *)(const void *)p;
+    lim -= Z7_CRC_NUM_TABLES_USE;
-    v =
+    do
-          (table + 0x300)[((v      ) & 0xFF)]
+    {
-        ^ (table + 0x200)[((v >>  8) & 0xFF)]
+      v ^= R(0);
-        ^ (table + 0x100)[((v >> 16) & 0xFF)]
+      {
-        ^ (table + 0x000)[((v >> 24))];
+#if Z7_CRC_NUM_TABLES_USE == 1 * 4
+        v = Q(0, v);
+#else
+#define U2(r, op) \
+        { d = R(r);  x op Q(Z7_CRC_NUM_TABLES_USE / 4 - 1 - (r), d); }
+        UInt32 d, x;
+        U2(1, =)
+#if Z7_CRC_NUM_TABLES_USE >= 3 * 4
+#define U(r)  U2(r, ^=)
+        U(2)
+#if Z7_CRC_NUM_TABLES_USE >= 4 * 4
+        U(3)
+#if Z7_CRC_NUM_TABLES_USE >= 5 * 4
+        U(4)
+#if Z7_CRC_NUM_TABLES_USE >= 6 * 4
+        U(5)
+#if Z7_CRC_NUM_TABLES_USE >= 7 * 4
+#error Stop_Compiling_Bad_Z7_CRC_NUM_TABLES
+#endif
+#endif
+#endif
+#endif
+#endif
+#undef U
+#undef U2
+        v = x ^ Q(Z7_CRC_NUM_TABLES_USE / 4 - 1, v);
+#endif
+      }
+      p += Z7_CRC_NUM_TABLES_USE;
+    }
+    while (p <= lim);
+    lim += Z7_CRC_NUM_TABLES_USE;
   }
-  for (; size > 0; size--, p++)
+  for (; p < lim; p++)
     v = CRC_UPDATE_BYTE_2(v, *p);
   return v;
 }
 
-UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table);
+#undef CRC_UPDATE_BYTE_2
-UInt32 MY_FAST_CALL CrcUpdateT8(UInt32 v, const void *data, size_t size, const UInt32 *table)
+#undef R
-{
+#undef Q
-  const Byte *p = (const Byte *)data;
+#undef CRC_FUNC_PRE_LE
-  for (; size > 0 && ((unsigned)(ptrdiff_t)p & 7) != 0; size--, p++)
+#undef CRC_FUNC_PRE_LE2
-    v = CRC_UPDATE_BYTE_2(v, *p);
-  for (; size >= 8; size -= 8, p += 8)
-  {
-    UInt32 d;
-    v ^= *(const UInt32 *)(const void *)p;
-    v =
-          (table + 0x700)[((v      ) & 0xFF)]
-        ^ (table + 0x600)[((v >>  8) & 0xFF)]
-        ^ (table + 0x500)[((v >> 16) & 0xFF)]
-        ^ (table + 0x400)[((v >> 24))];
-    d = *((const UInt32 *)(const void *)p + 1);
-    v ^=
-          (table + 0x300)[((d      ) & 0xFF)]
-        ^ (table + 0x200)[((d >>  8) & 0xFF)]
-        ^ (table + 0x100)[((d >> 16) & 0xFF)]
-        ^ (table + 0x000)[((d >> 24))];
-  }
-  for (; size > 0; size--, p++)
-    v = CRC_UPDATE_BYTE_2(v, *p);
-  return v;
-}
 
 #endif
 
 
+
+
 #ifndef MY_CPU_LE
 
-#define CRC_UINT32_SWAP(v) ((v >> 24) | ((v >> 8) & 0xFF00) | ((v << 8) & 0xFF0000) | (v << 24))
+#define CRC_UPDATE_BYTE_2_BE(crc, b)  (table[((crc) >> 24) ^ (b)] ^ ((crc) << 8))
 
-#define CRC_UPDATE_BYTE_2_BE(crc, b) (table[(((crc) >> 24) ^ (b))] ^ ((crc) << 8))
+#define Q(n, d) \
+    ( (table + ((n) * 4 + 0) * 0x100)[((d)) & 0xFF] \
+    ^ (table + ((n) * 4 + 1) * 0x100)[((d) >> 1 * 8) & 0xFF] \
+    ^ (table + ((n) * 4 + 2) * 0x100)[((d) >> 2 * 8) & 0xFF] \
+    ^ (table + ((n) * 4 + 3) * 0x100)[((d) >> 3 * 8)] )
 
-UInt32 MY_FAST_CALL CrcUpdateT1_BeT4(UInt32 v, const void *data, size_t size, const UInt32 *table)
+#ifdef Z7_CRC_DEBUG_BE
+  #define R(a)  GetBe32a((const UInt32 *)(const void *)p + (a))
+#else
+  #define R(a)         *((const UInt32 *)(const void *)p + (a))
+#endif
+
+
+#define CRC_FUNC_PRE_BE2(step) \
+UInt32 Z7_FASTCALL CrcUpdateT1_BeT ## step (UInt32 v, const void *data, size_t size, const UInt32 *table)
+
+#define CRC_FUNC_PRE_BE(step)   \
+        CRC_FUNC_PRE_BE2(step); \
+        CRC_FUNC_PRE_BE2(step)
+
+CRC_FUNC_PRE_BE(Z7_CRC_NUM_TABLES_USE)
 {
   const Byte *p = (const Byte *)data;
+  const Byte *lim;
   table += 0x100;
-  v = CRC_UINT32_SWAP(v);
+  v = Z7_BSWAP32(v);
-  for (; size > 0 && ((unsigned)(ptrdiff_t)p & 3) != 0; size--, p++)
+  for (; size && ((unsigned)(ptrdiff_t)p & (7 - (Z7_CRC_NUM_TABLES_USE & 4))) != 0; size--, p++)
     v = CRC_UPDATE_BYTE_2_BE(v, *p);
-  for (; size >= 4; size -= 4, p += 4)
+  lim = p + size;
+  if (size >= Z7_CRC_NUM_TABLES_USE)
   {
-    v ^= *(const UInt32 *)(const void *)p;
+    lim -= Z7_CRC_NUM_TABLES_USE;
-    v =
+    do
-          (table + 0x000)[((v      ) & 0xFF)]
+    {
-        ^ (table + 0x100)[((v >>  8) & 0xFF)]
+      v ^= R(0);
-        ^ (table + 0x200)[((v >> 16) & 0xFF)]
+      {
-        ^ (table + 0x300)[((v >> 24))];
+#if Z7_CRC_NUM_TABLES_USE == 1 * 4
+        v = Q(0, v);
+#else
+#define U2(r, op) \
+        { d = R(r);  x op Q(Z7_CRC_NUM_TABLES_USE / 4 - 1 - (r), d); }
+        UInt32 d, x;
+        U2(1, =)
+#if Z7_CRC_NUM_TABLES_USE >= 3 * 4
+#define U(r)  U2(r, ^=)
+        U(2)
+#if Z7_CRC_NUM_TABLES_USE >= 4 * 4
+        U(3)
+#if Z7_CRC_NUM_TABLES_USE >= 5 * 4
+        U(4)
+#if Z7_CRC_NUM_TABLES_USE >= 6 * 4
+        U(5)
+#if Z7_CRC_NUM_TABLES_USE >= 7 * 4
+#error Stop_Compiling_Bad_Z7_CRC_NUM_TABLES
+#endif
+#endif
+#endif
+#endif
+#endif
+#undef U
+#undef U2
+        v = x ^ Q(Z7_CRC_NUM_TABLES_USE / 4 - 1, v);
+#endif
+      }
+      p += Z7_CRC_NUM_TABLES_USE;
+    }
+    while (p <= lim);
+    lim += Z7_CRC_NUM_TABLES_USE;
   }
-  for (; size > 0; size--, p++)
+  for (; p < lim; p++)
     v = CRC_UPDATE_BYTE_2_BE(v, *p);
-  return CRC_UINT32_SWAP(v);
+  return Z7_BSWAP32(v);
 }
 
-UInt32 MY_FAST_CALL CrcUpdateT1_BeT8(UInt32 v, const void *data, size_t size, const UInt32 *table)
+#undef CRC_UPDATE_BYTE_2_BE
-{
+#undef R
-  const Byte *p = (const Byte *)data;
+#undef Q
-  table += 0x100;
+#undef CRC_FUNC_PRE_BE
-  v = CRC_UINT32_SWAP(v);
+#undef CRC_FUNC_PRE_BE2
-  for (; size > 0 && ((unsigned)(ptrdiff_t)p & 7) != 0; size--, p++)
-    v = CRC_UPDATE_BYTE_2_BE(v, *p);
-  for (; size >= 8; size -= 8, p += 8)
-  {
-    UInt32 d;
-    v ^= *(const UInt32 *)(const void *)p;
-    v =
-          (table + 0x400)[((v      ) & 0xFF)]
-        ^ (table + 0x500)[((v >>  8) & 0xFF)]
-        ^ (table + 0x600)[((v >> 16) & 0xFF)]
-        ^ (table + 0x700)[((v >> 24))];
-    d = *((const UInt32 *)(const void *)p + 1);
-    v ^=
-          (table + 0x000)[((d      ) & 0xFF)]
-        ^ (table + 0x100)[((d >>  8) & 0xFF)]
-        ^ (table + 0x200)[((d >> 16) & 0xFF)]
-        ^ (table + 0x300)[((d >> 24))];
-  }
-  for (; size > 0; size--, p++)
-    v = CRC_UPDATE_BYTE_2_BE(v, *p);
-  return CRC_UINT32_SWAP(v);
-}
 
 #endif
+#undef Z7_CRC_NUM_TABLES_USE
+#endif

C/7zDec.c

@@ -1,11 +1,11 @@
 /* 7zDec.c -- Decoding from 7z folder
-2021-02-09 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include <string.h>
 
-/* #define _7ZIP_PPMD_SUPPPORT */
+/* #define Z7_PPMD_SUPPORT */
 
 #include "7z.h"
 #include "7zCrc.h"
@@ -16,27 +16,50 @@
 #include "Delta.h"
 #include "LzmaDec.h"
 #include "Lzma2Dec.h"
-#ifdef _7ZIP_PPMD_SUPPPORT
+#ifdef Z7_PPMD_SUPPORT
 #include "Ppmd7.h"
 #endif
 
 #define k_Copy 0
-#ifndef _7Z_NO_METHOD_LZMA2
+#ifndef Z7_NO_METHOD_LZMA2
 #define k_LZMA2 0x21
 #endif
 #define k_LZMA  0x30101
 #define k_BCJ2  0x303011B
-#ifndef _7Z_NO_METHODS_FILTERS
+
+#if !defined(Z7_NO_METHODS_FILTERS)
+#define Z7_USE_BRANCH_FILTER
+#endif
+
+#if !defined(Z7_NO_METHODS_FILTERS) || \
+     defined(Z7_USE_NATIVE_BRANCH_FILTER) && defined(MY_CPU_ARM64)
+#define Z7_USE_FILTER_ARM64
+#ifndef Z7_USE_BRANCH_FILTER
+#define Z7_USE_BRANCH_FILTER
+#endif
+#define k_ARM64 0xa
+#endif
+
+#if !defined(Z7_NO_METHODS_FILTERS) || \
+     defined(Z7_USE_NATIVE_BRANCH_FILTER) && defined(MY_CPU_ARMT)
+#define Z7_USE_FILTER_ARMT
+#ifndef Z7_USE_BRANCH_FILTER
+#define Z7_USE_BRANCH_FILTER
+#endif
+#define k_ARMT  0x3030701
+#endif
+
+#ifndef Z7_NO_METHODS_FILTERS
 #define k_Delta 3
+#define k_RISCV 0xb
 #define k_BCJ   0x3030103
 #define k_PPC   0x3030205
 #define k_IA64  0x3030401
 #define k_ARM   0x3030501
-#define k_ARMT  0x3030701
 #define k_SPARC 0x3030805
 #endif
 
-#ifdef _7ZIP_PPMD_SUPPPORT
+#ifdef Z7_PPMD_SUPPORT
 
 #define k_PPMD 0x30401
 
@@ -49,12 +72,12 @@ typedef struct
   UInt64 processed;
   BoolInt extra;
   SRes res;
-  const ILookInStream *inStream;
+  ILookInStreamPtr inStream;
 } CByteInToLook;
 
-static Byte ReadByte(const IByteIn *pp)
+static Byte ReadByte(IByteInPtr pp)
 {
-  CByteInToLook *p = CONTAINER_FROM_VTBL(pp, CByteInToLook, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CByteInToLook)
   if (p->cur != p->end)
     return *p->cur++;
   if (p->res == SZ_OK)
@@ -67,13 +90,13 @@ static Byte ReadByte(const IByteIn *pp)
     p->cur = p->begin;
     p->end = p->begin + size;
     if (size != 0)
-      return *p->cur++;;
+      return *p->cur++;
   }
   p->extra = True;
   return 0;
 }
 
-static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, const ILookInStream *inStream,
+static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStreamPtr inStream,
     Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain)
 {
   CPpmd7 ppmd;
@@ -138,14 +161,14 @@ static SRes SzDecodePpmd(const Byte *props, unsigned propsSize, UInt64 inSize, c
 #endif
 
 
-static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStream *inStream,
+static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStreamPtr inStream,
     Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain)
 {
   CLzmaDec state;
   SRes res = SZ_OK;
 
-  LzmaDec_Construct(&state);
+  LzmaDec_CONSTRUCT(&state)
-  RINOK(LzmaDec_AllocateProbs(&state, props, propsSize, allocMain));
+  RINOK(LzmaDec_AllocateProbs(&state, props, propsSize, allocMain))
   state.dic = outBuffer;
   state.dicBufSize = outSize;
   LzmaDec_Init(&state);
@@ -196,18 +219,18 @@ static SRes SzDecodeLzma(const Byte *props, unsigned propsSize, UInt64 inSize, I
 }
 
 
-#ifndef _7Z_NO_METHOD_LZMA2
+#ifndef Z7_NO_METHOD_LZMA2
 
-static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStream *inStream,
+static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize, ILookInStreamPtr inStream,
     Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain)
 {
   CLzma2Dec state;
   SRes res = SZ_OK;
 
-  Lzma2Dec_Construct(&state);
+  Lzma2Dec_CONSTRUCT(&state)
   if (propsSize != 1)
     return SZ_ERROR_DATA;
-  RINOK(Lzma2Dec_AllocateProbs(&state, props[0], allocMain));
+  RINOK(Lzma2Dec_AllocateProbs(&state, props[0], allocMain))
   state.decoder.dic = outBuffer;
   state.decoder.dicBufSize = outSize;
   Lzma2Dec_Init(&state);
@@ -257,7 +280,7 @@ static SRes SzDecodeLzma2(const Byte *props, unsigned propsSize, UInt64 inSize,
 #endif
 
 
-static SRes SzDecodeCopy(UInt64 inSize, ILookInStream *inStream, Byte *outBuffer)
+static SRes SzDecodeCopy(UInt64 inSize, ILookInStreamPtr inStream, Byte *outBuffer)
 {
   while (inSize > 0)
   {
@@ -265,13 +288,13 @@ static SRes SzDecodeCopy(UInt64 inSize, ILookInStream *inStream, Byte *outBuffer
     size_t curSize = (1 << 18);
     if (curSize > inSize)
       curSize = (size_t)inSize;
-    RINOK(ILookInStream_Look(inStream, &inBuf, &curSize));
+    RINOK(ILookInStream_Look(inStream, &inBuf, &curSize))
     if (curSize == 0)
       return SZ_ERROR_INPUT_EOF;
     memcpy(outBuffer, inBuf, curSize);
     outBuffer += curSize;
     inSize -= curSize;
-    RINOK(ILookInStream_Skip(inStream, curSize));
+    RINOK(ILookInStream_Skip(inStream, curSize))
   }
   return SZ_OK;
 }
@@ -282,15 +305,16 @@ static BoolInt IS_MAIN_METHOD(UInt32 m)
   {
     case k_Copy:
     case k_LZMA:
-    #ifndef _7Z_NO_METHOD_LZMA2
+  #ifndef Z7_NO_METHOD_LZMA2
     case k_LZMA2:
-    #endif
+  #endif
-    #ifdef _7ZIP_PPMD_SUPPPORT
+  #ifdef Z7_PPMD_SUPPORT
     case k_PPMD:
-    #endif
+  #endif
       return True;
+    default:
+      return False;
   }
-  return False;
 }
 
 static BoolInt IS_SUPPORTED_CODER(const CSzCoderInfo *c)
@@ -317,7 +341,7 @@ static SRes CheckSupportedFolder(const CSzFolder *f)
   }
   
   
-  #ifndef _7Z_NO_METHODS_FILTERS
+  #if defined(Z7_USE_BRANCH_FILTER)
 
   if (f->NumCoders == 2)
   {
@@ -333,13 +357,21 @@ static SRes CheckSupportedFolder(const CSzFolder *f)
       return SZ_ERROR_UNSUPPORTED;
     switch ((UInt32)c->MethodID)
     {
+    #if !defined(Z7_NO_METHODS_FILTERS)
       case k_Delta:
       case k_BCJ:
       case k_PPC:
       case k_IA64:
       case k_SPARC:
       case k_ARM:
+      case k_RISCV:
+    #endif
+    #ifdef Z7_USE_FILTER_ARM64
+      case k_ARM64:
+    #endif
+    #ifdef Z7_USE_FILTER_ARMT
       case k_ARMT:
+    #endif
         break;
       default:
         return SZ_ERROR_UNSUPPORTED;
@@ -372,15 +404,16 @@ static SRes CheckSupportedFolder(const CSzFolder *f)
   return SZ_ERROR_UNSUPPORTED;
 }
 
-#ifndef _7Z_NO_METHODS_FILTERS
+
-#define CASE_BRA_CONV(isa) case k_ ## isa: isa ## _Convert(outBuffer, outSize, 0, 0); break;
+
-#endif
+
+
 
 static SRes SzFolder_Decode2(const CSzFolder *folder,
     const Byte *propsData,
     const UInt64 *unpackSizes,
     const UInt64 *packPositions,
-    ILookInStream *inStream, UInt64 startPos,
+    ILookInStreamPtr inStream, UInt64 startPos,
     Byte *outBuffer, SizeT outSize, ISzAllocPtr allocMain,
     Byte *tempBuf[])
 {
@@ -389,7 +422,7 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
   SizeT tempSize3 = 0;
   Byte *tempBuf3 = 0;
 
-  RINOK(CheckSupportedFolder(folder));
+  RINOK(CheckSupportedFolder(folder))
 
   for (ci = 0; ci < folder->NumCoders; ci++)
   {
@@ -404,8 +437,8 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
       SizeT outSizeCur = outSize;
       if (folder->NumCoders == 4)
       {
-        UInt32 indices[] = { 3, 2, 0 };
+        const UInt32 indices[] = { 3, 2, 0 };
-        UInt64 unpackSize = unpackSizes[ci];
+        const UInt64 unpackSize = unpackSizes[ci];
         si = indices[ci];
         if (ci < 2)
         {
@@ -431,37 +464,37 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
       }
       offset = packPositions[si];
       inSize = packPositions[(size_t)si + 1] - offset;
-      RINOK(LookInStream_SeekTo(inStream, startPos + offset));
+      RINOK(LookInStream_SeekTo(inStream, startPos + offset))
 
       if (coder->MethodID == k_Copy)
       {
         if (inSize != outSizeCur) /* check it */
           return SZ_ERROR_DATA;
-        RINOK(SzDecodeCopy(inSize, inStream, outBufCur));
+        RINOK(SzDecodeCopy(inSize, inStream, outBufCur))
       }
       else if (coder->MethodID == k_LZMA)
       {
-        RINOK(SzDecodeLzma(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain));
+        RINOK(SzDecodeLzma(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain))
       }
-      #ifndef _7Z_NO_METHOD_LZMA2
+    #ifndef Z7_NO_METHOD_LZMA2
       else if (coder->MethodID == k_LZMA2)
       {
-        RINOK(SzDecodeLzma2(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain));
+        RINOK(SzDecodeLzma2(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain))
       }
-      #endif
+    #endif
-      #ifdef _7ZIP_PPMD_SUPPPORT
+    #ifdef Z7_PPMD_SUPPORT
       else if (coder->MethodID == k_PPMD)
       {
-        RINOK(SzDecodePpmd(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain));
+        RINOK(SzDecodePpmd(propsData + coder->PropsOffset, coder->PropsSize, inSize, inStream, outBufCur, outSizeCur, allocMain))
       }
-      #endif
+    #endif
       else
         return SZ_ERROR_UNSUPPORTED;
     }
     else if (coder->MethodID == k_BCJ2)
     {
-      UInt64 offset = packPositions[1];
+      const UInt64 offset = packPositions[1];
-      UInt64 s3Size = packPositions[2] - offset;
+      const UInt64 s3Size = packPositions[2] - offset;
       
       if (ci != 3)
         return SZ_ERROR_UNSUPPORTED;
@@ -473,8 +506,8 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
       if (!tempBuf[2] && tempSizes[2] != 0)
         return SZ_ERROR_MEM;
       
-      RINOK(LookInStream_SeekTo(inStream, startPos + offset));
+      RINOK(LookInStream_SeekTo(inStream, startPos + offset))
-      RINOK(SzDecodeCopy(s3Size, inStream, tempBuf[2]));
+      RINOK(SzDecodeCopy(s3Size, inStream, tempBuf[2]))
 
       if ((tempSizes[0] & 3) != 0 ||
           (tempSizes[1] & 3) != 0 ||
@@ -493,26 +526,22 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
         p.destLim = outBuffer + outSize;
         
         Bcj2Dec_Init(&p);
-        RINOK(Bcj2Dec_Decode(&p));
+        RINOK(Bcj2Dec_Decode(&p))
 
         {
           unsigned i;
           for (i = 0; i < 4; i++)
             if (p.bufs[i] != p.lims[i])
               return SZ_ERROR_DATA;
-          
+          if (p.dest != p.destLim || !Bcj2Dec_IsMaybeFinished(&p))
-          if (!Bcj2Dec_IsFinished(&p))
-            return SZ_ERROR_DATA;
-
-          if (p.dest != p.destLim
-             || p.state != BCJ2_STREAM_MAIN)
             return SZ_ERROR_DATA;
         }
       }
     }
-    #ifndef _7Z_NO_METHODS_FILTERS
+#if defined(Z7_USE_BRANCH_FILTER)
     else if (ci == 1)
     {
+#if !defined(Z7_NO_METHODS_FILTERS)
       if (coder->MethodID == k_Delta)
       {
         if (coder->PropsSize != 1)
@@ -522,31 +551,75 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
           Delta_Init(state);
           Delta_Decode(state, (unsigned)(propsData[coder->PropsOffset]) + 1, outBuffer, outSize);
         }
+        continue;
       }
-      else
+#endif
+     
+#ifdef Z7_USE_FILTER_ARM64
+      if (coder->MethodID == k_ARM64)
+      {
+        UInt32 pc = 0;
+        if (coder->PropsSize == 4)
+        {
+          pc = GetUi32(propsData + coder->PropsOffset);
+          if (pc & 3)
+            return SZ_ERROR_UNSUPPORTED;
+        }
+        else if (coder->PropsSize != 0)
+          return SZ_ERROR_UNSUPPORTED;
+        z7_BranchConv_ARM64_Dec(outBuffer, outSize, pc);
+        continue;
+      }
+#endif
+
+#if !defined(Z7_NO_METHODS_FILTERS)
+      if (coder->MethodID == k_RISCV)
+      {
+        UInt32 pc = 0;
+        if (coder->PropsSize == 4)
+        {
+          pc = GetUi32(propsData + coder->PropsOffset);
+          if (pc & 1)
+            return SZ_ERROR_UNSUPPORTED;
+        }
+        else if (coder->PropsSize != 0)
+          return SZ_ERROR_UNSUPPORTED;
+        z7_BranchConv_RISCV_Dec(outBuffer, outSize, pc);
+        continue;
+      }
+#endif
+      
+#if !defined(Z7_NO_METHODS_FILTERS) || defined(Z7_USE_FILTER_ARMT)
       {
         if (coder->PropsSize != 0)
           return SZ_ERROR_UNSUPPORTED;
+       #define CASE_BRA_CONV(isa) case k_ ## isa: Z7_BRANCH_CONV_DEC(isa)(outBuffer, outSize, 0); break; // pc = 0;
         switch (coder->MethodID)
         {
+         #if !defined(Z7_NO_METHODS_FILTERS)
           case k_BCJ:
           {
-            UInt32 state;
+            UInt32 state = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL;
-            x86_Convert_Init(state);
+            z7_BranchConvSt_X86_Dec(outBuffer, outSize, 0, &state); // pc = 0
-            x86_Convert(outBuffer, outSize, 0, &state, 0);
             break;
           }
-          CASE_BRA_CONV(PPC)
+          case k_PPC: Z7_BRANCH_CONV_DEC_2(BranchConv_PPC)(outBuffer, outSize, 0); break; // pc = 0;
+          // CASE_BRA_CONV(PPC)
           CASE_BRA_CONV(IA64)
           CASE_BRA_CONV(SPARC)
           CASE_BRA_CONV(ARM)
+         #endif
+         #if !defined(Z7_NO_METHODS_FILTERS) || defined(Z7_USE_FILTER_ARMT)
           CASE_BRA_CONV(ARMT)
+         #endif
           default:
             return SZ_ERROR_UNSUPPORTED;
         }
+        continue;
       }
-    }
+#endif
-    #endif
+    } // (c == 1)
+#endif // Z7_USE_BRANCH_FILTER
     else
       return SZ_ERROR_UNSUPPORTED;
   }
@@ -556,7 +629,7 @@ static SRes SzFolder_Decode2(const CSzFolder *folder,
 
 
 SRes SzAr_DecodeFolder(const CSzAr *p, UInt32 folderIndex,
-    ILookInStream *inStream, UInt64 startPos,
+    ILookInStreamPtr inStream, UInt64 startPos,
     Byte *outBuffer, size_t outSize,
     ISzAllocPtr allocMain)
 {

C/7zFile.c

@@ -1,5 +1,5 @@
 /* 7zFile.c -- File IO
-2021-04-29 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -268,7 +268,7 @@ WRes File_Write(CSzFile *p, const void *data, size_t *size)
       return errno;
     if (processed == 0)
       break;
-    data = (void *)((Byte *)data + (size_t)processed);
+    data = (const void *)((const Byte *)data + (size_t)processed);
     originalSize -= (size_t)processed;
     *size += (size_t)processed;
   }
@@ -287,7 +287,8 @@ WRes File_Seek(CSzFile *p, Int64 *pos, ESzSeek origin)
   DWORD moveMethod;
   UInt32 low = (UInt32)*pos;
   LONG high = (LONG)((UInt64)*pos >> 16 >> 16); /* for case when UInt64 is 32-bit only */
-  switch (origin)
+  // (int) to eliminate clang warning
+  switch ((int)origin)
   {
     case SZ_SEEK_SET: moveMethod = FILE_BEGIN; break;
     case SZ_SEEK_CUR: moveMethod = FILE_CURRENT; break;
@@ -308,7 +309,7 @@ WRes File_Seek(CSzFile *p, Int64 *pos, ESzSeek origin)
   
   int moveMethod; // = origin;
 
-  switch (origin)
+  switch ((int)origin)
   {
     case SZ_SEEK_SET: moveMethod = SEEK_SET; break;
     case SZ_SEEK_CUR: moveMethod = SEEK_CUR; break;
@@ -387,10 +388,10 @@ WRes File_GetLength(CSzFile *p, UInt64 *length)
 
 /* ---------- FileSeqInStream ---------- */
 
-static SRes FileSeqInStream_Read(const ISeqInStream *pp, void *buf, size_t *size)
+static SRes FileSeqInStream_Read(ISeqInStreamPtr pp, void *buf, size_t *size)
 {
-  CFileSeqInStream *p = CONTAINER_FROM_VTBL(pp, CFileSeqInStream, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileSeqInStream)
-  WRes wres = File_Read(&p->file, buf, size);
+  const WRes wres = File_Read(&p->file, buf, size);
   p->wres = wres;
   return (wres == 0) ? SZ_OK : SZ_ERROR_READ;
 }
@@ -403,18 +404,18 @@ void FileSeqInStream_CreateVTable(CFileSeqInStream *p)
 
 /* ---------- FileInStream ---------- */
 
-static SRes FileInStream_Read(const ISeekInStream *pp, void *buf, size_t *size)
+static SRes FileInStream_Read(ISeekInStreamPtr pp, void *buf, size_t *size)
 {
-  CFileInStream *p = CONTAINER_FROM_VTBL(pp, CFileInStream, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileInStream)
-  WRes wres = File_Read(&p->file, buf, size);
+  const WRes wres = File_Read(&p->file, buf, size);
   p->wres = wres;
   return (wres == 0) ? SZ_OK : SZ_ERROR_READ;
 }
 
-static SRes FileInStream_Seek(const ISeekInStream *pp, Int64 *pos, ESzSeek origin)
+static SRes FileInStream_Seek(ISeekInStreamPtr pp, Int64 *pos, ESzSeek origin)
 {
-  CFileInStream *p = CONTAINER_FROM_VTBL(pp, CFileInStream, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileInStream)
-  WRes wres = File_Seek(&p->file, pos, origin);
+  const WRes wres = File_Seek(&p->file, pos, origin);
   p->wres = wres;
   return (wres == 0) ? SZ_OK : SZ_ERROR_READ;
 }
@@ -428,10 +429,10 @@ void FileInStream_CreateVTable(CFileInStream *p)
 
 /* ---------- FileOutStream ---------- */
 
-static size_t FileOutStream_Write(const ISeqOutStream *pp, const void *data, size_t size)
+static size_t FileOutStream_Write(ISeqOutStreamPtr pp, const void *data, size_t size)
 {
-  CFileOutStream *p = CONTAINER_FROM_VTBL(pp, CFileOutStream, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CFileOutStream)
-  WRes wres = File_Write(&p->file, data, &size);
+  const WRes wres = File_Write(&p->file, data, &size);
   p->wres = wres;
   return size;
 }

C/7zFile.h

@@ -1,8 +1,8 @@
 /* 7zFile.h -- File IO
-2021-02-15 : Igor Pavlov : Public domain */
+2023-03-05 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_FILE_H
+#ifndef ZIP7_INC_FILE_H
-#define __7Z_FILE_H
+#define ZIP7_INC_FILE_H
 
 #ifdef _WIN32
 #define USE_WINDOWS_FILE
@@ -10,7 +10,8 @@
 #endif
 
 #ifdef USE_WINDOWS_FILE
-#include <windows.h>
+#include "7zWindows.h"
+
 #else
 // note: USE_FOPEN mode is limited to 32-bit file size
 // #define USE_FOPEN

C/7zStream.c

@@ -1,5 +1,5 @@
 /* 7zStream.c -- 7z Stream functions
-2021-02-09 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -7,12 +7,33 @@
 
 #include "7zTypes.h"
 
-SRes SeqInStream_Read2(const ISeqInStream *stream, void *buf, size_t size, SRes errorType)
+
+SRes SeqInStream_ReadMax(ISeqInStreamPtr stream, void *buf, size_t *processedSize)
+{
+  size_t size = *processedSize;
+  *processedSize = 0;
+  while (size != 0)
+  {
+    size_t cur = size;
+    const SRes res = ISeqInStream_Read(stream, buf, &cur);
+    *processedSize += cur;
+    buf = (void *)((Byte *)buf + cur);
+    size -= cur;
+    if (res != SZ_OK)
+      return res;
+    if (cur == 0)
+      return SZ_OK;
+  }
+  return SZ_OK;
+}
+
+/*
+SRes SeqInStream_Read2(ISeqInStreamPtr stream, void *buf, size_t size, SRes errorType)
 {
   while (size != 0)
   {
     size_t processed = size;
-    RINOK(ISeqInStream_Read(stream, buf, &processed));
+    RINOK(ISeqInStream_Read(stream, buf, &processed))
     if (processed == 0)
       return errorType;
     buf = (void *)((Byte *)buf + processed);
@@ -21,42 +42,44 @@ SRes SeqInStream_Read2(const ISeqInStream *stream, void *buf, size_t size, SRes
   return SZ_OK;
 }
 
-SRes SeqInStream_Read(const ISeqInStream *stream, void *buf, size_t size)
+SRes SeqInStream_Read(ISeqInStreamPtr stream, void *buf, size_t size)
 {
   return SeqInStream_Read2(stream, buf, size, SZ_ERROR_INPUT_EOF);
 }
+*/
 
-SRes SeqInStream_ReadByte(const ISeqInStream *stream, Byte *buf)
+
+SRes SeqInStream_ReadByte(ISeqInStreamPtr stream, Byte *buf)
 {
   size_t processed = 1;
-  RINOK(ISeqInStream_Read(stream, buf, &processed));
+  RINOK(ISeqInStream_Read(stream, buf, &processed))
   return (processed == 1) ? SZ_OK : SZ_ERROR_INPUT_EOF;
 }
 
 
 
-SRes LookInStream_SeekTo(const ILookInStream *stream, UInt64 offset)
+SRes LookInStream_SeekTo(ILookInStreamPtr stream, UInt64 offset)
 {
   Int64 t = (Int64)offset;
   return ILookInStream_Seek(stream, &t, SZ_SEEK_SET);
 }
 
-SRes LookInStream_LookRead(const ILookInStream *stream, void *buf, size_t *size)
+SRes LookInStream_LookRead(ILookInStreamPtr stream, void *buf, size_t *size)
 {
   const void *lookBuf;
   if (*size == 0)
     return SZ_OK;
-  RINOK(ILookInStream_Look(stream, &lookBuf, size));
+  RINOK(ILookInStream_Look(stream, &lookBuf, size))
   memcpy(buf, lookBuf, *size);
   return ILookInStream_Skip(stream, *size);
 }
 
-SRes LookInStream_Read2(const ILookInStream *stream, void *buf, size_t size, SRes errorType)
+SRes LookInStream_Read2(ILookInStreamPtr stream, void *buf, size_t size, SRes errorType)
 {
   while (size != 0)
   {
     size_t processed = size;
-    RINOK(ILookInStream_Read(stream, buf, &processed));
+    RINOK(ILookInStream_Read(stream, buf, &processed))
     if (processed == 0)
       return errorType;
     buf = (void *)((Byte *)buf + processed);
@@ -65,16 +88,16 @@ SRes LookInStream_Read2(const ILookInStream *stream, void *buf, size_t size, SRe
   return SZ_OK;
 }
 
-SRes LookInStream_Read(const ILookInStream *stream, void *buf, size_t size)
+SRes LookInStream_Read(ILookInStreamPtr stream, void *buf, size_t size)
 {
   return LookInStream_Read2(stream, buf, size, SZ_ERROR_INPUT_EOF);
 }
 
 
 
-#define GET_LookToRead2 CLookToRead2 *p = CONTAINER_FROM_VTBL(pp, CLookToRead2, vt);
+#define GET_LookToRead2  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CLookToRead2)
 
-static SRes LookToRead2_Look_Lookahead(const ILookInStream *pp, const void **buf, size_t *size)
+static SRes LookToRead2_Look_Lookahead(ILookInStreamPtr pp, const void **buf, size_t *size)
 {
   SRes res = SZ_OK;
   GET_LookToRead2
@@ -93,7 +116,7 @@ static SRes LookToRead2_Look_Lookahead(const ILookInStream *pp, const void **buf
   return res;
 }
 
-static SRes LookToRead2_Look_Exact(const ILookInStream *pp, const void **buf, size_t *size)
+static SRes LookToRead2_Look_Exact(ILookInStreamPtr pp, const void **buf, size_t *size)
 {
   SRes res = SZ_OK;
   GET_LookToRead2
@@ -113,14 +136,14 @@ static SRes LookToRead2_Look_Exact(const ILookInStream *pp, const void **buf, si
   return res;
 }
 
-static SRes LookToRead2_Skip(const ILookInStream *pp, size_t offset)
+static SRes LookToRead2_Skip(ILookInStreamPtr pp, size_t offset)
 {
   GET_LookToRead2
   p->pos += offset;
   return SZ_OK;
 }
 
-static SRes LookToRead2_Read(const ILookInStream *pp, void *buf, size_t *size)
+static SRes LookToRead2_Read(ILookInStreamPtr pp, void *buf, size_t *size)
 {
   GET_LookToRead2
   size_t rem = p->size - p->pos;
@@ -134,7 +157,7 @@ static SRes LookToRead2_Read(const ILookInStream *pp, void *buf, size_t *size)
   return SZ_OK;
 }
 
-static SRes LookToRead2_Seek(const ILookInStream *pp, Int64 *pos, ESzSeek origin)
+static SRes LookToRead2_Seek(ILookInStreamPtr pp, Int64 *pos, ESzSeek origin)
 {
   GET_LookToRead2
   p->pos = p->size = 0;
@@ -153,9 +176,9 @@ void LookToRead2_CreateVTable(CLookToRead2 *p, int lookahead)
 
 
 
-static SRes SecToLook_Read(const ISeqInStream *pp, void *buf, size_t *size)
+static SRes SecToLook_Read(ISeqInStreamPtr pp, void *buf, size_t *size)
 {
-  CSecToLook *p = CONTAINER_FROM_VTBL(pp, CSecToLook, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSecToLook)
   return LookInStream_LookRead(p->realStream, buf, size);
 }
 
@@ -164,9 +187,9 @@ void SecToLook_CreateVTable(CSecToLook *p)
   p->vt.Read = SecToLook_Read;
 }
 
-static SRes SecToRead_Read(const ISeqInStream *pp, void *buf, size_t *size)
+static SRes SecToRead_Read(ISeqInStreamPtr pp, void *buf, size_t *size)
 {
-  CSecToRead *p = CONTAINER_FROM_VTBL(pp, CSecToRead, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CSecToRead)
   return ILookInStream_Read(p->realStream, buf, size);
 }
 

C/7zTypes.h

@@ -1,8 +1,8 @@
 /* 7zTypes.h -- Basic types
-2021-12-25 : Igor Pavlov : Public domain */
+2024-01-24 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_TYPES_H
+#ifndef ZIP7_7Z_TYPES_H
-#define __7Z_TYPES_H
+#define ZIP7_7Z_TYPES_H
 
 #ifdef _WIN32
 /* #include <windows.h> */
@@ -52,6 +52,11 @@ typedef int SRes;
     #define MY_ALIGN(n)
   #endif
 #else
+  /*
+  // C11/C++11:
+  #include <stdalign.h>
+  #define MY_ALIGN(n) alignas(n)
+  */
   #define MY_ALIGN(n) __attribute__ ((aligned(n)))
 #endif
 
@@ -62,7 +67,7 @@ typedef int SRes;
 typedef unsigned WRes;
 #define MY_SRes_HRESULT_FROM_WRes(x) HRESULT_FROM_WIN32(x)
 
-// #define MY_HRES_ERROR__INTERNAL_ERROR  MY_SRes_HRESULT_FROM_WRes(ERROR_INTERNAL_ERROR)
+// #define MY_HRES_ERROR_INTERNAL_ERROR  MY_SRes_HRESULT_FROM_WRes(ERROR_INTERNAL_ERROR)
 
 #else // _WIN32
 
@@ -70,13 +75,13 @@ typedef unsigned WRes;
 typedef int WRes;
 
 // (FACILITY_ERRNO = 0x800) is 7zip's FACILITY constant to represent (errno) errors in HRESULT
-#define MY__FACILITY_ERRNO  0x800
+#define MY_FACILITY_ERRNO  0x800
-#define MY__FACILITY_WIN32  7
+#define MY_FACILITY_WIN32  7
-#define MY__FACILITY__WRes  MY__FACILITY_ERRNO
+#define MY_FACILITY_WRes  MY_FACILITY_ERRNO
 
 #define MY_HRESULT_FROM_errno_CONST_ERROR(x) ((HRESULT)( \
           ( (HRESULT)(x) & 0x0000FFFF) \
-          | (MY__FACILITY__WRes << 16)  \
+          | (MY_FACILITY_WRes << 16)  \
           | (HRESULT)0x80000000 ))
 
 #define MY_SRes_HRESULT_FROM_WRes(x) \
@@ -120,23 +125,19 @@ typedef int WRes;
 #define ERROR_INVALID_REPARSE_DATA  ((HRESULT)0x80071128L)
 #define ERROR_REPARSE_TAG_INVALID   ((HRESULT)0x80071129L)
 
-// if (MY__FACILITY__WRes != FACILITY_WIN32),
+// if (MY_FACILITY_WRes != FACILITY_WIN32),
 // we use FACILITY_WIN32 for COM errors:
 #define E_OUTOFMEMORY               ((HRESULT)0x8007000EL)
 #define E_INVALIDARG                ((HRESULT)0x80070057L)
-#define MY__E_ERROR_NEGATIVE_SEEK   ((HRESULT)0x80070083L)
+#define MY_E_ERROR_NEGATIVE_SEEK    ((HRESULT)0x80070083L)
 
 /*
 // we can use FACILITY_ERRNO for some COM errors, that have errno equivalents:
 #define E_OUTOFMEMORY             MY_HRESULT_FROM_errno_CONST_ERROR(ENOMEM)
 #define E_INVALIDARG              MY_HRESULT_FROM_errno_CONST_ERROR(EINVAL)
-#define MY__E_ERROR_NEGATIVE_SEEK MY_HRESULT_FROM_errno_CONST_ERROR(EINVAL)
+#define MY_E_ERROR_NEGATIVE_SEEK  MY_HRESULT_FROM_errno_CONST_ERROR(EINVAL)
 */
 
-// gcc / clang : (sizeof(long) == sizeof(void*)) in 32/64 bits
-typedef          long INT_PTR;
-typedef unsigned long UINT_PTR;
-
 #define TEXT(quote) quote
 
 #define FILE_ATTRIBUTE_READONLY       0x0001
@@ -160,18 +161,18 @@ typedef unsigned long UINT_PTR;
 
 
 #ifndef RINOK
-#define RINOK(x) { int __result__ = (x); if (__result__ != 0) return __result__; }
+#define RINOK(x) { const int _result_ = (x); if (_result_ != 0) return _result_; }
 #endif
 
 #ifndef RINOK_WRes
-#define RINOK_WRes(x) { WRes __result__ = (x); if (__result__ != 0) return __result__; }
+#define RINOK_WRes(x) { const WRes _result_ = (x); if (_result_ != 0) return _result_; }
 #endif
 
 typedef unsigned char Byte;
 typedef short Int16;
 typedef unsigned short UInt16;
 
-#ifdef _LZMA_UINT32_IS_ULONG
+#ifdef Z7_DECL_Int32_AS_long
 typedef long Int32;
 typedef unsigned long UInt32;
 #else
@@ -210,37 +211,51 @@ typedef size_t SIZE_T;
 #endif //  _WIN32
 
 
-#define MY_HRES_ERROR__INTERNAL_ERROR  ((HRESULT)0x8007054FL)
+#define MY_HRES_ERROR_INTERNAL_ERROR  ((HRESULT)0x8007054FL)
 
 
-#ifdef _SZ_NO_INT_64
+#ifdef Z7_DECL_Int64_AS_long
-
-/* define _SZ_NO_INT_64, if your compiler doesn't support 64-bit integers.
-   NOTES: Some code will work incorrectly in that case! */
 
 typedef long Int64;
 typedef unsigned long UInt64;
 
 #else
 
-#if defined(_MSC_VER) || defined(__BORLANDC__)
+#if (defined(_MSC_VER) || defined(__BORLANDC__)) && !defined(__clang__)
 typedef __int64 Int64;
 typedef unsigned __int64 UInt64;
-#define UINT64_CONST(n) n
+#else
+#if defined(__clang__) || defined(__GNUC__)
+#include <stdint.h>
+typedef int64_t Int64;
+typedef uint64_t UInt64;
 #else
 typedef long long int Int64;
 typedef unsigned long long int UInt64;
-#define UINT64_CONST(n) n ## ULL
+// #define UINT64_CONST(n) n ## ULL
+#endif
 #endif
 
 #endif
 
-#ifdef _LZMA_NO_SYSTEM_SIZE_T
+#define UINT64_CONST(n) n
-typedef UInt32 SizeT;
+
+
+#ifdef Z7_DECL_SizeT_AS_unsigned_int
+typedef unsigned int SizeT;
 #else
 typedef size_t SizeT;
 #endif
 
+/*
+#if (defined(_MSC_VER) && _MSC_VER <= 1200)
+typedef size_t MY_uintptr_t;
+#else
+#include <stdint.h>
+typedef uintptr_t MY_uintptr_t;
+#endif
+*/
+
 typedef int BoolInt;
 /* typedef BoolInt Bool; */
 #define True 1
@@ -248,23 +263,23 @@ typedef int BoolInt;
 
 
 #ifdef _WIN32
-#define MY_STD_CALL __stdcall
+#define Z7_STDCALL __stdcall
 #else
-#define MY_STD_CALL
+#define Z7_STDCALL
 #endif
 
 #ifdef _MSC_VER
 
 #if _MSC_VER >= 1300
-#define MY_NO_INLINE __declspec(noinline)
+#define Z7_NO_INLINE __declspec(noinline)
 #else
-#define MY_NO_INLINE
+#define Z7_NO_INLINE
 #endif
 
-#define MY_FORCE_INLINE __forceinline
+#define Z7_FORCE_INLINE __forceinline
 
-#define MY_CDECL __cdecl
+#define Z7_CDECL      __cdecl
-#define MY_FAST_CALL __fastcall
+#define Z7_FASTCALL  __fastcall
 
 #else //  _MSC_VER
 
@@ -272,27 +287,25 @@ typedef int BoolInt;
     || (defined(__clang__) && (__clang_major__ >= 4)) \
     || defined(__INTEL_COMPILER) \
     || defined(__xlC__)
-#define MY_NO_INLINE __attribute__((noinline))
+#define Z7_NO_INLINE      __attribute__((noinline))
-// #define MY_FORCE_INLINE __attribute__((always_inline)) inline
+#define Z7_FORCE_INLINE   __attribute__((always_inline)) inline
 #else
-#define MY_NO_INLINE
+#define Z7_NO_INLINE
+#define Z7_FORCE_INLINE
 #endif
 
-#define MY_FORCE_INLINE
+#define Z7_CDECL
-
-
-#define MY_CDECL
 
 #if  defined(_M_IX86) \
   || defined(__i386__)
-// #define MY_FAST_CALL __attribute__((fastcall))
+// #define Z7_FASTCALL __attribute__((fastcall))
-// #define MY_FAST_CALL __attribute__((cdecl))
+// #define Z7_FASTCALL __attribute__((cdecl))
-#define MY_FAST_CALL
+#define Z7_FASTCALL
 #elif defined(MY_CPU_AMD64)
-// #define MY_FAST_CALL __attribute__((ms_abi))
+// #define Z7_FASTCALL __attribute__((ms_abi))
-#define MY_FAST_CALL
+#define Z7_FASTCALL
 #else
-#define MY_FAST_CALL
+#define Z7_FASTCALL
 #endif
 
 #endif //  _MSC_VER
@@ -300,41 +313,49 @@ typedef int BoolInt;
 
 /* The following interfaces use first parameter as pointer to structure */
 
-typedef struct IByteIn IByteIn;
+// #define Z7_C_IFACE_CONST_QUAL
-struct IByteIn
+#define Z7_C_IFACE_CONST_QUAL const
+
+#define Z7_C_IFACE_DECL(a) \
+  struct a ## _; \
+  typedef Z7_C_IFACE_CONST_QUAL struct a ## _ * a ## Ptr; \
+  typedef struct a ## _ a; \
+  struct a ## _
+
+
+Z7_C_IFACE_DECL (IByteIn)
 {
-  Byte (*Read)(const IByteIn *p); /* reads one byte, returns 0 in case of EOF or error */
+  Byte (*Read)(IByteInPtr p); /* reads one byte, returns 0 in case of EOF or error */
 };
 #define IByteIn_Read(p) (p)->Read(p)
 
 
-typedef struct IByteOut IByteOut;
+Z7_C_IFACE_DECL (IByteOut)
-struct IByteOut
 {
-  void (*Write)(const IByteOut *p, Byte b);
+  void (*Write)(IByteOutPtr p, Byte b);
 };
 #define IByteOut_Write(p, b) (p)->Write(p, b)
 
 
-typedef struct ISeqInStream ISeqInStream;
+Z7_C_IFACE_DECL (ISeqInStream)
-struct ISeqInStream
 {
-  SRes (*Read)(const ISeqInStream *p, void *buf, size_t *size);
+  SRes (*Read)(ISeqInStreamPtr p, void *buf, size_t *size);
     /* if (input(*size) != 0 && output(*size) == 0) means end_of_stream.
        (output(*size) < input(*size)) is allowed */
 };
 #define ISeqInStream_Read(p, buf, size) (p)->Read(p, buf, size)
 
+/* try to read as much as avail in stream and limited by (*processedSize) */
+SRes SeqInStream_ReadMax(ISeqInStreamPtr stream, void *buf, size_t *processedSize);
 /* it can return SZ_ERROR_INPUT_EOF */
-SRes SeqInStream_Read(const ISeqInStream *stream, void *buf, size_t size);
+// SRes SeqInStream_Read(ISeqInStreamPtr stream, void *buf, size_t size);
-SRes SeqInStream_Read2(const ISeqInStream *stream, void *buf, size_t size, SRes errorType);
+// SRes SeqInStream_Read2(ISeqInStreamPtr stream, void *buf, size_t size, SRes errorType);
-SRes SeqInStream_ReadByte(const ISeqInStream *stream, Byte *buf);
+SRes SeqInStream_ReadByte(ISeqInStreamPtr stream, Byte *buf);
 
 
-typedef struct ISeqOutStream ISeqOutStream;
+Z7_C_IFACE_DECL (ISeqOutStream)
-struct ISeqOutStream
 {
-  size_t (*Write)(const ISeqOutStream *p, const void *buf, size_t size);
+  size_t (*Write)(ISeqOutStreamPtr p, const void *buf, size_t size);
     /* Returns: result - the number of actually written bytes.
        (result < size) means error */
 };
@@ -348,29 +369,26 @@ typedef enum
 } ESzSeek;
 
 
-typedef struct ISeekInStream ISeekInStream;
+Z7_C_IFACE_DECL (ISeekInStream)
-struct ISeekInStream
 {
-  SRes (*Read)(const ISeekInStream *p, void *buf, size_t *size);  /* same as ISeqInStream::Read */
+  SRes (*Read)(ISeekInStreamPtr p, void *buf, size_t *size);  /* same as ISeqInStream::Read */
-  SRes (*Seek)(const ISeekInStream *p, Int64 *pos, ESzSeek origin);
+  SRes (*Seek)(ISeekInStreamPtr p, Int64 *pos, ESzSeek origin);
 };
 #define ISeekInStream_Read(p, buf, size)   (p)->Read(p, buf, size)
 #define ISeekInStream_Seek(p, pos, origin) (p)->Seek(p, pos, origin)
 
 
-typedef struct ILookInStream ILookInStream;
+Z7_C_IFACE_DECL (ILookInStream)
-struct ILookInStream
 {
-  SRes (*Look)(const ILookInStream *p, const void **buf, size_t *size);
+  SRes (*Look)(ILookInStreamPtr p, const void **buf, size_t *size);
     /* if (input(*size) != 0 && output(*size) == 0) means end_of_stream.
        (output(*size) > input(*size)) is not allowed
        (output(*size) < input(*size)) is allowed */
-  SRes (*Skip)(const ILookInStream *p, size_t offset);
+  SRes (*Skip)(ILookInStreamPtr p, size_t offset);
     /* offset must be <= output(*size) of Look */
-
+  SRes (*Read)(ILookInStreamPtr p, void *buf, size_t *size);
-  SRes (*Read)(const ILookInStream *p, void *buf, size_t *size);
     /* reads directly (without buffer). It's same as ISeqInStream::Read */
-  SRes (*Seek)(const ILookInStream *p, Int64 *pos, ESzSeek origin);
+  SRes (*Seek)(ILookInStreamPtr p, Int64 *pos, ESzSeek origin);
 };
 
 #define ILookInStream_Look(p, buf, size)   (p)->Look(p, buf, size)
@@ -379,19 +397,18 @@ struct ILookInStream
 #define ILookInStream_Seek(p, pos, origin) (p)->Seek(p, pos, origin)
 
 
-SRes LookInStream_LookRead(const ILookInStream *stream, void *buf, size_t *size);
+SRes LookInStream_LookRead(ILookInStreamPtr stream, void *buf, size_t *size);
-SRes LookInStream_SeekTo(const ILookInStream *stream, UInt64 offset);
+SRes LookInStream_SeekTo(ILookInStreamPtr stream, UInt64 offset);
 
 /* reads via ILookInStream::Read */
-SRes LookInStream_Read2(const ILookInStream *stream, void *buf, size_t size, SRes errorType);
+SRes LookInStream_Read2(ILookInStreamPtr stream, void *buf, size_t size, SRes errorType);
-SRes LookInStream_Read(const ILookInStream *stream, void *buf, size_t size);
+SRes LookInStream_Read(ILookInStreamPtr stream, void *buf, size_t size);
-
 
 
 typedef struct
 {
   ILookInStream vt;
-  const ISeekInStream *realStream;
+  ISeekInStreamPtr realStream;
  
   size_t pos;
   size_t size; /* it's data size */
@@ -403,13 +420,13 @@ typedef struct
 
 void LookToRead2_CreateVTable(CLookToRead2 *p, int lookahead);
 
-#define LookToRead2_Init(p) { (p)->pos = (p)->size = 0; }
+#define LookToRead2_INIT(p) { (p)->pos = (p)->size = 0; }
 
 
 typedef struct
 {
   ISeqInStream vt;
-  const ILookInStream *realStream;
+  ILookInStreamPtr realStream;
 } CSecToLook;
 
 void SecToLook_CreateVTable(CSecToLook *p);
@@ -419,20 +436,19 @@ void SecToLook_CreateVTable(CSecToLook *p);
 typedef struct
 {
   ISeqInStream vt;
-  const ILookInStream *realStream;
+  ILookInStreamPtr realStream;
 } CSecToRead;
 
 void SecToRead_CreateVTable(CSecToRead *p);
 
 
-typedef struct ICompressProgress ICompressProgress;
+Z7_C_IFACE_DECL (ICompressProgress)
-
-struct ICompressProgress
 {
-  SRes (*Progress)(const ICompressProgress *p, UInt64 inSize, UInt64 outSize);
+  SRes (*Progress)(ICompressProgressPtr p, UInt64 inSize, UInt64 outSize);
     /* Returns: result. (result != SZ_OK) means break.
        Value (UInt64)(Int64)-1 for size means unknown value. */
 };
+
 #define ICompressProgress_Progress(p, inSize, outSize) (p)->Progress(p, inSize, outSize)
 
 
@@ -470,13 +486,13 @@ struct ISzAlloc
 
 
 
-#ifndef MY_container_of
+#ifndef Z7_container_of
 
 /*
-#define MY_container_of(ptr, type, m) container_of(ptr, type, m)
+#define Z7_container_of(ptr, type, m) container_of(ptr, type, m)
-#define MY_container_of(ptr, type, m) CONTAINING_RECORD(ptr, type, m)
+#define Z7_container_of(ptr, type, m) CONTAINING_RECORD(ptr, type, m)
-#define MY_container_of(ptr, type, m) ((type *)((char *)(ptr) - offsetof(type, m)))
+#define Z7_container_of(ptr, type, m) ((type *)((char *)(ptr) - offsetof(type, m)))
-#define MY_container_of(ptr, type, m) (&((type *)0)->m == (ptr), ((type *)(((char *)(ptr)) - MY_offsetof(type, m))))
+#define Z7_container_of(ptr, type, m) (&((type *)0)->m == (ptr), ((type *)(((char *)(ptr)) - MY_offsetof(type, m))))
 */
 
 /*
@@ -485,24 +501,64 @@ struct ISzAlloc
     GCC 4.8.1 : classes with non-public variable members"
 */
 
-#define MY_container_of(ptr, type, m) ((type *)(void *)((char *)(void *)(1 ? (ptr) : &((type *)0)->m) - MY_offsetof(type, m)))
+#define Z7_container_of(ptr, type, m) \
+  ((type *)(void *)((char *)(void *) \
+  (1 ? (ptr) : &((type *)NULL)->m) - MY_offsetof(type, m)))
+
+#define Z7_container_of_CONST(ptr, type, m) \
+  ((const type *)(const void *)((const char *)(const void *) \
+  (1 ? (ptr) : &((type *)NULL)->m) - MY_offsetof(type, m)))
+
+/*
+#define Z7_container_of_NON_CONST_FROM_CONST(ptr, type, m) \
+  ((type *)(void *)(const void *)((const char *)(const void *) \
+  (1 ? (ptr) : &((type *)NULL)->m) - MY_offsetof(type, m)))
+*/
 
 #endif
 
-#define CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m) ((type *)(void *)(ptr))
+#define Z7_CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m) ((type *)(void *)(ptr))
 
+// #define Z7_CONTAINER_FROM_VTBL(ptr, type, m) Z7_CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m)
+#define Z7_CONTAINER_FROM_VTBL(ptr, type, m) Z7_container_of(ptr, type, m)
+// #define Z7_CONTAINER_FROM_VTBL(ptr, type, m) Z7_container_of_NON_CONST_FROM_CONST(ptr, type, m)
+
+#define Z7_CONTAINER_FROM_VTBL_CONST(ptr, type, m) Z7_container_of_CONST(ptr, type, m)
+
+#define Z7_CONTAINER_FROM_VTBL_CLS(ptr, type, m) Z7_CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m)
 /*
-#define CONTAINER_FROM_VTBL(ptr, type, m) CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m)
+#define Z7_CONTAINER_FROM_VTBL_CLS(ptr, type, m) Z7_CONTAINER_FROM_VTBL(ptr, type, m)
 */
-#define CONTAINER_FROM_VTBL(ptr, type, m) MY_container_of(ptr, type, m)
+#if defined (__clang__) || defined(__GNUC__)
+#define Z7_DIAGNOSTIC_IGNORE_BEGIN_CAST_QUAL \
+  _Pragma("GCC diagnostic push") \
+  _Pragma("GCC diagnostic ignored \"-Wcast-qual\"")
+#define Z7_DIAGNOSTIC_IGNORE_END_CAST_QUAL \
+  _Pragma("GCC diagnostic pop")
+#else
+#define Z7_DIAGNOSTIC_IGNORE_BEGIN_CAST_QUAL
+#define Z7_DIAGNOSTIC_IGNORE_END_CAST_QUAL
+#endif
 
-#define CONTAINER_FROM_VTBL_CLS(ptr, type, m) CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m)
+#define Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR(ptr, type, m, p) \
-/*
+  Z7_DIAGNOSTIC_IGNORE_BEGIN_CAST_QUAL \
-#define CONTAINER_FROM_VTBL_CLS(ptr, type, m) CONTAINER_FROM_VTBL(ptr, type, m)
+  type *p = Z7_CONTAINER_FROM_VTBL(ptr, type, m); \
-*/
+  Z7_DIAGNOSTIC_IGNORE_END_CAST_QUAL
+
+#define Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(type) \
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR(pp, type, vt, p)
 
 
-#define MY_memset_0_ARRAY(a) memset((a), 0, sizeof(a))
+// #define ZIP7_DECLARE_HANDLE(name)  typedef void *name;
+#define Z7_DECLARE_HANDLE(name)  struct name##_dummy{int unused;}; typedef struct name##_dummy *name;
+
+
+#define Z7_memset_0_ARRAY(a)  memset((a), 0, sizeof(a))
+
+#ifndef Z7_ARRAY_SIZE
+#define Z7_ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0]))
+#endif
+
 
 #ifdef _WIN32
 
@@ -520,6 +576,22 @@ struct ISzAlloc
 
 #endif
 
+#define k_PropVar_TimePrec_0        0
+#define k_PropVar_TimePrec_Unix     1
+#define k_PropVar_TimePrec_DOS      2
+#define k_PropVar_TimePrec_HighPrec 3
+#define k_PropVar_TimePrec_Base     16
+#define k_PropVar_TimePrec_100ns (k_PropVar_TimePrec_Base + 7)
+#define k_PropVar_TimePrec_1ns   (k_PropVar_TimePrec_Base + 9)
+
 EXTERN_C_END
 
 #endif
+
+/*
+#ifndef Z7_ST
+#ifdef _7ZIP_ST
+#define Z7_ST
+#endif
+#endif
+*/

C/7zVersion.h

@@ -1,7 +1,7 @@
-#define MY_VER_MAJOR 21
+#define MY_VER_MAJOR 25
-#define MY_VER_MINOR 07
+#define MY_VER_MINOR 1
 #define MY_VER_BUILD 0
-#define MY_VERSION_NUMBERS "21.07"
+#define MY_VERSION_NUMBERS "25.01"
 #define MY_VERSION MY_VERSION_NUMBERS
 
 #ifdef MY_CPU_NAME
@@ -10,12 +10,12 @@
   #define MY_VERSION_CPU MY_VERSION
 #endif
 
-#define MY_DATE "2021-12-26"
+#define MY_DATE "2025-08-03"
 #undef MY_COPYRIGHT
 #undef MY_VERSION_COPYRIGHT_DATE
 #define MY_AUTHOR_NAME "Igor Pavlov"
 #define MY_COPYRIGHT_PD "Igor Pavlov : Public domain"
-#define MY_COPYRIGHT_CR "Copyright (c) 1999-2021 Igor Pavlov"
+#define MY_COPYRIGHT_CR "Copyright (c) 1999-2025 Igor Pavlov"
 
 #ifdef USE_COPYRIGHT_CR
   #define MY_COPYRIGHT MY_COPYRIGHT_CR

C/7zWindows.h

@@ -0,0 +1,101 @@
+/* 7zWindows.h -- StdAfx
+2023-04-02 : Igor Pavlov : Public domain */
+
+#ifndef ZIP7_INC_7Z_WINDOWS_H
+#define ZIP7_INC_7Z_WINDOWS_H
+
+#ifdef _WIN32
+
+#if defined(__clang__)
+# pragma clang diagnostic push
+#endif
+
+#if defined(_MSC_VER)
+
+#pragma warning(push)
+#pragma warning(disable : 4668) // '_WIN32_WINNT' is not defined as a preprocessor macro, replacing with '0' for '#if/#elif'
+
+#if _MSC_VER == 1900
+// for old kit10 versions
+// #pragma warning(disable : 4255) // winuser.h(13979): warning C4255: 'GetThreadDpiAwarenessContext':
+#endif
+// win10 Windows Kit:
+#endif // _MSC_VER
+
+#if defined(_MSC_VER) && _MSC_VER <= 1200 && !defined(_WIN64)
+// for msvc6 without sdk2003
+#define RPC_NO_WINDOWS_H
+#endif
+
+#if defined(__MINGW32__) || defined(__MINGW64__)
+// #if defined(__GNUC__) && !defined(__clang__)
+#include <windows.h>
+#else
+#include <Windows.h>
+#endif
+// #include <basetsd.h>
+// #include <wtypes.h>
+
+// but if precompiled with clang-cl then we need
+// #include <windows.h>
+#if defined(_MSC_VER)
+#pragma warning(pop)
+#endif
+
+#if defined(__clang__)
+# pragma clang diagnostic pop
+#endif
+
+#if defined(_MSC_VER) && _MSC_VER <= 1200 && !defined(_WIN64)
+#ifndef _W64
+
+typedef long LONG_PTR, *PLONG_PTR;
+typedef unsigned long ULONG_PTR, *PULONG_PTR;
+typedef ULONG_PTR DWORD_PTR, *PDWORD_PTR;
+
+#define Z7_OLD_WIN_SDK
+#endif // _W64
+#endif // _MSC_VER == 1200
+
+#ifdef Z7_OLD_WIN_SDK
+
+#ifndef INVALID_FILE_ATTRIBUTES
+#define INVALID_FILE_ATTRIBUTES ((DWORD)-1)
+#endif
+#ifndef INVALID_SET_FILE_POINTER
+#define INVALID_SET_FILE_POINTER ((DWORD)-1)
+#endif
+#ifndef FILE_SPECIAL_ACCESS
+#define FILE_SPECIAL_ACCESS    (FILE_ANY_ACCESS)
+#endif
+
+// ShlObj.h:
+// #define BIF_NEWDIALOGSTYLE     0x0040
+
+#pragma warning(disable : 4201)
+// #pragma warning(disable : 4115)
+
+#undef  VARIANT_TRUE
+#define VARIANT_TRUE ((VARIANT_BOOL)-1)
+#endif
+
+#endif // Z7_OLD_WIN_SDK
+
+#ifdef UNDER_CE
+#undef  VARIANT_TRUE
+#define VARIANT_TRUE ((VARIANT_BOOL)-1)
+#endif
+
+
+#if defined(_MSC_VER)
+#if _MSC_VER >= 1400 && _MSC_VER <= 1600
+  // BaseTsd.h(148) : 'HandleToULong' : unreferenced inline function has been removed
+  // string.h
+  // #pragma warning(disable : 4514)
+#endif
+#endif
+
+
+/* #include "7zTypes.h" */
+
+#endif

C/7zip_gcc_c.mak

@@ -4,24 +4,57 @@ MY_ARCH_2 = $(MY_ARCH)
 MY_ASM = jwasm
 MY_ASM = asmc
 
+ifndef RC
+#RC=windres.exe --target=pe-x86-64
+#RC=windres.exe -F pe-i386
+RC=windres.exe
+endif
+
 PROGPATH = $(O)/$(PROG)
 PROGPATH_STATIC = $(O)/$(PROG)s
 
+ifneq ($(CC), xlc)
+CFLAGS_WARN_WALL = -Wall -Werror -Wextra
+endif
 
 # for object file
 CFLAGS_BASE_LIST = -c
 # for ASM file
 # CFLAGS_BASE_LIST = -S
-CFLAGS_BASE = $(MY_ARCH_2) -O2 $(CFLAGS_BASE_LIST) -Wall -Werror -Wextra $(CFLAGS_WARN) \
+
+FLAGS_FLTO = -flto
+FLAGS_FLTO =
+
+CFLAGS_BASE = $(MY_ARCH_2) -O2 $(CFLAGS_BASE_LIST) $(CFLAGS_WARN_WALL) $(CFLAGS_WARN) \
  -DNDEBUG -D_REENTRANT -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
 
 
-LDFLAGS_STATIC = -DNDEBUG
-# -static
-
 ifdef SystemDrive
 IS_MINGW = 1
+else
+ifdef SYSTEMDRIVE
+# ifdef OS
+IS_MINGW = 1
 endif
+endif
+
+ifdef IS_MINGW
+LDFLAGS_STATIC_2 = -static
+else
+ifndef DEF_FILE
+ifndef IS_NOT_STANDALONE
+ifndef MY_DYNAMIC_LINK
+ifneq ($(CC), clang)
+LDFLAGS_STATIC_2 =
+# -static
+# -static-libstdc++ -static-libgcc
+endif
+endif
+endif
+endif
+endif
+
+LDFLAGS_STATIC = -DNDEBUG $(LDFLAGS_STATIC_2)
 
 ifdef DEF_FILE
 
@@ -62,22 +95,29 @@ endif
 
 ifdef IS_MINGW
 
+ifdef MSYSTEM
+RM = rm -f
+MY_MKDIR=mkdir -p
+DEL_OBJ_EXE = -$(RM) $(PROGPATH) $(PROGPATH_STATIC) $(OBJS)
+else
 RM = del
 MY_MKDIR=mkdir
-LIB2 = -loleaut32 -luuid -ladvapi32 -lUser32
+DEL_OBJ_EXE = -$(RM) $(O)\*.o $(O)\$(PROG).exe $(O)\$(PROG).dll
+endif
 
 
-CXXFLAGS_EXTRA = -DUNICODE -D_UNICODE
+LIB2 = -lOle32 -loleaut32 -luuid -ladvapi32 -lUser32 -lShell32
+
+CFLAGS_EXTRA = -DUNICODE -D_UNICODE
 # -Wno-delete-non-virtual-dtor
 
-DEL_OBJ_EXE = -$(RM) $(O)\*.o $(O)\$(PROG).exe $(O)\$(PROG).dll
  
 else
 
 RM = rm -f
 MY_MKDIR=mkdir -p
-# CFLAGS_BASE := $(CFLAGS_BASE) -D_7ZIP_ST
+# CFLAGS_BASE := $(CFLAGS_BASE) -DZ7_ST
-# CXXFLAGS_EXTRA = -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
+# CFLAGS_EXTRA = -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
 
 # LOCAL_LIBS=-lpthread
 # LOCAL_LIBS_DLL=$(LOCAL_LIBS) -ldl
@@ -88,10 +128,6 @@ DEL_OBJ_EXE = -$(RM) $(PROGPATH) $(PROGPATH_STATIC) $(OBJS)
 endif
 
 
-
-CFLAGS = $(LOCAL_FLAGS) $(CFLAGS_BASE2) $(CFLAGS_BASE) $(CC_SHARED) -o $@
-
-
 ifdef IS_X64
 AFLAGS_ABI = -elf64 -DABI_LINUX
 else
@@ -102,12 +138,9 @@ AFLAGS_ABI = -elf -DABI_LINUX -DABI_CDECL
 endif
 AFLAGS = $(AFLAGS_ABI) -Fo$(O)/
 
+C_WARN_FLAGS =
 
-CXX_WARN_FLAGS =
+CFLAGS = $(LOCAL_FLAGS) $(CFLAGS_BASE2) $(CFLAGS_BASE) $(CFLAGS_EXTRA) $(C_WARN_FLAGS) $(FLAGS_FLTO) $(CC_SHARED) -o $@
-#-Wno-invalid-offsetof
-#-Wno-reorder
-
-CXXFLAGS = $(LOCAL_FLAGS) $(CXXFLAGS_BASE2) $(CFLAGS_BASE) $(CXXFLAGS_EXTRA) $(CC_SHARED) -o $@ $(CXX_WARN_FLAGS)
 
 STATIC_TARGET=
 ifdef COMPL_STATIC
@@ -120,18 +153,27 @@ all: $(O) $(PROGPATH) $(STATIC_TARGET)
 $(O):
 	$(MY_MKDIR) $(O)
 
-LFLAGS_ALL = -s $(MY_ARCH_2) $(LDFLAGS) $(LD_arch) $(OBJS) $(MY_LIBS) $(LIB2)
+ifneq ($(CC), $(CROSS_COMPILE)clang)
+LFLAGS_STRIP = -s
+endif
+
+LFLAGS_ALL = $(LFLAGS_STRIP) $(MY_ARCH_2) $(LDFLAGS) $(FLAGS_FLTO) $(LD_arch) $(OBJS) $(MY_LIBS) $(LIB2)
 $(PROGPATH): $(OBJS)
-	$(CXX) -o $(PROGPATH) $(LFLAGS_ALL)
+	$(CC) -o $(PROGPATH) $(LFLAGS_ALL)
 
 $(PROGPATH_STATIC): $(OBJS)
-	$(CXX) -static -o $(PROGPATH_STATIC) $(LFLAGS_ALL)
+	$(CC) -static -o $(PROGPATH_STATIC) $(LFLAGS_ALL)
 
 
 ifndef NO_DEFAULT_RES
+# old mingw without -FO
+# windres.exe $(RFLAGS) resource.rc $O/resource.o
 $O/resource.o: resource.rc
-	windres.exe $(RFLAGS) resource.rc $O/resource.o
+	$(RC) $(RFLAGS) resource.rc $(O)/resource.o
 endif
+# windres.exe $(RFLAGS) resource.rc $(O)\resource.o
+# windres.exe $(RFLAGS) resource.rc -FO $(O)/resource.o
+# $(RC) $(RFLAGS) resource.rc -FO $(O)/resource.o
 
 
 
@@ -229,10 +271,18 @@ $O/Sha256.o: ../../../C/Sha256.c
 	$(CC) $(CFLAGS) $<
 $O/Sort.o: ../../../C/Sort.c
 	$(CC) $(CFLAGS) $<
+$O/SwapBytes.o: ../../../C/SwapBytes.c
+	$(CC) $(CFLAGS) $<
 $O/Xz.o: ../../../C/Xz.c
 	$(CC) $(CFLAGS) $<
 $O/XzCrc64.o: ../../../C/XzCrc64.c
 	$(CC) $(CFLAGS) $<
+$O/XzDec.o: ../../../C/XzDec.c
+	$(CC) $(CFLAGS) $<
+$O/XzEnc.o: ../../../C/XzEnc.c
+	$(CC) $(CFLAGS) $<
+$O/XzIn.o: ../../../C/XzIn.c
+	$(CC) $(CFLAGS) $<
 
 
 ifdef USE_ASM
@@ -279,11 +329,11 @@ endif
 
 ifdef IS_ARM64
 $O/LzmaDecOpt.o: ../../../Asm/arm64/LzmaDecOpt.S ../../../Asm/arm64/7zAsm.S
-	$(CC) $(CFLAGS) $<
+	$(CC) $(CFLAGS) $(ASM_FLAGS) $<
 endif
 
 $O/LzmaDec.o: ../../LzmaDec.c
-	$(CC) $(CFLAGS) -D_LZMA_DEC_OPT $<
+	$(CC) $(CFLAGS) -DZ7_LZMA_DEC_OPT $<
 
 else
 
@@ -294,19 +344,16 @@ endif
 
 
 
-$O/XzDec.o: ../../../C/XzDec.c
-	$(CC) $(CFLAGS) $<
-$O/XzEnc.o: ../../../C/XzEnc.c
-	$(CC) $(CFLAGS) $<
-$O/XzIn.o: ../../../C/XzIn.c
-	$(CC) $(CFLAGS) $<
-
-
 $O/7zMain.o: ../../../C/Util/7z/7zMain.c
 	$(CC) $(CFLAGS) $<
+$O/7zipInstall.o: ../../../C/Util/7zipInstall/7zipInstall.c
+	$(CC) $(CFLAGS) $<
+$O/7zipUninstall.o: ../../../C/Util/7zipUninstall/7zipUninstall.c
+	$(CC) $(CFLAGS) $<
 $O/LzmaUtil.o: ../../../C/Util/Lzma/LzmaUtil.c
 	$(CC) $(CFLAGS) $<
-
+$O/XzUtil.o: ../../../C/Util/Xz/XzUtil.c
+	$(CC) $(CFLAGS) $<
 
 
 clean:

C/Aes.c

@@ -1,5 +1,5 @@
 /* Aes.c -- AES encryption / decryption
-2021-05-13 : Igor Pavlov : Public domain */
+2024-03-01 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -7,13 +7,15 @@
 #include "Aes.h"
 
 AES_CODE_FUNC g_AesCbc_Decode;
-#ifndef _SFX
+#ifndef Z7_SFX
 AES_CODE_FUNC g_AesCbc_Encode;
 AES_CODE_FUNC g_AesCtr_Code;
 UInt32 g_Aes_SupportedFunctions_Flags;
 #endif
 
+MY_ALIGN(64)
 static UInt32 T[256 * 4];
+MY_ALIGN(64)
 static const Byte Sbox[256] = {
   0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
   0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
@@ -33,7 +35,9 @@ static const Byte Sbox[256] = {
   0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16};
 
 
+MY_ALIGN(64)
 static UInt32 D[256 * 4];
+MY_ALIGN(64)
 static Byte InvS[256];
 
 #define xtime(x) ((((x) << 1) ^ (((x) & 0x80) != 0 ? 0x1B : 0)) & 0xFF)
@@ -51,32 +55,62 @@ static Byte InvS[256];
 #define DD(x) (D + (x << 8))
 
 
-// #define _SHOW_AES_STATUS
+// #define Z7_SHOW_AES_STATUS
 
 #ifdef MY_CPU_X86_OR_AMD64
-  #define USE_HW_AES
+
-#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE)
+  #if defined(__INTEL_COMPILER)
-  #if defined(__clang__)
+    #if (__INTEL_COMPILER >= 1110)
-    #if (__clang_major__ >= 8) // fix that check
-      #define USE_HW_AES
-    #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 6) // fix that check
       #define USE_HW_AES
+      #if (__INTEL_COMPILER >= 1900)
+        #define USE_HW_VAES
+      #endif
     #endif
+  #elif defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30800) \
+     || defined(Z7_GCC_VERSION)   && (Z7_GCC_VERSION   >= 40400)
+    #define USE_HW_AES
+      #if defined(__clang__) && (__clang_major__ >= 8) \
+          || defined(__GNUC__) && (__GNUC__ >= 8)
+        #define USE_HW_VAES
+      #endif
   #elif defined(_MSC_VER)
-    #if _MSC_VER >= 1910
+    #define USE_HW_AES
+    #define USE_HW_VAES
+  #endif
+
+#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE)
+  
+  #if   defined(__ARM_FEATURE_AES) \
+     || defined(__ARM_FEATURE_CRYPTO)
+    #define USE_HW_AES
+  #else
+    #if  defined(MY_CPU_ARM64) \
+      || defined(__ARM_ARCH) && (__ARM_ARCH >= 4) \
+      || defined(Z7_MSC_VER_ORIGINAL)
+    #if  defined(__ARM_FP) && \
+          (   defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30800) \
+           || defined(__GNUC__) && (__GNUC__ >= 6) \
+          ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1910)
+    #if  defined(MY_CPU_ARM64) \
+      || !defined(Z7_CLANG_VERSION) \
+      || defined(__ARM_NEON) && \
+          (Z7_CLANG_VERSION < 170000 || \
+           Z7_CLANG_VERSION > 170001)
       #define USE_HW_AES
     #endif
+    #endif
+    #endif
   #endif
 #endif
 
 #ifdef USE_HW_AES
-#ifdef _SHOW_AES_STATUS
+// #pragma message("=== Aes.c USE_HW_AES === ")
+#ifdef Z7_SHOW_AES_STATUS
 #include <stdio.h>
-#define _PRF(x) x
+#define PRF(x) x
 #else
-#define _PRF(x)
+#define PRF(x)
 #endif
 #endif
 
@@ -90,23 +124,23 @@ void AesGenTables(void)
   for (i = 0; i < 256; i++)
   {
     {
-      UInt32 a1 = Sbox[i];
+      const UInt32 a1 = Sbox[i];
-      UInt32 a2 = xtime(a1);
+      const UInt32 a2 = xtime(a1);
-      UInt32 a3 = a2 ^ a1;
+      const UInt32 a3 = a2 ^ a1;
       TT(0)[i] = Ui32(a2, a1, a1, a3);
       TT(1)[i] = Ui32(a3, a2, a1, a1);
       TT(2)[i] = Ui32(a1, a3, a2, a1);
       TT(3)[i] = Ui32(a1, a1, a3, a2);
     }
     {
-      UInt32 a1 = InvS[i];
+      const UInt32 a1 = InvS[i];
-      UInt32 a2 = xtime(a1);
+      const UInt32 a2 = xtime(a1);
-      UInt32 a4 = xtime(a2);
+      const UInt32 a4 = xtime(a2);
-      UInt32 a8 = xtime(a4);
+      const UInt32 a8 = xtime(a4);
-      UInt32 a9 = a8 ^ a1;
+      const UInt32 a9 = a8 ^ a1;
-      UInt32 aB = a8 ^ a2 ^ a1;
+      const UInt32 aB = a8 ^ a2 ^ a1;
-      UInt32 aD = a8 ^ a4 ^ a1;
+      const UInt32 aD = a8 ^ a4 ^ a1;
-      UInt32 aE = a8 ^ a4 ^ a2;
+      const UInt32 aE = a8 ^ a4 ^ a2;
       DD(0)[i] = Ui32(aE, a9, aD, aB);
       DD(1)[i] = Ui32(aB, aE, a9, aD);
       DD(2)[i] = Ui32(aD, aB, aE, a9);
@@ -116,7 +150,7 @@ void AesGenTables(void)
   
   {
   AES_CODE_FUNC d = AesCbc_Decode;
-  #ifndef _SFX
+  #ifndef Z7_SFX
   AES_CODE_FUNC e = AesCbc_Encode;
   AES_CODE_FUNC c = AesCtr_Code;
   UInt32 flags = 0;
@@ -126,31 +160,33 @@ void AesGenTables(void)
   if (CPU_IsSupported_AES())
   {
     // #pragma message ("AES HW")
-    _PRF(printf("\n===AES HW\n"));
+    PRF(printf("\n===AES HW\n"));
     d = AesCbc_Decode_HW;
 
-    #ifndef _SFX
+    #ifndef Z7_SFX
     e = AesCbc_Encode_HW;
     c = AesCtr_Code_HW;
     flags = k_Aes_SupportedFunctions_HW;
     #endif
 
     #ifdef MY_CPU_X86_OR_AMD64
+    #ifdef USE_HW_VAES
     if (CPU_IsSupported_VAES_AVX2())
     {
-      _PRF(printf("\n===vaes avx2\n"));
+      PRF(printf("\n===vaes avx2\n"));
       d = AesCbc_Decode_HW_256;
-      #ifndef _SFX
+      #ifndef Z7_SFX
       c = AesCtr_Code_HW_256;
       flags |= k_Aes_SupportedFunctions_HW_256;
       #endif
     }
     #endif
+    #endif
   }
   #endif
 
   g_AesCbc_Decode = d;
-  #ifndef _SFX
+  #ifndef Z7_SFX
   g_AesCbc_Encode = e;
   g_AesCtr_Code = c;
   g_Aes_SupportedFunctions_Flags = flags;
@@ -194,7 +230,7 @@ void AesGenTables(void)
 #define FD(i, x) InvS[gb(x, m[(i - x) & 3])]
 #define FD4(i) dest[i] = Ui32(FD(i, 0), FD(i, 1), FD(i, 2), FD(i, 3)) ^ w[i];
 
-void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize)
+void Z7_FASTCALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize)
 {
   unsigned i, m;
   const UInt32 *wLim;
@@ -230,7 +266,7 @@ void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize)
   while (++w != wLim);
 }
 
-void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize)
+void Z7_FASTCALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize)
 {
   unsigned i, num;
   Aes_SetKey_Enc(w, key, keySize);
@@ -251,7 +287,7 @@ void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize)
   src and dest are pointers to 4 UInt32 words.
   src and dest can point to same block */
 
-// MY_FORCE_INLINE
+// Z7_FORCE_INLINE
 static void Aes_Encode(const UInt32 *w, UInt32 *dest, const UInt32 *src)
 {
   UInt32 s[4];
@@ -265,17 +301,20 @@ static void Aes_Encode(const UInt32 *w, UInt32 *dest, const UInt32 *src)
   w += 4;
   for (;;)
   {
-    HT16(m, s, 0);
+    HT16(m, s, 0)
     if (--numRounds2 == 0)
       break;
-    HT16(s, m, 4);
+    HT16(s, m, 4)
     w += 8;
   }
   w += 4;
-  FT4(0); FT4(1); FT4(2); FT4(3);
+  FT4(0)
+  FT4(1)
+  FT4(2)
+  FT4(3)
 }
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
 static void Aes_Decode(const UInt32 *w, UInt32 *dest, const UInt32 *src)
 {
   UInt32 s[4];
@@ -289,12 +328,15 @@ static void Aes_Decode(const UInt32 *w, UInt32 *dest, const UInt32 *src)
   for (;;)
   {
     w -= 8;
-    HD16(m, s, 4);
+    HD16(m, s, 4)
     if (--numRounds2 == 0)
       break;
-    HD16(s, m, 0);
+    HD16(s, m, 0)
   }
-  FD4(0); FD4(1); FD4(2); FD4(3);
+  FD4(0)
+  FD4(1)
+  FD4(2)
+  FD4(3)
 }
 
 void AesCbc_Init(UInt32 *p, const Byte *iv)
@@ -304,7 +346,7 @@ void AesCbc_Init(UInt32 *p, const Byte *iv)
     p[i] = GetUi32(iv + i * 4);
 }
 
-void MY_FAST_CALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks)
+void Z7_FASTCALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks)
 {
   for (; numBlocks != 0; numBlocks--, data += AES_BLOCK_SIZE)
   {
@@ -315,14 +357,14 @@ void MY_FAST_CALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks)
     
     Aes_Encode(p + 4, p, p);
     
-    SetUi32(data,      p[0]);
+    SetUi32(data,      p[0])
-    SetUi32(data + 4,  p[1]);
+    SetUi32(data + 4,  p[1])
-    SetUi32(data + 8,  p[2]);
+    SetUi32(data + 8,  p[2])
-    SetUi32(data + 12, p[3]);
+    SetUi32(data + 12, p[3])
   }
 }
 
-void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks)
+void Z7_FASTCALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks)
 {
   UInt32 in[4], out[4];
   for (; numBlocks != 0; numBlocks--, data += AES_BLOCK_SIZE)
@@ -334,10 +376,10 @@ void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks)
 
     Aes_Decode(p + 4, out, in);
 
-    SetUi32(data,      p[0] ^ out[0]);
+    SetUi32(data,      p[0] ^ out[0])
-    SetUi32(data + 4,  p[1] ^ out[1]);
+    SetUi32(data + 4,  p[1] ^ out[1])
-    SetUi32(data + 8,  p[2] ^ out[2]);
+    SetUi32(data + 8,  p[2] ^ out[2])
-    SetUi32(data + 12, p[3] ^ out[3]);
+    SetUi32(data + 12, p[3] ^ out[3])
     
     p[0] = in[0];
     p[1] = in[1];
@@ -346,7 +388,7 @@ void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks)
   }
 }
 
-void MY_FAST_CALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks)
+void Z7_FASTCALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks)
 {
   for (; numBlocks != 0; numBlocks--)
   {
@@ -360,7 +402,7 @@ void MY_FAST_CALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks)
     
     for (i = 0; i < 4; i++, data += 4)
     {
-      UInt32 t = temp[i];
+      const UInt32 t = temp[i];
 
       #ifdef MY_CPU_LE_UNALIGN
         *((UInt32 *)(void *)data) ^= t;
@@ -373,3 +415,15 @@ void MY_FAST_CALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks)
     }
   }
 }
+
+#undef xtime
+#undef Ui32
+#undef gb0
+#undef gb1
+#undef gb2
+#undef gb3
+#undef gb
+#undef TT
+#undef DD
+#undef USE_HW_AES
+#undef PRF

C/Aes.h

@@ -1,8 +1,8 @@
 /* Aes.h -- AES encryption / decryption
-2018-04-28 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
-#ifndef __AES_H
+#ifndef ZIP7_INC_AES_H
-#define __AES_H
+#define ZIP7_INC_AES_H
 
 #include "7zTypes.h"
 
@@ -20,19 +20,19 @@ void AesGenTables(void);
 
 /* aes - 16-byte aligned pointer to keyMode+roundKeys sequence */
 /* keySize = 16 or 24 or 32 (bytes) */
-typedef void (MY_FAST_CALL *AES_SET_KEY_FUNC)(UInt32 *aes, const Byte *key, unsigned keySize);
+typedef void (Z7_FASTCALL *AES_SET_KEY_FUNC)(UInt32 *aes, const Byte *key, unsigned keySize);
-void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *aes, const Byte *key, unsigned keySize);
+void Z7_FASTCALL Aes_SetKey_Enc(UInt32 *aes, const Byte *key, unsigned keySize);
-void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *aes, const Byte *key, unsigned keySize);
+void Z7_FASTCALL Aes_SetKey_Dec(UInt32 *aes, const Byte *key, unsigned keySize);
 
 /* ivAes - 16-byte aligned pointer to iv+keyMode+roundKeys sequence: UInt32[AES_NUM_IVMRK_WORDS] */
 void AesCbc_Init(UInt32 *ivAes, const Byte *iv); /* iv size is AES_BLOCK_SIZE */
 
 /* data - 16-byte aligned pointer to data */
 /* numBlocks - the number of 16-byte blocks in data array */
-typedef void (MY_FAST_CALL *AES_CODE_FUNC)(UInt32 *ivAes, Byte *data, size_t numBlocks);
+typedef void (Z7_FASTCALL *AES_CODE_FUNC)(UInt32 *ivAes, Byte *data, size_t numBlocks);
 
 extern AES_CODE_FUNC g_AesCbc_Decode;
-#ifndef _SFX
+#ifndef Z7_SFX
 extern AES_CODE_FUNC g_AesCbc_Encode;
 extern AES_CODE_FUNC g_AesCtr_Code;
 #define k_Aes_SupportedFunctions_HW     (1 << 2)
@@ -41,19 +41,19 @@ extern UInt32 g_Aes_SupportedFunctions_Flags;
 #endif
 
 
-#define DECLARE__AES_CODE_FUNC(funcName) \
+#define Z7_DECLARE_AES_CODE_FUNC(funcName) \
-    void MY_FAST_CALL funcName(UInt32 *ivAes, Byte *data, size_t numBlocks);
+    void Z7_FASTCALL funcName(UInt32 *ivAes, Byte *data, size_t numBlocks);
 
-DECLARE__AES_CODE_FUNC (AesCbc_Encode)
+Z7_DECLARE_AES_CODE_FUNC (AesCbc_Encode)
-DECLARE__AES_CODE_FUNC (AesCbc_Decode)
+Z7_DECLARE_AES_CODE_FUNC (AesCbc_Decode)
-DECLARE__AES_CODE_FUNC (AesCtr_Code)
+Z7_DECLARE_AES_CODE_FUNC (AesCtr_Code)
 
-DECLARE__AES_CODE_FUNC (AesCbc_Encode_HW)
+Z7_DECLARE_AES_CODE_FUNC (AesCbc_Encode_HW)
-DECLARE__AES_CODE_FUNC (AesCbc_Decode_HW)
+Z7_DECLARE_AES_CODE_FUNC (AesCbc_Decode_HW)
-DECLARE__AES_CODE_FUNC (AesCtr_Code_HW)
+Z7_DECLARE_AES_CODE_FUNC (AesCtr_Code_HW)
 
-DECLARE__AES_CODE_FUNC (AesCbc_Decode_HW_256)
+Z7_DECLARE_AES_CODE_FUNC (AesCbc_Decode_HW_256)
-DECLARE__AES_CODE_FUNC (AesCtr_Code_HW_256)
+Z7_DECLARE_AES_CODE_FUNC (AesCtr_Code_HW_256)
 
 EXTERN_C_END
 

C/Alloc.c

@@ -1,38 +1,53 @@
 /* Alloc.c -- Memory allocation functions
-2021-07-13 : Igor Pavlov : Public domain */
+2024-02-18 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
-#include <stdio.h>
-
 #ifdef _WIN32
-#include <Windows.h>
+#include "7zWindows.h"
 #endif
 #include <stdlib.h>
 
 #include "Alloc.h"
 
-/* #define _SZ_ALLOC_DEBUG */
+#if defined(Z7_LARGE_PAGES) && defined(_WIN32) && \
+    (!defined(Z7_WIN32_WINNT_MIN) || Z7_WIN32_WINNT_MIN < 0x0502)  // < Win2003 (xp-64)
+  #define Z7_USE_DYN_GetLargePageMinimum
+#endif
 
-/* use _SZ_ALLOC_DEBUG to debug alloc/free operations */
+// for debug:
-#ifdef _SZ_ALLOC_DEBUG
+#if 0
+#if defined(__CHERI__) && defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 16)
+// #pragma message("=== Z7_ALLOC_NO_OFFSET_ALLOCATOR === ")
+#define Z7_ALLOC_NO_OFFSET_ALLOCATOR
+#endif
+#endif
 
+// #define SZ_ALLOC_DEBUG
+/* #define SZ_ALLOC_DEBUG */
+
+/* use SZ_ALLOC_DEBUG to debug alloc/free operations */
+#ifdef SZ_ALLOC_DEBUG
+
+#include <string.h>
 #include <stdio.h>
-int g_allocCount = 0;
+static int g_allocCount = 0;
-int g_allocCountMid = 0;
+#ifdef _WIN32
-int g_allocCountBig = 0;
+static int g_allocCountMid = 0;
+static int g_allocCountBig = 0;
+#endif
 
 
 #define CONVERT_INT_TO_STR(charType, tempSize) \
-  unsigned char temp[tempSize]; unsigned i = 0; \
+  char temp[tempSize]; unsigned i = 0; \
-  while (val >= 10) { temp[i++] = (unsigned char)('0' + (unsigned)(val % 10)); val /= 10; } \
+  while (val >= 10) { temp[i++] = (char)('0' + (unsigned)(val % 10)); val /= 10; } \
   *s++ = (charType)('0' + (unsigned)val); \
   while (i != 0) { i--; *s++ = temp[i]; } \
   *s = 0;
 
 static void ConvertUInt64ToString(UInt64 val, char *s)
 {
-  CONVERT_INT_TO_STR(char, 24);
+  CONVERT_INT_TO_STR(char, 24)
 }
 
 #define GET_HEX_CHAR(t) ((char)(((t < 10) ? ('0' + t) : ('A' + (t - 10)))))
@@ -77,7 +92,7 @@ static void PrintAligned(const char *s, size_t align)
   Print(s);
 }
 
-static void PrintLn()
+static void PrintLn(void)
 {
   Print("\n");
 }
@@ -89,10 +104,10 @@ static void PrintHex(UInt64 v, size_t align)
   PrintAligned(s, align);
 }
 
-static void PrintDec(UInt64 v, size_t align)
+static void PrintDec(int v, size_t align)
 {
   char s[32];
-  ConvertUInt64ToString(v, s);
+  ConvertUInt64ToString((unsigned)v, s);
   PrintAligned(s, align);
 }
 
@@ -102,12 +117,19 @@ static void PrintAddr(void *p)
 }
 
 
-#define PRINT_ALLOC(name, cnt, size, ptr) \
+#define PRINT_REALLOC(name, cnt, size, ptr) { \
+    Print(name " "); \
+    if (!ptr) PrintDec(cnt++, 10); \
+    PrintHex(size, 10); \
+    PrintAddr(ptr); \
+    PrintLn(); }
+
+#define PRINT_ALLOC(name, cnt, size, ptr) { \
     Print(name " "); \
     PrintDec(cnt++, 10); \
     PrintHex(size, 10); \
     PrintAddr(ptr); \
-    PrintLn();
+    PrintLn(); }
  
 #define PRINT_FREE(name, cnt, ptr) if (ptr) { \
     Print(name " "); \
@@ -117,26 +139,45 @@ static void PrintAddr(void *p)
  
 #else
 
+#ifdef _WIN32
 #define PRINT_ALLOC(name, cnt, size, ptr)
+#endif
 #define PRINT_FREE(name, cnt, ptr)
 #define Print(s)
 #define PrintLn()
+#ifndef Z7_ALLOC_NO_OFFSET_ALLOCATOR
 #define PrintHex(v, align)
+#endif
 #define PrintAddr(p)
 
 #endif
 
 
+/*
+by specification:
+  malloc(non_NULL, 0)   : returns NULL or a unique pointer value that can later be successfully passed to free()
+  realloc(NULL, size)   : the call is equivalent to malloc(size)
+  realloc(non_NULL, 0)  : the call is equivalent to free(ptr)
+
+in main compilers:
+  malloc(0)             : returns non_NULL
+  realloc(NULL,     0)  : returns non_NULL
+  realloc(non_NULL, 0)  : returns NULL
+*/
+
 
 void *MyAlloc(size_t size)
 {
   if (size == 0)
     return NULL;
-  PRINT_ALLOC("Alloc    ", g_allocCount, size, NULL);
+  // PRINT_ALLOC("Alloc    ", g_allocCount, size, NULL)
-  #ifdef _SZ_ALLOC_DEBUG
+  #ifdef SZ_ALLOC_DEBUG
   {
     void *p = malloc(size);
-    // PRINT_ALLOC("Alloc    ", g_allocCount, size, p);
+    if (p)
+    {
+      PRINT_ALLOC("Alloc    ", g_allocCount, size, p)
+    }
     return p;
   }
   #else
@@ -146,71 +187,107 @@ void *MyAlloc(size_t size)
 
 void MyFree(void *address)
 {
-  PRINT_FREE("Free    ", g_allocCount, address);
+  PRINT_FREE("Free    ", g_allocCount, address)
   
   free(address);
 }
 
+void *MyRealloc(void *address, size_t size)
+{
+  if (size == 0)
+  {
+    MyFree(address);
+    return NULL;
+  }
+  // PRINT_REALLOC("Realloc  ", g_allocCount, size, address)
+  #ifdef SZ_ALLOC_DEBUG
+  {
+    void *p = realloc(address, size);
+    if (p)
+    {
+      PRINT_REALLOC("Realloc    ", g_allocCount, size, address)
+    }
+    return p;
+  }
+  #else
+  return realloc(address, size);
+  #endif
+}
+
+
 #ifdef _WIN32
 
 void *MidAlloc(size_t size)
 {
   if (size == 0)
     return NULL;
-  
+  #ifdef SZ_ALLOC_DEBUG
-  PRINT_ALLOC("Alloc-Mid", g_allocCountMid, size, NULL);
+  {
-  
+    void *p = VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
+    if (p)
+    {
+      PRINT_ALLOC("Alloc-Mid", g_allocCountMid, size, p)
+    }
+    return p;
+  }
+  #else
   return VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
+  #endif
 }
 
 void MidFree(void *address)
 {
-  PRINT_FREE("Free-Mid", g_allocCountMid, address);
+  PRINT_FREE("Free-Mid", g_allocCountMid, address)
 
   if (!address)
     return;
   VirtualFree(address, 0, MEM_RELEASE);
 }
 
-#ifdef _7ZIP_LARGE_PAGES
+#ifdef Z7_LARGE_PAGES
 
 #ifdef MEM_LARGE_PAGES
-  #define MY__MEM_LARGE_PAGES  MEM_LARGE_PAGES
+  #define MY_MEM_LARGE_PAGES  MEM_LARGE_PAGES
 #else
-  #define MY__MEM_LARGE_PAGES  0x20000000
+  #define MY_MEM_LARGE_PAGES  0x20000000
 #endif
 
 extern
 SIZE_T g_LargePageSize;
 SIZE_T g_LargePageSize = 0;
-typedef SIZE_T (WINAPI *GetLargePageMinimumP)(VOID);
+typedef SIZE_T (WINAPI *Func_GetLargePageMinimum)(VOID);
 
-#endif // _7ZIP_LARGE_PAGES
+void SetLargePageSize(void)
-
-void SetLargePageSize()
 {
-  #ifdef _7ZIP_LARGE_PAGES
   SIZE_T size;
-  GetLargePageMinimumP largePageMinimum = (GetLargePageMinimumP)
+#ifdef Z7_USE_DYN_GetLargePageMinimum
-        GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), "GetLargePageMinimum");
+Z7_DIAGNOSTIC_IGNORE_CAST_FUNCTION
-  if (!largePageMinimum)
+
+  const
+   Func_GetLargePageMinimum fn =
+  (Func_GetLargePageMinimum) Z7_CAST_FUNC_C GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")),
+       "GetLargePageMinimum");
+  if (!fn)
     return;
-  size = largePageMinimum();
+  size = fn();
+#else
+  size = GetLargePageMinimum();
+#endif
   if (size == 0 || (size & (size - 1)) != 0)
     return;
   g_LargePageSize = size;
-  #endif
 }
 
+#endif // Z7_LARGE_PAGES
 
 void *BigAlloc(size_t size)
 {
   if (size == 0)
     return NULL;
 
-  PRINT_ALLOC("Alloc-Big", g_allocCountBig, size, NULL);
+  PRINT_ALLOC("Alloc-Big", g_allocCountBig, size, NULL)
 
-  #ifdef _7ZIP_LARGE_PAGES
+  #ifdef Z7_LARGE_PAGES
   {
     SIZE_T ps = g_LargePageSize;
     if (ps != 0 && ps <= (1 << 30) && size > (ps / 2))
@@ -220,56 +297,43 @@ void *BigAlloc(size_t size)
       size2 = (size + ps) & ~ps;
       if (size2 >= size)
       {
-        void *res = VirtualAlloc(NULL, size2, MEM_COMMIT | MY__MEM_LARGE_PAGES, PAGE_READWRITE);
+        void *p = VirtualAlloc(NULL, size2, MEM_COMMIT | MY_MEM_LARGE_PAGES, PAGE_READWRITE);
-        if (res)
+        if (p)
-          return res;
+        {
+          PRINT_ALLOC("Alloc-BM ", g_allocCountMid, size2, p)
+          return p;
+        }
       }
     }
   }
   #endif
 
-  return VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
+  return MidAlloc(size);
 }
 
 void BigFree(void *address)
 {
-  PRINT_FREE("Free-Big", g_allocCountBig, address);
+  PRINT_FREE("Free-Big", g_allocCountBig, address)
-  
+  MidFree(address);
-  if (!address)
-    return;
-  VirtualFree(address, 0, MEM_RELEASE);
 }
 
-#endif
+#endif // _WIN32
 
 
-static void *SzAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p); return MyAlloc(size); }
+static void *SzAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p)  return MyAlloc(size); }
-static void SzFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p); MyFree(address); }
+static void SzFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p)  MyFree(address); }
 const ISzAlloc g_Alloc = { SzAlloc, SzFree };
 
 #ifdef _WIN32
-static void *SzMidAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p); return MidAlloc(size); }
+static void *SzMidAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p)  return MidAlloc(size); }
-static void SzMidFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p); MidFree(address); }
+static void SzMidFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p)  MidFree(address); }
-static void *SzBigAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p); return BigAlloc(size); }
+static void *SzBigAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p)  return BigAlloc(size); }
-static void SzBigFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p); BigFree(address); }
+static void SzBigFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p)  BigFree(address); }
 const ISzAlloc g_MidAlloc = { SzMidAlloc, SzMidFree };
 const ISzAlloc g_BigAlloc = { SzBigAlloc, SzBigFree };
 #endif
 
-/*
+#ifndef Z7_ALLOC_NO_OFFSET_ALLOCATOR
-  uintptr_t : <stdint.h> C99 (optional)
-            : unsupported in VS6
-*/
-
-#ifdef _WIN32
-  typedef UINT_PTR UIntPtr;
-#else
-  /*
-  typedef uintptr_t UIntPtr;
-  */
-  typedef ptrdiff_t UIntPtr;
-#endif
-
 
 #define ADJUST_ALLOC_SIZE 0
 /*
@@ -280,14 +344,36 @@ const ISzAlloc g_BigAlloc = { SzBigAlloc, SzBigFree };
      MyAlloc() can return address that is NOT multiple of sizeof(void *).
 */
 
-
 /*
-#define MY_ALIGN_PTR_DOWN(p, align) ((void *)((char *)(p) - ((size_t)(UIntPtr)(p) & ((align) - 1))))
+  uintptr_t : <stdint.h> C99 (optional)
+            : unsupported in VS6
 */
-#define MY_ALIGN_PTR_DOWN(p, align) ((void *)((((UIntPtr)(p)) & ~((UIntPtr)(align) - 1))))
+typedef
+  #ifdef _WIN32
+    UINT_PTR
+  #elif 1
+    uintptr_t
+  #else
+    ptrdiff_t
+  #endif
+    MY_uintptr_t;
 
+#if 0 \
+    || (defined(__CHERI__) \
+    || defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ > 8))
+// for 128-bit pointers (cheri):
+#define MY_ALIGN_PTR_DOWN(p, align)  \
+    ((void *)((char *)(p) - ((size_t)(MY_uintptr_t)(p) & ((align) - 1))))
+#else
+#define MY_ALIGN_PTR_DOWN(p, align) \
+    ((void *)((((MY_uintptr_t)(p)) & ~((MY_uintptr_t)(align) - 1))))
+#endif
 
-#if !defined(_WIN32) && defined(_POSIX_C_SOURCE) && (_POSIX_C_SOURCE >= 200112L)
+#endif
+
+#if !defined(_WIN32) \
+    && (defined(Z7_ALLOC_NO_OFFSET_ALLOCATOR) \
+        || defined(_POSIX_C_SOURCE) && (_POSIX_C_SOURCE >= 200112L))
   #define USE_posix_memalign
 #endif
 
@@ -327,14 +413,13 @@ static int posix_memalign(void **ptr, size_t align, size_t size)
 
 #define ALLOC_ALIGN_SIZE ((size_t)1 << 7)
 
-static void *SzAlignedAlloc(ISzAllocPtr pp, size_t size)
+void *z7_AlignedAlloc(size_t size)
 {
-  #ifndef USE_posix_memalign
+#ifndef USE_posix_memalign
   
   void *p;
   void *pAligned;
   size_t newSize;
-  UNUSED_VAR(pp);
 
   /* also we can allocate additional dummy ALLOC_ALIGN_SIZE bytes after aligned
      block to prevent cache line sharing with another allocated blocks */
@@ -359,10 +444,9 @@ static void *SzAlignedAlloc(ISzAllocPtr pp, size_t size)
 
   return pAligned;
 
-  #else
+#else
 
   void *p;
-  UNUSED_VAR(pp);
   if (posix_memalign(&p, ALLOC_ALIGN_SIZE, size))
     return NULL;
 
@@ -371,19 +455,37 @@ static void *SzAlignedAlloc(ISzAllocPtr pp, size_t size)
 
   return p;
 
-  #endif
+#endif
+}
+
+
+void z7_AlignedFree(void *address)
+{
+#ifndef USE_posix_memalign
+  if (address)
+    MyFree(((void **)address)[-1]);
+#else
+  free(address);
+#endif
+}
+
+
+static void *SzAlignedAlloc(ISzAllocPtr pp, size_t size)
+{
+  UNUSED_VAR(pp)
+  return z7_AlignedAlloc(size);
 }
 
 
 static void SzAlignedFree(ISzAllocPtr pp, void *address)
 {
-  UNUSED_VAR(pp);
+  UNUSED_VAR(pp)
-  #ifndef USE_posix_memalign
+#ifndef USE_posix_memalign
   if (address)
     MyFree(((void **)address)[-1]);
-  #else
+#else
   free(address);
-  #endif
+#endif
 }
 
 
@@ -391,17 +493,45 @@ const ISzAlloc g_AlignedAlloc = { SzAlignedAlloc, SzAlignedFree };
 
 
 
-#define MY_ALIGN_PTR_DOWN_1(p) MY_ALIGN_PTR_DOWN(p, sizeof(void *))
-
 /* we align ptr to support cases where CAlignOffsetAlloc::offset is not multiply of sizeof(void *) */
-#define REAL_BLOCK_PTR_VAR(p) ((void **)MY_ALIGN_PTR_DOWN_1(p))[-1]
+#ifndef Z7_ALLOC_NO_OFFSET_ALLOCATOR
-/*
+#if 1
-#define REAL_BLOCK_PTR_VAR(p) ((void **)(p))[-1]
+  #define MY_ALIGN_PTR_DOWN_1(p)  MY_ALIGN_PTR_DOWN(p, sizeof(void *))
-*/
+  #define REAL_BLOCK_PTR_VAR(p)  ((void **)MY_ALIGN_PTR_DOWN_1(p))[-1]
+#else
+  // we can use this simplified code,
+  // if (CAlignOffsetAlloc::offset == (k * sizeof(void *))
+  #define REAL_BLOCK_PTR_VAR(p)  (((void **)(p))[-1])
+#endif
+#endif
+
+
+#if 0
+#ifndef Z7_ALLOC_NO_OFFSET_ALLOCATOR
+#include <stdio.h>
+static void PrintPtr(const char *s, const void *p)
+{
+  const Byte *p2 = (const Byte *)&p;
+  unsigned i;
+  printf("%s %p ", s, p);
+  for (i = sizeof(p); i != 0;)
+  {
+    i--;
+    printf("%02x", p2[i]);
+  }
+  printf("\n");
+}
+#endif
+#endif
+
 
 static void *AlignOffsetAlloc_Alloc(ISzAllocPtr pp, size_t size)
 {
-  CAlignOffsetAlloc *p = CONTAINER_FROM_VTBL(pp, CAlignOffsetAlloc, vt);
+#if defined(Z7_ALLOC_NO_OFFSET_ALLOCATOR)
+  UNUSED_VAR(pp)
+  return z7_AlignedAlloc(size);
+#else
+  const CAlignOffsetAlloc *p = Z7_CONTAINER_FROM_VTBL_CONST(pp, CAlignOffsetAlloc, vt);
   void *adr;
   void *pAligned;
   size_t newSize;
@@ -429,6 +559,12 @@ static void *AlignOffsetAlloc_Alloc(ISzAllocPtr pp, size_t size)
   pAligned = (char *)MY_ALIGN_PTR_DOWN((char *)adr +
       alignSize - p->offset + extra + ADJUST_ALLOC_SIZE, alignSize) + p->offset;
 
+#if 0
+  printf("\nalignSize = %6x, offset=%6x, size=%8x \n", (unsigned)alignSize, (unsigned)p->offset, (unsigned)size);
+  PrintPtr("base", adr);
+  PrintPtr("alig", pAligned);
+#endif
+
   PrintLn();
   Print("- Aligned: ");
   Print(" size="); PrintHex(size, 8);
@@ -440,19 +576,25 @@ static void *AlignOffsetAlloc_Alloc(ISzAllocPtr pp, size_t size)
   REAL_BLOCK_PTR_VAR(pAligned) = adr;
 
   return pAligned;
+#endif
 }
 
 
 static void AlignOffsetAlloc_Free(ISzAllocPtr pp, void *address)
 {
+#if defined(Z7_ALLOC_NO_OFFSET_ALLOCATOR)
+  UNUSED_VAR(pp)
+  z7_AlignedFree(address);
+#else
   if (address)
   {
-    CAlignOffsetAlloc *p = CONTAINER_FROM_VTBL(pp, CAlignOffsetAlloc, vt);
+    const CAlignOffsetAlloc *p = Z7_CONTAINER_FROM_VTBL_CONST(pp, CAlignOffsetAlloc, vt);
     PrintLn();
     Print("- Aligned Free: ");
     PrintLn();
     ISzAlloc_Free(p->baseAlloc, REAL_BLOCK_PTR_VAR(address));
   }
+#endif
 }
 
 

C/Alloc.h

@@ -1,31 +1,49 @@
 /* Alloc.h -- Memory allocation functions
-2021-07-13 : Igor Pavlov : Public domain */
+2024-01-22 : Igor Pavlov : Public domain */
 
-#ifndef __COMMON_ALLOC_H
+#ifndef ZIP7_INC_ALLOC_H
-#define __COMMON_ALLOC_H
+#define ZIP7_INC_ALLOC_H
 
 #include "7zTypes.h"
 
 EXTERN_C_BEGIN
 
+/*
+  MyFree(NULL)        : is allowed, as free(NULL)
+  MyAlloc(0)          : returns NULL : but malloc(0)        is allowed to return NULL or non_NULL
+  MyRealloc(NULL, 0)  : returns NULL : but realloc(NULL, 0) is allowed to return NULL or non_NULL
+MyRealloc() is similar to realloc() for the following cases:
+  MyRealloc(non_NULL, 0)         : returns NULL and always calls MyFree(ptr)
+  MyRealloc(NULL, non_ZERO)      : returns NULL, if allocation failed
+  MyRealloc(non_NULL, non_ZERO)  : returns NULL, if reallocation failed
+*/
+
 void *MyAlloc(size_t size);
 void MyFree(void *address);
+void *MyRealloc(void *address, size_t size);
+
+void *z7_AlignedAlloc(size_t size);
+void  z7_AlignedFree(void *p);
 
 #ifdef _WIN32
 
+#ifdef Z7_LARGE_PAGES
 void SetLargePageSize(void);
+#endif
 
 void *MidAlloc(size_t size);
 void MidFree(void *address);
 void *BigAlloc(size_t size);
 void BigFree(void *address);
 
+/* #define Z7_BIG_ALLOC_IS_ZERO_FILLED */
+
 #else
 
-#define MidAlloc(size) MyAlloc(size)
+#define MidAlloc(size)    z7_AlignedAlloc(size)
-#define MidFree(address) MyFree(address)
+#define MidFree(address)  z7_AlignedFree(address)
-#define BigAlloc(size) MyAlloc(size)
+#define BigAlloc(size)    z7_AlignedAlloc(size)
-#define BigFree(address) MyFree(address)
+#define BigFree(address)  z7_AlignedFree(address)
 
 #endif
 

C/Asm_c.mak

@@ -0,0 +1,12 @@
+!IFDEF ASM_OBJS
+!IF "$(PLATFORM)" == "arm64"
+$(ASM_OBJS): ../../../Asm/arm64/$(*B).S
+	$(COMPL_ASM_CLANG)
+!ELSEIF "$(PLATFORM)" == "arm"
+$(ASM_OBJS): ../../../Asm/arm/$(*B).asm
+	$(COMPL_ASM)
+!ELSEIF "$(PLATFORM)" != "ia64" && "$(PLATFORM)" != "mips"
+$(ASM_OBJS): ../../../Asm/x86/$(*B).asm
+	$(COMPL_ASM)
+!ENDIF
+!ENDIF

C/Bcj2.c

@@ -1,29 +1,24 @@
 /* Bcj2.c -- BCJ2 Decoder (Converter for x86 code)
-2021-02-09 : Igor Pavlov : Public domain */
+2023-03-01 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include "Bcj2.h"
 #include "CpuArch.h"
 
-#define CProb UInt16
-
 #define kTopValue ((UInt32)1 << 24)
-#define kNumModelBits 11
+#define kNumBitModelTotalBits 11
-#define kBitModelTotal (1 << kNumModelBits)
+#define kBitModelTotal (1 << kNumBitModelTotalBits)
 #define kNumMoveBits 5
 
-#define _IF_BIT_0 ttt = *prob; bound = (p->range >> kNumModelBits) * ttt; if (p->code < bound)
+// UInt32 bcj2_stats[256 + 2][2];
-#define _UPDATE_0 p->range = bound; *prob = (CProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits));
-#define _UPDATE_1 p->range -= bound; p->code -= bound; *prob = (CProb)(ttt - (ttt >> kNumMoveBits));
 
 void Bcj2Dec_Init(CBcj2Dec *p)
 {
   unsigned i;
-
+  p->state = BCJ2_STREAM_RC; // BCJ2_DEC_STATE_OK;
-  p->state = BCJ2_DEC_STATE_OK;
   p->ip = 0;
-  p->temp[3] = 0;
+  p->temp = 0;
   p->range = 0;
   p->code = 0;
   for (i = 0; i < sizeof(p->probs) / sizeof(p->probs[0]); i++)
@@ -32,217 +27,248 @@ void Bcj2Dec_Init(CBcj2Dec *p)
 
 SRes Bcj2Dec_Decode(CBcj2Dec *p)
 {
+  UInt32 v = p->temp;
+  // const Byte *src;
   if (p->range <= 5)
   {
-    p->state = BCJ2_DEC_STATE_OK;
+    UInt32 code = p->code;
+    p->state = BCJ2_DEC_STATE_ERROR; /* for case if we return SZ_ERROR_DATA; */
     for (; p->range != 5; p->range++)
     {
-      if (p->range == 1 && p->code != 0)
+      if (p->range == 1 && code != 0)
         return SZ_ERROR_DATA;
-      
       if (p->bufs[BCJ2_STREAM_RC] == p->lims[BCJ2_STREAM_RC])
       {
         p->state = BCJ2_STREAM_RC;
         return SZ_OK;
       }
-
+      code = (code << 8) | *(p->bufs[BCJ2_STREAM_RC])++;
-      p->code = (p->code << 8) | *(p->bufs[BCJ2_STREAM_RC])++;
+      p->code = code;
     }
-    
+    if (code == 0xffffffff)
-    if (p->code == 0xFFFFFFFF)
       return SZ_ERROR_DATA;
-    
+    p->range = 0xffffffff;
-    p->range = 0xFFFFFFFF;
   }
-  else if (p->state >= BCJ2_DEC_STATE_ORIG_0)
+  // else
   {
-    while (p->state <= BCJ2_DEC_STATE_ORIG_3)
+    unsigned state = p->state;
+    // we check BCJ2_IS_32BIT_STREAM() here instead of check in the main loop
+    if (BCJ2_IS_32BIT_STREAM(state))
+    {
+      const Byte *cur = p->bufs[state];
+      if (cur == p->lims[state])
+        return SZ_OK;
+      p->bufs[state] = cur + 4;
+      {
+        const UInt32 ip = p->ip + 4;
+        v = GetBe32a(cur) - ip;
+        p->ip = ip;
+      }
+      state = BCJ2_DEC_STATE_ORIG_0;
+    }
+    if ((unsigned)(state - BCJ2_DEC_STATE_ORIG_0) < 4)
     {
       Byte *dest = p->dest;
-      if (dest == p->destLim)
+      for (;;)
-        return SZ_OK;
-      *dest = p->temp[(size_t)p->state - BCJ2_DEC_STATE_ORIG_0];
-      p->state++;
-      p->dest = dest + 1;
-    }
-  }
-
-  /*
-  if (BCJ2_IS_32BIT_STREAM(p->state))
-  {
-    const Byte *cur = p->bufs[p->state];
-    if (cur == p->lims[p->state])
-      return SZ_OK;
-    p->bufs[p->state] = cur + 4;
-    
-    {
-      UInt32 val;
-      Byte *dest;
-      SizeT rem;
-      
-      p->ip += 4;
-      val = GetBe32(cur) - p->ip;
-      dest = p->dest;
-      rem = p->destLim - dest;
-      if (rem < 4)
       {
-        SizeT i;
+        if (dest == p->destLim)
-        SetUi32(p->temp, val);
+        {
-        for (i = 0; i < rem; i++)
+          p->state = state;
-          dest[i] = p->temp[i];
+          p->temp = v;
-        p->dest = dest + rem;
+          return SZ_OK;
-        p->state = BCJ2_DEC_STATE_ORIG_0 + (unsigned)rem;
+        }
-        return SZ_OK;
+        *dest++ = (Byte)v;
+        p->dest = dest;
+        if (++state == BCJ2_DEC_STATE_ORIG_3 + 1)
+          break;
+        v >>= 8;
       }
-      SetUi32(dest, val);
-      p->temp[3] = (Byte)(val >> 24);
-      p->dest = dest + 4;
-      p->state = BCJ2_DEC_STATE_OK;
     }
   }
-  */
 
+  // src = p->bufs[BCJ2_STREAM_MAIN];
   for (;;)
   {
+    /*
     if (BCJ2_IS_32BIT_STREAM(p->state))
       p->state = BCJ2_DEC_STATE_OK;
     else
+    */
     {
       if (p->range < kTopValue)
       {
         if (p->bufs[BCJ2_STREAM_RC] == p->lims[BCJ2_STREAM_RC])
         {
           p->state = BCJ2_STREAM_RC;
+          p->temp = v;
           return SZ_OK;
         }
         p->range <<= 8;
         p->code = (p->code << 8) | *(p->bufs[BCJ2_STREAM_RC])++;
       }
-
       {
         const Byte *src = p->bufs[BCJ2_STREAM_MAIN];
         const Byte *srcLim;
-        Byte *dest;
+        Byte *dest = p->dest;
-        SizeT num = (SizeT)(p->lims[BCJ2_STREAM_MAIN] - src);
-        
-        if (num == 0)
         {
-          p->state = BCJ2_STREAM_MAIN;
+          const SizeT rem = (SizeT)(p->lims[BCJ2_STREAM_MAIN] - src);
-          return SZ_OK;
+          SizeT num = (SizeT)(p->destLim - dest);
+          if (num >= rem)
+            num = rem;
+        #define NUM_ITERS 4
+        #if (NUM_ITERS & (NUM_ITERS - 1)) == 0
+          num &= ~((SizeT)NUM_ITERS - 1);   // if (NUM_ITERS == (1 << x))
+        #else
+          num -= num % NUM_ITERS; // if (NUM_ITERS != (1 << x))
+        #endif
+          srcLim = src + num;
         }
 
-        dest = p->dest;
+        #define NUM_SHIFT_BITS  24
-        if (num > (SizeT)(p->destLim - dest))
+        #define ONE_ITER(indx) { \
-        {
+          const unsigned b = src[indx]; \
-          num = (SizeT)(p->destLim - dest);
+          *dest++ = (Byte)b; \
-          if (num == 0)
+          v = (v << NUM_SHIFT_BITS) | b; \
-          {
+          if (((b + (0x100 - 0xe8)) & 0xfe) == 0) break; \
-            p->state = BCJ2_DEC_STATE_ORIG;
+          if (((v - (((UInt32)0x0f << (NUM_SHIFT_BITS)) + 0x80)) & \
-            return SZ_OK;
+              ((((UInt32)1 << (4 + NUM_SHIFT_BITS)) - 0x1) << 4)) == 0) break; \
-          }
+            /* ++dest */; /* v = b; */ }
-        }
           
-        srcLim = src + num;
+        if (src != srcLim)
-
+        for (;;)
-        if (p->temp[3] == 0x0F && (src[0] & 0xF0) == 0x80)
-          *dest = src[0];
-        else for (;;)
         {
-          Byte b = *src;
+            /* The dependency chain of 2-cycle for (v) calculation is not big problem here.
-          *dest = b;
+               But we can remove dependency chain with v = b in the end of loop. */
-          if (b != 0x0F)
+          ONE_ITER(0)
-          {
+          #if (NUM_ITERS > 1)
-            if ((b & 0xFE) == 0xE8)
+            ONE_ITER(1)
-              break;
+          #if (NUM_ITERS > 2)
-            dest++;
+            ONE_ITER(2)
-            if (++src != srcLim)
+          #if (NUM_ITERS > 3)
-              continue;
+            ONE_ITER(3)
+          #if (NUM_ITERS > 4)
+            ONE_ITER(4)
+          #if (NUM_ITERS > 5)
+            ONE_ITER(5)
+          #if (NUM_ITERS > 6)
+            ONE_ITER(6)
+          #if (NUM_ITERS > 7)
+            ONE_ITER(7)
+          #endif
+          #endif
+          #endif
+          #endif
+          #endif
+          #endif
+          #endif
+          
+          src += NUM_ITERS;
+          if (src == srcLim)
             break;
-          }
-          dest++;
-          if (++src == srcLim)
-            break;
-          if ((*src & 0xF0) != 0x80)
-            continue;
-          *dest = *src;
-          break;
         }
 
-        num = (SizeT)(src - p->bufs[BCJ2_STREAM_MAIN]);
-        
         if (src == srcLim)
+      #if (NUM_ITERS > 1)
+        for (;;)
+      #endif
         {
-          p->temp[3] = src[-1];
+        #if (NUM_ITERS > 1)
-          p->bufs[BCJ2_STREAM_MAIN] = src;
+          if (src == p->lims[BCJ2_STREAM_MAIN] || dest == p->destLim)
-          p->ip += (UInt32)num;
+        #endif
-          p->dest += num;
+          {
-          p->state =
+            const SizeT num = (SizeT)(src - p->bufs[BCJ2_STREAM_MAIN]);
-            p->bufs[BCJ2_STREAM_MAIN] ==
+            p->bufs[BCJ2_STREAM_MAIN] = src;
-            p->lims[BCJ2_STREAM_MAIN] ?
+            p->dest = dest;
-              (unsigned)BCJ2_STREAM_MAIN :
+            p->ip += (UInt32)num;
-              (unsigned)BCJ2_DEC_STATE_ORIG;
+            /* state BCJ2_STREAM_MAIN has more priority than BCJ2_STATE_ORIG */
-          return SZ_OK;
+            p->state =
+              src == p->lims[BCJ2_STREAM_MAIN] ?
+                (unsigned)BCJ2_STREAM_MAIN :
+                (unsigned)BCJ2_DEC_STATE_ORIG;
+            p->temp = v;
+            return SZ_OK;
+          }
+        #if (NUM_ITERS > 1)
+          ONE_ITER(0)
+          src++;
+        #endif
         }
 
         {
-          UInt32 bound, ttt;
+          const SizeT num = (SizeT)(dest - p->dest);
-          CProb *prob;
+          p->dest = dest; // p->dest += num;
-          Byte b = src[0];
+          p->bufs[BCJ2_STREAM_MAIN] += num; // = src;
-          Byte prev = (Byte)(num == 0 ? p->temp[3] : src[-1]);
-          
-          p->temp[3] = b;
-          p->bufs[BCJ2_STREAM_MAIN] = src + 1;
-          num++;
           p->ip += (UInt32)num;
-          p->dest += num;
+        }
-          
+        {
-          prob = p->probs + (unsigned)(b == 0xE8 ? 2 + (unsigned)prev : (b == 0xE9 ? 1 : 0));
+          UInt32 bound, ttt;
-          
+          CBcj2Prob *prob; // unsigned index;
-          _IF_BIT_0
+          /*
+          prob = p->probs + (unsigned)((Byte)v == 0xe8 ?
+              2 + (Byte)(v >> 8) :
+              ((v >> 5) & 1));  // ((Byte)v < 0xe8 ? 0 : 1));
+          */
           {
-            _UPDATE_0
+            const unsigned c = ((v + 0x17) >> 6) & 1;
+            prob = p->probs + (unsigned)
+                (((0 - c) & (Byte)(v >> NUM_SHIFT_BITS)) + c + ((v >> 5) & 1));
+                // (Byte)
+                // 8x->0     : e9->1     : xxe8->xx+2
+                // 8x->0x100 : e9->0x101 : xxe8->xx
+                // (((0x100 - (e & ~v)) & (0x100 | (v >> 8))) + (e & v));
+                // (((0x101 + (~e | v)) & (0x100 | (v >> 8))) + (e & v));
+          }
+          ttt = *prob;
+          bound = (p->range >> kNumBitModelTotalBits) * ttt;
+          if (p->code < bound)
+          {
+            // bcj2_stats[prob - p->probs][0]++;
+            p->range = bound;
+            *prob = (CBcj2Prob)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits));
             continue;
           }
-          _UPDATE_1
+          {
-            
+            // bcj2_stats[prob - p->probs][1]++;
+            p->range -= bound;
+            p->code -= bound;
+            *prob = (CBcj2Prob)(ttt - (ttt >> kNumMoveBits));
+          }
         }
       }
     }
-
     {
-      UInt32 val;
+      /* (v == 0xe8 ? 0 : 1) uses setcc instruction with additional zero register usage in x64 MSVC. */
-      unsigned cj = (p->temp[3] == 0xE8) ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP;
+      // const unsigned cj = ((Byte)v == 0xe8) ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP;
+      const unsigned cj = (((v + 0x57) >> 6) & 1) + BCJ2_STREAM_CALL;
       const Byte *cur = p->bufs[cj];
       Byte *dest;
       SizeT rem;
-      
       if (cur == p->lims[cj])
       {
         p->state = cj;
         break;
       }
-      
+      v = GetBe32a(cur);
-      val = GetBe32(cur);
       p->bufs[cj] = cur + 4;
-
+      {
-      p->ip += 4;
+        const UInt32 ip = p->ip + 4;
-      val -= p->ip;
+        v -= ip;
+        p->ip = ip;
+      }
       dest = p->dest;
       rem = (SizeT)(p->destLim - dest);
-      
       if (rem < 4)
       {
-        p->temp[0] = (Byte)val; if (rem > 0) dest[0] = (Byte)val; val >>= 8;
+        if ((unsigned)rem > 0) { dest[0] = (Byte)v;  v >>= 8;
-        p->temp[1] = (Byte)val; if (rem > 1) dest[1] = (Byte)val; val >>= 8;
+        if ((unsigned)rem > 1) { dest[1] = (Byte)v;  v >>= 8;
-        p->temp[2] = (Byte)val; if (rem > 2) dest[2] = (Byte)val; val >>= 8;
+        if ((unsigned)rem > 2) { dest[2] = (Byte)v;  v >>= 8; }}}
-        p->temp[3] = (Byte)val;
+        p->temp = v;
         p->dest = dest + rem;
         p->state = BCJ2_DEC_STATE_ORIG_0 + (unsigned)rem;
         break;
       }
-      
+      SetUi32(dest, v)
-      SetUi32(dest, val);
+      v >>= 24;
-      p->temp[3] = (Byte)(val >> 24);
       p->dest = dest + 4;
     }
   }
@@ -252,6 +278,13 @@ SRes Bcj2Dec_Decode(CBcj2Dec *p)
     p->range <<= 8;
     p->code = (p->code << 8) | *(p->bufs[BCJ2_STREAM_RC])++;
   }
-
   return SZ_OK;
 }
+
+#undef NUM_ITERS
+#undef ONE_ITER
+#undef NUM_SHIFT_BITS
+#undef kTopValue
+#undef kNumBitModelTotalBits
+#undef kBitModelTotal
+#undef kNumMoveBits

C/Bcj2.h

@@ -1,8 +1,8 @@
-/* Bcj2.h -- BCJ2 Converter for x86 code
+/* Bcj2.h -- BCJ2 converter for x86 code (Branch CALL/JUMP variant2)
-2014-11-10 : Igor Pavlov : Public domain */
+2023-03-02 : Igor Pavlov : Public domain */
 
-#ifndef __BCJ2_H
+#ifndef ZIP7_INC_BCJ2_H
-#define __BCJ2_H
+#define ZIP7_INC_BCJ2_H
 
 #include "7zTypes.h"
 
@@ -26,37 +26,68 @@ enum
   BCJ2_DEC_STATE_ORIG_3,
   
   BCJ2_DEC_STATE_ORIG,
-  BCJ2_DEC_STATE_OK
+  BCJ2_DEC_STATE_ERROR     /* after detected data error */
 };
 
 enum
 {
   BCJ2_ENC_STATE_ORIG = BCJ2_NUM_STREAMS,
-  BCJ2_ENC_STATE_OK
+  BCJ2_ENC_STATE_FINISHED  /* it's state after fully encoded stream */
 };
 
 
-#define BCJ2_IS_32BIT_STREAM(s) ((s) == BCJ2_STREAM_CALL || (s) == BCJ2_STREAM_JUMP)
+/* #define BCJ2_IS_32BIT_STREAM(s) ((s) == BCJ2_STREAM_CALL || (s) == BCJ2_STREAM_JUMP) */
+#define BCJ2_IS_32BIT_STREAM(s) ((unsigned)((unsigned)(s) - (unsigned)BCJ2_STREAM_CALL) < 2)
 
 /*
 CBcj2Dec / CBcj2Enc
 bufs sizes:
   BUF_SIZE(n) = lims[n] - bufs[n]
-bufs sizes for BCJ2_STREAM_CALL and BCJ2_STREAM_JUMP must be mutliply of 4:
+bufs sizes for BCJ2_STREAM_CALL and BCJ2_STREAM_JUMP must be multiply of 4:
     (BUF_SIZE(BCJ2_STREAM_CALL) & 3) == 0
     (BUF_SIZE(BCJ2_STREAM_JUMP) & 3) == 0
 */
 
+// typedef UInt32 CBcj2Prob;
+typedef UInt16 CBcj2Prob;
+
+/*
+BCJ2 encoder / decoder internal requirements:
+  - If last bytes of stream contain marker (e8/e8/0f8x), then
+    there is also encoded symbol (0 : no conversion) in RC stream.
+  - One case of overlapped instructions is supported,
+    if last byte of converted instruction is (0f) and next byte is (8x):
+      marker [xx xx xx 0f] 8x
+    then the pair (0f 8x) is treated as marker.
+*/
+
+/* ---------- BCJ2 Decoder ---------- */
+
 /*
 CBcj2Dec:
-dest is allowed to overlap with bufs[BCJ2_STREAM_MAIN], with the following conditions:
+(dest) is allowed to overlap with bufs[BCJ2_STREAM_MAIN], with the following conditions:
   bufs[BCJ2_STREAM_MAIN] >= dest &&
-  bufs[BCJ2_STREAM_MAIN] - dest >= tempReserv +
+  bufs[BCJ2_STREAM_MAIN] - dest >=
         BUF_SIZE(BCJ2_STREAM_CALL) +
         BUF_SIZE(BCJ2_STREAM_JUMP)
-     tempReserv = 0 : for first call of Bcj2Dec_Decode
+  reserve = bufs[BCJ2_STREAM_MAIN] - dest -
-     tempReserv = 4 : for any other calls of Bcj2Dec_Decode
+      ( BUF_SIZE(BCJ2_STREAM_CALL) +
-  overlap with offset = 1 is not allowed
+        BUF_SIZE(BCJ2_STREAM_JUMP) )
+  and additional conditions:
+  if (it's first call of Bcj2Dec_Decode() after Bcj2Dec_Init())
+  {
+    (reserve != 1) : if (ver <  v23.00)
+  }
+  else // if there are more than one calls of Bcj2Dec_Decode() after Bcj2Dec_Init())
+  {
+    (reserve >= 6) : if (ver <  v23.00)
+    (reserve >= 4) : if (ver >= v23.00)
+    We need that (reserve) because after first call of Bcj2Dec_Decode(),
+    CBcj2Dec::temp can contain up to 4 bytes for writing to (dest).
+  }
+  (reserve == 0) is allowed, if we decode full stream via single call of Bcj2Dec_Decode().
+  (reserve == 0) also is allowed in case of multi-call, if we use fixed buffers,
+     and (reserve) is calculated from full (final) sizes of all streams before first call.
 */
 
 typedef struct
@@ -68,22 +99,66 @@ typedef struct
 
   unsigned state; /* BCJ2_STREAM_MAIN has more priority than BCJ2_STATE_ORIG */
 
-  UInt32 ip;
+  UInt32 ip;      /* property of starting base for decoding */
-  Byte temp[4];
+  UInt32 temp;    /* Byte temp[4]; */
   UInt32 range;
   UInt32 code;
-  UInt16 probs[2 + 256];
+  CBcj2Prob probs[2 + 256];
 } CBcj2Dec;
 
+
+/* Note:
+   Bcj2Dec_Init() sets (CBcj2Dec::ip = 0)
+   if (ip != 0) property is required, the caller must set CBcj2Dec::ip after Bcj2Dec_Init()
+*/
 void Bcj2Dec_Init(CBcj2Dec *p);
 
-/* Returns: SZ_OK or SZ_ERROR_DATA */
+
+/* Bcj2Dec_Decode():
+   returns:
+     SZ_OK
+     SZ_ERROR_DATA : if data in 5 starting bytes of BCJ2_STREAM_RC stream are not correct
+*/
 SRes Bcj2Dec_Decode(CBcj2Dec *p);
 
-#define Bcj2Dec_IsFinished(_p_) ((_p_)->code == 0)
+/* To check that decoding was finished you can compare
+   sizes of processed streams with sizes known from another sources.
+   You must do at least one mandatory check from the two following options:
+      - the check for size of processed output (ORIG) stream.
+      - the check for size of processed input  (MAIN) stream.
+   additional optional checks:
+      - the checks for processed sizes of all input streams (MAIN, CALL, JUMP, RC)
+      - the checks Bcj2Dec_IsMaybeFinished*()
+   also before actual decoding you can check that the
+   following condition is met for stream sizes:
+     ( size(ORIG) == size(MAIN) + size(CALL) + size(JUMP) )
+*/
+
+/* (state == BCJ2_STREAM_MAIN) means that decoder is ready for
+      additional input data in BCJ2_STREAM_MAIN stream.
+   Note that (state == BCJ2_STREAM_MAIN) is allowed for non-finished decoding.
+*/
+#define Bcj2Dec_IsMaybeFinished_state_MAIN(_p_) ((_p_)->state == BCJ2_STREAM_MAIN)
+
+/* if the stream decoding was finished correctly, then range decoder
+   part of CBcj2Dec also was finished, and then (CBcj2Dec::code == 0).
+   Note that (CBcj2Dec::code == 0) is allowed for non-finished decoding.
+*/
+#define Bcj2Dec_IsMaybeFinished_code(_p_) ((_p_)->code == 0)
+
+/* use Bcj2Dec_IsMaybeFinished() only as additional check
+    after at least one mandatory check from the two following options:
+      - the check for size of processed output (ORIG) stream.
+      - the check for size of processed input  (MAIN) stream.
+*/
+#define Bcj2Dec_IsMaybeFinished(_p_) ( \
+        Bcj2Dec_IsMaybeFinished_state_MAIN(_p_) && \
+        Bcj2Dec_IsMaybeFinished_code(_p_))
 
 
 
+/* ---------- BCJ2 Encoder ---------- */
+
 typedef enum
 {
   BCJ2_ENC_FINISH_MODE_CONTINUE,
@@ -91,6 +166,91 @@ typedef enum
   BCJ2_ENC_FINISH_MODE_END_STREAM
 } EBcj2Enc_FinishMode;
 
+/*
+  BCJ2_ENC_FINISH_MODE_CONTINUE:
+     process non finished encoding.
+     It notifies the encoder that additional further calls
+     can provide more input data (src) than provided by current call.
+     In  that case the CBcj2Enc encoder still can move (src) pointer
+     up to (srcLim), but CBcj2Enc encoder can store some of the last
+     processed bytes (up to 4 bytes) from src to internal CBcj2Enc::temp[] buffer.
+   at return:
+       (CBcj2Enc::src will point to position that includes
+       processed data and data copied to (temp[]) buffer)
+       That data from (temp[]) buffer will be used in further calls.
+  
+  BCJ2_ENC_FINISH_MODE_END_BLOCK:
+     finish encoding of current block (ended at srcLim) without RC flushing.
+   at return: if (CBcj2Enc::state == BCJ2_ENC_STATE_ORIG) &&
+                  CBcj2Enc::src == CBcj2Enc::srcLim)
+        :  it shows that block encoding was finished. And the encoder is
+           ready for new (src) data or for stream finish operation.
+     finished block means
+     {
+       CBcj2Enc has completed block encoding up to (srcLim).
+       (1 + 4 bytes) or (2 + 4 bytes) CALL/JUMP cortages will
+       not cross block boundary at (srcLim).
+       temporary CBcj2Enc buffer for (ORIG) src data is empty.
+       3 output uncompressed streams (MAIN, CALL, JUMP) were flushed.
+       RC stream was not flushed. And RC stream will cross block boundary.
+     }
+     Note: some possible implementation of BCJ2 encoder could
+     write branch marker (e8/e8/0f8x) in one call of Bcj2Enc_Encode(),
+     and it could calculate symbol for RC in another call of Bcj2Enc_Encode().
+     BCJ2 encoder uses ip/fileIp/fileSize/relatLimit values to calculate RC symbol.
+     And these CBcj2Enc variables can have different values in different Bcj2Enc_Encode() calls.
+     So caller must finish each block with BCJ2_ENC_FINISH_MODE_END_BLOCK
+     to ensure that RC symbol is calculated and written in proper block.
+    
+  BCJ2_ENC_FINISH_MODE_END_STREAM
+     finish encoding of stream (ended at srcLim) fully including RC flushing.
+   at return: if (CBcj2Enc::state == BCJ2_ENC_STATE_FINISHED)
+        : it shows that stream encoding was finished fully,
+          and all output streams were flushed fully.
+     also Bcj2Enc_IsFinished() can be called.
+*/
+
+
+/*
+  32-bit relative offset in JUMP/CALL commands is
+    - (mod 4 GiB)  for 32-bit x86 code
+    - signed Int32 for 64-bit x86-64 code
+  BCJ2 encoder also does internal relative to absolute address conversions.
+  And there are 2 possible ways to do it:
+    before v23: we used 32-bit variables and (mod 4 GiB) conversion
+    since  v23: we use  64-bit variables and (signed Int32 offset) conversion.
+  The absolute address condition for conversion in v23:
+    ((UInt64)((Int64)ip64 - (Int64)fileIp64 + 5 + (Int32)offset) < (UInt64)fileSize64)
+  note that if (fileSize64 > 2 GiB). there is difference between
+  old (mod 4 GiB) way (v22) and new (signed Int32 offset) way (v23).
+  And new (v23) way is more suitable to encode 64-bit x86-64 code for (fileSize64 > 2 GiB) cases.
+*/
+
+/*
+// for old (v22) way for conversion:
+typedef UInt32 CBcj2Enc_ip_unsigned;
+typedef  Int32 CBcj2Enc_ip_signed;
+#define BCJ2_ENC_FileSize_MAX ((UInt32)1 << 31)
+*/
+typedef UInt64 CBcj2Enc_ip_unsigned;
+typedef  Int64 CBcj2Enc_ip_signed;
+
+/* maximum size of file that can be used for conversion condition */
+#define BCJ2_ENC_FileSize_MAX             ((CBcj2Enc_ip_unsigned)0 - 2)
+
+/* default value of fileSize64_minus1 variable that means
+   that absolute address limitation will not be used */
+#define BCJ2_ENC_FileSizeField_UNLIMITED  ((CBcj2Enc_ip_unsigned)0 - 1)
+
+/* calculate value that later can be set to CBcj2Enc::fileSize64_minus1 */
+#define BCJ2_ENC_GET_FileSizeField_VAL_FROM_FileSize(fileSize) \
+    ((CBcj2Enc_ip_unsigned)(fileSize) - 1)
+
+/* set CBcj2Enc::fileSize64_minus1 variable from size of file */
+#define Bcj2Enc_SET_FileSize(p, fileSize) \
+    (p)->fileSize64_minus1 = BCJ2_ENC_GET_FileSizeField_VAL_FROM_FileSize(fileSize);
+
+
 typedef struct
 {
   Byte *bufs[BCJ2_NUM_STREAMS];
@@ -101,45 +261,71 @@ typedef struct
   unsigned state;
   EBcj2Enc_FinishMode finishMode;
 
-  Byte prevByte;
+  Byte context;
+  Byte flushRem;
+  Byte isFlushState;
 
   Byte cache;
   UInt32 range;
   UInt64 low;
   UInt64 cacheSize;
   
-  UInt32 ip;
+  // UInt32 context;  // for marker version, it can include marker flag.
 
-  /* 32-bit ralative offset in JUMP/CALL commands is
+  /* (ip64) and (fileIp64) correspond to virtual source stream position
-       - (mod 4 GB)   in 32-bit mode
+     that doesn't include data in temp[] */
-       - signed Int32 in 64-bit mode
+  CBcj2Enc_ip_unsigned ip64;         /* current (ip) position */
-     We use (mod 4 GB) check for fileSize.
+  CBcj2Enc_ip_unsigned fileIp64;     /* start (ip) position of current file */
-     Use fileSize up to 2 GB, if you want to support 32-bit and 64-bit code conversion. */
+  CBcj2Enc_ip_unsigned fileSize64_minus1;   /* size of current file (for conversion limitation) */
-  UInt32 fileIp;
+  UInt32 relatLimit;  /* (relatLimit <= ((UInt32)1 << 31)) : 0 means disable_conversion */
-  UInt32 fileSize;    /* (fileSize <= ((UInt32)1 << 31)), 0 means no_limit */
+  // UInt32 relatExcludeBits;
-  UInt32 relatLimit;  /* (relatLimit <= ((UInt32)1 << 31)), 0 means desable_conversion */
 
   UInt32 tempTarget;
-  unsigned tempPos;
+  unsigned tempPos; /* the number of bytes that were copied to temp[] buffer
-  Byte temp[4 * 2];
+                       (tempPos <= 4) outside of Bcj2Enc_Encode() */
-
+  // Byte temp[4]; // for marker version
-  unsigned flushPos;
+  Byte temp[8];
-  
+  CBcj2Prob probs[2 + 256];
-  UInt16 probs[2 + 256];
 } CBcj2Enc;
 
 void Bcj2Enc_Init(CBcj2Enc *p);
+
+
+/*
+Bcj2Enc_Encode(): at exit:
+  p->State <  BCJ2_NUM_STREAMS    : we need more buffer space for output stream
+                                    (bufs[p->State] == lims[p->State])
+  p->State == BCJ2_ENC_STATE_ORIG : we need more data in input src stream
+                                    (src == srcLim)
+  p->State == BCJ2_ENC_STATE_FINISHED : after fully encoded stream
+*/
 void Bcj2Enc_Encode(CBcj2Enc *p);
 
-#define Bcj2Enc_Get_InputData_Size(p) ((SizeT)((p)->srcLim - (p)->src) + (p)->tempPos)
+/* Bcj2Enc encoder can look ahead for up 4 bytes of source stream.
-#define Bcj2Enc_IsFinished(p) ((p)->flushPos == 5)
+   CBcj2Enc::tempPos : is the number of bytes that were copied from input stream to temp[] buffer.
+   (CBcj2Enc::src) after Bcj2Enc_Encode() is starting position after
+   fully processed data and after data copied to temp buffer.
+   So if the caller needs to get real number of fully processed input
+   bytes (without look ahead data in temp buffer),
+   the caller must subtruct (CBcj2Enc::tempPos) value from processed size
+   value that is calculated based on current (CBcj2Enc::src):
+     cur_processed_pos = Calc_Big_Processed_Pos(enc.src)) -
+        Bcj2Enc_Get_AvailInputSize_in_Temp(&enc);
+*/
+/* get the size of input data that was stored in temp[] buffer: */
+#define Bcj2Enc_Get_AvailInputSize_in_Temp(p) ((p)->tempPos)
 
+#define Bcj2Enc_IsFinished(p) ((p)->flushRem == 0)
 
-#define BCJ2_RELAT_LIMIT_NUM_BITS 26
+/* Note : the decoder supports overlapping of marker (0f 80).
-#define BCJ2_RELAT_LIMIT ((UInt32)1 << BCJ2_RELAT_LIMIT_NUM_BITS)
+   But we can eliminate such overlapping cases by setting
-
+   the limit for relative offset conversion as
-/* limit for CBcj2Enc::fileSize variable */
+     CBcj2Enc::relatLimit <= (0x0f << 24) == (240 MiB)
-#define BCJ2_FileSize_MAX ((UInt32)1 << 31)
+*/
+/* default value for CBcj2Enc::relatLimit */
+#define BCJ2_ENC_RELAT_LIMIT_DEFAULT  ((UInt32)0x0f << 24)
+#define BCJ2_ENC_RELAT_LIMIT_MAX      ((UInt32)1 << 31)
+// #define BCJ2_RELAT_EXCLUDE_NUM_BITS 5
 
 EXTERN_C_END
 

C/Bcj2Enc.c

@@ -1,60 +1,62 @@
-/* Bcj2Enc.c -- BCJ2 Encoder (Converter for x86 code)
+/* Bcj2Enc.c -- BCJ2 Encoder converter for x86 code (Branch CALL/JUMP variant2)
-2021-02-09 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 /* #define SHOW_STAT */
-
 #ifdef SHOW_STAT
 #include <stdio.h>
-#define PRF(x) x
+#define PRF2(s) printf("%s ip=%8x  tempPos=%d  src= %8x\n", s, (unsigned)p->ip64, p->tempPos, (unsigned)(p->srcLim - p->src));
 #else
-#define PRF(x)
+#define PRF2(s)
 #endif
 
-#include <string.h>
-
 #include "Bcj2.h"
 #include "CpuArch.h"
 
-#define CProb UInt16
-
 #define kTopValue ((UInt32)1 << 24)
-#define kNumModelBits 11
+#define kNumBitModelTotalBits 11
-#define kBitModelTotal (1 << kNumModelBits)
+#define kBitModelTotal (1 << kNumBitModelTotalBits)
 #define kNumMoveBits 5
 
 void Bcj2Enc_Init(CBcj2Enc *p)
 {
   unsigned i;
-
+  p->state = BCJ2_ENC_STATE_ORIG;
-  p->state = BCJ2_ENC_STATE_OK;
   p->finishMode = BCJ2_ENC_FINISH_MODE_CONTINUE;
-
+  p->context = 0;
-  p->prevByte = 0;
+  p->flushRem = 5;
-
+  p->isFlushState = 0;
   p->cache = 0;
-  p->range = 0xFFFFFFFF;
+  p->range = 0xffffffff;
   p->low = 0;
   p->cacheSize = 1;
-
+  p->ip64 = 0;
-  p->ip = 0;
+  p->fileIp64 = 0;
-
+  p->fileSize64_minus1 = BCJ2_ENC_FileSizeField_UNLIMITED;
-  p->fileIp = 0;
+  p->relatLimit = BCJ2_ENC_RELAT_LIMIT_DEFAULT;
-  p->fileSize = 0;
+  // p->relatExcludeBits = 0;
-  p->relatLimit = BCJ2_RELAT_LIMIT;
-
   p->tempPos = 0;
-
-  p->flushPos = 0;
-
   for (i = 0; i < sizeof(p->probs) / sizeof(p->probs[0]); i++)
     p->probs[i] = kBitModelTotal >> 1;
 }
 
-static BoolInt MY_FAST_CALL RangeEnc_ShiftLow(CBcj2Enc *p)
+// Z7_NO_INLINE
+Z7_FORCE_INLINE
+static BoolInt Bcj2_RangeEnc_ShiftLow(CBcj2Enc *p)
 {
-  if ((UInt32)p->low < (UInt32)0xFF000000 || (UInt32)(p->low >> 32) != 0)
+  const UInt32 low = (UInt32)p->low;
+  const unsigned high = (unsigned)
+    #if defined(Z7_MSC_VER_ORIGINAL) \
+        && defined(MY_CPU_X86) \
+        && defined(MY_CPU_LE) \
+        && !defined(MY_CPU_64BIT)
+      // we try to rid of __aullshr() call in MSVS-x86
+      (((const UInt32 *)&p->low)[1]); // [1] : for little-endian only
+    #else
+      (p->low >> 32);
+    #endif
+  if (low < (UInt32)0xff000000 || high != 0)
   {
     Byte *buf = p->bufs[BCJ2_STREAM_RC];
     do
@@ -65,247 +67,440 @@ static BoolInt MY_FAST_CALL RangeEnc_ShiftLow(CBcj2Enc *p)
         p->bufs[BCJ2_STREAM_RC] = buf;
         return True;
       }
-      *buf++ = (Byte)(p->cache + (Byte)(p->low >> 32));
+      *buf++ = (Byte)(p->cache + high);
-      p->cache = 0xFF;
+      p->cache = 0xff;
     }
     while (--p->cacheSize);
     p->bufs[BCJ2_STREAM_RC] = buf;
-    p->cache = (Byte)((UInt32)p->low >> 24);
+    p->cache = (Byte)(low >> 24);
   }
   p->cacheSize++;
-  p->low = (UInt32)p->low << 8;
+  p->low = low << 8;
   return False;
 }
 
+
+/*
+We can use 2 alternative versions of code:
+1) non-marker version:
+  Byte CBcj2Enc::context
+  Byte temp[8];
+  Last byte of marker (e8/e9/[0f]8x) can be written to temp[] buffer.
+  Encoder writes last byte of marker (e8/e9/[0f]8x) to dest, only in conjunction
+  with writing branch symbol to range coder in same Bcj2Enc_Encode_2() call.
+
+2) marker version:
+  UInt32 CBcj2Enc::context
+  Byte CBcj2Enc::temp[4];
+  MARKER_FLAG in CBcj2Enc::context shows that CBcj2Enc::context contains finded marker.
+  it's allowed that
+    one call of Bcj2Enc_Encode_2() writes last byte of marker (e8/e9/[0f]8x) to dest,
+    and another call of Bcj2Enc_Encode_2() does offset conversion.
+    So different values of (fileIp) and (fileSize) are possible
+    in these different Bcj2Enc_Encode_2() calls.
+
+Also marker version requires additional if((v & MARKER_FLAG) == 0) check in main loop.
+So we use non-marker version.
+*/
+
+/*
+  Corner cases with overlap in multi-block.
+  before v23: there was one corner case, where converted instruction
+    could start in one sub-stream and finish in next sub-stream.
+  If multi-block (solid) encoding is used,
+    and BCJ2_ENC_FINISH_MODE_END_BLOCK is used for each sub-stream.
+    and (0f) is last byte of previous sub-stream
+    and (8x) is first byte of current sub-stream
+  then (0f 8x) pair is treated as marker by BCJ2 encoder and decoder.
+  BCJ2 encoder can converts 32-bit offset for that (0f 8x) cortage,
+  if that offset meets limit requirements.
+  If encoder allows 32-bit offset conversion for such overlap case,
+  then the data in 3 uncompressed BCJ2 streams for some sub-stream
+  can depend from data of previous sub-stream.
+  That corner case is not big problem, and it's rare case.
+  Since v23.00 we do additional check to prevent conversions in such overlap cases.
+*/
+
+/*
+  Bcj2Enc_Encode_2() output variables at exit:
+  {
+    if (Bcj2Enc_Encode_2() exits with (p->state == BCJ2_ENC_STATE_ORIG))
+    {
+      it means that encoder needs more input data.
+      if (p->srcLim == p->src) at exit, then
+      {
+        (p->finishMode != BCJ2_ENC_FINISH_MODE_END_STREAM)
+        all input data were read and processed, and we are ready for
+        new input data.
+      }
+      else
+      {
+        (p->srcLim != p->src)
+        (p->finishMode == BCJ2_ENC_FINISH_MODE_CONTINUE)
+          The encoder have found e8/e9/0f_8x marker,
+          and p->src points to last byte of that marker,
+          Bcj2Enc_Encode_2() needs more input data to get totally
+          5 bytes (last byte of marker and 32-bit branch offset)
+          as continuous array starting from p->src.
+        (p->srcLim - p->src < 5) requirement is met after exit.
+          So non-processed resedue from p->src to p->srcLim is always less than 5 bytes.
+      }
+    }
+  }
+*/
+
+Z7_NO_INLINE
 static void Bcj2Enc_Encode_2(CBcj2Enc *p)
 {
-  if (BCJ2_IS_32BIT_STREAM(p->state))
+  if (!p->isFlushState)
   {
-    Byte *cur = p->bufs[p->state];
+    const Byte *src;
-    if (cur == p->lims[p->state])
+    UInt32 v;
-      return;
-    SetBe32(cur, p->tempTarget);
-    p->bufs[p->state] = cur + 4;
-  }
-
-  p->state = BCJ2_ENC_STATE_ORIG;
-
-  for (;;)
-  {
-    if (p->range < kTopValue)
-    {
-      if (RangeEnc_ShiftLow(p))
-        return;
-      p->range <<= 8;
-    }
-
     {
+      const unsigned state = p->state;
+      if (BCJ2_IS_32BIT_STREAM(state))
+      {
+        Byte *cur = p->bufs[state];
+        if (cur == p->lims[state])
+          return;
+        SetBe32a(cur, p->tempTarget)
+        p->bufs[state] = cur + 4;
+      }
+    }
+    p->state = BCJ2_ENC_STATE_ORIG; // for main reason of exit
+    src = p->src;
+    v = p->context;
+    
+    // #define WRITE_CONTEXT  p->context = v; // for marker version
+    #define WRITE_CONTEXT           p->context = (Byte)v;
+    #define WRITE_CONTEXT_AND_SRC   p->src = src;  WRITE_CONTEXT
+
+    for (;;)
+    {
+      // const Byte *src;
+      // UInt32 v;
+      CBcj2Enc_ip_unsigned ip;
+      if (p->range < kTopValue)
+      {
+        // to reduce register pressure and code size: we save and restore local variables.
+        WRITE_CONTEXT_AND_SRC
+        if (Bcj2_RangeEnc_ShiftLow(p))
+          return;
+        p->range <<= 8;
+        src = p->src;
+        v = p->context;
+      }
+      // src = p->src;
+      // #define MARKER_FLAG  ((UInt32)1 << 17)
+      // if ((v & MARKER_FLAG) == 0) // for marker version
       {
-        const Byte *src = p->src;
         const Byte *srcLim;
-        Byte *dest;
+        Byte *dest = p->bufs[BCJ2_STREAM_MAIN];
-        SizeT num = (SizeT)(p->srcLim - src);
-
-        if (p->finishMode == BCJ2_ENC_FINISH_MODE_CONTINUE)
         {
-          if (num <= 4)
+          const SizeT remSrc = (SizeT)(p->srcLim - src);
-            return;
+          SizeT rem = (SizeT)(p->lims[BCJ2_STREAM_MAIN] - dest);
-          num -= 4;
+          if (rem >= remSrc)
+            rem = remSrc;
+          srcLim = src + rem;
         }
-        else if (num == 0)
+        /* p->context contains context of previous byte:
-          break;
+           bits [0 : 7]  : src[-1], if (src) was changed in this call
+           bits [8 : 31] : are undefined for non-marker version
+        */
+        // v = p->context;
+        #define NUM_SHIFT_BITS  24
+        #define CONV_FLAG  ((UInt32)1 << 16)
+        #define ONE_ITER { \
+          b = src[0]; \
+          *dest++ = (Byte)b; \
+          v = (v << NUM_SHIFT_BITS) | b; \
+          if (((b + (0x100 - 0xe8)) & 0xfe) == 0) break; \
+          if (((v - (((UInt32)0x0f << (NUM_SHIFT_BITS)) + 0x80)) & \
+              ((((UInt32)1 << (4 + NUM_SHIFT_BITS)) - 0x1) << 4)) == 0) break; \
+          src++; if (src == srcLim) { break; } }
 
-        dest = p->bufs[BCJ2_STREAM_MAIN];
+        if (src != srcLim)
-        if (num > (SizeT)(p->lims[BCJ2_STREAM_MAIN] - dest))
+        for (;;)
         {
-          num = (SizeT)(p->lims[BCJ2_STREAM_MAIN] - dest);
+          /* clang can generate ineffective code with setne instead of two jcc instructions.
-          if (num == 0)
+             we can use 2 iterations and external (unsigned b) to avoid that ineffective code genaration. */
+          unsigned b;
+          ONE_ITER
+          ONE_ITER
+        }
+        
+        ip = p->ip64 + (CBcj2Enc_ip_unsigned)(SizeT)(dest - p->bufs[BCJ2_STREAM_MAIN]);
+        p->bufs[BCJ2_STREAM_MAIN] = dest;
+        p->ip64 = ip;
+
+        if (src == srcLim)
+        {
+          WRITE_CONTEXT_AND_SRC
+          if (src != p->srcLim)
           {
             p->state = BCJ2_STREAM_MAIN;
             return;
           }
-        }
+          /* (p->src == p->srcLim)
-       
+          (p->state == BCJ2_ENC_STATE_ORIG) */
-        srcLim = src + num;
+          if (p->finishMode != BCJ2_ENC_FINISH_MODE_END_STREAM)
-
+            return;
-        if (p->prevByte == 0x0F && (src[0] & 0xF0) == 0x80)
+          /* (p->finishMode == BCJ2_ENC_FINISH_MODE_END_STREAM */
-          *dest = src[0];
+          // (p->flushRem == 5);
-        else for (;;)
+          p->isFlushState = 1;
-        {
-          Byte b = *src;
-          *dest = b;
-          if (b != 0x0F)
-          {
-            if ((b & 0xFE) == 0xE8)
-              break;
-            dest++;
-            if (++src != srcLim)
-              continue;
-            break;
-          }
-          dest++;
-          if (++src == srcLim)
-            break;
-          if ((*src & 0xF0) != 0x80)
-            continue;
-          *dest = *src;
           break;
         }
-        
+        src++;
-        num = (SizeT)(src - p->src);
+        // p->src = src;
-        
+      }
-        if (src == srcLim)
+      // ip = p->ip; // for marker version
+      /* marker was found */
+      /* (v) contains marker that was found:
+           bits [NUM_SHIFT_BITS : NUM_SHIFT_BITS + 7]
+                         : value of src[-2] : xx/xx/0f
+           bits [0 : 7]  : value of src[-1] : e8/e9/8x
+      */
+      {
         {
-          p->prevByte = src[-1];
+        #if NUM_SHIFT_BITS != 24
-          p->bufs[BCJ2_STREAM_MAIN] = dest;
+          v &= ~(UInt32)CONV_FLAG;
-          p->src = src;
+        #endif
-          p->ip += (UInt32)num;
+          // UInt32 relat = 0;
-          continue;
-        }
- 
-        {
-          Byte context = (Byte)(num == 0 ? p->prevByte : src[-1]);
-          BoolInt needConvert;
-
-          p->bufs[BCJ2_STREAM_MAIN] = dest + 1;
-          p->ip += (UInt32)num + 1;
-          src++;
-          
-          needConvert = False;
-
           if ((SizeT)(p->srcLim - src) >= 4)
           {
-            UInt32 relatVal = GetUi32(src);
+            /*
-            if ((p->fileSize == 0 || (UInt32)(p->ip + 4 + relatVal - p->fileIp) < p->fileSize)
+            if (relat != 0 || (Byte)v != 0xe8)
-                && ((relatVal + p->relatLimit) >> 1) < p->relatLimit)
+            BoolInt isBigOffset = True;
-              needConvert = True;
+            */
-          }
+            const UInt32 relat = GetUi32(src);
-
+            /*
-          {
+            #define EXCLUDE_FLAG  ((UInt32)1 << 4)
-            UInt32 bound;
+            #define NEED_CONVERT(rel) ((((rel) + EXCLUDE_FLAG) & (0 - EXCLUDE_FLAG * 2)) != 0)
-            unsigned ttt;
+            if (p->relatExcludeBits != 0)
-            Byte b = src[-1];
-            CProb *prob = p->probs + (unsigned)(b == 0xE8 ? 2 + (unsigned)context : (b == 0xE9 ? 1 : 0));
-
-            ttt = *prob;
-            bound = (p->range >> kNumModelBits) * ttt;
-            
-            if (!needConvert)
             {
+              const UInt32 flag = (UInt32)1 << (p->relatExcludeBits - 1);
+              isBigOffset = (((relat + flag) & (0 - flag * 2)) != 0);
+            }
+            // isBigOffset = False; // for debug
+            */
+            ip -= p->fileIp64;
+            // Use the following if check, if (ip) is 64-bit:
+            if (ip > (((v + 0x20) >> 5) & 1))  // 23.00 : we eliminate milti-block overlap for (Of 80) and (e8/e9)
+            if ((CBcj2Enc_ip_unsigned)((CBcj2Enc_ip_signed)ip + 4 + (Int32)relat) <= p->fileSize64_minus1)
+            if (((UInt32)(relat + p->relatLimit) >> 1) < p->relatLimit)
+              v |= CONV_FLAG;
+          }
+          else if (p->finishMode == BCJ2_ENC_FINISH_MODE_CONTINUE)
+          {
+            // (p->srcLim - src < 4)
+            // /*
+            // for non-marker version
+            p->ip64--; // p->ip = ip - 1;
+            p->bufs[BCJ2_STREAM_MAIN]--;
+            src--;
+            v >>= NUM_SHIFT_BITS;
+            // (0 < p->srcLim - p->src <= 4)
+            // */
+            // v |= MARKER_FLAG; // for marker version
+            /* (p->state == BCJ2_ENC_STATE_ORIG) */
+            WRITE_CONTEXT_AND_SRC
+            return;
+          }
+          {
+            const unsigned c = ((v + 0x17) >> 6) & 1;
+            CBcj2Prob *prob = p->probs + (unsigned)
+                (((0 - c) & (Byte)(v >> NUM_SHIFT_BITS)) + c + ((v >> 5) & 1));
+            /*
+                ((Byte)v == 0xe8 ? 2 + ((Byte)(v >> 8)) :
+                ((Byte)v < 0xe8 ? 0 : 1));  // ((v >> 5) & 1));
+            */
+            const unsigned ttt = *prob;
+            const UInt32 bound = (p->range >> kNumBitModelTotalBits) * ttt;
+            if ((v & CONV_FLAG) == 0)
+            {
+              // static int yyy = 0; yyy++; printf("\n!needConvert = %d\n", yyy);
+              // v = (Byte)v; // for marker version
               p->range = bound;
-              *prob = (CProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits));
+              *prob = (CBcj2Prob)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits));
-              p->src = src;
+              // WRITE_CONTEXT_AND_SRC
-              p->prevByte = b;
               continue;
             }
-            
             p->low += bound;
             p->range -= bound;
-            *prob = (CProb)(ttt - (ttt >> kNumMoveBits));
+            *prob = (CBcj2Prob)(ttt - (ttt >> kNumMoveBits));
-
+          }
+          // p->context = src[3];
+          {
+            // const unsigned cj = ((Byte)v == 0xe8 ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP);
+            const unsigned cj = (((v + 0x57) >> 6) & 1) + BCJ2_STREAM_CALL;
+            ip = p->ip64;
+            v = GetUi32(src); // relat
+            ip += 4;
+            p->ip64 = ip;
+            src += 4;
+            // p->src = src;
             {
-              UInt32 relatVal = GetUi32(src);
+              const UInt32 absol = (UInt32)ip + v;
-              UInt32 absVal;
+              Byte *cur = p->bufs[cj];
-              p->ip += 4;
+              v >>= 24;
-              absVal = p->ip + relatVal;
+              // WRITE_CONTEXT
-              p->prevByte = src[3];
+              if (cur == p->lims[cj])
-              src += 4;
-              p->src = src;
               {
-                unsigned cj = (b == 0xE8) ? BCJ2_STREAM_CALL : BCJ2_STREAM_JUMP;
+                p->state = cj;
-                Byte *cur = p->bufs[cj];
+                p->tempTarget = absol;
-                if (cur == p->lims[cj])
+                WRITE_CONTEXT_AND_SRC
-                {
+                return;
-                  p->state = cj;
-                  p->tempTarget = absVal;
-                  return;
-                }
-                SetBe32(cur, absVal);
-                p->bufs[cj] = cur + 4;
               }
+              SetBe32a(cur, absol)
+              p->bufs[cj] = cur + 4;
             }
           }
         }
       }
-    }
+    } // end of loop
   }
 
-  if (p->finishMode != BCJ2_ENC_FINISH_MODE_END_STREAM)
+  for (; p->flushRem != 0; p->flushRem--)
-    return;
+    if (Bcj2_RangeEnc_ShiftLow(p))
-
-  for (; p->flushPos < 5; p->flushPos++)
-    if (RangeEnc_ShiftLow(p))
       return;
-  p->state = BCJ2_ENC_STATE_OK;
+  p->state = BCJ2_ENC_STATE_FINISHED;
 }
 
 
+/*
+BCJ2 encoder needs look ahead for up to 4 bytes in (src) buffer.
+So base function Bcj2Enc_Encode_2()
+  in BCJ2_ENC_FINISH_MODE_CONTINUE mode can return with
+  (p->state == BCJ2_ENC_STATE_ORIG && p->src < p->srcLim)
+Bcj2Enc_Encode() solves that look ahead problem by using p->temp[] buffer.
+  so if (p->state == BCJ2_ENC_STATE_ORIG) after Bcj2Enc_Encode(),
+    then (p->src == p->srcLim).
+  And the caller's code is simpler with Bcj2Enc_Encode().
+*/
+
+Z7_NO_INLINE
 void Bcj2Enc_Encode(CBcj2Enc *p)
 {
-  PRF(printf("\n"));
+  PRF2("\n----")
-  PRF(printf("---- ip = %8d   tempPos = %8d   src = %8d\n", p->ip, p->tempPos, p->srcLim - p->src));
-
   if (p->tempPos != 0)
   {
+    /* extra: number of bytes that were copied from (src) to (temp) buffer in this call */
     unsigned extra = 0;
-   
+    /* We will touch only minimal required number of bytes in input (src) stream.
+       So we will add input bytes from (src) stream to temp[] with step of 1 byte.
+       We don't add new bytes to temp[] before Bcj2Enc_Encode_2() call
+         in first loop iteration because
+         - previous call of Bcj2Enc_Encode() could use another (finishMode),
+         - previous call could finish with (p->state != BCJ2_ENC_STATE_ORIG).
+       the case with full temp[] buffer (p->tempPos == 4) is possible here.
+    */
     for (;;)
     {
+      // (0 < p->tempPos <= 5) // in non-marker version
+      /* p->src : the current src data position including extra bytes
+                  that were copied to temp[] buffer in this call */
       const Byte *src = p->src;
       const Byte *srcLim = p->srcLim;
-      EBcj2Enc_FinishMode finishMode = p->finishMode;
+      const EBcj2Enc_FinishMode finishMode = p->finishMode;
-      
+      if (src != srcLim)
+      {
+        /* if there are some src data after the data copied to temp[],
+           then we use MODE_CONTINUE for temp data */
+        p->finishMode = BCJ2_ENC_FINISH_MODE_CONTINUE;
+      }
       p->src = p->temp;
       p->srcLim = p->temp + p->tempPos;
-      if (src != srcLim)
+      PRF2("    ")
-        p->finishMode = BCJ2_ENC_FINISH_MODE_CONTINUE;
-      
-      PRF(printf("     ip = %8d   tempPos = %8d   src = %8d\n", p->ip, p->tempPos, p->srcLim - p->src));
-
       Bcj2Enc_Encode_2(p);
-      
       {
-        unsigned num = (unsigned)(p->src - p->temp);
+        const unsigned num = (unsigned)(p->src - p->temp);
-        unsigned tempPos = p->tempPos - num;
+        const unsigned tempPos = p->tempPos - num;
         unsigned i;
         p->tempPos = tempPos;
         for (i = 0; i < tempPos; i++)
-          p->temp[i] = p->temp[(size_t)i + num];
+          p->temp[i] = p->temp[(SizeT)i + num];
-      
+        // tempPos : number of bytes in temp buffer
         p->src = src;
         p->srcLim = srcLim;
         p->finishMode = finishMode;
-        
+        if (p->state != BCJ2_ENC_STATE_ORIG)
-        if (p->state != BCJ2_ENC_STATE_ORIG || src == srcLim)
+        {
+          // (p->tempPos <= 4) // in non-marker version
+          /* if (the reason of exit from Bcj2Enc_Encode_2()
+                 is not BCJ2_ENC_STATE_ORIG),
+             then we exit from Bcj2Enc_Encode() with same reason */
+          // optional code begin : we rollback (src) and tempPos, if it's possible:
+          if (extra >= tempPos)
+            extra = tempPos;
+          p->src = src - extra;
+          p->tempPos = tempPos - extra;
+          // optional code end : rollback of (src) and tempPos
           return;
-        
+        }
+        /* (p->tempPos <= 4)
+           (p->state == BCJ2_ENC_STATE_ORIG)
+             so encoder needs more data than in temp[] */
+        if (src == srcLim)
+          return; // src buffer has no more input data.
+        /* (src != srcLim)
+           so we can provide more input data from src for Bcj2Enc_Encode_2() */
         if (extra >= tempPos)
         {
-          p->src = src - tempPos;
+          /* (extra >= tempPos) means that temp buffer contains
+             only data from src buffer of this call.
+             So now we can encode without temp buffer */
+          p->src = src - tempPos; // rollback (src)
           p->tempPos = 0;
           break;
         }
-        
+        // we append one additional extra byte from (src) to temp[] buffer:
-        p->temp[tempPos] = src[0];
+        p->temp[tempPos] = *src;
         p->tempPos = tempPos + 1;
+        // (0 < p->tempPos <= 5) // in non-marker version
         p->src = src + 1;
         extra++;
       }
     }
   }
 
-  PRF(printf("++++ ip = %8d   tempPos = %8d   src = %8d\n", p->ip, p->tempPos, p->srcLim - p->src));
+  PRF2("++++")
-
+  // (p->tempPos == 0)
   Bcj2Enc_Encode_2(p);
+  PRF2("====")
   
   if (p->state == BCJ2_ENC_STATE_ORIG)
   {
     const Byte *src = p->src;
-    unsigned rem = (unsigned)(p->srcLim - src);
+    const Byte *srcLim = p->srcLim;
-    unsigned i;
+    const unsigned rem = (unsigned)(srcLim - src);
-    for (i = 0; i < rem; i++)
+    /* (rem <= 4) here.
-      p->temp[i] = src[i];
+       if (p->src != p->srcLim), then
-    p->tempPos = rem;
+         - we copy non-processed bytes from (p->src) to temp[] buffer,
-    p->src = src + rem;
+         - we set p->src equal to p->srcLim.
+    */
+    if (rem)
+    {
+      unsigned i = 0;
+      p->src = srcLim;
+      p->tempPos = rem;
+      // (0 < p->tempPos <= 4)
+      do
+        p->temp[i] = src[i];
+      while (++i != rem);
+    }
+    // (p->tempPos <= 4)
+    // (p->src == p->srcLim)
   }
 }
+
+#undef PRF2
+#undef CONV_FLAG
+#undef MARKER_FLAG
+#undef WRITE_CONTEXT
+#undef WRITE_CONTEXT_AND_SRC
+#undef ONE_ITER
+#undef NUM_SHIFT_BITS
+#undef kTopValue
+#undef kNumBitModelTotalBits
+#undef kBitModelTotal
+#undef kNumMoveBits

C/Blake2.h

@@ -1,47 +1,104 @@
-/* Blake2.h -- BLAKE2 Hash
+/* Blake2.h -- BLAKE2sp Hash
-2015-06-30 : Igor Pavlov : Public domain
+2024-01-17 : Igor Pavlov : Public domain */
-2015 : Samuel Neves : Public domain */
 
-#ifndef __BLAKE2_H
+#ifndef ZIP7_INC_BLAKE2_H
-#define __BLAKE2_H
+#define ZIP7_INC_BLAKE2_H
 
 #include "7zTypes.h"
 
+#if 0
+#include "Compiler.h"
+#include "CpuArch.h"
+#if defined(MY_CPU_X86_OR_AMD64)
+#if defined(__SSE2__) \
+    || defined(_MSC_VER) && _MSC_VER > 1200 \
+    || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 30300) \
+    || defined(__clang__) \
+    || defined(__INTEL_COMPILER)
+#include <emmintrin.h> // SSE2
+#endif
+
+#if defined(__AVX2__) \
+    || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40900) \
+    || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 40600) \
+    || defined(Z7_LLVM_CLANG_VERSION) && (Z7_LLVM_CLANG_VERSION >= 30100) \
+    || defined(Z7_MSC_VER_ORIGINAL) && (Z7_MSC_VER_ORIGINAL >= 1800) \
+    || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1400)
+#include <immintrin.h>
+#if defined(__clang__)
+#include <avxintrin.h>
+#include <avx2intrin.h>
+#endif
+#endif // avx2
+#endif // MY_CPU_X86_OR_AMD64
+#endif // 0
+
 EXTERN_C_BEGIN
 
-#define BLAKE2S_BLOCK_SIZE 64
+#define Z7_BLAKE2S_BLOCK_SIZE         64
-#define BLAKE2S_DIGEST_SIZE 32
+#define Z7_BLAKE2S_DIGEST_SIZE        32
-#define BLAKE2SP_PARALLEL_DEGREE 8
+#define Z7_BLAKE2SP_PARALLEL_DEGREE   8
+#define Z7_BLAKE2SP_NUM_STRUCT_WORDS  16
 
+#if 1 || defined(Z7_BLAKE2SP_USE_FUNCTIONS)
+typedef void (Z7_FASTCALL *Z7_BLAKE2SP_FUNC_COMPRESS)(UInt32 *states, const Byte *data, const Byte *end);
+typedef void (Z7_FASTCALL *Z7_BLAKE2SP_FUNC_INIT)(UInt32 *states);
+#endif
+
+// it's required that CBlake2sp is aligned for 32-bytes,
+// because the code can use unaligned access with sse and avx256.
+// but 64-bytes alignment can be better.
+MY_ALIGN(64)
 typedef struct
 {
-  UInt32 h[8];
+  union
-  UInt32 t[2];
+  {
-  UInt32 f[2];
+#if 0
-  Byte buf[BLAKE2S_BLOCK_SIZE];
+#if defined(MY_CPU_X86_OR_AMD64)
-  UInt32 bufPos;
+#if defined(__SSE2__) \
-  UInt32 lastNode_f1;
+    || defined(_MSC_VER) && _MSC_VER > 1200 \
-  UInt32 dummy[2]; /* for sizeof(CBlake2s) alignment */
+    || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 30300) \
-} CBlake2s;
+    || defined(__clang__) \
+    || defined(__INTEL_COMPILER)
+    __m128i _pad_align_128bit[4];
+#endif // sse2
+#if defined(__AVX2__) \
+    || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40900) \
+    || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 40600) \
+    || defined(Z7_LLVM_CLANG_VERSION) && (Z7_LLVM_CLANG_VERSION >= 30100) \
+    || defined(Z7_MSC_VER_ORIGINAL) && (Z7_MSC_VER_ORIGINAL >= 1800) \
+    || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1400)
+    __m256i _pad_align_256bit[2];
+#endif // avx2
+#endif // x86
+#endif // 0
 
-/* You need to xor CBlake2s::h[i] with input parameter block after Blake2s_Init0() */
+    void * _pad_align_ptr[8];
-/*
+    UInt32 _pad_align_32bit[16];
-void Blake2s_Init0(CBlake2s *p);
+    struct
-void Blake2s_Update(CBlake2s *p, const Byte *data, size_t size);
+    {
-void Blake2s_Final(CBlake2s *p, Byte *digest);
+      unsigned cycPos;
-*/
+      unsigned _pad_unused;
-
+#if 1 || defined(Z7_BLAKE2SP_USE_FUNCTIONS)
-
+      Z7_BLAKE2SP_FUNC_COMPRESS func_Compress_Fast;
-typedef struct
+      Z7_BLAKE2SP_FUNC_COMPRESS func_Compress_Single;
-{
+      Z7_BLAKE2SP_FUNC_INIT func_Init;
-  CBlake2s S[BLAKE2SP_PARALLEL_DEGREE];
+      Z7_BLAKE2SP_FUNC_INIT func_Final;
-  unsigned bufPos;
+#endif
+    } header;
+  } u;
+  // MY_ALIGN(64)
+  UInt32 states[Z7_BLAKE2SP_PARALLEL_DEGREE * Z7_BLAKE2SP_NUM_STRUCT_WORDS];
+  // MY_ALIGN(64)
+  UInt32  buf32[Z7_BLAKE2SP_PARALLEL_DEGREE * Z7_BLAKE2SP_NUM_STRUCT_WORDS * 2];
 } CBlake2sp;
 
-
+BoolInt Blake2sp_SetFunction(CBlake2sp *p, unsigned algo);
 void Blake2sp_Init(CBlake2sp *p);
+void Blake2sp_InitState(CBlake2sp *p);
 void Blake2sp_Update(CBlake2sp *p, const Byte *data, size_t size);
 void Blake2sp_Final(CBlake2sp *p, Byte *digest);
+void z7_Black2sp_Prepare(void);
 
 EXTERN_C_END
 

C/Bra.c

@@ -1,230 +1,709 @@
-/* Bra.c -- Converters for RISC code
+/* Bra.c -- Branch converters for RISC code
-2021-02-09 : Igor Pavlov : Public domain */
+2024-01-20 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
-#include "CpuArch.h"
 #include "Bra.h"
+#include "RotateDefs.h"
+#include "CpuArch.h"
 
-SizeT ARM_Convert(Byte *data, SizeT size, UInt32 ip, int encoding)
+#if defined(MY_CPU_SIZEOF_POINTER) \
+    && ( MY_CPU_SIZEOF_POINTER == 4 \
+      || MY_CPU_SIZEOF_POINTER == 8)
+  #define BR_CONV_USE_OPT_PC_PTR
+#endif
+
+#ifdef BR_CONV_USE_OPT_PC_PTR
+#define BR_PC_INIT  pc -= (UInt32)(SizeT)p;
+#define BR_PC_GET   (pc + (UInt32)(SizeT)p)
+#else
+#define BR_PC_INIT  pc += (UInt32)size;
+#define BR_PC_GET   (pc - (UInt32)(SizeT)(lim - p))
+// #define BR_PC_INIT
+// #define BR_PC_GET   (pc + (UInt32)(SizeT)(p - data))
+#endif
+
+#define BR_CONVERT_VAL(v, c) if (encoding) v += c; else v -= c;
+// #define BR_CONVERT_VAL(v, c) if (!encoding) c = (UInt32)0 - c; v += c;
+
+#define Z7_BRANCH_CONV(name) z7_ ## name
+
+#define Z7_BRANCH_FUNC_MAIN(name) \
+static \
+Z7_FORCE_INLINE \
+Z7_ATTRIB_NO_VECTOR \
+Byte *Z7_BRANCH_CONV(name)(Byte *p, SizeT size, UInt32 pc, int encoding)
+
+#define Z7_BRANCH_FUNC_IMP(name, m, encoding) \
+Z7_NO_INLINE \
+Z7_ATTRIB_NO_VECTOR \
+Byte *m(name)(Byte *data, SizeT size, UInt32 pc) \
+  { return Z7_BRANCH_CONV(name)(data, size, pc, encoding); } \
+
+#ifdef Z7_EXTRACT_ONLY
+#define Z7_BRANCH_FUNCS_IMP(name) \
+  Z7_BRANCH_FUNC_IMP(name, Z7_BRANCH_CONV_DEC_2, 0)
+#else
+#define Z7_BRANCH_FUNCS_IMP(name) \
+  Z7_BRANCH_FUNC_IMP(name, Z7_BRANCH_CONV_DEC_2, 0) \
+  Z7_BRANCH_FUNC_IMP(name, Z7_BRANCH_CONV_ENC_2, 1)
+#endif
+
+#if defined(__clang__)
+#define BR_EXTERNAL_FOR
+#define BR_NEXT_ITERATION  continue;
+#else
+#define BR_EXTERNAL_FOR    for (;;)
+#define BR_NEXT_ITERATION  break;
+#endif
+
+#if defined(__clang__) && (__clang_major__ >= 8) \
+  || defined(__GNUC__) && (__GNUC__ >= 1000) \
+  // GCC is not good for __builtin_expect() here
+  /* || defined(_MSC_VER) && (_MSC_VER >= 1920) */
+  // #define Z7_unlikely [[unlikely]]
+  // #define Z7_LIKELY(x)   (__builtin_expect((x), 1))
+  #define Z7_UNLIKELY(x) (__builtin_expect((x), 0))
+  // #define Z7_likely [[likely]]
+#else
+  // #define Z7_LIKELY(x)   (x)
+  #define Z7_UNLIKELY(x) (x)
+  // #define Z7_likely
+#endif
+
+
+Z7_BRANCH_FUNC_MAIN(BranchConv_ARM64)
 {
-  Byte *p;
+  // Byte *p = data;
   const Byte *lim;
-  size &= ~(size_t)3;
+  const UInt32 flag = (UInt32)1 << (24 - 4);
-  ip += 4;
+  const UInt32 mask = ((UInt32)1 << 24) - (flag << 1);
-  p = data;
+  size &= ~(SizeT)3;
-  lim = data + size;
+  // if (size == 0) return p;
+  lim = p + size;
+  BR_PC_INIT
+  pc -= 4;  // because (p) will point to next instruction
   
-  if (encoding)
+  BR_EXTERNAL_FOR
-
-  for (;;)
   {
+    // Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE
     for (;;)
     {
-      if (p >= lim)
+      UInt32 v;
-        return (SizeT)(p - data);
+      if Z7_UNLIKELY(p == lim)
+        return p;
+      v = GetUi32a(p);
       p += 4;
-      if (p[-1] == 0xEB)
+      if Z7_UNLIKELY(((v - 0x94000000) & 0xfc000000) == 0)
-        break;
+      {
-    }
+        UInt32 c = BR_PC_GET >> 2;
-    {
+        BR_CONVERT_VAL(v, c)
-      UInt32 v = GetUi32(p - 4);
+        v &= 0x03ffffff;
-      v <<= 2;
+        v |= 0x94000000;
-        v += ip + (UInt32)(p - data);
+        SetUi32a(p - 4, v)
-      v >>= 2;
+        BR_NEXT_ITERATION
-      v &= 0x00FFFFFF;
+      }
-      v |= 0xEB000000;
+      // v = rotlFixed(v, 8);  v += (flag << 8) - 0x90;  if Z7_UNLIKELY((v & ((mask << 8) + 0x9f)) == 0)
-      SetUi32(p - 4, v);
+      v -= 0x90000000;  if Z7_UNLIKELY((v & 0x9f000000) == 0)
-    }
+      {
-  }
+        UInt32 z, c;
-
+        // v = rotrFixed(v, 8);
-  for (;;)
+        v += flag; if Z7_UNLIKELY(v & mask) continue;
-  {
+        z = (v & 0xffffffe0) | (v >> 26);
-    for (;;)
+        c = (BR_PC_GET >> (12 - 3)) & ~(UInt32)7;
-    {
+        BR_CONVERT_VAL(z, c)
-      if (p >= lim)
+        v &= 0x1f;
-        return (SizeT)(p - data);
+        v |= 0x90000000;
-      p += 4;
+        v |= z << 26;
-      if (p[-1] == 0xEB)
+        v |= 0x00ffffe0 & ((z & (((flag << 1) - 1))) - flag);
-        break;
+        SetUi32a(p - 4, v)
-    }
+      }
-    {
-      UInt32 v = GetUi32(p - 4);
-      v <<= 2;
-        v -= ip + (UInt32)(p - data);
-      v >>= 2;
-      v &= 0x00FFFFFF;
-      v |= 0xEB000000;
-      SetUi32(p - 4, v);
     }
   }
 }
+Z7_BRANCH_FUNCS_IMP(BranchConv_ARM64)
 
 
-SizeT ARMT_Convert(Byte *data, SizeT size, UInt32 ip, int encoding)
+Z7_BRANCH_FUNC_MAIN(BranchConv_ARM)
 {
-  Byte *p;
+  // Byte *p = data;
   const Byte *lim;
-  size &= ~(size_t)1;
+  size &= ~(SizeT)3;
-  p = data;
+  lim = p + size;
-  lim = data + size - 4;
+  BR_PC_INIT
-
+  /* in ARM: branch offset is relative to the +2 instructions from current instruction.
-  if (encoding)
+     (p) will point to next instruction */
+  pc += 8 - 4;
   
   for (;;)
   {
-    UInt32 b1;
     for (;;)
     {
-      UInt32 b3;
+      if Z7_UNLIKELY(p >= lim) { return p; }  p += 4;  if Z7_UNLIKELY(p[-1] == 0xeb) break;
-      if (p > lim)
+      if Z7_UNLIKELY(p >= lim) { return p; }  p += 4;  if Z7_UNLIKELY(p[-1] == 0xeb) break;
-        return (SizeT)(p - data);
-      b1 = p[1];
-      b3 = p[3];
-      p += 2;
-      b1 ^= 8;
-      if ((b3 & b1) >= 0xF8)
-        break;
     }
     {
-      UInt32 v =
+      UInt32 v = GetUi32a(p - 4);
-             ((UInt32)b1 << 19)
+      UInt32 c = BR_PC_GET >> 2;
-          + (((UInt32)p[1] & 0x7) << 8)
+      BR_CONVERT_VAL(v, c)
-          + (((UInt32)p[-2] << 11))
+      v &= 0x00ffffff;
-          + (p[0]);
+      v |= 0xeb000000;
-
+      SetUi32a(p - 4, v)
-      p += 2;
-      {
-        UInt32 cur = (ip + (UInt32)(p - data)) >> 1;
-          v += cur;
-      }
-
-      p[-4] = (Byte)(v >> 11);
-      p[-3] = (Byte)(0xF0 | ((v >> 19) & 0x7));
-      p[-2] = (Byte)v;
-      p[-1] = (Byte)(0xF8 | (v >> 8));
-    }
-  }
-  
-  for (;;)
-  {
-    UInt32 b1;
-    for (;;)
-    {
-      UInt32 b3;
-      if (p > lim)
-        return (SizeT)(p - data);
-      b1 = p[1];
-      b3 = p[3];
-      p += 2;
-      b1 ^= 8;
-      if ((b3 & b1) >= 0xF8)
-        break;
-    }
-    {
-      UInt32 v =
-             ((UInt32)b1 << 19)
-          + (((UInt32)p[1] & 0x7) << 8)
-          + (((UInt32)p[-2] << 11))
-          + (p[0]);
-
-      p += 2;
-      {
-        UInt32 cur = (ip + (UInt32)(p - data)) >> 1;
-          v -= cur;
-      }
-
-      /*
-      SetUi16(p - 4, (UInt16)(((v >> 11) & 0x7FF) | 0xF000));
-      SetUi16(p - 2, (UInt16)(v | 0xF800));
-      */
-      
-      p[-4] = (Byte)(v >> 11);
-      p[-3] = (Byte)(0xF0 | ((v >> 19) & 0x7));
-      p[-2] = (Byte)v;
-      p[-1] = (Byte)(0xF8 | (v >> 8));
     }
   }
 }
+Z7_BRANCH_FUNCS_IMP(BranchConv_ARM)
 
 
-SizeT PPC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding)
+Z7_BRANCH_FUNC_MAIN(BranchConv_PPC)
 {
-  Byte *p;
+  // Byte *p = data;
   const Byte *lim;
-  size &= ~(size_t)3;
+  size &= ~(SizeT)3;
-  ip -= 4;
+  lim = p + size;
-  p = data;
+  BR_PC_INIT
-  lim = data + size;
+  pc -= 4;  // because (p) will point to next instruction
   
   for (;;)
   {
+    UInt32 v;
     for (;;)
     {
-      if (p >= lim)
+      if Z7_UNLIKELY(p == lim)
-        return (SizeT)(p - data);
+        return p;
+      // v = GetBe32a(p);
+      v = *(UInt32 *)(void *)p;
       p += 4;
-      /* if ((v & 0xFC000003) == 0x48000001) */
+      // if ((v & 0xfc000003) == 0x48000001) break;
-      if ((p[-4] & 0xFC) == 0x48 && (p[-1] & 3) == 1)
+      // if ((p[-4] & 0xFC) == 0x48 && (p[-1] & 3) == 1) break;
-        break;
+      if Z7_UNLIKELY(
+          ((v - Z7_CONV_BE_TO_NATIVE_CONST32(0x48000001))
+              & Z7_CONV_BE_TO_NATIVE_CONST32(0xfc000003)) == 0) break;
     }
     {
-      UInt32 v = GetBe32(p - 4);
+      v = Z7_CONV_NATIVE_TO_BE_32(v);
-      if (encoding)
+      {
-        v += ip + (UInt32)(p - data);
+        UInt32 c = BR_PC_GET;
-      else
+        BR_CONVERT_VAL(v, c)
-        v -= ip + (UInt32)(p - data);
+      }
-      v &= 0x03FFFFFF;
+      v &= 0x03ffffff;
       v |= 0x48000000;
-      SetBe32(p - 4, v);
+      SetBe32a(p - 4, v)
     }
   }
 }
+Z7_BRANCH_FUNCS_IMP(BranchConv_PPC)
 
 
-SizeT SPARC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding)
+#ifdef Z7_CPU_FAST_ROTATE_SUPPORTED
+#define BR_SPARC_USE_ROTATE
+#endif
+
+Z7_BRANCH_FUNC_MAIN(BranchConv_SPARC)
 {
-  Byte *p;
+  // Byte *p = data;
   const Byte *lim;
-  size &= ~(size_t)3;
+  const UInt32 flag = (UInt32)1 << 22;
-  ip -= 4;
+  size &= ~(SizeT)3;
-  p = data;
+  lim = p + size;
-  lim = data + size;
+  BR_PC_INIT
+  pc -= 4;  // because (p) will point to next instruction
+  for (;;)
+  {
+    UInt32 v;
+    for (;;)
+    {
+      if Z7_UNLIKELY(p == lim)
+        return p;
+      /* // the code without GetBe32a():
+      { const UInt32 v = GetUi16a(p) & 0xc0ff; p += 4; if (v == 0x40 || v == 0xc07f) break; }
+      */
+      v = GetBe32a(p);
+      p += 4;
+    #ifdef BR_SPARC_USE_ROTATE
+      v = rotlFixed(v, 2);
+      v += (flag << 2) - 1;
+      if Z7_UNLIKELY((v & (3 - (flag << 3))) == 0)
+    #else
+      v += (UInt32)5 << 29;
+      v ^= (UInt32)7 << 29;
+      v += flag;
+      if Z7_UNLIKELY((v & (0 - (flag << 1))) == 0)
+    #endif
+        break;
+    }
+    {
+      // UInt32 v = GetBe32a(p - 4);
+    #ifndef BR_SPARC_USE_ROTATE
+      v <<= 2;
+    #endif
+      {
+        UInt32 c = BR_PC_GET;
+        BR_CONVERT_VAL(v, c)
+      }
+      v &= (flag << 3) - 1;
+    #ifdef BR_SPARC_USE_ROTATE
+      v -= (flag << 2) - 1;
+      v = rotrFixed(v, 2);
+    #else
+      v -= (flag << 2);
+      v >>= 2;
+      v |= (UInt32)1 << 30;
+    #endif
+      SetBe32a(p - 4, v)
+    }
+  }
+}
+Z7_BRANCH_FUNCS_IMP(BranchConv_SPARC)
+
+
+Z7_BRANCH_FUNC_MAIN(BranchConv_ARMT)
+{
+  // Byte *p = data;
+  Byte *lim;
+  size &= ~(SizeT)1;
+  // if (size == 0) return p;
+  if (size <= 2) return p;
+  size -= 2;
+  lim = p + size;
+  BR_PC_INIT
+  /* in ARM: branch offset is relative to the +2 instructions from current instruction.
+     (p) will point to the +2 instructions from current instruction */
+  // pc += 4 - 4;
+  // if (encoding) pc -= 0xf800 << 1; else pc += 0xf800 << 1;
+  // #define ARMT_TAIL_PROC { goto armt_tail; }
+  #define ARMT_TAIL_PROC { return p; }
+  
+  do
+  {
+    /* in MSVC 32-bit x86 compilers:
+       UInt32 version : it loads value from memory with movzx
+       Byte   version : it loads value to 8-bit register (AL/CL)
+       movzx version is slightly faster in some cpus
+    */
+    unsigned b1;
+    // Byte / unsigned
+    b1 = p[1];
+    // optimized version to reduce one (p >= lim) check:
+    // unsigned a1 = p[1];  b1 = p[3];  p += 2;  if Z7_LIKELY((b1 & (a1 ^ 8)) < 0xf8)
+    for (;;)
+    {
+      unsigned b3; // Byte / UInt32
+      /* (Byte)(b3) normalization can use low byte computations in MSVC.
+         It gives smaller code, and no loss of speed in some compilers/cpus.
+         But new MSVC 32-bit x86 compilers use more slow load
+         from memory to low byte register in that case.
+         So we try to use full 32-bit computations for faster code.
+      */
+      // if (p >= lim) { ARMT_TAIL_PROC }  b3 = b1 + 8;  b1 = p[3];  p += 2;  if ((b3 & b1) >= 0xf8) break;
+      if Z7_UNLIKELY(p >= lim) { ARMT_TAIL_PROC }  b3 = p[3];  p += 2;  if Z7_UNLIKELY((b3 & (b1 ^ 8)) >= 0xf8) break;
+      if Z7_UNLIKELY(p >= lim) { ARMT_TAIL_PROC }  b1 = p[3];  p += 2;  if Z7_UNLIKELY((b1 & (b3 ^ 8)) >= 0xf8) break;
+    }
+    {
+      /* we can adjust pc for (0xf800) to rid of (& 0x7FF) operation.
+         But gcc/clang for arm64 can use bfi instruction for full code here */
+      UInt32 v =
+          ((UInt32)GetUi16a(p - 2) << 11) |
+          ((UInt32)GetUi16a(p) & 0x7FF);
+      /*
+      UInt32 v =
+            ((UInt32)p[1 - 2] << 19)
+          + (((UInt32)p[1] & 0x7) << 8)
+          + (((UInt32)p[-2] << 11))
+          + (p[0]);
+      */
+      p += 2;
+      {
+        UInt32 c = BR_PC_GET >> 1;
+        BR_CONVERT_VAL(v, c)
+      }
+      SetUi16a(p - 4, (UInt16)(((v >> 11) & 0x7ff) | 0xf000))
+      SetUi16a(p - 2, (UInt16)(v | 0xf800))
+      /*
+      p[-4] = (Byte)(v >> 11);
+      p[-3] = (Byte)(0xf0 | ((v >> 19) & 0x7));
+      p[-2] = (Byte)v;
+      p[-1] = (Byte)(0xf8 | (v >> 8));
+      */
+    }
+  }
+  while (p < lim);
+  return p;
+  // armt_tail:
+  // if ((Byte)((lim[1] & 0xf8)) != 0xf0) { lim += 2; }  return lim;
+  // return (Byte *)(lim + ((Byte)((lim[1] ^ 0xf0) & 0xf8) == 0 ? 0 : 2));
+  // return (Byte *)(lim + (((lim[1] ^ ~0xfu) & ~7u) == 0 ? 0 : 2));
+  // return (Byte *)(lim + 2 - (((((unsigned)lim[1] ^ 8) + 8) >> 7) & 2));
+}
+Z7_BRANCH_FUNCS_IMP(BranchConv_ARMT)
+
+
+// #define BR_IA64_NO_INLINE
+
+Z7_BRANCH_FUNC_MAIN(BranchConv_IA64)
+{
+  // Byte *p = data;
+  const Byte *lim;
+  size &= ~(SizeT)15;
+  lim = p + size;
+  pc -= 1 << 4;
+  pc >>= 4 - 1;
+  // pc -= 1 << 1;
   
   for (;;)
   {
+    unsigned m;
     for (;;)
     {
-      if (p >= lim)
+      if Z7_UNLIKELY(p == lim)
-        return (SizeT)(p - data);
+        return p;
-      /*
+      m = (unsigned)((UInt32)0x334b0000 >> (*p & 0x1e));
-      v = GetBe32(p);
+      p += 16;
-      p += 4;
+      pc += 1 << 1;
-      m = v + ((UInt32)5 << 29);
+      if (m &= 3)
-      m ^= (UInt32)7 << 29;
-      m += (UInt32)1 << 22;
-      if ((m & ((UInt32)0x1FF << 23)) == 0)
-        break;
-      */
-      p += 4;
-      if ((p[-4] == 0x40 && (p[-3] & 0xC0) == 0) ||
-          (p[-4] == 0x7F && (p[-3] >= 0xC0)))
         break;
     }
     {
-      UInt32 v = GetBe32(p - 4);
+      p += (ptrdiff_t)m * 5 - 20; // negative value is expected here.
-      v <<= 2;
+      do
-      if (encoding)
+      {
-        v += ip + (UInt32)(p - data);
+        const UInt32 t =
-      else
+          #if defined(MY_CPU_X86_OR_AMD64)
-        v -= ip + (UInt32)(p - data);
+            // we use 32-bit load here to reduce code size on x86:
-      
+            GetUi32(p);
-      v &= 0x01FFFFFF;
+          #else
-      v -= (UInt32)1 << 24;
+            GetUi16(p);
-      v ^= 0xFF000000;
+          #endif
-      v >>= 2;
+        UInt32 z = GetUi32(p + 1) >> m;
-      v |= 0x40000000;
+        p += 5;
-      SetBe32(p - 4, v);
+        if (((t >> m) & (0x70 << 1)) == 0
+            && ((z - (0x5000000 << 1)) & (0xf000000 << 1)) == 0)
+        {
+          UInt32 v = (UInt32)((0x8fffff << 1) | 1) & z;
+          z ^= v;
+        #ifdef BR_IA64_NO_INLINE
+          v |= (v & ((UInt32)1 << (23 + 1))) >> 3;
+          {
+            UInt32 c = pc;
+            BR_CONVERT_VAL(v, c)
+          }
+          v &= (0x1fffff << 1) | 1;
+        #else
+          {
+            if (encoding)
+            {
+              // pc &= ~(0xc00000 << 1); // we just need to clear at least 2 bits
+              pc &= (0x1fffff << 1) | 1;
+              v += pc;
+            }
+            else
+            {
+              // pc |= 0xc00000 << 1; // we need to set at least 2 bits
+              pc |= ~(UInt32)((0x1fffff << 1) | 1);
+              v -= pc;
+            }
+          }
+          v &= ~(UInt32)(0x600000 << 1);
+        #endif
+          v += (0x700000 << 1);
+          v &= (0x8fffff << 1) | 1;
+          z |= v;
+          z <<= m;
+          SetUi32(p + 1 - 5, z)
+        }
+        m++;
+      }
+      while (m &= 3); // while (m < 4);
     }
   }
 }
+Z7_BRANCH_FUNCS_IMP(BranchConv_IA64)
+
+
+#define BR_CONVERT_VAL_ENC(v)  v += BR_PC_GET;
+#define BR_CONVERT_VAL_DEC(v)  v -= BR_PC_GET;
+
+#if 1 && defined(MY_CPU_LE_UNALIGN)
+  #define RISCV_USE_UNALIGNED_LOAD
+#endif
+
+#ifdef RISCV_USE_UNALIGNED_LOAD
+  #define RISCV_GET_UI32(p)      GetUi32(p)
+  #define RISCV_SET_UI32(p, v)   { SetUi32(p, v) }
+#else
+  #define RISCV_GET_UI32(p) \
+    ((UInt32)GetUi16a(p) + \
+    ((UInt32)GetUi16a((p) + 2) << 16))
+  #define RISCV_SET_UI32(p, v) { \
+    SetUi16a(p, (UInt16)(v)) \
+    SetUi16a((p) + 2, (UInt16)(v >> 16)) }
+#endif
+
+#if 1 && defined(MY_CPU_LE)
+  #define RISCV_USE_16BIT_LOAD
+#endif
+
+#ifdef RISCV_USE_16BIT_LOAD
+  #define RISCV_LOAD_VAL(p)  GetUi16a(p)
+#else
+  #define RISCV_LOAD_VAL(p)  (*(p))
+#endif
+
+#define RISCV_INSTR_SIZE  2
+#define RISCV_STEP_1      (4 + RISCV_INSTR_SIZE)
+#define RISCV_STEP_2      4
+#define RISCV_REG_VAL     (2 << 7)
+#define RISCV_CMD_VAL     3
+#if 1
+  // for code size optimization:
+  #define RISCV_DELTA_7F  0x7f
+#else
+  #define RISCV_DELTA_7F  0
+#endif
+
+#define RISCV_CHECK_1(v, b) \
+    (((((b) - RISCV_CMD_VAL) ^ ((v) << 8)) & (0xf8000 + RISCV_CMD_VAL)) == 0)
+
+#if 1
+  #define RISCV_CHECK_2(v, r) \
+    ((((v) - ((RISCV_CMD_VAL << 12) | RISCV_REG_VAL | 8)) \
+           << 18) \
+     < ((r) & 0x1d))
+#else
+  // this branch gives larger code, because
+  // compilers generate larger code for big constants.
+  #define RISCV_CHECK_2(v, r) \
+    ((((v) - ((RISCV_CMD_VAL << 12) | RISCV_REG_VAL)) \
+           & ((RISCV_CMD_VAL << 12) | RISCV_REG_VAL)) \
+     < ((r) & 0x1d))
+#endif
+
+
+#define RISCV_SCAN_LOOP \
+  Byte *lim; \
+  size &= ~(SizeT)(RISCV_INSTR_SIZE - 1); \
+  if (size <= 6) return p; \
+  size -= 6; \
+  lim = p + size; \
+  BR_PC_INIT \
+  for (;;) \
+  { \
+    UInt32 a, v; \
+    /* Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE */ \
+    for (;;) \
+    { \
+      if Z7_UNLIKELY(p >= lim) { return p; } \
+      a = (RISCV_LOAD_VAL(p) ^ 0x10u) + 1; \
+      if ((a & 0x77) == 0) break; \
+      a = (RISCV_LOAD_VAL(p + RISCV_INSTR_SIZE) ^ 0x10u) + 1; \
+      p += RISCV_INSTR_SIZE * 2; \
+      if ((a & 0x77) == 0) \
+      { \
+        p -= RISCV_INSTR_SIZE; \
+        if Z7_UNLIKELY(p >= lim) { return p; } \
+        break; \
+      } \
+    }
+// (xx6f ^ 10) + 1 = xx7f + 1 = xx80       : JAL
+// (xxef ^ 10) + 1 = xxff + 1 = xx00 + 100 : JAL
+// (xx17 ^ 10) + 1 = xx07 + 1 = xx08       : AUIPC
+// (xx97 ^ 10) + 1 = xx87 + 1 = xx88       : AUIPC
+
+Byte * Z7_BRANCH_CONV_ENC(RISCV)(Byte *p, SizeT size, UInt32 pc)
+{
+  RISCV_SCAN_LOOP
+    v = a;
+    a = RISCV_GET_UI32(p);
+#ifndef RISCV_USE_16BIT_LOAD
+    v += (UInt32)p[1] << 8;
+#endif
+
+    if ((v & 8) == 0) // JAL
+    {
+      if ((v - (0x100 /* - RISCV_DELTA_7F */)) & 0xd80)
+      {
+        p += RISCV_INSTR_SIZE;
+        continue;
+      }
+      {
+        v = ((a &    1u << 31) >> 11)
+          | ((a & 0x3ff << 21) >> 20)
+          | ((a &     1 << 20) >> 9)
+          |  (a &  0xff << 12);
+        BR_CONVERT_VAL_ENC(v)
+        // ((v & 1) == 0)
+        // v: bits [1 : 20] contain offset bits
+#if 0 && defined(RISCV_USE_UNALIGNED_LOAD)
+        a &= 0xfff;
+        a |= ((UInt32)(v << 23))
+          |  ((UInt32)(v <<  7) & ((UInt32)0xff << 16))
+          |  ((UInt32)(v >>  5) & ((UInt32)0xf0 << 8));
+        RISCV_SET_UI32(p, a)
+#else // aligned
+#if 0
+        SetUi16a(p, (UInt16)(((v >> 5) & 0xf000) | (a & 0xfff)))
+#else
+        p[1] = (Byte)(((v >> 13) & 0xf0) | ((a >> 8) & 0xf));
+#endif
+
+#if 1 && defined(Z7_CPU_FAST_BSWAP_SUPPORTED) && defined(MY_CPU_LE)
+        v <<= 15;
+        v = Z7_BSWAP32(v);
+        SetUi16a(p + 2, (UInt16)v)
+#else
+        p[2] = (Byte)(v >> 9);
+        p[3] = (Byte)(v >> 1);
+#endif
+#endif // aligned
+      }
+      p += 4;
+      continue;
+    } // JAL
+
+    {
+      // AUIPC
+      if (v & 0xe80)  // (not x0) and (not x2)
+      {
+        const UInt32 b = RISCV_GET_UI32(p + 4);
+        if (RISCV_CHECK_1(v, b))
+        {
+          {
+            const UInt32 temp = (b << 12) | (0x17 + RISCV_REG_VAL);
+            RISCV_SET_UI32(p, temp)
+          }
+          a &= 0xfffff000;
+          {
+#if 1
+          const int t = -1 >> 1;
+          if (t != -1)
+            a += (b >> 20) - ((b >> 19) & 0x1000); // arithmetic right shift emulation
+          else
+#endif
+            a += (UInt32)((Int32)b >> 20); // arithmetic right shift (sign-extension).
+          }
+          BR_CONVERT_VAL_ENC(a)
+#if 1 && defined(Z7_CPU_FAST_BSWAP_SUPPORTED) && defined(MY_CPU_LE)
+          a = Z7_BSWAP32(a);
+          RISCV_SET_UI32(p + 4, a)
+#else
+          SetBe32(p + 4, a)
+#endif
+          p += 8;
+        }
+        else
+          p += RISCV_STEP_1;
+      }
+      else
+      {
+        UInt32 r = a >> 27;
+        if (RISCV_CHECK_2(v, r))
+        {
+          v = RISCV_GET_UI32(p + 4);
+          r = (r << 7) + 0x17 + (v & 0xfffff000);
+          a = (a >> 12) | (v << 20);
+          RISCV_SET_UI32(p, r)
+          RISCV_SET_UI32(p + 4, a)
+          p += 8;
+        }
+        else
+          p += RISCV_STEP_2;
+      }
+    }
+  } // for
+}
+
+
+Byte * Z7_BRANCH_CONV_DEC(RISCV)(Byte *p, SizeT size, UInt32 pc)
+{
+  RISCV_SCAN_LOOP
+#ifdef RISCV_USE_16BIT_LOAD
+    if ((a & 8) == 0)
+    {
+#else
+    v = a;
+    a += (UInt32)p[1] << 8;
+    if ((v & 8) == 0)
+    {
+#endif
+      // JAL
+      a -= 0x100 - RISCV_DELTA_7F;
+      if (a & 0xd80)
+      {
+        p += RISCV_INSTR_SIZE;
+        continue;
+      }
+      {
+        const UInt32 a_old = (a + (0xef - RISCV_DELTA_7F)) & 0xfff;
+#if 0 // unaligned
+        a = GetUi32(p);
+        v = (UInt32)(a >> 23) & ((UInt32)0xff << 1)
+          | (UInt32)(a >>  7) & ((UInt32)0xff << 9)
+#elif 1 && defined(Z7_CPU_FAST_BSWAP_SUPPORTED) && defined(MY_CPU_LE)
+        v = GetUi16a(p + 2);
+        v = Z7_BSWAP32(v) >> 15
+#else
+        v = (UInt32)p[3] << 1
+          | (UInt32)p[2] << 9
+#endif
+          | (UInt32)((a & 0xf000) << 5);
+        BR_CONVERT_VAL_DEC(v)
+        a = a_old
+          | (v << 11 &    1u << 31)
+          | (v << 20 & 0x3ff << 21)
+          | (v <<  9 &     1 << 20)
+          | (v       &  0xff << 12);
+        RISCV_SET_UI32(p, a)
+      }
+      p += 4;
+      continue;
+    } // JAL
+
+    {
+      // AUIPC
+      v = a;
+#if 1 && defined(RISCV_USE_UNALIGNED_LOAD)
+      a = GetUi32(p);
+#else
+      a |= (UInt32)GetUi16a(p + 2) << 16;
+#endif
+      if ((v & 0xe80) == 0)  // x0/x2
+      {
+        const UInt32 r = a >> 27;
+        if (RISCV_CHECK_2(v, r))
+        {
+          UInt32 b;
+#if 1 && defined(Z7_CPU_FAST_BSWAP_SUPPORTED) && defined(MY_CPU_LE)
+          b = RISCV_GET_UI32(p + 4);
+          b = Z7_BSWAP32(b);
+#else
+          b = GetBe32(p + 4);
+#endif
+          v = a >> 12;
+          BR_CONVERT_VAL_DEC(b)
+          a = (r << 7) + 0x17;
+          a += (b + 0x800) & 0xfffff000;
+          v |= b << 20;
+          RISCV_SET_UI32(p, a)
+          RISCV_SET_UI32(p + 4, v)
+          p += 8;
+        }
+        else
+          p += RISCV_STEP_2;
+      }
+      else
+      {
+        const UInt32 b = RISCV_GET_UI32(p + 4);
+        if (!RISCV_CHECK_1(v, b))
+          p += RISCV_STEP_1;
+        else
+        {
+          v = (a & 0xfffff000) | (b >> 20);
+          a = (b << 12) | (0x17 + RISCV_REG_VAL);
+          RISCV_SET_UI32(p, a)
+          RISCV_SET_UI32(p + 4, v)
+          p += 8;
+        }
+      }
+    }
+  } // for
+}

C/Bra.h

@@ -1,64 +1,105 @@
 /* Bra.h -- Branch converters for executables
-2013-01-18 : Igor Pavlov : Public domain */
+2024-01-20 : Igor Pavlov : Public domain */
 
-#ifndef __BRA_H
+#ifndef ZIP7_INC_BRA_H
-#define __BRA_H
+#define ZIP7_INC_BRA_H
 
 #include "7zTypes.h"
 
 EXTERN_C_BEGIN
 
+/* #define PPC BAD_PPC_11 // for debug */
+
+#define Z7_BRANCH_CONV_DEC_2(name)  z7_ ## name ## _Dec
+#define Z7_BRANCH_CONV_ENC_2(name)  z7_ ## name ## _Enc
+#define Z7_BRANCH_CONV_DEC(name)    Z7_BRANCH_CONV_DEC_2(BranchConv_ ## name)
+#define Z7_BRANCH_CONV_ENC(name)    Z7_BRANCH_CONV_ENC_2(BranchConv_ ## name)
+#define Z7_BRANCH_CONV_ST_DEC(name) z7_BranchConvSt_ ## name ## _Dec
+#define Z7_BRANCH_CONV_ST_ENC(name) z7_BranchConvSt_ ## name ## _Enc
+
+#define Z7_BRANCH_CONV_DECL(name)    Byte * name(Byte *data, SizeT size, UInt32 pc)
+#define Z7_BRANCH_CONV_ST_DECL(name) Byte * name(Byte *data, SizeT size, UInt32 pc, UInt32 *state)
+
+typedef Z7_BRANCH_CONV_DECL(   (*z7_Func_BranchConv));
+typedef Z7_BRANCH_CONV_ST_DECL((*z7_Func_BranchConvSt));
+
+#define Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL 0
+Z7_BRANCH_CONV_ST_DECL (Z7_BRANCH_CONV_ST_DEC(X86));
+Z7_BRANCH_CONV_ST_DECL (Z7_BRANCH_CONV_ST_ENC(X86));
+
+#define Z7_BRANCH_FUNCS_DECL(name) \
+Z7_BRANCH_CONV_DECL (Z7_BRANCH_CONV_DEC_2(name)); \
+Z7_BRANCH_CONV_DECL (Z7_BRANCH_CONV_ENC_2(name));
+
+Z7_BRANCH_FUNCS_DECL (BranchConv_ARM64)
+Z7_BRANCH_FUNCS_DECL (BranchConv_ARM)
+Z7_BRANCH_FUNCS_DECL (BranchConv_ARMT)
+Z7_BRANCH_FUNCS_DECL (BranchConv_PPC)
+Z7_BRANCH_FUNCS_DECL (BranchConv_SPARC)
+Z7_BRANCH_FUNCS_DECL (BranchConv_IA64)
+Z7_BRANCH_FUNCS_DECL (BranchConv_RISCV)
+
 /*
-These functions convert relative addresses to absolute addresses
+These functions convert data that contain CPU instructions.
-in CALL instructions to increase the compression ratio.
+Each such function converts relative addresses to absolute addresses in some
+branch instructions: CALL (in all converters) and JUMP (X86 converter only).
+Such conversion allows to increase compression ratio, if we compress that data.
 
-  In:
+There are 2 types of converters:
-    data     - data buffer
+  Byte * Conv_RISC (Byte *data, SizeT size, UInt32 pc);
-    size     - size of data
+  Byte * ConvSt_X86(Byte *data, SizeT size, UInt32 pc, UInt32 *state);
-    ip       - current virtual Instruction Pinter (IP) value
+Each Converter supports 2 versions: one for encoding
-    state    - state variable for x86 converter
+and one for decoding (_Enc/_Dec postfixes in function name).
-    encoding - 0 (for decoding), 1 (for encoding)
 
-  Out:
+In params:
-    state    - state variable for x86 converter
+  data  : data buffer
+  size  : size of data
+  pc    : current virtual Program Counter (Instruction Pointer) value
+In/Out param:
+  state : pointer to state variable (for X86 converter only)
 
-  Returns:
+Return:
-    The number of processed bytes. If you call these functions with multiple calls,
+  The pointer to position in (data) buffer after last byte that was processed.
-    you must start next call with first byte after block of processed bytes.
+  If the caller calls converter again, it must call it starting with that position.
+  But the caller is allowed to move data in buffer. So pointer to
+  current processed position also will be changed for next call.
+  Also the caller must increase internal (pc) value for next call.
   
+Each converter has some characteristics: Endian, Alignment, LookAhead.
   Type   Endian  Alignment  LookAhead
   
-  x86    little      1          4
+  X86    little      1          4
   ARMT   little      2          2
+  RISCV  little      2          6
   ARM    little      4          0
+  ARM64  little      4          0
   PPC     big        4          0
   SPARC   big        4          0
   IA64   little     16          0
 
-  size must be >= Alignment + LookAhead, if it's not last block.
+  (data) must be aligned for (Alignment).
-  If (size < Alignment + LookAhead), converter returns 0.
+  processed size can be calculated as:
+    SizeT processed = Conv(data, size, pc) - data;
+  if (processed == 0)
+    it means that converter needs more data for processing.
+  If (size < Alignment + LookAhead)
+    then (processed == 0) is allowed.
 
-  Example:
+Example code for conversion in loop:
-
+  UInt32 pc = 0;
-    UInt32 ip = 0;
+  size = 0;
-    for ()
+  for (;;)
-    {
+  {
-      ; size must be >= Alignment + LookAhead, if it's not last block
+    size += Load_more_input_data(data + size);
-      SizeT processed = Convert(data, size, ip, 1);
+    SizeT processed = Conv(data, size, pc) - data;
-      data += processed;
+    if (processed == 0 && no_more_input_data_after_size)
-      size -= processed;
+      break; // we stop convert loop
-      ip += processed;
+    data += processed;
-    }
+    size -= processed;
+    pc += processed;
+  }
 */
 
-#define x86_Convert_Init(state) { state = 0; }
-SizeT x86_Convert(Byte *data, SizeT size, UInt32 ip, UInt32 *state, int encoding);
-SizeT ARM_Convert(Byte *data, SizeT size, UInt32 ip, int encoding);
-SizeT ARMT_Convert(Byte *data, SizeT size, UInt32 ip, int encoding);
-SizeT PPC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding);
-SizeT SPARC_Convert(Byte *data, SizeT size, UInt32 ip, int encoding);
-SizeT IA64_Convert(Byte *data, SizeT size, UInt32 ip, int encoding);
-
 EXTERN_C_END
 
 #endif

C/Bra86.c

@@ -1,82 +1,187 @@
-/* Bra86.c -- Converter for x86 code (BCJ)
+/* Bra86.c -- Branch converter for X86 code (BCJ)
-2021-02-09 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include "Bra.h"
+#include "CpuArch.h"
 
-#define Test86MSByte(b) ((((b) + 1) & 0xFE) == 0)
 
-SizeT x86_Convert(Byte *data, SizeT size, UInt32 ip, UInt32 *state, int encoding)
+#if defined(MY_CPU_SIZEOF_POINTER) \
+    && ( MY_CPU_SIZEOF_POINTER == 4 \
+      || MY_CPU_SIZEOF_POINTER == 8)
+  #define BR_CONV_USE_OPT_PC_PTR
+#endif
+
+#ifdef BR_CONV_USE_OPT_PC_PTR
+#define BR_PC_INIT  pc -= (UInt32)(SizeT)p; // (MY_uintptr_t)
+#define BR_PC_GET   (pc + (UInt32)(SizeT)p)
+#else
+#define BR_PC_INIT  pc += (UInt32)size;
+#define BR_PC_GET   (pc - (UInt32)(SizeT)(lim - p))
+// #define BR_PC_INIT
+// #define BR_PC_GET   (pc + (UInt32)(SizeT)(p - data))
+#endif
+
+#define BR_CONVERT_VAL(v, c) if (encoding) v += c; else v -= c;
+// #define BR_CONVERT_VAL(v, c) if (!encoding) c = (UInt32)0 - c; v += c;
+
+#define Z7_BRANCH_CONV_ST(name) z7_BranchConvSt_ ## name
+
+#define BR86_NEED_CONV_FOR_MS_BYTE(b) ((((b) + 1) & 0xfe) == 0)
+
+#ifdef MY_CPU_LE_UNALIGN
+  #define BR86_PREPARE_BCJ_SCAN  const UInt32 v = GetUi32(p) ^ 0xe8e8e8e8;
+  #define BR86_IS_BCJ_BYTE(n)    ((v & ((UInt32)0xfe << (n) * 8)) == 0)
+#else
+  #define BR86_PREPARE_BCJ_SCAN
+  // bad for MSVC X86 (partial write to byte reg):
+  #define BR86_IS_BCJ_BYTE(n)    ((p[n - 4] & 0xfe) == 0xe8)
+  // bad for old MSVC (partial write to byte reg):
+  // #define BR86_IS_BCJ_BYTE(n)    (((*p ^ 0xe8) & 0xfe) == 0)
+#endif
+ 
+static
+Z7_FORCE_INLINE
+Z7_ATTRIB_NO_VECTOR
+Byte *Z7_BRANCH_CONV_ST(X86)(Byte *p, SizeT size, UInt32 pc, UInt32 *state, int encoding)
 {
-  SizeT pos = 0;
-  UInt32 mask = *state & 7;
   if (size < 5)
-    return 0;
+    return p;
-  size -= 4;
+ {
-  ip += 5;
+  // Byte *p = data;
+  const Byte *lim = p + size - 4;
+  unsigned mask = (unsigned)*state;  // & 7;
+#ifdef BR_CONV_USE_OPT_PC_PTR
+  /* if BR_CONV_USE_OPT_PC_PTR is defined: we need to adjust (pc) for (+4),
+        because call/jump offset is relative to the next instruction.
+     if BR_CONV_USE_OPT_PC_PTR is not defined : we don't need to adjust (pc) for (+4),
+         because  BR_PC_GET uses (pc - (lim - p)), and lim was adjusted for (-4) before.
+  */
+  pc += 4;
+#endif
+  BR_PC_INIT
+  goto start;
 
-  for (;;)
+  for (;; mask |= 4)
   {
-    Byte *p = data + pos;
+    // cont: mask |= 4;
-    const Byte *limit = data + size;
+  start:
-    for (; p < limit; p++)
+    if (p >= lim)
-      if ((*p & 0xFE) == 0xE8)
+      goto fin;
-        break;
-
     {
-      SizeT d = (SizeT)(p - data) - pos;
+      BR86_PREPARE_BCJ_SCAN
-      pos = (SizeT)(p - data);
+      p += 4;
-      if (p >= limit)
+      if (BR86_IS_BCJ_BYTE(0))  { goto m0; }  mask >>= 1;
-      {
+      if (BR86_IS_BCJ_BYTE(1))  { goto m1; }  mask >>= 1;
-        *state = (d > 2 ? 0 : mask >> (unsigned)d);
+      if (BR86_IS_BCJ_BYTE(2))  { goto m2; }  mask = 0;
-        return pos;
+      if (BR86_IS_BCJ_BYTE(3))  { goto a3; }
-      }
-      if (d > 2)
-        mask = 0;
-      else
-      {
-        mask >>= (unsigned)d;
-        if (mask != 0 && (mask > 4 || mask == 3 || Test86MSByte(p[(size_t)(mask >> 1) + 1])))
-        {
-          mask = (mask >> 1) | 4;
-          pos++;
-          continue;
-        }
-      }
     }
+    goto main_loop;
 
-    if (Test86MSByte(p[4]))
+  m0: p--;
+  m1: p--;
+  m2: p--;
+    if (mask == 0)
+      goto a3;
+    if (p > lim)
+      goto fin_p;
+   
+    // if (((0x17u >> mask) & 1) == 0)
+    if (mask > 4 || mask == 3)
     {
-      UInt32 v = ((UInt32)p[4] << 24) | ((UInt32)p[3] << 16) | ((UInt32)p[2] << 8) | ((UInt32)p[1]);
+      mask >>= 1;
-      UInt32 cur = ip + (UInt32)pos;
+      continue; // goto cont;
-      pos += 5;
+    }
-      if (encoding)
+    mask >>= 1;
-        v += cur;
+    if (BR86_NEED_CONV_FOR_MS_BYTE(p[mask]))
-      else
+      continue; // goto cont;
-        v -= cur;
+    // if (!BR86_NEED_CONV_FOR_MS_BYTE(p[3])) continue; // goto cont;
-      if (mask != 0)
+    {
+      UInt32 v = GetUi32(p);
+      UInt32 c;
+      v += (1 << 24);  if (v & 0xfe000000) continue; // goto cont;
+      c = BR_PC_GET;
+      BR_CONVERT_VAL(v, c)
       {
-        unsigned sh = (mask & 6) << 2;
+        mask <<= 3;
-        if (Test86MSByte((Byte)(v >> sh)))
+        if (BR86_NEED_CONV_FOR_MS_BYTE(v >> mask))
         {
-          v ^= (((UInt32)0x100 << sh) - 1);
+          v ^= (((UInt32)0x100 << mask) - 1);
-          if (encoding)
+          #ifdef MY_CPU_X86
-            v += cur;
+          // for X86 : we can recalculate (c) to reduce register pressure
-          else
+            c = BR_PC_GET;
-            v -= cur;
+          #endif
+          BR_CONVERT_VAL(v, c)
         }
         mask = 0;
       }
-      p[1] = (Byte)v;
+      // v = (v & ((1 << 24) - 1)) - (v & (1 << 24));
-      p[2] = (Byte)(v >> 8);
+      v &= (1 << 25) - 1;  v -= (1 << 24);
-      p[3] = (Byte)(v >> 16);
+      SetUi32(p, v)
-      p[4] = (Byte)(0 - ((v >> 24) & 1));
+      p += 4;
+      goto main_loop;
     }
-    else
+
+  main_loop:
+    if (p >= lim)
+      goto fin;
+    for (;;)
     {
-      mask = (mask >> 1) | 4;
+      BR86_PREPARE_BCJ_SCAN
-      pos++;
+      p += 4;
+      if (BR86_IS_BCJ_BYTE(0))  { goto a0; }
+      if (BR86_IS_BCJ_BYTE(1))  { goto a1; }
+      if (BR86_IS_BCJ_BYTE(2))  { goto a2; }
+      if (BR86_IS_BCJ_BYTE(3))  { goto a3; }
+      if (p >= lim)
+        goto fin;
+    }
+  
+  a0: p--;
+  a1: p--;
+  a2: p--;
+  a3:
+    if (p > lim)
+      goto fin_p;
+    // if (!BR86_NEED_CONV_FOR_MS_BYTE(p[3])) continue; // goto cont;
+    {
+      UInt32 v = GetUi32(p);
+      UInt32 c;
+      v += (1 << 24);  if (v & 0xfe000000) continue; // goto cont;
+      c = BR_PC_GET;
+      BR_CONVERT_VAL(v, c)
+      // v = (v & ((1 << 24) - 1)) - (v & (1 << 24));
+      v &= (1 << 25) - 1;  v -= (1 << 24);
+      SetUi32(p, v)
+      p += 4;
+      goto main_loop;
     }
   }
+
+fin_p:
+  p--;
+fin:
+  // the following processing for tail is optional and can be commented
+  /*
+  lim += 4;
+  for (; p < lim; p++, mask >>= 1)
+    if ((*p & 0xfe) == 0xe8)
+      break;
+  */
+  *state = (UInt32)mask;
+  return p;
+ }
 }
+
+
+#define Z7_BRANCH_CONV_ST_FUNC_IMP(name, m, encoding) \
+Z7_NO_INLINE \
+Z7_ATTRIB_NO_VECTOR \
+Byte *m(name)(Byte *data, SizeT size, UInt32 pc, UInt32 *state) \
+  { return Z7_BRANCH_CONV_ST(name)(data, size, pc, state, encoding); }
+
+Z7_BRANCH_CONV_ST_FUNC_IMP(X86, Z7_BRANCH_CONV_ST_DEC, 0)
+#ifndef Z7_EXTRACT_ONLY
+Z7_BRANCH_CONV_ST_FUNC_IMP(X86, Z7_BRANCH_CONV_ST_ENC, 1)
+#endif

C/BraIA64.c

@@ -1,53 +1,14 @@
 /* BraIA64.c -- Converter for IA-64 code
-2017-01-26 : Igor Pavlov : Public domain */
+2023-02-20 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
-#include "CpuArch.h"
+// the code was moved to Bra.c
-#include "Bra.h"
 
-SizeT IA64_Convert(Byte *data, SizeT size, UInt32 ip, int encoding)
+#ifdef _MSC_VER
-{
+#pragma warning(disable : 4206) // nonstandard extension used : translation unit is empty
-  SizeT i;
+#endif
-  if (size < 16)
-    return 0;
-  size -= 16;
-  i = 0;
-  do
-  {
-    unsigned m = ((UInt32)0x334B0000 >> (data[i] & 0x1E)) & 3;
-    if (m)
-    {
-      m++;
-      do
-      {
-        Byte *p = data + (i + (size_t)m * 5 - 8);
-        if (((p[3] >> m) & 15) == 5
-            && (((p[-1] | ((UInt32)p[0] << 8)) >> m) & 0x70) == 0)
-        {
-          unsigned raw = GetUi32(p);
-          unsigned v = raw >> m;
-          v = (v & 0xFFFFF) | ((v & (1 << 23)) >> 3);
 
-          v <<= 4;
+#if defined(__clang__)
-          if (encoding)
+#pragma GCC diagnostic ignored "-Wempty-translation-unit"
-            v += ip + (UInt32)i;
+#endif
-          else
-            v -= ip + (UInt32)i;
-          v >>= 4;
-          
-          v &= 0x1FFFFF;
-          v += 0x700000;
-          v &= 0x8FFFFF;
-          raw &= ~((UInt32)0x8FFFFF << m);
-          raw |= (v << m);
-          SetUi32(p, raw);
-        }
-      }
-      while (++m <= 4);
-    }
-    i += 16;
-  }
-  while (i <= size);
-  return i;
-}

C/BwtSort.c

@@ -1,5 +1,5 @@
 /* BwtSort.c -- BWT block sorting
-2021-04-01 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -7,8 +7,44 @@
 #include "Sort.h"
 
 /* #define BLOCK_SORT_USE_HEAP_SORT */
+// #define BLOCK_SORT_USE_HEAP_SORT
+
+#ifdef BLOCK_SORT_USE_HEAP_SORT
+
+#define HeapSortRefDown(p, vals, n, size, temp) \
+  { size_t k = n; UInt32 val = vals[temp]; for (;;) { \
+    size_t s = k << 1; \
+    if (s > size) break; \
+    if (s < size && vals[p[s + 1]] > vals[p[s]]) s++; \
+    if (val >= vals[p[s]]) break; \
+    p[k] = p[s]; k = s; \
+  } p[k] = temp; }
+
+void HeapSortRef(UInt32 *p, UInt32 *vals, size_t size)
+{
+  if (size <= 1)
+    return;
+  p--;
+  {
+    size_t i = size / 2;
+    do
+    {
+      UInt32 temp = p[i];
+      HeapSortRefDown(p, vals, i, size, temp);
+    }
+    while (--i != 0);
+  }
+  do
+  {
+    UInt32 temp = p[size];
+    p[size--] = p[1];
+    HeapSortRefDown(p, vals, 1, size, temp);
+  }
+  while (size > 1);
+}
+
+#endif // BLOCK_SORT_USE_HEAP_SORT
 
-#define NO_INLINE MY_FAST_CALL
 
 /* Don't change it !!! */
 #define kNumHashBytes 2
@@ -29,26 +65,27 @@
 
 #else
 
-#define kNumBitsMax 20
+#define kNumBitsMax     20
-#define kIndexMask ((1 << kNumBitsMax) - 1)
+#define kIndexMask      (((UInt32)1 << kNumBitsMax) - 1)
-#define kNumExtraBits (32 - kNumBitsMax)
+#define kNumExtraBits   (32 - kNumBitsMax)
-#define kNumExtra0Bits (kNumExtraBits - 2)
+#define kNumExtra0Bits  (kNumExtraBits - 2)
-#define kNumExtra0Mask ((1 << kNumExtra0Bits) - 1)
+#define kNumExtra0Mask  ((1 << kNumExtra0Bits) - 1)
 
 #define SetFinishedGroupSize(p, size) \
-  {  *(p) |= ((((size) - 1) & kNumExtra0Mask) << kNumBitsMax); \
+  {  *(p) |= ((((UInt32)(size) - 1) & kNumExtra0Mask) << kNumBitsMax); \
     if ((size) > (1 << kNumExtra0Bits)) { \
-    *(p) |= 0x40000000;  *((p) + 1) |= ((((size) - 1)>> kNumExtra0Bits) << kNumBitsMax); } } \
+      *(p) |= 0x40000000; \
+      *((p) + 1) |= (((UInt32)(size) - 1) >> kNumExtra0Bits) << kNumBitsMax; } } \
 
-static void SetGroupSize(UInt32 *p, UInt32 size)
+static void SetGroupSize(UInt32 *p, size_t size)
 {
   if (--size == 0)
     return;
-  *p |= 0x80000000 | ((size & kNumExtra0Mask) << kNumBitsMax);
+  *p |= 0x80000000 | (((UInt32)size & kNumExtra0Mask) << kNumBitsMax);
   if (size >= (1 << kNumExtra0Bits))
   {
     *p |= 0x40000000;
-    p[1] |= ((size >> kNumExtra0Bits) << kNumBitsMax);
+    p[1] |= (((UInt32)size >> kNumExtra0Bits) << kNumBitsMax);
   }
 }
 
@@ -60,10 +97,15 @@ SortGroup - is recursive Range-Sort function with HeapSort optimization for smal
 returns: 1 - if there are groups, 0 - no more groups
 */
 
-static UInt32 NO_INLINE SortGroup(UInt32 BlockSize, UInt32 NumSortedBytes, UInt32 groupOffset, UInt32 groupSize, int NumRefBits, UInt32 *Indices
+static
-  #ifndef BLOCK_SORT_USE_HEAP_SORT
+unsigned
-  , UInt32 left, UInt32 range
+Z7_FASTCALL
-  #endif
+SortGroup(size_t BlockSize, size_t NumSortedBytes,
+    size_t groupOffset, size_t groupSize,
+    unsigned NumRefBits, UInt32 *Indices
+#ifndef BLOCK_SORT_USE_HEAP_SORT
+    , size_t left, size_t range
+#endif
   )
 {
   UInt32 *ind2 = Indices + groupOffset;
@@ -72,96 +114,99 @@ static UInt32 NO_INLINE SortGroup(UInt32 BlockSize, UInt32 NumSortedBytes, UInt3
   {
     /*
     #ifndef BLOCK_SORT_EXTERNAL_FLAGS
-    SetFinishedGroupSize(ind2, 1);
+    SetFinishedGroupSize(ind2, 1)
     #endif
     */
     return 0;
   }
   Groups = Indices + BlockSize + BS_TEMP_SIZE;
-  if (groupSize <= ((UInt32)1 << NumRefBits)
+  if (groupSize <= ((size_t)1 << NumRefBits)
-      #ifndef BLOCK_SORT_USE_HEAP_SORT
+#ifndef BLOCK_SORT_USE_HEAP_SORT
       && groupSize <= range
-      #endif
+#endif
       )
   {
     UInt32 *temp = Indices + BlockSize;
-    UInt32 j;
+    size_t j, group;
-    UInt32 mask, thereAreGroups, group, cg;
+    UInt32 mask, cg;
+    unsigned thereAreGroups;
     {
       UInt32 gPrev;
       UInt32 gRes = 0;
       {
-        UInt32 sp = ind2[0] + NumSortedBytes;
+        size_t sp = ind2[0] + NumSortedBytes;
-        if (sp >= BlockSize) sp -= BlockSize;
+        if (sp >= BlockSize)
+            sp -= BlockSize;
         gPrev = Groups[sp];
-        temp[0] = (gPrev << NumRefBits);
+        temp[0] = gPrev << NumRefBits;
       }
       
       for (j = 1; j < groupSize; j++)
       {
-        UInt32 sp = ind2[j] + NumSortedBytes;
+        size_t sp = ind2[j] + NumSortedBytes;
         UInt32 g;
-        if (sp >= BlockSize) sp -= BlockSize;
+        if (sp >= BlockSize)
+            sp -= BlockSize;
         g = Groups[sp];
-        temp[j] = (g << NumRefBits) | j;
+        temp[j] = (g << NumRefBits) | (UInt32)j;
         gRes |= (gPrev ^ g);
       }
       if (gRes == 0)
       {
-        #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
         SetGroupSize(ind2, groupSize);
-        #endif
+#endif
         return 1;
       }
     }
     
     HeapSort(temp, groupSize);
-    mask = (((UInt32)1 << NumRefBits) - 1);
+    mask = ((UInt32)1 << NumRefBits) - 1;
     thereAreGroups = 0;
     
     group = groupOffset;
-    cg = (temp[0] >> NumRefBits);
+    cg = temp[0] >> NumRefBits;
     temp[0] = ind2[temp[0] & mask];
 
     {
-    #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
     UInt32 *Flags = Groups + BlockSize;
-    #else
+#else
-    UInt32 prevGroupStart = 0;
+    size_t prevGroupStart = 0;
-    #endif
+#endif
     
     for (j = 1; j < groupSize; j++)
     {
-      UInt32 val = temp[j];
+      const UInt32 val = temp[j];
-      UInt32 cgCur = (val >> NumRefBits);
+      const UInt32 cgCur = val >> NumRefBits;
       
       if (cgCur != cg)
       {
         cg = cgCur;
         group = groupOffset + j;
 
-        #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
         {
-        UInt32 t = group - 1;
+          const size_t t = group - 1;
-        Flags[t >> kNumFlagsBits] &= ~(1 << (t & kFlagsMask));
+          Flags[t >> kNumFlagsBits] &= ~((UInt32)1 << (t & kFlagsMask));
         }
-        #else
+#else
         SetGroupSize(temp + prevGroupStart, j - prevGroupStart);
         prevGroupStart = j;
-        #endif
+#endif
       }
       else
         thereAreGroups = 1;
       {
-      UInt32 ind = ind2[val & mask];
+        const UInt32 ind = ind2[val & mask];
-      temp[j] = ind;
+        temp[j] = ind;
-      Groups[ind] = group;
+        Groups[ind] = (UInt32)group;
       }
     }
 
-    #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
     SetGroupSize(temp + prevGroupStart, j - prevGroupStart);
-    #endif
+#endif
     }
 
     for (j = 0; j < groupSize; j++)
@@ -171,37 +216,42 @@ static UInt32 NO_INLINE SortGroup(UInt32 BlockSize, UInt32 NumSortedBytes, UInt3
 
   /* Check that all strings are in one group (cannot sort) */
   {
-    UInt32 group, j;
+    UInt32 group;
-    UInt32 sp = ind2[0] + NumSortedBytes; if (sp >= BlockSize) sp -= BlockSize;
+    size_t j;
+    size_t sp = ind2[0] + NumSortedBytes;
+    if (sp >= BlockSize)
+        sp -= BlockSize;
     group = Groups[sp];
     for (j = 1; j < groupSize; j++)
     {
-      sp = ind2[j] + NumSortedBytes; if (sp >= BlockSize) sp -= BlockSize;
+      sp = ind2[j] + NumSortedBytes;
+      if (sp >= BlockSize)
+          sp -= BlockSize;
       if (Groups[sp] != group)
         break;
     }
     if (j == groupSize)
     {
-      #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
       SetGroupSize(ind2, groupSize);
-      #endif
+#endif
       return 1;
     }
   }
 
-  #ifndef BLOCK_SORT_USE_HEAP_SORT
+#ifndef BLOCK_SORT_USE_HEAP_SORT
   {
   /* ---------- Range Sort ---------- */
-  UInt32 i;
+  size_t i;
-  UInt32 mid;
+  size_t mid;
   for (;;)
   {
-    UInt32 j;
+    size_t j;
     if (range <= 1)
     {
-      #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
       SetGroupSize(ind2, groupSize);
-      #endif
+#endif
       return 1;
     }
     mid = left + ((range + 1) >> 1);
@@ -209,7 +259,7 @@ static UInt32 NO_INLINE SortGroup(UInt32 BlockSize, UInt32 NumSortedBytes, UInt3
     i = 0;
     do
     {
-      UInt32 sp = ind2[i] + NumSortedBytes; if (sp >= BlockSize) sp -= BlockSize;
+      size_t sp = ind2[i] + NumSortedBytes; if (sp >= BlockSize) sp -= BlockSize;
       if (Groups[sp] >= mid)
       {
         for (j--; j > i; j--)
@@ -237,51 +287,53 @@ static UInt32 NO_INLINE SortGroup(UInt32 BlockSize, UInt32 NumSortedBytes, UInt3
       break;
   }
 
-  #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
   {
-    UInt32 t = (groupOffset + i - 1);
+    const size_t t = groupOffset + i - 1;
     UInt32 *Flags = Groups + BlockSize;
-    Flags[t >> kNumFlagsBits] &= ~(1 << (t & kFlagsMask));
+    Flags[t >> kNumFlagsBits] &= ~((UInt32)1 << (t & kFlagsMask));
   }
-  #endif
+#endif
 
   {
-    UInt32 j;
+    size_t j;
     for (j = i; j < groupSize; j++)
-      Groups[ind2[j]] = groupOffset + i;
+      Groups[ind2[j]] = (UInt32)(groupOffset + i);
   }
 
   {
-  UInt32 res = SortGroup(BlockSize, NumSortedBytes, groupOffset, i, NumRefBits, Indices, left, mid - left);
+    unsigned res = SortGroup(BlockSize, NumSortedBytes, groupOffset, i, NumRefBits, Indices, left, mid - left);
-  return res | SortGroup(BlockSize, NumSortedBytes, groupOffset + i, groupSize - i, NumRefBits, Indices, mid, range - (mid - left));
+    return   res | SortGroup(BlockSize, NumSortedBytes, groupOffset + i, groupSize - i, NumRefBits, Indices, mid, range - (mid - left));
   }
 
   }
 
-  #else
+#else // BLOCK_SORT_USE_HEAP_SORT
 
   /* ---------- Heap Sort ---------- */
 
   {
-    UInt32 j;
+    size_t j;
     for (j = 0; j < groupSize; j++)
     {
-      UInt32 sp = ind2[j] + NumSortedBytes; if (sp >= BlockSize) sp -= BlockSize;
+      size_t sp = ind2[j] + NumSortedBytes;
-      ind2[j] = sp;
+      if (sp >= BlockSize)
+          sp -= BlockSize;
+      ind2[j] = (UInt32)sp;
     }
 
     HeapSortRef(ind2, Groups, groupSize);
 
     /* Write Flags */
     {
-    UInt32 sp = ind2[0];
+    size_t sp = ind2[0];
     UInt32 group = Groups[sp];
 
-    #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
     UInt32 *Flags = Groups + BlockSize;
-    #else
+#else
-    UInt32 prevGroupStart = 0;
+    size_t prevGroupStart = 0;
-    #endif
+#endif
 
     for (j = 1; j < groupSize; j++)
     {
@@ -289,149 +341,210 @@ static UInt32 NO_INLINE SortGroup(UInt32 BlockSize, UInt32 NumSortedBytes, UInt3
       if (Groups[sp] != group)
       {
         group = Groups[sp];
-        #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
         {
-        UInt32 t = groupOffset + j - 1;
+        const size_t t = groupOffset + j - 1;
         Flags[t >> kNumFlagsBits] &= ~(1 << (t & kFlagsMask));
         }
-        #else
+#else
         SetGroupSize(ind2 + prevGroupStart, j - prevGroupStart);
         prevGroupStart = j;
-        #endif
+#endif
       }
     }
 
-    #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
     SetGroupSize(ind2 + prevGroupStart, j - prevGroupStart);
-    #endif
+#endif
     }
     {
     /* Write new Groups values and Check that there are groups */
-    UInt32 thereAreGroups = 0;
+    unsigned thereAreGroups = 0;
     for (j = 0; j < groupSize; j++)
     {
-      UInt32 group = groupOffset + j;
+      size_t group = groupOffset + j;
-      #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
       UInt32 subGroupSize = ((ind2[j] & ~0xC0000000) >> kNumBitsMax);
-      if ((ind2[j] & 0x40000000) != 0)
+      if (ind2[j] & 0x40000000)
         subGroupSize += ((ind2[(size_t)j + 1] >> kNumBitsMax) << kNumExtra0Bits);
       subGroupSize++;
       for (;;)
       {
-        UInt32 original = ind2[j];
+        const UInt32 original = ind2[j];
-        UInt32 sp = original & kIndexMask;
+        size_t sp = original & kIndexMask;
-        if (sp < NumSortedBytes) sp += BlockSize; sp -= NumSortedBytes;
+        if (sp < NumSortedBytes)
-        ind2[j] = sp | (original & ~kIndexMask);
+          sp += BlockSize;
-        Groups[sp] = group;
+        sp -= NumSortedBytes;
+        ind2[j] = (UInt32)sp | (original & ~kIndexMask);
+        Groups[sp] = (UInt32)group;
         if (--subGroupSize == 0)
           break;
         j++;
         thereAreGroups = 1;
       }
-      #else
+#else
       UInt32 *Flags = Groups + BlockSize;
       for (;;)
       {
-        UInt32 sp = ind2[j]; if (sp < NumSortedBytes) sp += BlockSize; sp -= NumSortedBytes;
+        size_t sp = ind2[j];
-        ind2[j] = sp;
+        if (sp < NumSortedBytes)
-        Groups[sp] = group;
+          sp += BlockSize;
+        sp -= NumSortedBytes;
+        ind2[j] = (UInt32)sp;
+        Groups[sp] = (UInt32)group;
         if ((Flags[(groupOffset + j) >> kNumFlagsBits] & (1 << ((groupOffset + j) & kFlagsMask))) == 0)
           break;
         j++;
         thereAreGroups = 1;
       }
-      #endif
+#endif
     }
     return thereAreGroups;
     }
   }
-  #endif
+#endif // BLOCK_SORT_USE_HEAP_SORT
 }
 
+
 /* conditions: blockSize > 0 */
-UInt32 BlockSort(UInt32 *Indices, const Byte *data, UInt32 blockSize)
+UInt32 BlockSort(UInt32 *Indices, const Byte *data, size_t blockSize)
 {
   UInt32 *counters = Indices + blockSize;
-  UInt32 i;
+  size_t i;
   UInt32 *Groups;
-  #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
   UInt32 *Flags;
-  #endif
+#endif
 
-  /* Radix-Sort for 2 bytes */
+/* Radix-Sort for 2 bytes */
+// { UInt32 yyy; for (yyy = 0; yyy < 100; yyy++) {
   for (i = 0; i < kNumHashValues; i++)
     counters[i] = 0;
-  for (i = 0; i < blockSize - 1; i++)
+  {
-    counters[((UInt32)data[i] << 8) | data[(size_t)i + 1]]++;
+    const Byte *data2 = data;
-  counters[((UInt32)data[i] << 8) | data[0]]++;
+    size_t a = data[(size_t)blockSize - 1];
+    const Byte *data_lim = data + blockSize;
+    if (blockSize >= 4)
+    {
+      data_lim -= 3;
+      do
+      {
+        size_t b;
+        b = data2[0]; counters[(a << 8) | b]++;
+        a = data2[1]; counters[(b << 8) | a]++;
+        b = data2[2]; counters[(a << 8) | b]++;
+        a = data2[3]; counters[(b << 8) | a]++;
+        data2 += 4;
+      }
+      while (data2 < data_lim);
+      data_lim += 3;
+    }
+    while (data2 != data_lim)
+    {
+      size_t b = *data2++;
+      counters[(a << 8) | b]++;
+      a = b;
+    }
+  }
+// }}
 
   Groups = counters + BS_TEMP_SIZE;
-  #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
   Flags = Groups + blockSize;
-    {
+  {
-      UInt32 numWords = (blockSize + kFlagsMask) >> kNumFlagsBits;
+    const size_t numWords = (blockSize + kFlagsMask) >> kNumFlagsBits;
-      for (i = 0; i < numWords; i++)
+    for (i = 0; i < numWords; i++)
-        Flags[i] = kAllFlags;
+      Flags[i] = kAllFlags;
-    }
+  }
-  #endif
+#endif
 
   {
     UInt32 sum = 0;
     for (i = 0; i < kNumHashValues; i++)
     {
-      UInt32 groupSize = counters[i];
+      const UInt32 groupSize = counters[i];
-      if (groupSize > 0)
+      counters[i] = sum;
+      sum += groupSize;
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
+      if (groupSize)
       {
-        #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+        const UInt32 t = sum - 1;
-        UInt32 t = sum + groupSize - 1;
+        Flags[t >> kNumFlagsBits] &= ~((UInt32)1 << (t & kFlagsMask));
-        Flags[t >> kNumFlagsBits] &= ~(1 << (t & kFlagsMask));
-        #endif
-        sum += groupSize;
       }
-      counters[i] = sum - groupSize;
+#endif
+    }
+  }
+
+  for (i = 0; i < blockSize - 1; i++)
+    Groups[i] = counters[((unsigned)data[i] << 8) | data[(size_t)i + 1]];
+  Groups[i] = counters[((unsigned)data[i] << 8) | data[0]];
+  
+  {
+#define SET_Indices(a, b, i)  \
+    { UInt32 c; \
+      a = (a << 8) | (b); \
+      c = counters[a]; \
+      Indices[c] = (UInt32)i++; \
+      counters[a] = c + 1; \
     }
 
-    for (i = 0; i < blockSize - 1; i++)
+    size_t a = data[0];
-      Groups[i] = counters[((UInt32)data[i] << 8) | data[(size_t)i + 1]];
+    const Byte *data_ptr = data + 1;
-    Groups[i] = counters[((UInt32)data[i] << 8) | data[0]];
+    i = 0;
-
+    if (blockSize >= 3)
-    for (i = 0; i < blockSize - 1; i++)
-      Indices[counters[((UInt32)data[i] << 8) | data[(size_t)i + 1]]++] = i;
-    Indices[counters[((UInt32)data[i] << 8) | data[0]]++] = i;
-    
-    #ifndef BLOCK_SORT_EXTERNAL_FLAGS
     {
+      blockSize -= 2;
+      do
+      {
+        size_t b;
+        b = data_ptr[0];  SET_Indices(a, b, i)
+        a = data_ptr[1];  SET_Indices(b, a, i)
+        data_ptr += 2;
+      }
+      while (i < blockSize);
+      blockSize += 2;
+    }
+    if (i < blockSize - 1)
+    {
+      SET_Indices(a, data[(size_t)i + 1], i)
+      a = (Byte)a;
+    }
+    SET_Indices(a, data[0], i)
+  }
+  
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
+  {
     UInt32 prev = 0;
     for (i = 0; i < kNumHashValues; i++)
     {
-      UInt32 prevGroupSize = counters[i] - prev;
+      const UInt32 prevGroupSize = counters[i] - prev;
       if (prevGroupSize == 0)
         continue;
       SetGroupSize(Indices + prev, prevGroupSize);
       prev = counters[i];
     }
-    }
-    #endif
   }
+#endif
 
   {
-  int NumRefBits;
+  unsigned NumRefBits;
-  UInt32 NumSortedBytes;
+  size_t NumSortedBytes;
-  for (NumRefBits = 0; ((blockSize - 1) >> NumRefBits) != 0; NumRefBits++);
+  for (NumRefBits = 0; ((blockSize - 1) >> NumRefBits) != 0; NumRefBits++)
+  {}
   NumRefBits = 32 - NumRefBits;
   if (NumRefBits > kNumRefBitsMax)
-    NumRefBits = kNumRefBitsMax;
+      NumRefBits = kNumRefBitsMax;
 
   for (NumSortedBytes = kNumHashBytes; ; NumSortedBytes <<= 1)
   {
-    #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
-    UInt32 finishedGroupSize = 0;
+    size_t finishedGroupSize = 0;
-    #endif
+#endif
-    UInt32 newLimit = 0;
+    size_t newLimit = 0;
     for (i = 0; i < blockSize;)
     {
-      UInt32 groupSize;
+      size_t groupSize;
-      #ifdef BLOCK_SORT_EXTERNAL_FLAGS
+#ifdef BLOCK_SORT_EXTERNAL_FLAGS
 
       if ((Flags[i >> kNumFlagsBits] & (1 << (i & kFlagsMask))) == 0)
       {
@@ -440,56 +553,56 @@ UInt32 BlockSort(UInt32 *Indices, const Byte *data, UInt32 blockSize)
       }
       for (groupSize = 1;
         (Flags[(i + groupSize) >> kNumFlagsBits] & (1 << ((i + groupSize) & kFlagsMask))) != 0;
-        groupSize++);
+        groupSize++)
-      
+        {}
       groupSize++;
 
-      #else
+#else
 
-      groupSize = ((Indices[i] & ~0xC0000000) >> kNumBitsMax);
+      groupSize = (Indices[i] & ~0xC0000000) >> kNumBitsMax;
       {
-      BoolInt finishedGroup = ((Indices[i] & 0x80000000) == 0);
+        const BoolInt finishedGroup = ((Indices[i] & 0x80000000) == 0);
-      if ((Indices[i] & 0x40000000) != 0)
+        if (Indices[i] & 0x40000000)
-      {
-        groupSize += ((Indices[(size_t)i + 1] >> kNumBitsMax) << kNumExtra0Bits);
-        Indices[(size_t)i + 1] &= kIndexMask;
-      }
-      Indices[i] &= kIndexMask;
-      groupSize++;
-      if (finishedGroup || groupSize == 1)
-      {
-        Indices[i - finishedGroupSize] &= kIndexMask;
-        if (finishedGroupSize > 1)
-          Indices[(size_t)(i - finishedGroupSize) + 1] &= kIndexMask;
         {
-        UInt32 newGroupSize = groupSize + finishedGroupSize;
+          groupSize += ((Indices[(size_t)i + 1] >> kNumBitsMax) << kNumExtra0Bits);
-        SetFinishedGroupSize(Indices + i - finishedGroupSize, newGroupSize);
+          Indices[(size_t)i + 1] &= kIndexMask;
-        finishedGroupSize = newGroupSize;
         }
-        i += groupSize;
+        Indices[i] &= kIndexMask;
-        continue;
+        groupSize++;
-      }
+        if (finishedGroup || groupSize == 1)
-      finishedGroupSize = 0;
+        {
+          Indices[i - finishedGroupSize] &= kIndexMask;
+          if (finishedGroupSize > 1)
+            Indices[(size_t)(i - finishedGroupSize) + 1] &= kIndexMask;
+          {
+            const size_t newGroupSize = groupSize + finishedGroupSize;
+            SetFinishedGroupSize(Indices + i - finishedGroupSize, newGroupSize)
+            finishedGroupSize = newGroupSize;
+          }
+          i += groupSize;
+          continue;
+        }
+        finishedGroupSize = 0;
       }
 
-      #endif
+#endif
       
       if (NumSortedBytes >= blockSize)
       {
-        UInt32 j;
+        size_t j;
         for (j = 0; j < groupSize; j++)
         {
-          UInt32 t = (i + j);
+          size_t t = i + j;
           /* Flags[t >> kNumFlagsBits] &= ~(1 << (t & kFlagsMask)); */
-          Groups[Indices[t]] = t;
+          Groups[Indices[t]] = (UInt32)t;
         }
       }
       else
         if (SortGroup(blockSize, NumSortedBytes, i, groupSize, NumRefBits, Indices
-          #ifndef BLOCK_SORT_USE_HEAP_SORT
+            #ifndef BLOCK_SORT_USE_HEAP_SORT
-          , 0, blockSize
+              , 0, blockSize
-          #endif
+            #endif
-          ) != 0)
+            ))
           newLimit = i + groupSize;
       i += groupSize;
     }
@@ -497,19 +610,19 @@ UInt32 BlockSort(UInt32 *Indices, const Byte *data, UInt32 blockSize)
       break;
   }
   }
-  #ifndef BLOCK_SORT_EXTERNAL_FLAGS
+#ifndef BLOCK_SORT_EXTERNAL_FLAGS
   for (i = 0; i < blockSize;)
   {
-    UInt32 groupSize = ((Indices[i] & ~0xC0000000) >> kNumBitsMax);
+    size_t groupSize = (Indices[i] & ~0xC0000000) >> kNumBitsMax;
-    if ((Indices[i] & 0x40000000) != 0)
+    if (Indices[i] & 0x40000000)
     {
-      groupSize += ((Indices[(size_t)i + 1] >> kNumBitsMax) << kNumExtra0Bits);
+      groupSize += (Indices[(size_t)i + 1] >> kNumBitsMax) << kNumExtra0Bits;
       Indices[(size_t)i + 1] &= kIndexMask;
     }
     Indices[i] &= kIndexMask;
     groupSize++;
     i += groupSize;
   }
-  #endif
+#endif
   return Groups[0];
 }

C/BwtSort.h

@@ -1,8 +1,8 @@
 /* BwtSort.h -- BWT block sorting
-2013-01-18 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __BWT_SORT_H
+#ifndef ZIP7_INC_BWT_SORT_H
-#define __BWT_SORT_H
+#define ZIP7_INC_BWT_SORT_H
 
 #include "7zTypes.h"
 
@@ -10,16 +10,17 @@ EXTERN_C_BEGIN
 
 /* use BLOCK_SORT_EXTERNAL_FLAGS if blockSize can be > 1M */
 /* #define BLOCK_SORT_EXTERNAL_FLAGS */
+// #define BLOCK_SORT_EXTERNAL_FLAGS
 
 #ifdef BLOCK_SORT_EXTERNAL_FLAGS
-#define BLOCK_SORT_EXTERNAL_SIZE(blockSize) ((((blockSize) + 31) >> 5))
+#define BLOCK_SORT_EXTERNAL_SIZE(blockSize) (((blockSize) + 31) >> 5)
 #else
 #define BLOCK_SORT_EXTERNAL_SIZE(blockSize) 0
 #endif
 
 #define BLOCK_SORT_BUF_SIZE(blockSize) ((blockSize) * 2 + BLOCK_SORT_EXTERNAL_SIZE(blockSize) + (1 << 16))
 
-UInt32 BlockSort(UInt32 *indices, const Byte *data, UInt32 blockSize);
+UInt32 BlockSort(UInt32 *indices, const Byte *data, size_t blockSize);
 
 EXTERN_C_END
 

C/Compiler.h

@@ -1,12 +1,105 @@
-/* Compiler.h
+/* Compiler.h : Compiler specific defines and pragmas
-2021-01-05 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __7Z_COMPILER_H
+#ifndef ZIP7_INC_COMPILER_H
-#define __7Z_COMPILER_H
+#define ZIP7_INC_COMPILER_H
 
-  #ifdef __clang__
+#if defined(__clang__)
-    #pragma clang diagnostic ignored "-Wunused-private-field"
+# define Z7_CLANG_VERSION  (__clang_major__ * 10000 + __clang_minor__ * 100 + __clang_patchlevel__)
+#endif
+#if defined(__clang__) && defined(__apple_build_version__)
+# define Z7_APPLE_CLANG_VERSION   Z7_CLANG_VERSION
+#elif defined(__clang__)
+# define Z7_LLVM_CLANG_VERSION    Z7_CLANG_VERSION
+#elif defined(__GNUC__)
+# define Z7_GCC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__)
+#endif
+
+#ifdef _MSC_VER
+#if !defined(__clang__) && !defined(__GNUC__)
+#define Z7_MSC_VER_ORIGINAL _MSC_VER
+#endif
+#endif
+
+#if defined(__MINGW32__) || defined(__MINGW64__)
+#define Z7_MINGW
+#endif
+
+#if defined(__LCC__) && (defined(__MCST__) || defined(__e2k__))
+#define Z7_MCST_LCC
+#define Z7_MCST_LCC_VERSION (__LCC__ * 100 + __LCC_MINOR__)
+#endif
+
+/*
+#if defined(__AVX2__) \
+    || defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40900) \
+    || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 40600) \
+    || defined(Z7_LLVM_CLANG_VERSION) && (Z7_LLVM_CLANG_VERSION >= 30100) \
+    || defined(Z7_MSC_VER_ORIGINAL) && (Z7_MSC_VER_ORIGINAL >= 1800) \
+    || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1400)
+    #define Z7_COMPILER_AVX2_SUPPORTED
   #endif
+#endif
+*/
+
+// #pragma GCC diagnostic ignored "-Wunknown-pragmas"
+
+#ifdef __clang__
+// padding size of '' with 4 bytes to alignment boundary
+#pragma GCC diagnostic ignored "-Wpadded"
+
+#if defined(Z7_LLVM_CLANG_VERSION) && (__clang_major__ == 13) \
+  && defined(__FreeBSD__)
+// freebsd:
+#pragma GCC diagnostic ignored "-Wexcess-padding"
+#endif
+
+#if __clang_major__ >= 16
+#pragma GCC diagnostic ignored "-Wunsafe-buffer-usage"
+#endif
+
+#if __clang_major__ == 13
+#if defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 16)
+// cheri
+#pragma GCC diagnostic ignored "-Wcapability-to-integer-cast"
+#endif
+#endif
+
+#if __clang_major__ == 13
+  // for <arm_neon.h>
+  #pragma GCC diagnostic ignored "-Wreserved-identifier"
+#endif
+
+#endif // __clang__
+
+#if defined(_WIN32) && defined(__clang__) && __clang_major__ >= 16
+// #pragma GCC diagnostic ignored "-Wcast-function-type-strict"
+#define Z7_DIAGNOSTIC_IGNORE_CAST_FUNCTION \
+  _Pragma("GCC diagnostic ignored \"-Wcast-function-type-strict\"")
+#else
+#define Z7_DIAGNOSTIC_IGNORE_CAST_FUNCTION
+#endif
+
+typedef void (*Z7_void_Function)(void);
+#if defined(__clang__) || defined(__GNUC__)
+#define Z7_CAST_FUNC_C  (Z7_void_Function)
+#elif defined(_MSC_VER) && _MSC_VER > 1920
+#define Z7_CAST_FUNC_C  (void *)
+// #pragma warning(disable : 4191) // 'type cast': unsafe conversion from 'FARPROC' to 'void (__cdecl *)()'
+#else
+#define Z7_CAST_FUNC_C
+#endif
+/*
+#if (defined(__GNUC__) && (__GNUC__ >= 8)) || defined(__clang__)
+  // #pragma GCC diagnostic ignored "-Wcast-function-type"
+#endif
+*/
+#ifdef __GNUC__
+#if defined(Z7_GCC_VERSION) && (Z7_GCC_VERSION >= 40000) && (Z7_GCC_VERSION < 70000)
+#pragma GCC diagnostic ignored "-Wstrict-aliasing"
+#endif
+#endif
+
 
 #ifdef _MSC_VER
 
@@ -17,24 +110,134 @@
     #pragma warning(disable : 4214) // nonstandard extension used : bit field types other than int
   #endif
 
-  #if _MSC_VER >= 1300
+#if defined(_MSC_VER) && _MSC_VER >= 1800
-    #pragma warning(disable : 4996) // This function or variable may be unsafe
+#pragma warning(disable : 4464) // relative include path contains '..'
-  #else
+#endif
-    #pragma warning(disable : 4511) // copy constructor could not be generated
-    #pragma warning(disable : 4512) // assignment operator could not be generated
-    #pragma warning(disable : 4514) // unreferenced inline function has been removed
-    #pragma warning(disable : 4702) // unreachable code
-    #pragma warning(disable : 4710) // not inlined
-    #pragma warning(disable : 4714) // function marked as __forceinline not inlined
-    #pragma warning(disable : 4786) // identifier was truncated to '255' characters in the debug information
-  #endif
 
-  #ifdef __clang__
+// == 1200 : -O1 : for __forceinline
-    #pragma clang diagnostic ignored "-Wdeprecated-declarations"
+// >= 1900 : -O1 : for printf
-    #pragma clang diagnostic ignored "-Wmicrosoft-exception-spec"
+#pragma warning(disable : 4710) // function not inlined
-    // #pragma clang diagnostic ignored "-Wreserved-id-macro"
-  #endif
 
+#if _MSC_VER < 1900
+// winnt.h: 'Int64ShllMod32'
+#pragma warning(disable : 4514) // unreferenced inline function has been removed
+#endif
+    
+#if _MSC_VER < 1300
+// #pragma warning(disable : 4702) // unreachable code
+// Bra.c : -O1:
+#pragma warning(disable : 4714) // function marked as __forceinline not inlined
+#endif
+
+/*
+#if _MSC_VER > 1400 && _MSC_VER <= 1900
+// strcat: This function or variable may be unsafe
+// sysinfoapi.h: kit10: GetVersion was declared deprecated
+#pragma warning(disable : 4996)
+#endif
+*/
+
+#if _MSC_VER > 1200
+// -Wall warnings
+
+#pragma warning(disable : 4711) // function selected for automatic inline expansion
+#pragma warning(disable : 4820) // '2' bytes padding added after data member
+
+#if _MSC_VER >= 1400 && _MSC_VER < 1920
+// 1400: string.h: _DBG_MEMCPY_INLINE_
+// 1600 - 191x : smmintrin.h __cplusplus'
+// is not defined as a preprocessor macro, replacing with '0' for '#if/#elif'
+#pragma warning(disable : 4668)
+
+// 1400 - 1600 : WinDef.h : 'FARPROC' :
+// 1900 - 191x : immintrin.h: _readfsbase_u32
+// no function prototype given : converting '()' to '(void)'
+#pragma warning(disable : 4255)
+#endif
+
+#if _MSC_VER >= 1914
+// Compiler will insert Spectre mitigation for memory load if /Qspectre switch specified
+#pragma warning(disable : 5045)
+#endif
+
+#endif // _MSC_VER > 1200
+#endif // _MSC_VER
+
+
+#if defined(__clang__) && (__clang_major__ >= 4)
+  #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE \
+    _Pragma("clang loop unroll(disable)") \
+    _Pragma("clang loop vectorize(disable)")
+  #define Z7_ATTRIB_NO_VECTORIZE
+#elif defined(__GNUC__) && (__GNUC__ >= 5) \
+    && (!defined(Z7_MCST_LCC_VERSION) || (Z7_MCST_LCC_VERSION >= 12610))
+  #define Z7_ATTRIB_NO_VECTORIZE __attribute__((optimize("no-tree-vectorize")))
+  // __attribute__((optimize("no-unroll-loops")));
+  #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE
+#elif defined(_MSC_VER) && (_MSC_VER >= 1920)
+  #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE \
+    _Pragma("loop( no_vector )")
+  #define Z7_ATTRIB_NO_VECTORIZE
+#else
+  #define Z7_PRAGMA_OPT_DISABLE_LOOP_UNROLL_VECTORIZE
+  #define Z7_ATTRIB_NO_VECTORIZE
+#endif
+
+#if defined(Z7_MSC_VER_ORIGINAL) && (Z7_MSC_VER_ORIGINAL >= 1920)
+  #define Z7_PRAGMA_OPTIMIZE_FOR_CODE_SIZE _Pragma("optimize ( \"s\", on )")
+  #define Z7_PRAGMA_OPTIMIZE_DEFAULT       _Pragma("optimize ( \"\", on )")
+#else
+  #define Z7_PRAGMA_OPTIMIZE_FOR_CODE_SIZE
+  #define Z7_PRAGMA_OPTIMIZE_DEFAULT
+#endif
+
+
+
+#if defined(MY_CPU_X86_OR_AMD64) && ( \
+       defined(__clang__) && (__clang_major__ >= 4) \
+    || defined(__GNUC__) && (__GNUC__ >= 5))
+  #define Z7_ATTRIB_NO_SSE  __attribute__((__target__("no-sse")))
+#else
+  #define Z7_ATTRIB_NO_SSE
+#endif
+
+#define Z7_ATTRIB_NO_VECTOR \
+  Z7_ATTRIB_NO_VECTORIZE \
+  Z7_ATTRIB_NO_SSE
+
+
+#if defined(__clang__) && (__clang_major__ >= 8) \
+  || defined(__GNUC__) && (__GNUC__ >= 1000) \
+  /* || defined(_MSC_VER) && (_MSC_VER >= 1920) */
+  // GCC is not good for __builtin_expect()
+  #define Z7_LIKELY(x)   (__builtin_expect((x), 1))
+  #define Z7_UNLIKELY(x) (__builtin_expect((x), 0))
+  // #define Z7_unlikely [[unlikely]]
+  // #define Z7_likely [[likely]]
+#else
+  #define Z7_LIKELY(x)   (x)
+  #define Z7_UNLIKELY(x) (x)
+  // #define Z7_likely
+#endif
+
+
+#if (defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30600))
+
+#if (Z7_CLANG_VERSION < 130000)
+#define Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER \
+  _Pragma("GCC diagnostic push") \
+  _Pragma("GCC diagnostic ignored \"-Wreserved-id-macro\"")
+#else
+#define Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER \
+  _Pragma("GCC diagnostic push") \
+  _Pragma("GCC diagnostic ignored \"-Wreserved-macro-identifier\"")
+#endif
+
+#define Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER \
+  _Pragma("GCC diagnostic pop")
+#else
+#define Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+#define Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
 #endif
 
 #define UNUSED_VAR(x) (void)x;

C/CpuArch.h

@@ -1,8 +1,8 @@
 /* CpuArch.h -- CPU specific code
-2021-07-13 : Igor Pavlov : Public domain */
+Igor Pavlov : Public domain */
 
-#ifndef __CPU_ARCH_H
+#ifndef ZIP7_INC_CPU_ARCH_H
-#define __CPU_ARCH_H
+#define ZIP7_INC_CPU_ARCH_H
 
 #include "7zTypes.h"
 
@@ -20,6 +20,7 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
   MY_CPU_64BIT doesn't mean that (sizeof(void *) == 8)
 */
 
+#if !defined(_M_ARM64EC)
 #if  defined(_M_X64) \
   || defined(_M_AMD64) \
   || defined(__x86_64__) \
@@ -35,6 +36,7 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
   #endif
   #define MY_CPU_64BIT
 #endif
+#endif
 
 
 #if  defined(_M_IX86) \
@@ -45,13 +47,34 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
   #define MY_CPU_SIZEOF_POINTER 4
 #endif
 
+#if defined(__SSE2__) \
+    || defined(MY_CPU_AMD64) \
+    || defined(_M_IX86_FP) && (_M_IX86_FP >= 2)
+#define MY_CPU_SSE2
+#endif
+
 
 #if  defined(_M_ARM64) \
+  || defined(_M_ARM64EC) \
   || defined(__AARCH64EL__) \
   || defined(__AARCH64EB__) \
   || defined(__aarch64__)
   #define MY_CPU_ARM64
-  #define MY_CPU_NAME "arm64"
+#if   defined(__ILP32__) \
+   || defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 4)
+    #define MY_CPU_NAME "arm64-32"
+    #define MY_CPU_SIZEOF_POINTER 4
+#elif defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 16)
+    #define MY_CPU_NAME "arm64-128"
+    #define MY_CPU_SIZEOF_POINTER 16
+#else
+#if defined(_M_ARM64EC)
+    #define MY_CPU_NAME "arm64ec"
+#else
+    #define MY_CPU_NAME "arm64"
+#endif
+    #define MY_CPU_SIZEOF_POINTER 8
+#endif
   #define MY_CPU_64BIT
 #endif
 
@@ -68,8 +91,10 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
   #define MY_CPU_ARM
 
   #if defined(__thumb__) || defined(__THUMBEL__) || defined(_M_ARMT)
+    #define MY_CPU_ARMT
     #define MY_CPU_NAME "armt"
   #else
+    #define MY_CPU_ARM32
     #define MY_CPU_NAME "arm"
   #endif
   /* #define MY_CPU_32BIT */
@@ -103,6 +128,8 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
   || defined(__PPC__) \
   || defined(_POWER)
 
+#define MY_CPU_PPC_OR_PPC64
+
 #if  defined(__ppc64__) \
   || defined(__powerpc64__) \
   || defined(_LP64) \
@@ -123,12 +150,76 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
 #endif
 
 
-#if  defined(__sparc64__)
+#if   defined(__sparc__) \
-  #define MY_CPU_NAME "sparc64"
+   || defined(__sparc)
+  #define MY_CPU_SPARC
+  #if  defined(__LP64__) \
+    || defined(_LP64) \
+    || defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 8)
+    #define MY_CPU_NAME "sparcv9"
+    #define MY_CPU_SIZEOF_POINTER 8
+    #define MY_CPU_64BIT
+  #elif defined(__sparc_v9__) \
+     || defined(__sparcv9)
+    #define MY_CPU_64BIT
+    #if defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 4)
+      #define MY_CPU_NAME "sparcv9-32"
+    #else
+      #define MY_CPU_NAME "sparcv9m"
+    #endif
+  #elif defined(__sparc_v8__) \
+     || defined(__sparcv8)
+    #define MY_CPU_NAME "sparcv8"
+    #define MY_CPU_SIZEOF_POINTER 4
+  #else
+    #define MY_CPU_NAME "sparc"
+  #endif
+#endif
+
+
+#if  defined(__riscv) \
+  || defined(__riscv__)
+    #define MY_CPU_RISCV
+  #if __riscv_xlen == 32
+    #define MY_CPU_NAME "riscv32"
+  #elif __riscv_xlen == 64
+    #define MY_CPU_NAME "riscv64"
+  #else
+    #define MY_CPU_NAME "riscv"
+  #endif
+#endif
+
+
+#if defined(__loongarch__)
+  #define MY_CPU_LOONGARCH
+  #if defined(__loongarch64) || defined(__loongarch_grlen) && (__loongarch_grlen == 64)
+  #define MY_CPU_64BIT
+  #endif
+  #if defined(__loongarch64)
+  #define MY_CPU_NAME "loongarch64"
+  #define MY_CPU_LOONGARCH64
+  #else
+  #define MY_CPU_NAME "loongarch"
+  #endif
+#endif
+
+
+// #undef MY_CPU_NAME
+// #undef MY_CPU_SIZEOF_POINTER
+// #define __e2k__
+// #define __SIZEOF_POINTER__ 4
+#if  defined(__e2k__)
+  #define MY_CPU_E2K
+  #if defined(__ILP32__) || defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 4)
+    #define MY_CPU_NAME "e2k-32"
+    #define MY_CPU_SIZEOF_POINTER 4
+  #else
+    #define MY_CPU_NAME "e2k"
+    #if defined(__LP64__) || defined(__SIZEOF_POINTER__) && (__SIZEOF_POINTER__ == 8)
+      #define MY_CPU_SIZEOF_POINTER 8
+    #endif
+  #endif
   #define MY_CPU_64BIT
-#elif defined(__sparc__)
-  #define MY_CPU_NAME "sparc"
-  /* #define MY_CPU_32BIT */
 #endif
 
 
@@ -162,6 +253,7 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
     || defined(MY_CPU_ARM_LE) \
     || defined(MY_CPU_ARM64_LE) \
     || defined(MY_CPU_IA64_LE) \
+    || defined(_LITTLE_ENDIAN) \
     || defined(__LITTLE_ENDIAN__) \
     || defined(__ARMEL__) \
     || defined(__THUMBEL__) \
@@ -194,6 +286,9 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
   #error Stop_Compiling_Bad_Endian
 #endif
 
+#if !defined(MY_CPU_LE) && !defined(MY_CPU_BE)
+  #error Stop_Compiling_CPU_ENDIAN_must_be_detected_at_compile_time
+#endif
 
 #if defined(MY_CPU_32BIT) && defined(MY_CPU_64BIT)
   #error Stop_Compiling_Bad_32_64_BIT
@@ -235,6 +330,7 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
 
 
 #ifndef MY_CPU_NAME
+  // #define MY_CPU_IS_UNKNOWN
   #ifdef MY_CPU_LE
     #define MY_CPU_NAME "LE"
   #elif defined(MY_CPU_BE)
@@ -250,15 +346,121 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
 
 
 
+#ifdef __has_builtin
+  #define Z7_has_builtin(x)  __has_builtin(x)
+#else
+  #define Z7_has_builtin(x)  0
+#endif
+
+
+#define Z7_BSWAP32_CONST(v) \
+       ( (((UInt32)(v) << 24)                   ) \
+       | (((UInt32)(v) <<  8) & (UInt32)0xff0000) \
+       | (((UInt32)(v) >>  8) & (UInt32)0xff00  ) \
+       | (((UInt32)(v) >> 24)                   ))
+
+
+#if defined(_MSC_VER) && (_MSC_VER >= 1300)
+
+#include <stdlib.h>
+
+/* Note: these macros will use bswap instruction (486), that is unsupported in 386 cpu */
+
+#pragma intrinsic(_byteswap_ushort)
+#pragma intrinsic(_byteswap_ulong)
+#pragma intrinsic(_byteswap_uint64)
+
+#define Z7_BSWAP16(v)  _byteswap_ushort(v)
+#define Z7_BSWAP32(v)  _byteswap_ulong (v)
+#define Z7_BSWAP64(v)  _byteswap_uint64(v)
+#define Z7_CPU_FAST_BSWAP_SUPPORTED
+
+/* GCC can generate slow code that calls function for __builtin_bswap32() for:
+     - GCC for RISCV, if Zbb/XTHeadBb extension is not used.
+     - GCC for SPARC.
+   The code from CLANG for SPARC also is not fastest.
+   So we don't define Z7_CPU_FAST_BSWAP_SUPPORTED in some cases.
+*/
+#elif (!defined(MY_CPU_RISCV) || defined (__riscv_zbb) || defined(__riscv_xtheadbb)) \
+    && !defined(MY_CPU_SPARC) \
+    && ( \
+       (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3))) \
+    || (defined(__clang__) && Z7_has_builtin(__builtin_bswap16)) \
+    )
+
+#define Z7_BSWAP16(v)  __builtin_bswap16(v)
+#define Z7_BSWAP32(v)  __builtin_bswap32(v)
+#define Z7_BSWAP64(v)  __builtin_bswap64(v)
+#define Z7_CPU_FAST_BSWAP_SUPPORTED
+
+#else
+
+#define Z7_BSWAP16(v) ((UInt16) \
+       ( ((UInt32)(v) << 8) \
+       | ((UInt32)(v) >> 8) \
+       ))
+
+#define Z7_BSWAP32(v) Z7_BSWAP32_CONST(v)
+
+#define Z7_BSWAP64(v) \
+       ( ( ( (UInt64)(v)                           ) << 8 * 7 ) \
+       | ( ( (UInt64)(v) & ((UInt32)0xff << 8 * 1) ) << 8 * 5 ) \
+       | ( ( (UInt64)(v) & ((UInt32)0xff << 8 * 2) ) << 8 * 3 ) \
+       | ( ( (UInt64)(v) & ((UInt32)0xff << 8 * 3) ) << 8 * 1 ) \
+       | ( ( (UInt64)(v) >> 8 * 1 ) & ((UInt32)0xff << 8 * 3) ) \
+       | ( ( (UInt64)(v) >> 8 * 3 ) & ((UInt32)0xff << 8 * 2) ) \
+       | ( ( (UInt64)(v) >> 8 * 5 ) & ((UInt32)0xff << 8 * 1) ) \
+       | ( ( (UInt64)(v) >> 8 * 7 )                           ) \
+       )
+
+#endif
+
+
+
 #ifdef MY_CPU_LE
   #if defined(MY_CPU_X86_OR_AMD64) \
-      || defined(MY_CPU_ARM64)
+      || defined(MY_CPU_ARM64) \
+      || defined(MY_CPU_RISCV) && defined(__riscv_misaligned_fast) \
+      || defined(MY_CPU_E2K) && defined(__iset__) && (__iset__ >= 6)
     #define MY_CPU_LE_UNALIGN
     #define MY_CPU_LE_UNALIGN_64
   #elif defined(__ARM_FEATURE_UNALIGNED)
-    /* gcc9 for 32-bit arm can use LDRD instruction that requires 32-bit alignment.
+/* === ALIGNMENT on 32-bit arm and LDRD/STRD/LDM/STM instructions.
-       So we can't use unaligned 64-bit operations. */
+  Description of problems:
-    #define MY_CPU_LE_UNALIGN
+problem-1 : 32-bit ARM architecture:
+  multi-access (pair of 32-bit accesses) instructions (LDRD/STRD/LDM/STM)
+  require 32-bit (WORD) alignment (by 32-bit ARM architecture).
+  So there is "Alignment fault exception", if data is not aligned for 32-bit.
+
+problem-2 : 32-bit kernels and arm64 kernels:
+  32-bit linux kernels provide fixup for these "paired" instruction "Alignment fault exception".
+  So unaligned paired-access instructions work via exception handler in kernel in 32-bit linux.
+ 
+  But some arm64 kernels do not handle these faults in 32-bit programs.
+  So we have unhandled exception for such instructions.
+  Probably some new arm64 kernels have fixed it, and unaligned
+  paired-access instructions work in new kernels?
+
+problem-3 : compiler for 32-bit arm:
+  Compilers use LDRD/STRD/LDM/STM for UInt64 accesses
+  and for another cases where two 32-bit accesses are fused
+  to one multi-access instruction.
+  So UInt64 variables must be aligned for 32-bit, and each
+  32-bit access must be aligned for 32-bit, if we want to
+  avoid "Alignment fault" exception (handled or unhandled).
+
+problem-4 : performace:
+  Even if unaligned access is handled by kernel, it will be slow.
+  So if we allow unaligned access, we can get fast unaligned
+  single-access, and slow unaligned paired-access.
+
+  We don't allow unaligned access on 32-bit arm, because compiler
+  genarates paired-access instructions that require 32-bit alignment,
+  and some arm64 kernels have no handler for these instructions.
+  Also unaligned paired-access instructions will be slow, if kernel handles them.
+*/
+    // it must be disabled:
+    // #define MY_CPU_LE_UNALIGN
   #endif
 #endif
 
@@ -269,13 +471,11 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
 #define GetUi32(p) (*(const UInt32 *)(const void *)(p))
 #ifdef MY_CPU_LE_UNALIGN_64
 #define GetUi64(p) (*(const UInt64 *)(const void *)(p))
+#define SetUi64(p, v) { *(UInt64 *)(void *)(p) = (v); }
 #endif
 
 #define SetUi16(p, v) { *(UInt16 *)(void *)(p) = (v); }
 #define SetUi32(p, v) { *(UInt32 *)(void *)(p) = (v); }
-#ifdef MY_CPU_LE_UNALIGN_64
-#define SetUi64(p, v) { *(UInt64 *)(void *)(p) = (v); }
-#endif
 
 #else
 
@@ -302,50 +502,33 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
 #endif
 
 
-#ifndef MY_CPU_LE_UNALIGN_64
+#ifndef GetUi64
-
 #define GetUi64(p) (GetUi32(p) | ((UInt64)GetUi32(((const Byte *)(p)) + 4) << 32))
+#endif
 
+#ifndef SetUi64
 #define SetUi64(p, v) { Byte *_ppp2_ = (Byte *)(p); UInt64 _vvv2_ = (v); \
-    SetUi32(_ppp2_    , (UInt32)_vvv2_); \
+    SetUi32(_ppp2_    , (UInt32)_vvv2_) \
-    SetUi32(_ppp2_ + 4, (UInt32)(_vvv2_ >> 32)); }
+    SetUi32(_ppp2_ + 4, (UInt32)(_vvv2_ >> 32)) }
-
 #endif
 
 
+#if defined(MY_CPU_LE_UNALIGN) && defined(Z7_CPU_FAST_BSWAP_SUPPORTED)
 
-
+#if 0
-#ifdef __has_builtin
+// Z7_BSWAP16 can be slow for x86-msvc
-  #define MY__has_builtin(x) __has_builtin(x)
+#define GetBe16_to32(p)  (Z7_BSWAP16 (*(const UInt16 *)(const void *)(p)))
 #else
-  #define MY__has_builtin(x) 0
+#define GetBe16_to32(p)  (Z7_BSWAP32 (*(const UInt16 *)(const void *)(p)) >> 16)
 #endif
 
-#if defined(MY_CPU_LE_UNALIGN) && /* defined(_WIN64) && */ defined(_MSC_VER) && (_MSC_VER >= 1300)
+#define GetBe32(p)  Z7_BSWAP32 (*(const UInt32 *)(const void *)(p))
+#define SetBe32(p, v) { (*(UInt32 *)(void *)(p)) = Z7_BSWAP32(v); }
 
-/* Note: we use bswap instruction, that is unsupported in 386 cpu */
+#if defined(MY_CPU_LE_UNALIGN_64)
-
+#define GetBe64(p)  Z7_BSWAP64 (*(const UInt64 *)(const void *)(p))
-#include <stdlib.h>
+#define SetBe64(p, v) { (*(UInt64 *)(void *)(p)) = Z7_BSWAP64(v); }
-
+#endif
-#pragma intrinsic(_byteswap_ushort)
-#pragma intrinsic(_byteswap_ulong)
-#pragma intrinsic(_byteswap_uint64)
-
-/* #define GetBe16(p) _byteswap_ushort(*(const UInt16 *)(const Byte *)(p)) */
-#define GetBe32(p) _byteswap_ulong (*(const UInt32 *)(const void *)(p))
-#define GetBe64(p) _byteswap_uint64(*(const UInt64 *)(const void *)(p))
-
-#define SetBe32(p, v) (*(UInt32 *)(void *)(p)) = _byteswap_ulong(v)
-
-#elif defined(MY_CPU_LE_UNALIGN) && ( \
-       (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3))) \
-    || (defined(__clang__) && MY__has_builtin(__builtin_bswap16)) )
-
-/* #define GetBe16(p) __builtin_bswap16(*(const UInt16 *)(const void *)(p)) */
-#define GetBe32(p) __builtin_bswap32(*(const UInt32 *)(const void *)(p))
-#define GetBe64(p) __builtin_bswap64(*(const UInt64 *)(const void *)(p))
-
-#define SetBe32(p, v) (*(UInt32 *)(void *)(p)) = __builtin_bswap32(v)
 
 #else
 
@@ -355,8 +538,6 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
     ((UInt32)((const Byte *)(p))[2] <<  8) | \
              ((const Byte *)(p))[3] )
 
-#define GetBe64(p) (((UInt64)GetBe32(p) << 32) | GetBe32(((const Byte *)(p)) + 4))
-
 #define SetBe32(p, v) { Byte *_ppp_ = (Byte *)(p); UInt32 _vvv_ = (v); \
     _ppp_[0] = (Byte)(_vvv_ >> 24); \
     _ppp_[1] = (Byte)(_vvv_ >> 16); \
@@ -365,53 +546,115 @@ MY_CPU_64BIT means that processor can work with 64-bit registers.
 
 #endif
 
+#ifndef GetBe64
+#define GetBe64(p) (((UInt64)GetBe32(p) << 32) | GetBe32(((const Byte *)(p)) + 4))
+#endif
+
+#ifndef SetBe64
+#define SetBe64(p, v) { Byte *_ppp_ = (Byte *)(p); UInt64 _vvv_ = (v); \
+    _ppp_[0] = (Byte)(_vvv_ >> 56); \
+    _ppp_[1] = (Byte)(_vvv_ >> 48); \
+    _ppp_[2] = (Byte)(_vvv_ >> 40); \
+    _ppp_[3] = (Byte)(_vvv_ >> 32); \
+    _ppp_[4] = (Byte)(_vvv_ >> 24); \
+    _ppp_[5] = (Byte)(_vvv_ >> 16); \
+    _ppp_[6] = (Byte)(_vvv_ >> 8); \
+    _ppp_[7] = (Byte)_vvv_; }
+#endif
 
 #ifndef GetBe16
-
+#ifdef GetBe16_to32
+#define GetBe16(p) ( (UInt16) GetBe16_to32(p))
+#else
 #define GetBe16(p) ( (UInt16) ( \
     ((UInt16)((const Byte *)(p))[0] << 8) | \
              ((const Byte *)(p))[1] ))
-
+#endif
 #endif
 
 
+#if defined(MY_CPU_BE)
+#define Z7_CONV_BE_TO_NATIVE_CONST32(v)  (v)
+#define Z7_CONV_LE_TO_NATIVE_CONST32(v)  Z7_BSWAP32_CONST(v)
+#define Z7_CONV_NATIVE_TO_BE_32(v)       (v)
+// #define Z7_GET_NATIVE16_FROM_2_BYTES(b0, b1)  ((b1) | ((b0) << 8))
+#elif defined(MY_CPU_LE)
+#define Z7_CONV_BE_TO_NATIVE_CONST32(v)  Z7_BSWAP32_CONST(v)
+#define Z7_CONV_LE_TO_NATIVE_CONST32(v)  (v)
+#define Z7_CONV_NATIVE_TO_BE_32(v)       Z7_BSWAP32(v)
+// #define Z7_GET_NATIVE16_FROM_2_BYTES(b0, b1)  ((b0) | ((b1) << 8))
+#else
+#error Stop_Compiling_Unknown_Endian_CONV
+#endif
+
+
+#if defined(MY_CPU_BE)
+
+#define GetBe64a(p)      (*(const UInt64 *)(const void *)(p))
+#define GetBe32a(p)      (*(const UInt32 *)(const void *)(p))
+#define GetBe16a(p)      (*(const UInt16 *)(const void *)(p))
+#define SetBe32a(p, v)   { *(UInt32 *)(void *)(p) = (v); }
+#define SetBe16a(p, v)   { *(UInt16 *)(void *)(p) = (v); }
+
+#define GetUi64a(p)      GetUi64(p)
+#define GetUi32a(p)      GetUi32(p)
+#define GetUi16a(p)      GetUi16(p)
+#define SetUi32a(p, v)   SetUi32(p, v)
+#define SetUi16a(p, v)   SetUi16(p, v)
+
+#elif defined(MY_CPU_LE)
+
+#define GetUi64a(p)      (*(const UInt64 *)(const void *)(p))
+#define GetUi32a(p)      (*(const UInt32 *)(const void *)(p))
+#define GetUi16a(p)      (*(const UInt16 *)(const void *)(p))
+#define SetUi32a(p, v)   { *(UInt32 *)(void *)(p) = (v); }
+#define SetUi16a(p, v)   { *(UInt16 *)(void *)(p) = (v); }
+
+#define GetBe64a(p)      GetBe64(p)
+#define GetBe32a(p)      GetBe32(p)
+#define GetBe16a(p)      GetBe16(p)
+#define SetBe32a(p, v)   SetBe32(p, v)
+#define SetBe16a(p, v)   SetBe16(p, v)
+
+#else
+#error Stop_Compiling_Unknown_Endian_CPU_a
+#endif
+
+
+#ifndef GetBe16_to32
+#define GetBe16_to32(p) GetBe16(p)
+#endif
+
+
+#if defined(MY_CPU_X86_OR_AMD64) \
+  || defined(MY_CPU_ARM_OR_ARM64) \
+  || defined(MY_CPU_PPC_OR_PPC64)
+  #define Z7_CPU_FAST_ROTATE_SUPPORTED
+#endif
+
 
 #ifdef MY_CPU_X86_OR_AMD64
 
-typedef struct
+void Z7_FASTCALL z7_x86_cpuid(UInt32 a[4], UInt32 function);
-{
+UInt32 Z7_FASTCALL z7_x86_cpuid_GetMaxFunc(void);
-  UInt32 maxFunc;
+#if defined(MY_CPU_AMD64)
-  UInt32 vendor[3];
+#define Z7_IF_X86_CPUID_SUPPORTED
-  UInt32 ver;
+#else
-  UInt32 b;
+#define Z7_IF_X86_CPUID_SUPPORTED if (z7_x86_cpuid_GetMaxFunc())
-  UInt32 c;
+#endif
-  UInt32 d;
-} Cx86cpuid;
-
-enum
-{
-  CPU_FIRM_INTEL,
-  CPU_FIRM_AMD,
-  CPU_FIRM_VIA
-};
-
-void MyCPUID(UInt32 function, UInt32 *a, UInt32 *b, UInt32 *c, UInt32 *d);
-
-BoolInt x86cpuid_CheckAndRead(Cx86cpuid *p);
-int x86cpuid_GetFirm(const Cx86cpuid *p);
-
-#define x86cpuid_GetFamily(ver) (((ver >> 16) & 0xFF0) | ((ver >> 8) & 0xF))
-#define x86cpuid_GetModel(ver)  (((ver >> 12) &  0xF0) | ((ver >> 4) & 0xF))
-#define x86cpuid_GetStepping(ver) (ver & 0xF)
-
-BoolInt CPU_Is_InOrder(void);
 
 BoolInt CPU_IsSupported_AES(void);
+BoolInt CPU_IsSupported_AVX(void);
 BoolInt CPU_IsSupported_AVX2(void);
+BoolInt CPU_IsSupported_AVX512F_AVX512VL(void);
 BoolInt CPU_IsSupported_VAES_AVX2(void);
+BoolInt CPU_IsSupported_CMOV(void);
+BoolInt CPU_IsSupported_SSE(void);
+BoolInt CPU_IsSupported_SSE2(void);
 BoolInt CPU_IsSupported_SSSE3(void);
 BoolInt CPU_IsSupported_SSE41(void);
 BoolInt CPU_IsSupported_SHA(void);
+BoolInt CPU_IsSupported_SHA512(void);
 BoolInt CPU_IsSupported_PageGB(void);
 
 #elif defined(MY_CPU_ARM_OR_ARM64)
@@ -429,12 +672,13 @@ BoolInt CPU_IsSupported_SHA1(void);
 BoolInt CPU_IsSupported_SHA2(void);
 BoolInt CPU_IsSupported_AES(void);
 #endif
+BoolInt CPU_IsSupported_SHA512(void);
 
 #endif
 
 #if defined(__APPLE__)
-int My_sysctlbyname_Get(const char *name, void *buf, size_t *bufSize);
+int z7_sysctlbyname_Get(const char *name, void *buf, size_t *bufSize);
-int My_sysctlbyname_Get_UInt32(const char *name, UInt32 *val);
+int z7_sysctlbyname_Get_UInt32(const char *name, UInt32 *val);
 #endif
 
 EXTERN_C_END

C/Delta.h

@@ -1,8 +1,8 @@
 /* Delta.h -- Delta converter
-2013-01-18 : Igor Pavlov : Public domain */
+2023-03-03 : Igor Pavlov : Public domain */
 
-#ifndef __DELTA_H
+#ifndef ZIP7_INC_DELTA_H
-#define __DELTA_H
+#define ZIP7_INC_DELTA_H
 
 #include "7zTypes.h"
 

C/DllSecur.c

@@ -1,110 +1,99 @@
 /* DllSecur.c -- DLL loading security
-2021-12-25 : Igor Pavlov : Public domain */
+2023-12-03 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #ifdef _WIN32
 
-#include <Windows.h>
+#include "7zWindows.h"
 
 #include "DllSecur.h"
 
 #ifndef UNDER_CE
 
+Z7_DIAGNOSTIC_IGNORE_CAST_FUNCTION
+
 typedef BOOL (WINAPI *Func_SetDefaultDllDirectories)(DWORD DirectoryFlags);
 
 #define MY_LOAD_LIBRARY_SEARCH_USER_DIRS 0x400
 #define MY_LOAD_LIBRARY_SEARCH_SYSTEM32  0x800
 
+#define DELIM "\0"
+
 static const char * const g_Dlls =
+         "userenv"
+  DELIM  "setupapi"
+  DELIM  "apphelp"
+  DELIM  "propsys"
+  DELIM  "dwmapi"
+  DELIM  "cryptbase"
+  DELIM  "oleacc"
+  DELIM  "clbcatq"
+  DELIM  "version"
   #ifndef _CONSOLE
-  "UXTHEME\0"
+  DELIM  "uxtheme"
   #endif
-  "USERENV\0"
+  DELIM;
-  "SETUPAPI\0"
-  "APPHELP\0"
-  "PROPSYS\0"
-  "DWMAPI\0"
-  "CRYPTBASE\0"
-  "OLEACC\0"
-  "CLBCATQ\0"
-  "VERSION\0"
-  ;
 
 #endif
 
-// #define MY_CAST_FUNC  (void(*)())
+#ifdef __clang__
-#define MY_CAST_FUNC  
+  #pragma GCC diagnostic ignored "-Wdeprecated-declarations"
+#endif
+#if defined (_MSC_VER) && _MSC_VER >= 1900
+// sysinfoapi.h: kit10: GetVersion was declared deprecated
+#pragma warning(disable : 4996)
+#endif
 
-void My_SetDefaultDllDirectories()
+#define IF_NON_VISTA_SET_DLL_DIRS_AND_RETURN \
+    if ((UInt16)GetVersion() != 6) { \
+      const \
+       Func_SetDefaultDllDirectories setDllDirs = \
+      (Func_SetDefaultDllDirectories) Z7_CAST_FUNC_C GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), \
+           "SetDefaultDllDirectories"); \
+      if (setDllDirs) if (setDllDirs(MY_LOAD_LIBRARY_SEARCH_SYSTEM32 | MY_LOAD_LIBRARY_SEARCH_USER_DIRS)) return; }
+
+void My_SetDefaultDllDirectories(void)
 {
   #ifndef UNDER_CE
-  
+  IF_NON_VISTA_SET_DLL_DIRS_AND_RETURN
-    OSVERSIONINFO vi;
-    vi.dwOSVersionInfoSize = sizeof(vi);
-    if (!GetVersionEx(&vi) || vi.dwMajorVersion != 6 || vi.dwMinorVersion != 0)
-    {
-      Func_SetDefaultDllDirectories setDllDirs = (Func_SetDefaultDllDirectories)
-          MY_CAST_FUNC GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), "SetDefaultDllDirectories");
-      if (setDllDirs)
-        if (setDllDirs(MY_LOAD_LIBRARY_SEARCH_SYSTEM32 | MY_LOAD_LIBRARY_SEARCH_USER_DIRS))
-          return;
-    }
-
   #endif
 }
 
 
-void LoadSecurityDlls()
+void LoadSecurityDlls(void)
 {
   #ifndef UNDER_CE
-  
+  // at Vista (ver 6.0) : CoCreateInstance(CLSID_ShellLink, ...) doesn't work after SetDefaultDllDirectories() : Check it ???
-  wchar_t buf[MAX_PATH + 100];
+  IF_NON_VISTA_SET_DLL_DIRS_AND_RETURN
-
-  {
-    // at Vista (ver 6.0) : CoCreateInstance(CLSID_ShellLink, ...) doesn't work after SetDefaultDllDirectories() : Check it ???
-    OSVERSIONINFO vi;
-    vi.dwOSVersionInfoSize = sizeof(vi);
-    if (!GetVersionEx(&vi) || vi.dwMajorVersion != 6 || vi.dwMinorVersion != 0)
-    {
-      Func_SetDefaultDllDirectories setDllDirs = (Func_SetDefaultDllDirectories)
-          MY_CAST_FUNC GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), "SetDefaultDllDirectories");
-      if (setDllDirs)
-        if (setDllDirs(MY_LOAD_LIBRARY_SEARCH_SYSTEM32 | MY_LOAD_LIBRARY_SEARCH_USER_DIRS))
-          return;
-    }
-  }
-
-  {
-    unsigned len = GetSystemDirectoryW(buf, MAX_PATH + 2);
-    if (len == 0 || len > MAX_PATH)
-      return;
-  }
   {
+    wchar_t buf[MAX_PATH + 100];
     const char *dll;
-    unsigned pos = (unsigned)lstrlenW(buf);
+    unsigned pos = GetSystemDirectoryW(buf, MAX_PATH + 2);
-
+    if (pos == 0 || pos > MAX_PATH)
+      return;
     if (buf[pos - 1] != '\\')
       buf[pos++] = '\\';
-    
+    for (dll = g_Dlls; *dll != 0;)
-    for (dll = g_Dlls; dll[0] != 0;)
     {
-      unsigned k = 0;
+      wchar_t *dest = &buf[pos];
       for (;;)
       {
-        char c = *dll++;
+        const char c = *dll++;
-        buf[pos + k] = (Byte)c;
-        k++;
         if (c == 0)
           break;
+        *dest++ = (Byte)c;
       }
-
+      dest[0] = '.';
-      lstrcatW(buf, L".dll");
+      dest[1] = 'd';
+      dest[2] = 'l';
+      dest[3] = 'l';
+      dest[4] = 0;
+      // lstrcatW(buf, L".dll");
       LoadLibraryExW(buf, NULL, LOAD_WITH_ALTERED_SEARCH_PATH);
     }
   }
-  
   #endif
 }
 
-#endif
+#endif // _WIN32

C/DllSecur.h

@@ -1,8 +1,8 @@
 /* DllSecur.h -- DLL loading for security
-2018-02-19 : Igor Pavlov : Public domain */
+2023-03-03 : Igor Pavlov : Public domain */
 
-#ifndef __DLL_SECUR_H
+#ifndef ZIP7_INC_DLL_SECUR_H
-#define __DLL_SECUR_H
+#define ZIP7_INC_DLL_SECUR_H
 
 #include "7zTypes.h"
 

C/HuffEnc.c

@@ -1,60 +1,125 @@
 /* HuffEnc.c -- functions for Huffman encoding
-2021-02-09 : Igor Pavlov : Public domain */
+Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
+#include <string.h>
+
 #include "HuffEnc.h"
 #include "Sort.h"
+#include "CpuArch.h"
 
-#define kMaxLen 16
+#define kMaxLen       Z7_HUFFMAN_LEN_MAX
-#define NUM_BITS 10
+#define NUM_BITS      10
-#define MASK (((unsigned)1 << NUM_BITS) - 1)
+#define MASK          ((1u << NUM_BITS) - 1)
+#define FREQ_MASK     (~(UInt32)MASK)
+#define NUM_COUNTERS  (48 * 2)
 
-#define NUM_COUNTERS 64
+#if 1 && (defined(MY_CPU_LE) || defined(MY_CPU_BE))
+#if defined(MY_CPU_LE)
+  #define HI_HALF_OFFSET 1
+#else
+  #define HI_HALF_OFFSET 0
+#endif
+#define LOAD_PARENT(p)                  ((unsigned)*((const UInt16 *)(p) + HI_HALF_OFFSET))
+#define STORE_PARENT(p, fb, val)         *((UInt16 *)(p) + HI_HALF_OFFSET) = (UInt16)(val);
+#define STORE_PARENT_DIRECT(p, fb, hi)  STORE_PARENT(p, fb, hi)
+#define UPDATE_E(eHi)                   eHi++;
+#else
+#define LOAD_PARENT(p)                  ((unsigned)(*(p) >> NUM_BITS))
+#define STORE_PARENT_DIRECT(p, fb, hi)  *(p) = ((fb) & MASK) | (hi); // set parent field
+#define STORE_PARENT(p, fb, val)        STORE_PARENT_DIRECT(p, fb, ((UInt32)(val) << NUM_BITS))
+#define UPDATE_E(eHi)                   eHi += 1 << NUM_BITS;
+#endif
 
-#define HUFFMAN_SPEED_OPT
+void Huffman_Generate(const UInt32 *freqs, UInt32 *p, Byte *lens, unsigned numSymbols, unsigned maxLen)
-
-void Huffman_Generate(const UInt32 *freqs, UInt32 *p, Byte *lens, UInt32 numSymbols, UInt32 maxLen)
 {
-  UInt32 num = 0;
+#if NUM_COUNTERS > 2
-  /* if (maxLen > 10) maxLen = 10; */
+  unsigned counters[NUM_COUNTERS];
+#endif
+#if 1 && NUM_COUNTERS > (kMaxLen + 4) * 2
+  #define lenCounters   (counters)
+  #define codes         (counters + kMaxLen + 4)
+#else
+  unsigned lenCounters[kMaxLen + 1];
+  UInt32 codes[kMaxLen + 1];
+#endif
+
+  unsigned num;
   {
-    UInt32 i;
+    unsigned i;
+    // UInt32 sum = 0;
 
-    #ifdef HUFFMAN_SPEED_OPT
+#if NUM_COUNTERS > 2
+    
+#define CTR_ITEM_FOR_FREQ(freq) \
+    counters[(freq) >= NUM_COUNTERS - 1 ? NUM_COUNTERS - 1 : (unsigned)(freq)]
 
-    UInt32 counters[NUM_COUNTERS];
     for (i = 0; i < NUM_COUNTERS; i++)
       counters[i] = 0;
-    for (i = 0; i < numSymbols; i++)
+    memset(lens, 0, numSymbols);
     {
-      UInt32 freq = freqs[i];
+      const UInt32 *fp = freqs + numSymbols;
-      counters[(freq < NUM_COUNTERS - 1) ? freq : NUM_COUNTERS - 1]++;
+#define NUM_UNROLLS 1
+#if NUM_UNROLLS > 1 // use 1 if odd (numSymbols) is possisble
+      if (numSymbols & 1)
+      {
+        UInt32 f;
+        f = *--fp;  CTR_ITEM_FOR_FREQ(f)++;
+        // sum += f;
+      }
+#endif
+      do
+      {
+        UInt32 f;
+        fp -= NUM_UNROLLS;
+        f = fp[0];  CTR_ITEM_FOR_FREQ(f)++;
+        // sum += f;
+#if NUM_UNROLLS > 1
+        f = fp[1];  CTR_ITEM_FOR_FREQ(f)++;
+        // sum += f;
+#endif
+      }
+      while (fp != freqs);
     }
+#if 0
+    printf("\nsum=%8u numSymbols =%3u ctrs:", sum, numSymbols);
+    {
+      unsigned k = 0;
+      for (k = 0; k < NUM_COUNTERS; k++)
+        printf(" %u", counters[k]);
+    }
+#endif
         
-    for (i = 1; i < NUM_COUNTERS; i++)
+    num = counters[1];
+    counters[1] = 0;
+    for (i = 2; i != NUM_COUNTERS; i += 2)
     {
-      UInt32 temp = counters[i];
+      unsigned c;
-      counters[i] = num;
+      c = (counters    )[i];  (counters    )[i] = num;  num += c;
-      num += temp;
+      c = (counters + 1)[i];  (counters + 1)[i] = num;  num += c;
     }
-
+    counters[0] = num; // we want to write (freq==0) symbols to the end of (p) array
-    for (i = 0; i < numSymbols; i++)
     {
-      UInt32 freq = freqs[i];
+      i = 0;
-      if (freq == 0)
+      do
-        lens[i] = 0;
+      {
-      else
+        const UInt32 f = freqs[i];
-        p[counters[((freq < NUM_COUNTERS - 1) ? freq : NUM_COUNTERS - 1)]++] = i | (freq << NUM_BITS);
+#if 0
+        if (f == 0) lens[i] = 0; else
+#endif
+          p[CTR_ITEM_FOR_FREQ(f)++] = i | (f << NUM_BITS);
+      }
+      while (++i != numSymbols);
     }
-    counters[0] = 0;
     HeapSort(p + counters[NUM_COUNTERS - 2], counters[NUM_COUNTERS - 1] - counters[NUM_COUNTERS - 2]);
     
-    #else
+#else // NUM_COUNTERS <= 2
     
+    num = 0;
     for (i = 0; i < numSymbols; i++)
     {
-      UInt32 freq = freqs[i];
+      const UInt32 freq = freqs[i];
       if (freq == 0)
         lens[i] = 0;
       else
@@ -62,17 +127,27 @@ void Huffman_Generate(const UInt32 *freqs, UInt32 *p, Byte *lens, UInt32 numSymb
     }
     HeapSort(p, num);
 
-    #endif
+#endif
   }
 
-  if (num < 2)
+  if (num <= 2)
   {
     unsigned minCode = 0;
     unsigned maxCode = 1;
-    if (num == 1)
+    if (num)
     {
-      maxCode = (unsigned)p[0] & MASK;
+      maxCode = (unsigned)p[(size_t)num - 1] & MASK;
-      if (maxCode == 0)
+      if (num == 2)
+      {
+        minCode = (unsigned)p[0] & MASK;
+        if (minCode > maxCode)
+        {
+          const unsigned temp = minCode;
+          minCode = maxCode;
+          maxCode = temp;
+        }
+      }
+      else if (maxCode == 0)
         maxCode++;
     }
     p[minCode] = 0;
@@ -80,69 +155,206 @@ void Huffman_Generate(const UInt32 *freqs, UInt32 *p, Byte *lens, UInt32 numSymb
     lens[minCode] = lens[maxCode] = 1;
     return;
   }
-  
   {
-    UInt32 b, e, i;
+    unsigned i;
-  
+    for (i = 0; i <= kMaxLen; i++)
-    i = b = e = 0;
+      lenCounters[i] = 0;
-    do
+    lenCounters[1] = 2;  // by default root node has 2 child leaves at level 1.
+  }
+  // if (num != 2)
+  {
+    // num > 2
+    // the binary tree will contain (num - 1) internal nodes.
+    // p[num - 2] will be root node of binary tree.
+    UInt32 *b;
+    UInt32 *n;
+    // first node will have two leaf childs: p[0] and p[1]:
+    // p[0] += p[1] & FREQ_MASK; // set frequency sum of child leafs
+    // if (pi == n) exit(0);
+    // if (pi != n)
     {
-      UInt32 n, m, freq;
+      UInt32 fb = (p[1] & FREQ_MASK) + p[0];
-      n = (i != num && (b == e || (p[i] >> NUM_BITS) <= (p[b] >> NUM_BITS))) ? i++ : b++;
+      UInt32 f = p[2] & FREQ_MASK;
-      freq = (p[n] & ~MASK);
+      const UInt32 *pi = p + 2;
-      p[n] = (p[n] & MASK) | (e << NUM_BITS);
+      UInt32 *e = p;
-      m = (i != num && (b == e || (p[i] >> NUM_BITS) <= (p[b] >> NUM_BITS))) ? i++ : b++;
+      UInt32 eHi = 0;
-      freq += (p[m] & ~MASK);
+      n = p + num;
-      p[m] = (p[m] & MASK) | (e << NUM_BITS);
+      b = p;
-      p[e] = (p[e] & MASK) | freq;
+      // p[0] = fb;
-      e++;
+      for (;;)
+      {
+        // (b <= e)
+        UInt32 sum;
+        e++;
+        UPDATE_E(eHi)
+
+        // (b < e)
+
+        // p range : high bits
+        // [0, b)  : parent :     processed nodes that have parent and childs
+        // [b, e)  : FREQ   : non-processed nodes that have no parent but have childs
+        // [e, pi) : FREQ   :     processed leaves for which parent node was     created
+        // [pi, n) : FREQ   : non-processed leaves for which parent node was not created
+
+        // first child
+        // note : (*b < f) is same result as ((*b & FREQ_MASK) < f)
+        if (fb < f)
+        {
+          // node freq is smaller
+          sum = fb & FREQ_MASK;
+          STORE_PARENT_DIRECT (b, fb, eHi)
+          b++;
+          fb = *b;
+          if (b == e)
+          {
+            if (++pi == n)
+              break;
+            sum += f;
+            fb &= MASK;
+            fb |= sum;
+            *e = fb;
+            f = *pi & FREQ_MASK;
+            continue;
+          }
+        }
+        else if (++pi == n)
+        {
+          STORE_PARENT_DIRECT (b, fb, eHi)
+          b++;
+          break;
+        }
+        else
+        {
+          sum = f;
+          f = *pi & FREQ_MASK;
+        }
+
+        // (b < e)
+
+        // second child
+        if (fb < f)
+        {
+          sum += fb;
+          sum &= FREQ_MASK;
+          STORE_PARENT_DIRECT (b, fb, eHi)
+          b++;
+          *e = (*e & MASK) | sum; // set frequency sum
+          // (b <= e) is possible here
+          fb = *b;
+        }
+        else if (++pi == n)
+          break;
+        else
+        {
+          sum += f;
+          f = *pi & FREQ_MASK;
+          *e = (*e & MASK) | sum; // set frequency sum
+        }
+      }
     }
-    while (num - e > 1);
     
+    // printf("\nnum-e=%3u, numSymbols=%3u, num=%3u, b=%3u", n - e, numSymbols, n - p, b - p);
     {
-      UInt32 lenCounters[kMaxLen + 1];
+      n -= 2;
-      for (i = 0; i <= kMaxLen; i++)
+      *n &= MASK; // root node : we clear high bits (zero bits mean level == 0)
-        lenCounters[i] = 0;
+      if (n != b)
-      
-      p[--e] &= MASK;
-      lenCounters[1] = 2;
-      while (e > 0)
       {
-        UInt32 len = (p[p[--e] >> NUM_BITS] >> NUM_BITS) + 1;
+        // We go here, if we have some number of non-created nodes up to root.
-        p[e] = (p[e] & MASK) | (len << NUM_BITS);
+        // We process them in simplified code:
+        // position of parent for each pair of nodes is known.
+        // n[-2], n[-1] : current pair of child nodes
+        // (p1) : parent node for current pair.
+        UInt32 *p1 = n;
+        do
+        {
+          const unsigned len = LOAD_PARENT(p1) + 1;
+          p1--;
+          (lenCounters    )[len] -= 2;                  // we remove 2 leaves from level (len)
+          (lenCounters + 1)[len] += 2 * 2;  // we add 4 leaves at level (len + 1)
+          n -= 2;
+          STORE_PARENT (n    , n[0], len)
+          STORE_PARENT (n + 1, n[1], len)
+        }
+        while (n != b);
+      }
+    }
+    
+    if (b != p)
+    {
+      // we detect level of each node (realtive to root),
+      // and update lenCounters[].
+      // We process only intermediate nodes and we don't process leaves.
+      do
+      {
+        // if (ii <  b) : parent_bits_of (p[ii]) == index of parent node : ii < (p[ii])
+        // if (ii >= b) : parent_bits_of (p[ii]) == level of this (ii) node in tree
+        unsigned len;
+        b--;
+        len = (unsigned)LOAD_PARENT(p + LOAD_PARENT(b)) + 1;
+        STORE_PARENT (b, *b, len)
         if (len >= maxLen)
-          for (len = maxLen - 1; lenCounters[len] == 0; len--);
+        {
-        lenCounters[len]--;
+          // We are not allowed to create node at level (maxLen) and higher,
-        lenCounters[(size_t)len + 1] += 2;
+          // because all leaves must be placed to level (maxLen) or lower.
+          // We find nearest allowed leaf and place current node to level of that leaf:
+          for (len = maxLen - 1; lenCounters[len] == 0; len--) {}
+        }
+        lenCounters[len]--;                 // we remove 1 leaf from level (len)
+        (lenCounters + 1)[len] += 2;  // we add 2 leaves at level (len + 1)
       }
-      
+      while (b != p);
+    }
+  }
+  {
+    {
+      unsigned len = maxLen;
+      const UInt32 *p2 = p;
+      do
       {
-        UInt32 len;
+        unsigned k = lenCounters[len];
-        i = 0;
+        if (k)
-        for (len = maxLen; len != 0; len--)
+          do
-        {
+            lens[(unsigned)*p2++ & MASK] = (Byte)len;
-          UInt32 k;
+          while (--k);
-          for (k = lenCounters[len]; k != 0; k--)
-            lens[p[i++] & MASK] = (Byte)len;
-        }
       }
-      
+      while (--len);
+    }
+    codes[0] = 0; // we don't want garbage values to be written to p[] array.
+    // codes[1] = 0;
+    {
+      UInt32 code = 0;
+      unsigned len;
+      for (len = 0; len < kMaxLen; len++)
+        (codes + 1)[len] = code = (code + lenCounters[len]) << 1;
+    }
+    /* if (code + lenCounters[kMaxLen] - 1 != (1 << kMaxLen) - 1) throw 1; */
+    {
+      const Byte * const limit = lens + numSymbols;
+      do
       {
-        UInt32 nextCodes[kMaxLen + 1];
+        unsigned len;
-        {
+        UInt32 c;
-          UInt32 code = 0;
+        len = lens[0];  c = codes[len];  p[0] = c;  codes[len] = c + 1;
-          UInt32 len;
+        // len = lens[1];  c = codes[len];  p[1] = c;  codes[len] = c + 1;
-          for (len = 1; len <= kMaxLen; len++)
+        p += 1;
-            nextCodes[len] = code = (code + lenCounters[(size_t)len - 1]) << 1;
+        lens += 1;
-        }
-        /* if (code + lenCounters[kMaxLen] - 1 != (1 << kMaxLen) - 1) throw 1; */
-
-        {
-          UInt32 k;
-          for (k = 0; k < numSymbols; k++)
-            p[k] = nextCodes[lens[k]]++;
-        }
       }
+      while (lens != limit);
     }
   }
 }
+
+#undef kMaxLen
+#undef NUM_BITS
+#undef MASK
+#undef FREQ_MASK
+#undef NUM_COUNTERS
+#undef CTR_ITEM_FOR_FREQ
+#undef LOAD_PARENT
+#undef STORE_PARENT
+#undef STORE_PARENT_DIRECT
+#undef UPDATE_E
+#undef HI_HALF_OFFSET
+#undef NUM_UNROLLS
+#undef lenCounters
+#undef codes

C/HuffEnc.h

@@ -1,21 +1,21 @@
 /* HuffEnc.h -- Huffman encoding
-2013-01-18 : Igor Pavlov : Public domain */
+Igor Pavlov : Public domain */
 
-#ifndef __HUFF_ENC_H
+#ifndef ZIP7_INC_HUFF_ENC_H
-#define __HUFF_ENC_H
+#define ZIP7_INC_HUFF_ENC_H
 
 #include "7zTypes.h"
 
 EXTERN_C_BEGIN
 
+#define Z7_HUFFMAN_LEN_MAX 16
 /*
 Conditions:
-  num <= 1024 = 2 ^ NUM_BITS
+  2 <= num <= 1024 = 2 ^ NUM_BITS
   Sum(freqs) < 4M = 2 ^ (32 - NUM_BITS)
-  maxLen <= 16 = kMaxLen
+  1 <= maxLen <= 16 = Z7_HUFFMAN_LEN_MAX
   Num_Items(p) >= HUFFMAN_TEMP_SIZE(num)
 */
- 
 void Huffman_Generate(const UInt32 *freqs, UInt32 *p, Byte *lens, UInt32 num, UInt32 maxLen);
 
 EXTERN_C_END

C/LzFind.h

@@ -1,8 +1,8 @@
 /* LzFind.h -- Match finder for LZ algorithms
-2021-07-13 : Igor Pavlov : Public domain */
+2024-01-22 : Igor Pavlov : Public domain */
 
-#ifndef __LZ_FIND_H
+#ifndef ZIP7_INC_LZ_FIND_H
-#define __LZ_FIND_H
+#define ZIP7_INC_LZ_FIND_H
 
 #include "7zTypes.h"
 
@@ -10,9 +10,9 @@ EXTERN_C_BEGIN
 
 typedef UInt32 CLzRef;
 
-typedef struct _CMatchFinder
+typedef struct
 {
-  Byte *buffer;
+  const Byte *buffer;
   UInt32 pos;
   UInt32 posLimit;
   UInt32 streamPos;  /* wrap over Zero is allowed (streamPos < pos). Use (UInt32)(streamPos - pos) */
@@ -32,8 +32,8 @@ typedef struct _CMatchFinder
   UInt32 hashMask;
   UInt32 cutValue;
 
-  Byte *bufferBase;
+  Byte *bufBase;
-  ISeqInStream *stream;
+  ISeqInStreamPtr stream;
   
   UInt32 blockSize;
   UInt32 keepSizeBefore;
@@ -43,7 +43,9 @@ typedef struct _CMatchFinder
   size_t directInputRem;
   UInt32 historySize;
   UInt32 fixedHashSize;
-  UInt32 hashSizeSum;
+  Byte numHashBytes_Min;
+  Byte numHashOutBits;
+  Byte _pad2_[2];
   SRes result;
   UInt32 crc[256];
   size_t numRefs;
@@ -69,24 +71,45 @@ void MatchFinder_ReadIfRequired(CMatchFinder *p);
 
 void MatchFinder_Construct(CMatchFinder *p);
 
-/* Conditions:
+/* (directInput = 0) is default value.
-     historySize <= 3 GB
+   It's required to provide correct (directInput) value
-     keepAddBufferBefore + matchMaxLen + keepAddBufferAfter < 511MB
+   before calling MatchFinder_Create().
+   You can set (directInput) by any of the following calls:
+     - MatchFinder_SET_DIRECT_INPUT_BUF()
+     - MatchFinder_SET_STREAM()
+     - MatchFinder_SET_STREAM_MODE()
 */
+
+#define MatchFinder_SET_DIRECT_INPUT_BUF(p, _src_, _srcLen_) { \
+  (p)->stream = NULL; \
+  (p)->directInput = 1; \
+  (p)->buffer = (_src_); \
+  (p)->directInputRem = (_srcLen_); }
+
+/*
+#define MatchFinder_SET_STREAM_MODE(p) { \
+  (p)->directInput = 0; }
+*/
+
+#define MatchFinder_SET_STREAM(p, _stream_) { \
+  (p)->stream = _stream_; \
+  (p)->directInput = 0; }
+  
+
 int MatchFinder_Create(CMatchFinder *p, UInt32 historySize,
     UInt32 keepAddBufferBefore, UInt32 matchMaxLen, UInt32 keepAddBufferAfter,
     ISzAllocPtr alloc);
 void MatchFinder_Free(CMatchFinder *p, ISzAllocPtr alloc);
 void MatchFinder_Normalize3(UInt32 subValue, CLzRef *items, size_t numItems);
-// void MatchFinder_ReduceOffsets(CMatchFinder *p, UInt32 subValue);
 
 /*
-#define Inline_MatchFinder_InitPos(p, val) \
+#define MatchFinder_INIT_POS(p, val) \
     (p)->pos = (val); \
     (p)->streamPos = (val);
 */
 
-#define Inline_MatchFinder_ReduceOffsets(p, subValue) \
+// void MatchFinder_ReduceOffsets(CMatchFinder *p, UInt32 subValue);
+#define MatchFinder_REDUCE_OFFSETS(p, subValue) \
     (p)->pos -= (subValue); \
     (p)->streamPos -= (subValue);
 
@@ -107,7 +130,7 @@ typedef const Byte * (*Mf_GetPointerToCurrentPos_Func)(void *object);
 typedef UInt32 * (*Mf_GetMatches_Func)(void *object, UInt32 *distances);
 typedef void (*Mf_Skip_Func)(void *object, UInt32);
 
-typedef struct _IMatchFinder
+typedef struct
 {
   Mf_Init_Func Init;
   Mf_GetNumAvailableBytes_Func GetNumAvailableBytes;
@@ -121,7 +144,8 @@ void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder2 *vTable);
 void MatchFinder_Init_LowHash(CMatchFinder *p);
 void MatchFinder_Init_HighHash(CMatchFinder *p);
 void MatchFinder_Init_4(CMatchFinder *p);
-void MatchFinder_Init(CMatchFinder *p);
+// void MatchFinder_Init(CMatchFinder *p);
+void MatchFinder_Init(void *p);
 
 UInt32* Bt3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances);
 UInt32* Hc3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances);

C/LzFindMt.c

@@ -1,5 +1,5 @@
 /* LzFindMt.c -- multithreaded Match finder for LZ algorithms
-2021-12-21 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -69,7 +69,7 @@ extern UInt64 g_NumIters_Bytes;
   UInt32 temp = p->crc[cur[0]] ^ cur[1]; \
   h3 = (temp ^ ((UInt32)cur[2] << 8)) & (kHash3Size - 1); }
 
-#define __MT_HASH4_CALC { \
+#define MT_HASH4_CALC { \
   UInt32 temp = p->crc[cur[0]] ^ cur[1]; \
   h2 = temp & (kHash2Size - 1); \
   temp ^= ((UInt32)cur[2] << 8); \
@@ -79,14 +79,16 @@ extern UInt64 g_NumIters_Bytes;
 */
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static void MtSync_Construct(CMtSync *p)
 {
+  p->affinityGroup = -1;
+  p->affinityInGroup = 0;
   p->affinity = 0;
   p->wasCreated = False;
   p->csWasInitialized = False;
   p->csWasEntered = False;
-  Thread_Construct(&p->thread);
+  Thread_CONSTRUCT(&p->thread)
   Event_Construct(&p->canStart);
   Event_Construct(&p->wasStopped);
   Semaphore_Construct(&p->freeSemaphore);
@@ -94,7 +96,7 @@ static void MtSync_Construct(CMtSync *p)
 }
 
 
-#define DEBUG_BUFFER_LOCK   // define it to debug lock state
+// #define DEBUG_BUFFER_LOCK   // define it to debug lock state
 
 #ifdef DEBUG_BUFFER_LOCK
 #include <stdlib.h>
@@ -116,7 +118,7 @@ static void MtSync_Construct(CMtSync *p)
     (p)->csWasEntered = False; }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static UInt32 MtSync_GetNextBlock(CMtSync *p)
 {
   UInt32 numBlocks = 0;
@@ -140,14 +142,14 @@ static UInt32 MtSync_GetNextBlock(CMtSync *p)
 
   // buffer is UNLOCKED here
   Semaphore_Wait(&p->filledSemaphore);
-  LOCK_BUFFER(p);
+  LOCK_BUFFER(p)
   return numBlocks;
 }
 
 
 /* if Writing (Processing) thread was started, we must call MtSync_StopWriting() */
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static void MtSync_StopWriting(CMtSync *p)
 {
   if (!Thread_WasCreated(&p->thread) || p->needStart)
@@ -185,7 +187,7 @@ static void MtSync_StopWriting(CMtSync *p)
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static void MtSync_Destruct(CMtSync *p)
 {
     PRF(printf("\nMtSync_Destruct %p\n", p));
@@ -220,11 +222,11 @@ static void MtSync_Destruct(CMtSync *p)
 
 // #define RINOK_THREAD(x) { if ((x) != 0) return SZ_ERROR_THREAD; }
 // we want to get real system error codes here instead of SZ_ERROR_THREAD
-#define RINOK_THREAD(x) RINOK(x)
+#define RINOK_THREAD(x)  RINOK_WRes(x)
 
 
 // call it before each new file (when new starting is required):
-MY_NO_INLINE
+Z7_NO_INLINE
 static SRes MtSync_Init(CMtSync *p, UInt32 numBlocks)
 {
   WRes wres;
@@ -245,12 +247,12 @@ static WRes MtSync_Create_WRes(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *
   if (p->wasCreated)
     return SZ_OK;
 
-  RINOK_THREAD(CriticalSection_Init(&p->cs));
+  RINOK_THREAD(CriticalSection_Init(&p->cs))
   p->csWasInitialized = True;
   p->csWasEntered = False;
 
-  RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->canStart));
+  RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->canStart))
-  RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->wasStopped));
+  RINOK_THREAD(AutoResetEvent_CreateNotSignaled(&p->wasStopped))
 
   p->needStart = True;
   p->exit = True;  /* p->exit is unused before (canStart) Event.
@@ -259,18 +261,24 @@ static WRes MtSync_Create_WRes(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *
   // return ERROR_TOO_MANY_POSTS; // for debug
   // return EINVAL; // for debug
 
+#ifdef _WIN32
+  if (p->affinityGroup >= 0)
+    wres = Thread_Create_With_Group(&p->thread, startAddress, obj,
+        (unsigned)(UInt32)p->affinityGroup, (CAffinityMask)p->affinityInGroup);
+  else
+#endif
   if (p->affinity != 0)
     wres = Thread_Create_With_Affinity(&p->thread, startAddress, obj, (CAffinityMask)p->affinity);
   else
     wres = Thread_Create(&p->thread, startAddress, obj);
 
-  RINOK_THREAD(wres);
+  RINOK_THREAD(wres)
   p->wasCreated = True;
   return SZ_OK;
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static SRes MtSync_Create(CMtSync *p, THREAD_FUNC_TYPE startAddress, void *obj)
 {
   const WRes wres = MtSync_Create_WRes(p, startAddress, obj);
@@ -519,7 +527,7 @@ static void HashThreadFunc(CMatchFinderMt *mt)
             if (mf->pos > (UInt32)kMtMaxValForNormalize - num)
             {
               const UInt32 subValue = (mf->pos - mf->historySize - 1); // & ~(UInt32)(kNormalizeAlign - 1);
-              Inline_MatchFinder_ReduceOffsets(mf, subValue);
+              MatchFinder_REDUCE_OFFSETS(mf, subValue)
               MatchFinder_Normalize3(subValue, mf->hash + mf->fixedHashSize, (size_t)mf->hashMask + 1);
             }
 
@@ -560,7 +568,7 @@ static void HashThreadFunc(CMatchFinderMt *mt)
 */
 
 
-UInt32 * MY_FAST_CALL  GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
+UInt32 * Z7_FASTCALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
     UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size,
     size_t _cyclicBufferPos, UInt32 _cyclicBufferSize,
     UInt32 *posRes);
@@ -749,7 +757,7 @@ static void BtFillBlock(CMatchFinderMt *p, UInt32 globalBlockIndex)
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static void BtThreadFunc(CMatchFinderMt *mt)
 {
   CMtSync *p = &mt->btSync;
@@ -864,21 +872,22 @@ SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddB
   if (!MatchFinder_Create(mf, historySize, keepAddBufferBefore, matchMaxLen, keepAddBufferAfter, alloc))
     return SZ_ERROR_MEM;
 
-  RINOK(MtSync_Create(&p->hashSync, HashThreadFunc2, p));
+  RINOK(MtSync_Create(&p->hashSync, HashThreadFunc2, p))
-  RINOK(MtSync_Create(&p->btSync, BtThreadFunc2, p));
+  RINOK(MtSync_Create(&p->btSync, BtThreadFunc2, p))
   return SZ_OK;
 }
 
 
 SRes MatchFinderMt_InitMt(CMatchFinderMt *p)
 {
-  RINOK(MtSync_Init(&p->hashSync, kMtHashNumBlocks));
+  RINOK(MtSync_Init(&p->hashSync, kMtHashNumBlocks))
   return MtSync_Init(&p->btSync, kMtBtNumBlocks);
 }
 
 
-static void MatchFinderMt_Init(CMatchFinderMt *p)
+static void MatchFinderMt_Init(void *_p)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   CMatchFinder *mf = MF(p);
   
   p->btBufPos =
@@ -941,7 +950,7 @@ void MatchFinderMt_ReleaseStream(CMatchFinderMt *p)
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static UInt32 MatchFinderMt_GetNextBlock_Bt(CMatchFinderMt *p)
 {
   if (p->failure_LZ_BT)
@@ -981,8 +990,9 @@ static UInt32 MatchFinderMt_GetNextBlock_Bt(CMatchFinderMt *p)
 
 
 
-static const Byte * MatchFinderMt_GetPointerToCurrentPos(CMatchFinderMt *p)
+static const Byte * MatchFinderMt_GetPointerToCurrentPos(void *_p)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   return p->pointerToCurPos;
 }
 
@@ -990,8 +1000,9 @@ static const Byte * MatchFinderMt_GetPointerToCurrentPos(CMatchFinderMt *p)
 #define GET_NEXT_BLOCK_IF_REQUIRED if (p->btBufPos == p->btBufPosLimit) MatchFinderMt_GetNextBlock_Bt(p);
 
 
-static UInt32 MatchFinderMt_GetNumAvailableBytes(CMatchFinderMt *p)
+static UInt32 MatchFinderMt_GetNumAvailableBytes(void *_p)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   if (p->btBufPos != p->btBufPosLimit)
     return p->btNumAvailBytes;
   return MatchFinderMt_GetNextBlock_Bt(p);
@@ -1163,7 +1174,7 @@ UInt32* MatchFinderMt_GetMatches_Bt4(CMatchFinderMt *p, UInt32 *d)
 */
 
 
-static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
+static UInt32 * MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
 {
   UInt32 h2, h3, /* h4, */ c2, c3 /* , c4 */;
   UInt32 *hash = p->hash;
@@ -1179,9 +1190,8 @@ static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
   (hash + kFix3HashSize)[h3] = m;
   // (hash + kFix4HashSize)[h4] = m;
 
-  #define _USE_H2
+  // #define BT5_USE_H2
-
+  // #ifdef BT5_USE_H2
-  #ifdef _USE_H2
   if (c2 >= matchMinPos && cur[(ptrdiff_t)c2 - (ptrdiff_t)m] == cur[0])
   {
     d[1] = m - c2 - 1;
@@ -1198,7 +1208,7 @@ static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
       d[0] = 3;
       d += 2;
 
-      #ifdef _USE_H4
+      #ifdef BT5_USE_H4
       if (c4 >= matchMinPos)
         if (
           cur[(ptrdiff_t)c4 - (ptrdiff_t)m]     == cur[0] &&
@@ -1214,7 +1224,7 @@ static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
     d[0] = 2;
     d += 2;
   }
-  #endif
+  // #endif
   
   if (c3 >= matchMinPos && cur[(ptrdiff_t)c3 - (ptrdiff_t)m] == cur[0])
   {
@@ -1228,7 +1238,7 @@ static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
     d += 2;
   }
 
-  #ifdef _USE_H4
+  #ifdef BT5_USE_H4
   if (c4 >= matchMinPos)
     if (
       cur[(ptrdiff_t)c4 - (ptrdiff_t)m]     == cur[0] &&
@@ -1244,8 +1254,9 @@ static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *d)
 }
 
 
-static UInt32* MatchFinderMt2_GetMatches(CMatchFinderMt *p, UInt32 *d)
+static UInt32 * MatchFinderMt2_GetMatches(void *_p, UInt32 *d)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   const UInt32 *bt = p->btBufPos;
   const UInt32 len = *bt++;
   const UInt32 *btLim = bt + len;
@@ -1268,8 +1279,9 @@ static UInt32* MatchFinderMt2_GetMatches(CMatchFinderMt *p, UInt32 *d)
 
 
 
-static UInt32* MatchFinderMt_GetMatches(CMatchFinderMt *p, UInt32 *d)
+static UInt32 * MatchFinderMt_GetMatches(void *_p, UInt32 *d)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   const UInt32 *bt = p->btBufPos;
   UInt32 len = *bt++;
   const UInt32 avail = p->btNumAvailBytes - 1;
@@ -1316,14 +1328,16 @@ static UInt32* MatchFinderMt_GetMatches(CMatchFinderMt *p, UInt32 *d)
 #define SKIP_HEADER_MT(n) SKIP_HEADER2_MT if (p->btNumAvailBytes-- >= (n)) { const Byte *cur = p->pointerToCurPos; UInt32 *hash = p->hash;
 #define SKIP_FOOTER_MT } INCREASE_LZ_POS p->btBufPos += (size_t)*p->btBufPos + 1; } while (--num != 0);
 
-static void MatchFinderMt0_Skip(CMatchFinderMt *p, UInt32 num)
+static void MatchFinderMt0_Skip(void *_p, UInt32 num)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   SKIP_HEADER2_MT { p->btNumAvailBytes--;
   SKIP_FOOTER_MT
 }
 
-static void MatchFinderMt2_Skip(CMatchFinderMt *p, UInt32 num)
+static void MatchFinderMt2_Skip(void *_p, UInt32 num)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   SKIP_HEADER_MT(2)
       UInt32 h2;
       MT_HASH2_CALC
@@ -1331,8 +1345,9 @@ static void MatchFinderMt2_Skip(CMatchFinderMt *p, UInt32 num)
   SKIP_FOOTER_MT
 }
 
-static void MatchFinderMt3_Skip(CMatchFinderMt *p, UInt32 num)
+static void MatchFinderMt3_Skip(void *_p, UInt32 num)
 {
+  CMatchFinderMt *p = (CMatchFinderMt *)_p;
   SKIP_HEADER_MT(3)
       UInt32 h2, h3;
       MT_HASH3_CALC
@@ -1362,39 +1377,46 @@ static void MatchFinderMt4_Skip(CMatchFinderMt *p, UInt32 num)
 
 void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder2 *vTable)
 {
-  vTable->Init = (Mf_Init_Func)MatchFinderMt_Init;
+  vTable->Init = MatchFinderMt_Init;
-  vTable->GetNumAvailableBytes = (Mf_GetNumAvailableBytes_Func)MatchFinderMt_GetNumAvailableBytes;
+  vTable->GetNumAvailableBytes = MatchFinderMt_GetNumAvailableBytes;
-  vTable->GetPointerToCurrentPos = (Mf_GetPointerToCurrentPos_Func)MatchFinderMt_GetPointerToCurrentPos;
+  vTable->GetPointerToCurrentPos = MatchFinderMt_GetPointerToCurrentPos;
-  vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt_GetMatches;
+  vTable->GetMatches = MatchFinderMt_GetMatches;
   
   switch (MF(p)->numHashBytes)
   {
     case 2:
       p->GetHeadsFunc = GetHeads2;
-      p->MixMatchesFunc = (Mf_Mix_Matches)NULL;
+      p->MixMatchesFunc = NULL;
-      vTable->Skip = (Mf_Skip_Func)MatchFinderMt0_Skip;
+      vTable->Skip = MatchFinderMt0_Skip;
-      vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt2_GetMatches;
+      vTable->GetMatches = MatchFinderMt2_GetMatches;
       break;
     case 3:
       p->GetHeadsFunc = MF(p)->bigHash ? GetHeads3b : GetHeads3;
-      p->MixMatchesFunc = (Mf_Mix_Matches)MixMatches2;
+      p->MixMatchesFunc = MixMatches2;
-      vTable->Skip = (Mf_Skip_Func)MatchFinderMt2_Skip;
+      vTable->Skip = MatchFinderMt2_Skip;
       break;
     case 4:
       p->GetHeadsFunc = MF(p)->bigHash ? GetHeads4b : GetHeads4;
 
       // it's fast inline version of GetMatches()
-      // vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt_GetMatches_Bt4;
+      // vTable->GetMatches = MatchFinderMt_GetMatches_Bt4;
 
-      p->MixMatchesFunc = (Mf_Mix_Matches)MixMatches3;
+      p->MixMatchesFunc = MixMatches3;
-      vTable->Skip = (Mf_Skip_Func)MatchFinderMt3_Skip;
+      vTable->Skip = MatchFinderMt3_Skip;
       break;
     default:
       p->GetHeadsFunc = MF(p)->bigHash ? GetHeads5b : GetHeads5;
-      p->MixMatchesFunc = (Mf_Mix_Matches)MixMatches4;
+      p->MixMatchesFunc = MixMatches4;
       vTable->Skip =
-          (Mf_Skip_Func)MatchFinderMt3_Skip;
+          MatchFinderMt3_Skip;
-          // (Mf_Skip_Func)MatchFinderMt4_Skip;
+          // MatchFinderMt4_Skip;
       break;
   }
 }
+
+#undef RINOK_THREAD
+#undef PRF
+#undef MF
+#undef GetUi24hi_from32
+#undef LOCK_BUFFER
+#undef UNLOCK_BUFFER

C/LzFindMt.h

@@ -1,19 +1,21 @@
 /* LzFindMt.h -- multithreaded Match finder for LZ algorithms
-2021-07-12 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __LZ_FIND_MT_H
+#ifndef ZIP7_INC_LZ_FIND_MT_H
-#define __LZ_FIND_MT_H
+#define ZIP7_INC_LZ_FIND_MT_H
 
 #include "LzFind.h"
 #include "Threads.h"
 
 EXTERN_C_BEGIN
 
-typedef struct _CMtSync
+typedef struct
 {
   UInt32 numProcessedBlocks;
-  CThread thread;
+  Int32 affinityGroup;
+  UInt64 affinityInGroup;
   UInt64 affinity;
+  CThread thread;
 
   BoolInt wasCreated;
   BoolInt needStart;
@@ -31,7 +33,10 @@ typedef struct _CMtSync
   // UInt32 numBlocks_Sent;
 } CMtSync;
 
-typedef UInt32 * (*Mf_Mix_Matches)(void *p, UInt32 matchMinPos, UInt32 *distances);
+
+struct CMatchFinderMt_;
+
+typedef UInt32 * (*Mf_Mix_Matches)(struct CMatchFinderMt_ *p, UInt32 matchMinPos, UInt32 *distances);
 
 /* kMtCacheLineDummy must be >= size_of_CPU_cache_line */
 #define kMtCacheLineDummy 128
@@ -39,7 +44,7 @@ typedef UInt32 * (*Mf_Mix_Matches)(void *p, UInt32 matchMinPos, UInt32 *distance
 typedef void (*Mf_GetHeads)(const Byte *buffer, UInt32 pos,
   UInt32 *hash, UInt32 hashMask, UInt32 *heads, UInt32 numHeads, const UInt32 *crc);
 
-typedef struct _CMatchFinderMt
+typedef struct CMatchFinderMt_
 {
   /* LZ */
   const Byte *pointerToCurPos;

C/LzFindOpt.c

@@ -1,5 +1,5 @@
 /* LzFindOpt.c -- multithreaded Match finder for LZ algorithms
-2021-07-13 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -41,8 +41,8 @@ UInt64 g_NumIters_Bytes;
 // #define CYC_TO_POS_OFFSET 1 // for debug
 
 /*
-MY_NO_INLINE
+Z7_NO_INLINE
-UInt32 * MY_FAST_CALL GetMatchesSpecN_1(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
+UInt32 * Z7_FASTCALL GetMatchesSpecN_1(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
     UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size, UInt32 *posRes)
 {
   do
@@ -214,13 +214,13 @@ else
   to eliminate "movsx" BUG in old MSVC x64 compiler.
 */
 
-UInt32 * MY_FAST_CALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
+UInt32 * Z7_FASTCALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
     UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size,
     size_t _cyclicBufferPos, UInt32 _cyclicBufferSize,
     UInt32 *posRes);
 
-MY_NO_INLINE
+Z7_NO_INLINE
-UInt32 * MY_FAST_CALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
+UInt32 * Z7_FASTCALL GetMatchesSpecN_2(const Byte *lenLimit, size_t pos, const Byte *cur, CLzRef *son,
     UInt32 _cutValue, UInt32 *d, size_t _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size,
     size_t _cyclicBufferPos, UInt32 _cyclicBufferSize,
     UInt32 *posRes)
@@ -404,7 +404,7 @@ else
 /*
 typedef UInt32 uint32plus; // size_t
 
-UInt32 * MY_FAST_CALL GetMatchesSpecN_3(uint32plus lenLimit, size_t pos, const Byte *cur, CLzRef *son,
+UInt32 * Z7_FASTCALL GetMatchesSpecN_3(uint32plus lenLimit, size_t pos, const Byte *cur, CLzRef *son,
     UInt32 _cutValue, UInt32 *d, uint32plus _maxLen, const UInt32 *hash, const UInt32 *limit, const UInt32 *size,
     size_t _cyclicBufferPos, UInt32 _cyclicBufferSize,
     UInt32 *posRes)

C/LzHash.h

@@ -1,8 +1,8 @@
-/* LzHash.h -- HASH functions for LZ algorithms
+/* LzHash.h -- HASH constants for LZ algorithms
-2019-10-30 : Igor Pavlov : Public domain */
+2023-03-05 : Igor Pavlov : Public domain */
 
-#ifndef __LZ_HASH_H
+#ifndef ZIP7_INC_LZ_HASH_H
-#define __LZ_HASH_H
+#define ZIP7_INC_LZ_HASH_H
 
 /*
   (kHash2Size >= (1 <<  8)) : Required

C/Lzma2Dec.c

@@ -1,5 +1,5 @@
 /* Lzma2Dec.c -- LZMA2 Decoder
-2021-02-09 : Igor Pavlov : Public domain */
+2024-03-01 : Igor Pavlov : Public domain */
 
 /* #define SHOW_DEBUG_INFO */
 
@@ -71,14 +71,14 @@ static SRes Lzma2Dec_GetOldProps(Byte prop, Byte *props)
 SRes Lzma2Dec_AllocateProbs(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc)
 {
   Byte props[LZMA_PROPS_SIZE];
-  RINOK(Lzma2Dec_GetOldProps(prop, props));
+  RINOK(Lzma2Dec_GetOldProps(prop, props))
   return LzmaDec_AllocateProbs(&p->decoder, props, LZMA_PROPS_SIZE, alloc);
 }
 
 SRes Lzma2Dec_Allocate(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc)
 {
   Byte props[LZMA_PROPS_SIZE];
-  RINOK(Lzma2Dec_GetOldProps(prop, props));
+  RINOK(Lzma2Dec_GetOldProps(prop, props))
   return LzmaDec_Allocate(&p->decoder, props, LZMA_PROPS_SIZE, alloc);
 }
 
@@ -157,8 +157,10 @@ static unsigned Lzma2Dec_UpdateState(CLzma2Dec *p, Byte b)
       p->decoder.prop.lp = (Byte)lp;
       return LZMA2_STATE_DATA;
     }
+    
+    default:
+      return LZMA2_STATE_ERROR;
   }
-  return LZMA2_STATE_ERROR;
 }
 
 static void LzmaDec_UpdateWithUncompressed(CLzmaDec *p, const Byte *src, SizeT size)
@@ -474,8 +476,8 @@ SRes Lzma2Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen,
   SizeT outSize = *destLen, inSize = *srcLen;
   *destLen = *srcLen = 0;
   *status = LZMA_STATUS_NOT_SPECIFIED;
-  Lzma2Dec_Construct(&p);
+  Lzma2Dec_CONSTRUCT(&p)
-  RINOK(Lzma2Dec_AllocateProbs(&p, prop, alloc));
+  RINOK(Lzma2Dec_AllocateProbs(&p, prop, alloc))
   p.decoder.dic = dest;
   p.decoder.dicBufSize = outSize;
   Lzma2Dec_Init(&p);
@@ -487,3 +489,5 @@ SRes Lzma2Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen,
   Lzma2Dec_FreeProbs(&p, alloc);
   return res;
 }
+
+#undef PRF

C/Lzma2Dec.h

@@ -1,8 +1,8 @@
 /* Lzma2Dec.h -- LZMA2 Decoder
-2018-02-19 : Igor Pavlov : Public domain */
+2023-03-03 : Igor Pavlov : Public domain */
 
-#ifndef __LZMA2_DEC_H
+#ifndef ZIP7_INC_LZMA2_DEC_H
-#define __LZMA2_DEC_H
+#define ZIP7_INC_LZMA2_DEC_H
 
 #include "LzmaDec.h"
 
@@ -22,9 +22,10 @@ typedef struct
   CLzmaDec decoder;
 } CLzma2Dec;
 
-#define Lzma2Dec_Construct(p) LzmaDec_Construct(&(p)->decoder)
+#define Lzma2Dec_CONSTRUCT(p)  LzmaDec_CONSTRUCT(&(p)->decoder)
-#define Lzma2Dec_FreeProbs(p, alloc) LzmaDec_FreeProbs(&(p)->decoder, alloc)
+#define Lzma2Dec_Construct(p)  Lzma2Dec_CONSTRUCT(p)
-#define Lzma2Dec_Free(p, alloc) LzmaDec_Free(&(p)->decoder, alloc)
+#define Lzma2Dec_FreeProbs(p, alloc)  LzmaDec_FreeProbs(&(p)->decoder, alloc)
+#define Lzma2Dec_Free(p, alloc)  LzmaDec_Free(&(p)->decoder, alloc)
 
 SRes Lzma2Dec_AllocateProbs(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc);
 SRes Lzma2Dec_Allocate(CLzma2Dec *p, Byte prop, ISzAllocPtr alloc);
@@ -90,7 +91,7 @@ Lzma2Dec_GetUnpackExtra() returns the value that shows
     at current input positon.
 */
 
-#define Lzma2Dec_GetUnpackExtra(p) ((p)->isExtraMode ? (p)->unpackSize : 0);
+#define Lzma2Dec_GetUnpackExtra(p)  ((p)->isExtraMode ? (p)->unpackSize : 0)
 
 
 /* ---------- One Call Interface ---------- */

C/Lzma2DecMt.c

@@ -1,44 +1,44 @@
 /* Lzma2DecMt.c -- LZMA2 Decoder Multi-thread
-2021-04-01 : Igor Pavlov : Public domain */
+2023-04-13 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 // #define SHOW_DEBUG_INFO
-
+// #define Z7_ST
-// #define _7ZIP_ST
 
 #ifdef SHOW_DEBUG_INFO
 #include <stdio.h>
 #endif
 
-#ifndef _7ZIP_ST
-#ifdef SHOW_DEBUG_INFO
-#define PRF(x) x
-#else
-#define PRF(x)
-#endif
-#define PRF_STR(s) PRF(printf("\n" s "\n"))
-#define PRF_STR_INT_2(s, d1, d2) PRF(printf("\n" s " %d %d\n", (unsigned)d1, (unsigned)d2))
-#endif
-
 #include "Alloc.h"
 
 #include "Lzma2Dec.h"
 #include "Lzma2DecMt.h"
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 #include "MtDec.h"
 
 #define LZMA2DECMT_OUT_BLOCK_MAX_DEFAULT (1 << 28)
 #endif
 
 
+#ifndef Z7_ST
+#ifdef SHOW_DEBUG_INFO
+#define PRF(x) x
+#else
+#define PRF(x)
+#endif
+#define PRF_STR(s) PRF(printf("\n" s "\n");)
+#define PRF_STR_INT_2(s, d1, d2) PRF(printf("\n" s " %d %d\n", (unsigned)d1, (unsigned)d2);)
+#endif
+
+
 void Lzma2DecMtProps_Init(CLzma2DecMtProps *p)
 {
   p->inBufSize_ST = 1 << 20;
   p->outStep_ST = 1 << 20;
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   p->numThreads = 1;
   p->inBufSize_MT = 1 << 18;
   p->outBlockMax = LZMA2DECMT_OUT_BLOCK_MAX_DEFAULT;
@@ -48,7 +48,7 @@ void Lzma2DecMtProps_Init(CLzma2DecMtProps *p)
 
 
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
 /* ---------- CLzma2DecMtThread ---------- */
 
@@ -81,7 +81,7 @@ typedef struct
 
 /* ---------- CLzma2DecMt ---------- */
 
-typedef struct
+struct CLzma2DecMt
 {
   // ISzAllocPtr alloc;
   ISzAllocPtr allocMid;
@@ -90,9 +90,9 @@ typedef struct
   CLzma2DecMtProps props;
   Byte prop;
   
-  ISeqInStream *inStream;
+  ISeqInStreamPtr inStream;
-  ISeqOutStream *outStream;
+  ISeqOutStreamPtr outStream;
-  ICompressProgress *progress;
+  ICompressProgressPtr progress;
 
   BoolInt finishMode;
   BoolInt outSize_Defined;
@@ -111,14 +111,13 @@ typedef struct
   size_t inPos;
   size_t inLim;
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   UInt64 outProcessed_Parse;
   BoolInt mtc_WasConstructed;
   CMtDec mtc;
-  CLzma2DecMtThread coders[MTDEC__THREADS_MAX];
+  CLzma2DecMtThread coders[MTDEC_THREADS_MAX];
   #endif
-
+};
-} CLzma2DecMt;
 
 
 
@@ -142,11 +141,11 @@ CLzma2DecMtHandle Lzma2DecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid)
 
   // Lzma2DecMtProps_Init(&p->props);
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   p->mtc_WasConstructed = False;
   {
     unsigned i;
-    for (i = 0; i < MTDEC__THREADS_MAX; i++)
+    for (i = 0; i < MTDEC_THREADS_MAX; i++)
     {
       CLzma2DecMtThread *t = &p->coders[i];
       t->dec_created = False;
@@ -156,16 +155,16 @@ CLzma2DecMtHandle Lzma2DecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid)
   }
   #endif
 
-  return p;
+  return (CLzma2DecMtHandle)(void *)p;
 }
 
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
 static void Lzma2DecMt_FreeOutBufs(CLzma2DecMt *p)
 {
   unsigned i;
-  for (i = 0; i < MTDEC__THREADS_MAX; i++)
+  for (i = 0; i < MTDEC_THREADS_MAX; i++)
   {
     CLzma2DecMtThread *t = &p->coders[i];
     if (t->outBuf)
@@ -196,13 +195,15 @@ static void Lzma2DecMt_FreeSt(CLzma2DecMt *p)
 }
 
 
-void Lzma2DecMt_Destroy(CLzma2DecMtHandle pp)
+// #define GET_CLzma2DecMt_p CLzma2DecMt *p = (CLzma2DecMt *)(void *)pp;
+
+void Lzma2DecMt_Destroy(CLzma2DecMtHandle p)
 {
-  CLzma2DecMt *p = (CLzma2DecMt *)pp;
+  // GET_CLzma2DecMt_p
 
   Lzma2DecMt_FreeSt(p);
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
 
   if (p->mtc_WasConstructed)
   {
@@ -211,7 +212,7 @@ void Lzma2DecMt_Destroy(CLzma2DecMtHandle pp)
   }
   {
     unsigned i;
-    for (i = 0; i < MTDEC__THREADS_MAX; i++)
+    for (i = 0; i < MTDEC_THREADS_MAX; i++)
     {
       CLzma2DecMtThread *t = &p->coders[i];
       if (t->dec_created)
@@ -226,19 +227,19 @@ void Lzma2DecMt_Destroy(CLzma2DecMtHandle pp)
 
   #endif
 
-  ISzAlloc_Free(p->alignOffsetAlloc.baseAlloc, pp);
+  ISzAlloc_Free(p->alignOffsetAlloc.baseAlloc, p);
 }
 
 
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
 static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCallbackInfo *cc)
 {
   CLzma2DecMt *me = (CLzma2DecMt *)obj;
   CLzma2DecMtThread *t = &me->coders[coderIndex];
 
-  PRF_STR_INT_2("Parse", coderIndex, cc->srcSize);
+  PRF_STR_INT_2("Parse", coderIndex, cc->srcSize)
 
   cc->state = MTDEC_PARSE_CONTINUE;
 
@@ -246,7 +247,7 @@ static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCa
   {
     if (!t->dec_created)
     {
-      Lzma2Dec_Construct(&t->dec);
+      Lzma2Dec_CONSTRUCT(&t->dec)
       t->dec_created = True;
       AlignOffsetAlloc_CreateVTable(&t->alloc);
       {
@@ -297,7 +298,7 @@ static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCa
     // that must be finished at position <= outBlockMax.
 
     {
-      const SizeT srcOrig = cc->srcSize;
+      const size_t srcOrig = cc->srcSize;
       SizeT srcSize_Point = 0;
       SizeT dicPos_Point = 0;
       
@@ -306,10 +307,10 @@ static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCa
 
       for (;;)
       {
-        SizeT srcCur = srcOrig - cc->srcSize;
+        SizeT srcCur = (SizeT)(srcOrig - cc->srcSize);
         
         status = Lzma2Dec_Parse(&t->dec,
-            limit - t->dec.decoder.dicPos,
+            (SizeT)limit - t->dec.decoder.dicPos,
             cc->src + cc->srcSize, &srcCur,
             checkFinishBlock);
 
@@ -333,7 +334,7 @@ static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCa
           if (t->dec.decoder.dicPos >= (1 << 14))
             break;
           dicPos_Point = t->dec.decoder.dicPos;
-          srcSize_Point = cc->srcSize;
+          srcSize_Point = (SizeT)cc->srcSize;
           continue;
         }
 
@@ -391,7 +392,7 @@ static void Lzma2DecMt_MtCallback_Parse(void *obj, unsigned coderIndex, CMtDecCa
             if (unpackRem != 0)
             {
               /* we also reserve space for max possible number of output bytes of current LZMA chunk */
-              SizeT rem = limit - dicPos;
+              size_t rem = limit - dicPos;
               if (rem > unpackRem)
                 rem = unpackRem;
               dicPos += rem;
@@ -444,7 +445,7 @@ static SRes Lzma2DecMt_MtCallback_PreCode(void *pp, unsigned coderIndex)
   }
 
   t->dec.decoder.dic = dest;
-  t->dec.decoder.dicBufSize = t->outPreSize;
+  t->dec.decoder.dicBufSize = (SizeT)t->outPreSize;
 
   t->needInit = True;
 
@@ -462,7 +463,7 @@ static SRes Lzma2DecMt_MtCallback_Code(void *pp, unsigned coderIndex,
 
   UNUSED_VAR(srcFinished)
 
-  PRF_STR_INT_2("Code", coderIndex, srcSize);
+  PRF_STR_INT_2("Code", coderIndex, srcSize)
 
   *inCodePos = t->inCodeSize;
   *outCodePos = 0;
@@ -476,13 +477,13 @@ static SRes Lzma2DecMt_MtCallback_Code(void *pp, unsigned coderIndex,
 
   {
     ELzmaStatus status;
-    size_t srcProcessed = srcSize;
+    SizeT srcProcessed = (SizeT)srcSize;
     BoolInt blockWasFinished =
         ((int)t->parseStatus == LZMA_STATUS_FINISHED_WITH_MARK
         || t->parseStatus == LZMA2_PARSE_STATUS_NEW_BLOCK);
     
     SRes res = Lzma2Dec_DecodeToDic(&t->dec,
-        t->outPreSize,
+        (SizeT)t->outPreSize,
         src, &srcProcessed,
         blockWasFinished ? LZMA_FINISH_END : LZMA_FINISH_ANY,
         &status);
@@ -540,7 +541,7 @@ static SRes Lzma2DecMt_MtCallback_Write(void *pp, unsigned coderIndex,
   UNUSED_VAR(srcSize)
   UNUSED_VAR(isCross)
 
-  PRF_STR_INT_2("Write", coderIndex, srcSize);
+  PRF_STR_INT_2("Write", coderIndex, srcSize)
 
   *needContinue = False;
   *canRecode = True;
@@ -588,7 +589,7 @@ static SRes Lzma2DecMt_MtCallback_Write(void *pp, unsigned coderIndex,
         *needContinue = needContinue2;
         return SZ_OK;
       }
-      RINOK(MtProgress_ProgressAdd(&me->mtc.mtProgress, 0, 0));
+      RINOK(MtProgress_ProgressAdd(&me->mtc.mtProgress, 0, 0))
     }
   }
   
@@ -611,11 +612,11 @@ static SRes Lzma2Dec_Prepare_ST(CLzma2DecMt *p)
 {
   if (!p->dec_created)
   {
-    Lzma2Dec_Construct(&p->dec);
+    Lzma2Dec_CONSTRUCT(&p->dec)
     p->dec_created = True;
   }
 
-  RINOK(Lzma2Dec_Allocate(&p->dec, p->prop, &p->alignOffsetAlloc.vt));
+  RINOK(Lzma2Dec_Allocate(&p->dec, p->prop, &p->alignOffsetAlloc.vt))
 
   if (!p->inBuf || p->inBufSize != p->props.inBufSize_ST)
   {
@@ -634,7 +635,7 @@ static SRes Lzma2Dec_Prepare_ST(CLzma2DecMt *p)
 
 
 static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
-    #ifndef _7ZIP_ST
+    #ifndef Z7_ST
     , BoolInt tMode
     #endif
     )
@@ -646,7 +647,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
 
   CLzma2Dec *dec;
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   if (tMode)
   {
     Lzma2DecMt_FreeOutBufs(p);
@@ -654,7 +655,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
   }
   #endif
 
-  RINOK(Lzma2Dec_Prepare_ST(p));
+  RINOK(Lzma2Dec_Prepare_ST(p))
 
   dec = &p->dec;
 
@@ -681,7 +682,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
 
     if (inPos == inLim)
     {
-      #ifndef _7ZIP_ST
+      #ifndef Z7_ST
       if (tMode)
       {
         inData = MtDec_Read(&p->mtc, &inLim);
@@ -710,7 +711,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
     {
       SizeT next = dec->decoder.dicBufSize;
       if (next - wrPos > p->props.outStep_ST)
-        next = wrPos + p->props.outStep_ST;
+        next = wrPos + (SizeT)p->props.outStep_ST;
       size = next - dicPos;
     }
 
@@ -726,7 +727,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
       }
     }
 
-    inProcessed = inLim - inPos;
+    inProcessed = (SizeT)(inLim - inPos);
     
     res = Lzma2Dec_DecodeToDic(dec, dicPos + size, inData + inPos, &inProcessed, finishMode, &status);
 
@@ -755,7 +756,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
         dec->decoder.dicPos = 0;
       wrPos = dec->decoder.dicPos;
 
-      RINOK(res2);
+      RINOK(res2)
 
       if (needStop)
       {
@@ -788,7 +789,7 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
       UInt64 outDelta = p->outProcessed - outPrev;
       if (inDelta >= (1 << 22) || outDelta >= (1 << 22))
       {
-        RINOK(ICompressProgress_Progress(p->progress, p->inProcessed, p->outProcessed));
+        RINOK(ICompressProgress_Progress(p->progress, p->inProcessed, p->outProcessed))
         inPrev = p->inProcessed;
         outPrev = p->outProcessed;
       }
@@ -798,20 +799,20 @@ static SRes Lzma2Dec_Decode_ST(CLzma2DecMt *p
 
 
 
-SRes Lzma2DecMt_Decode(CLzma2DecMtHandle pp,
+SRes Lzma2DecMt_Decode(CLzma2DecMtHandle p,
     Byte prop,
     const CLzma2DecMtProps *props,
-    ISeqOutStream *outStream, const UInt64 *outDataSize, int finishMode,
+    ISeqOutStreamPtr outStream, const UInt64 *outDataSize, int finishMode,
     // Byte *outBuf, size_t *outBufSize,
-    ISeqInStream *inStream,
+    ISeqInStreamPtr inStream,
     // const Byte *inData, size_t inDataSize,
     UInt64 *inProcessed,
     // UInt64 *outProcessed,
     int *isMT,
-    ICompressProgress *progress)
+    ICompressProgressPtr progress)
 {
-  CLzma2DecMt *p = (CLzma2DecMt *)pp;
+  // GET_CLzma2DecMt_p
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   BoolInt tMode;
   #endif
 
@@ -845,7 +846,7 @@ SRes Lzma2DecMt_Decode(CLzma2DecMtHandle pp,
   *isMT = False;
 
   
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
 
   tMode = False;
 
@@ -939,7 +940,7 @@ SRes Lzma2DecMt_Decode(CLzma2DecMtHandle pp,
       p->readWasFinished = p->mtc.readWasFinished;
       p->inProcessed = p->mtc.inProcessed;
       
-      PRF_STR("----- decoding ST -----");
+      PRF_STR("----- decoding ST -----")
     }
   }
 
@@ -950,7 +951,7 @@ SRes Lzma2DecMt_Decode(CLzma2DecMtHandle pp,
 
   {
     SRes res = Lzma2Dec_Decode_ST(p
-        #ifndef _7ZIP_ST
+        #ifndef Z7_ST
         , tMode
         #endif
         );
@@ -967,7 +968,7 @@ SRes Lzma2DecMt_Decode(CLzma2DecMtHandle pp,
       res = p->readRes;
     
     /*
-    #ifndef _7ZIP_ST
+    #ifndef Z7_ST
     if (res == SZ_OK && tMode && p->mtc.parseRes != SZ_OK)
       res = p->mtc.parseRes;
     #endif
@@ -980,13 +981,13 @@ SRes Lzma2DecMt_Decode(CLzma2DecMtHandle pp,
 
 /* ---------- Read from CLzma2DecMtHandle Interface ---------- */
 
-SRes Lzma2DecMt_Init(CLzma2DecMtHandle pp,
+SRes Lzma2DecMt_Init(CLzma2DecMtHandle p,
     Byte prop,
     const CLzma2DecMtProps *props,
     const UInt64 *outDataSize, int finishMode,
-    ISeqInStream *inStream)
+    ISeqInStreamPtr inStream)
 {
-  CLzma2DecMt *p = (CLzma2DecMt *)pp;
+  // GET_CLzma2DecMt_p
 
   if (prop > 40)
     return SZ_ERROR_UNSUPPORTED;
@@ -1015,11 +1016,11 @@ SRes Lzma2DecMt_Init(CLzma2DecMtHandle pp,
 }
 
 
-SRes Lzma2DecMt_Read(CLzma2DecMtHandle pp,
+SRes Lzma2DecMt_Read(CLzma2DecMtHandle p,
     Byte *data, size_t *outSize,
     UInt64 *inStreamProcessed)
 {
-  CLzma2DecMt *p = (CLzma2DecMt *)pp;
+  // GET_CLzma2DecMt_p
   ELzmaFinishMode finishMode;
   SRes readRes;
   size_t size = *outSize;
@@ -1055,8 +1056,8 @@ SRes Lzma2DecMt_Read(CLzma2DecMtHandle pp,
       readRes = ISeqInStream_Read(p->inStream, p->inBuf, &p->inLim);
     }
 
-    inCur = p->inLim - p->inPos;
+    inCur = (SizeT)(p->inLim - p->inPos);
-    outCur = size;
+    outCur = (SizeT)size;
 
     res = Lzma2Dec_DecodeToBuf(&p->dec, data, &outCur,
         p->inBuf + p->inPos, &inCur, finishMode, &status);
@@ -1088,3 +1089,7 @@ SRes Lzma2DecMt_Read(CLzma2DecMtHandle pp,
       return readRes;
   }
 }
+
+#undef PRF
+#undef PRF_STR
+#undef PRF_STR_INT_2

C/Lzma2DecMt.h

@@ -1,8 +1,8 @@
 /* Lzma2DecMt.h -- LZMA2 Decoder Multi-thread
-2018-02-17 : Igor Pavlov : Public domain */
+2023-04-13 : Igor Pavlov : Public domain */
 
-#ifndef __LZMA2_DEC_MT_H
+#ifndef ZIP7_INC_LZMA2_DEC_MT_H
-#define __LZMA2_DEC_MT_H
+#define ZIP7_INC_LZMA2_DEC_MT_H
 
 #include "7zTypes.h"
 
@@ -13,7 +13,7 @@ typedef struct
   size_t inBufSize_ST;
   size_t outStep_ST;
   
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   unsigned numThreads;
   size_t inBufSize_MT;
   size_t outBlockMax;
@@ -38,7 +38,9 @@ SRes:
   SZ_ERROR_THREAD - error in multithreading functions (only for Mt version)
 */
 
-typedef void * CLzma2DecMtHandle;
+typedef struct CLzma2DecMt CLzma2DecMt;
+typedef CLzma2DecMt * CLzma2DecMtHandle;
+// Z7_DECLARE_HANDLE(CLzma2DecMtHandle)
 
 CLzma2DecMtHandle Lzma2DecMt_Create(ISzAllocPtr alloc, ISzAllocPtr allocMid);
 void Lzma2DecMt_Destroy(CLzma2DecMtHandle p);
@@ -46,11 +48,11 @@ void Lzma2DecMt_Destroy(CLzma2DecMtHandle p);
 SRes Lzma2DecMt_Decode(CLzma2DecMtHandle p,
     Byte prop,
     const CLzma2DecMtProps *props,
-    ISeqOutStream *outStream,
+    ISeqOutStreamPtr outStream,
     const UInt64 *outDataSize, // NULL means undefined
     int finishMode,            // 0 - partial unpacking is allowed, 1 - if lzma2 stream must be finished
     // Byte *outBuf, size_t *outBufSize,
-    ISeqInStream *inStream,
+    ISeqInStreamPtr inStream,
     // const Byte *inData, size_t inDataSize,
     
     // out variables:
@@ -58,7 +60,7 @@ SRes Lzma2DecMt_Decode(CLzma2DecMtHandle p,
     int *isMT,  /* out: (*isMT == 0), if single thread decoding was used */
 
     // UInt64 *outProcessed,
-    ICompressProgress *progress);
+    ICompressProgressPtr progress);
 
 
 /* ---------- Read from CLzma2DecMtHandle Interface ---------- */
@@ -67,7 +69,7 @@ SRes Lzma2DecMt_Init(CLzma2DecMtHandle pp,
     Byte prop,
     const CLzma2DecMtProps *props,
     const UInt64 *outDataSize, int finishMode,
-    ISeqInStream *inStream);
+    ISeqInStreamPtr inStream);
 
 SRes Lzma2DecMt_Read(CLzma2DecMtHandle pp,
     Byte *data, size_t *outSize,

C/Lzma2Enc.c

@@ -1,18 +1,18 @@
 /* Lzma2Enc.c -- LZMA2 Encoder
-2021-02-09 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include <string.h>
 
-/* #define _7ZIP_ST */
+/* #define Z7_ST */
 
 #include "Lzma2Enc.h"
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 #include "MtCoder.h"
 #else
-#define MTCODER__THREADS_MAX 1
+#define MTCODER_THREADS_MAX 1
 #endif
 
 #define LZMA2_CONTROL_LZMA (1 << 7)
@@ -40,7 +40,7 @@
 typedef struct
 {
   ISeqInStream vt;
-  ISeqInStream *realStream;
+  ISeqInStreamPtr realStream;
   UInt64 limit;
   UInt64 processed;
   int finished;
@@ -53,15 +53,15 @@ static void LimitedSeqInStream_Init(CLimitedSeqInStream *p)
   p->finished = 0;
 }
 
-static SRes LimitedSeqInStream_Read(const ISeqInStream *pp, void *data, size_t *size)
+static SRes LimitedSeqInStream_Read(ISeqInStreamPtr pp, void *data, size_t *size)
 {
-  CLimitedSeqInStream *p = CONTAINER_FROM_VTBL(pp, CLimitedSeqInStream, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CLimitedSeqInStream)
   size_t size2 = *size;
   SRes res = SZ_OK;
   
   if (p->limit != (UInt64)(Int64)-1)
   {
-    UInt64 rem = p->limit - p->processed;
+    const UInt64 rem = p->limit - p->processed;
     if (size2 > rem)
       size2 = (size_t)rem;
   }
@@ -95,8 +95,8 @@ static SRes Lzma2EncInt_InitStream(CLzma2EncInt *p, const CLzma2EncProps *props)
   {
     SizeT propsSize = LZMA_PROPS_SIZE;
     Byte propsEncoded[LZMA_PROPS_SIZE];
-    RINOK(LzmaEnc_SetProps(p->enc, &props->lzmaProps));
+    RINOK(LzmaEnc_SetProps(p->enc, &props->lzmaProps))
-    RINOK(LzmaEnc_WriteProperties(p->enc, propsEncoded, &propsSize));
+    RINOK(LzmaEnc_WriteProperties(p->enc, propsEncoded, &propsSize))
     p->propsByte = propsEncoded[0];
     p->propsAreSet = True;
   }
@@ -111,23 +111,23 @@ static void Lzma2EncInt_InitBlock(CLzma2EncInt *p)
 }
 
 
-SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle pp, ISeqInStream *inStream, UInt32 keepWindowSize,
+SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle p, ISeqInStreamPtr inStream, UInt32 keepWindowSize,
     ISzAllocPtr alloc, ISzAllocPtr allocBig);
-SRes LzmaEnc_MemPrepare(CLzmaEncHandle pp, const Byte *src, SizeT srcLen,
+SRes LzmaEnc_MemPrepare(CLzmaEncHandle p, const Byte *src, SizeT srcLen,
     UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig);
-SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, BoolInt reInit,
+SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle p, BoolInt reInit,
     Byte *dest, size_t *destLen, UInt32 desiredPackSize, UInt32 *unpackSize);
-const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle pp);
+const Byte *LzmaEnc_GetCurBuf(CLzmaEncHandle p);
-void LzmaEnc_Finish(CLzmaEncHandle pp);
+void LzmaEnc_Finish(CLzmaEncHandle p);
-void LzmaEnc_SaveState(CLzmaEncHandle pp);
+void LzmaEnc_SaveState(CLzmaEncHandle p);
-void LzmaEnc_RestoreState(CLzmaEncHandle pp);
+void LzmaEnc_RestoreState(CLzmaEncHandle p);
 
 /*
-UInt32 LzmaEnc_GetNumAvailableBytes(CLzmaEncHandle pp);
+UInt32 LzmaEnc_GetNumAvailableBytes(CLzmaEncHandle p);
 */
 
 static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf,
-    size_t *packSizeRes, ISeqOutStream *outStream)
+    size_t *packSizeRes, ISeqOutStreamPtr outStream)
 {
   size_t packSizeLimit = *packSizeRes;
   size_t packSize = packSizeLimit;
@@ -167,7 +167,7 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf,
 
     while (unpackSize > 0)
     {
-      UInt32 u = (unpackSize < LZMA2_COPY_CHUNK_SIZE) ? unpackSize : LZMA2_COPY_CHUNK_SIZE;
+      const UInt32 u = (unpackSize < LZMA2_COPY_CHUNK_SIZE) ? unpackSize : LZMA2_COPY_CHUNK_SIZE;
       if (packSizeLimit - destPos < u + 3)
         return SZ_ERROR_OUTPUT_EOF;
       outBuf[destPos++] = (Byte)(p->srcPos == 0 ? LZMA2_CONTROL_COPY_RESET_DIC : LZMA2_CONTROL_COPY_NO_RESET);
@@ -196,9 +196,9 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf,
 
   {
     size_t destPos = 0;
-    UInt32 u = unpackSize - 1;
+    const UInt32 u = unpackSize - 1;
-    UInt32 pm = (UInt32)(packSize - 1);
+    const UInt32 pm = (UInt32)(packSize - 1);
-    unsigned mode = (p->srcPos == 0) ? 3 : (p->needInitState ? (p->needInitProp ? 2 : 1) : 0);
+    const unsigned mode = (p->srcPos == 0) ? 3 : (p->needInitState ? (p->needInitProp ? 2 : 1) : 0);
 
     PRF(printf("               "));
 
@@ -231,10 +231,11 @@ static SRes Lzma2EncInt_EncodeSubblock(CLzma2EncInt *p, Byte *outBuf,
 void Lzma2EncProps_Init(CLzma2EncProps *p)
 {
   LzmaEncProps_Init(&p->lzmaProps);
-  p->blockSize = LZMA2_ENC_PROPS__BLOCK_SIZE__AUTO;
+  p->blockSize = LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO;
   p->numBlockThreads_Reduced = -1;
   p->numBlockThreads_Max = -1;
   p->numTotalThreads = -1;
+  p->numThreadGroups = 0;
 }
 
 void Lzma2EncProps_Normalize(CLzma2EncProps *p)
@@ -251,8 +252,8 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p)
   t2 = p->numBlockThreads_Max;
   t3 = p->numTotalThreads;
 
-  if (t2 > MTCODER__THREADS_MAX)
+  if (t2 > MTCODER_THREADS_MAX)
-    t2 = MTCODER__THREADS_MAX;
+    t2 = MTCODER_THREADS_MAX;
 
   if (t3 <= 0)
   {
@@ -268,8 +269,8 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p)
       t1 = 1;
       t2 = t3;
     }
-    if (t2 > MTCODER__THREADS_MAX)
+    if (t2 > MTCODER_THREADS_MAX)
-      t2 = MTCODER__THREADS_MAX;
+      t2 = MTCODER_THREADS_MAX;
   }
   else if (t1 <= 0)
   {
@@ -286,8 +287,8 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p)
 
   fileSize = p->lzmaProps.reduceSize;
 
-  if (   p->blockSize != LZMA2_ENC_PROPS__BLOCK_SIZE__SOLID
+  if (   p->blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID
-      && p->blockSize != LZMA2_ENC_PROPS__BLOCK_SIZE__AUTO
+      && p->blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO
       && (p->blockSize < fileSize || fileSize == (UInt64)(Int64)-1))
     p->lzmaProps.reduceSize = p->blockSize;
 
@@ -297,19 +298,19 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p)
 
   t1 = p->lzmaProps.numThreads;
 
-  if (p->blockSize == LZMA2_ENC_PROPS__BLOCK_SIZE__SOLID)
+  if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID)
   {
     t2r = t2 = 1;
     t3 = t1;
   }
-  else if (p->blockSize == LZMA2_ENC_PROPS__BLOCK_SIZE__AUTO && t2 <= 1)
+  else if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO && t2 <= 1)
   {
     /* if there is no block multi-threading, we use SOLID block */
-    p->blockSize = LZMA2_ENC_PROPS__BLOCK_SIZE__SOLID;
+    p->blockSize = LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID;
   }
   else
   {
-    if (p->blockSize == LZMA2_ENC_PROPS__BLOCK_SIZE__AUTO)
+    if (p->blockSize == LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO)
     {
       const UInt32 kMinSize = (UInt32)1 << 20;
       const UInt32 kMaxSize = (UInt32)1 << 28;
@@ -344,7 +345,7 @@ void Lzma2EncProps_Normalize(CLzma2EncProps *p)
 }
 
 
-static SRes Progress(ICompressProgress *p, UInt64 inSize, UInt64 outSize)
+static SRes Progress(ICompressProgressPtr p, UInt64 inSize, UInt64 outSize)
 {
   return (p && ICompressProgress_Progress(p, inSize, outSize) != SZ_OK) ? SZ_ERROR_PROGRESS : SZ_OK;
 }
@@ -352,7 +353,7 @@ static SRes Progress(ICompressProgress *p, UInt64 inSize, UInt64 outSize)
 
 /* ---------- Lzma2 ---------- */
 
-typedef struct
+struct CLzma2Enc
 {
   Byte propEncoded;
   CLzma2EncProps props;
@@ -363,23 +364,22 @@ typedef struct
   ISzAllocPtr alloc;
   ISzAllocPtr allocBig;
 
-  CLzma2EncInt coders[MTCODER__THREADS_MAX];
+  CLzma2EncInt coders[MTCODER_THREADS_MAX];
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   
-  ISeqOutStream *outStream;
+  ISeqOutStreamPtr outStream;
   Byte *outBuf;
   size_t outBuf_Rem;   /* remainder in outBuf */
 
   size_t outBufSize;   /* size of allocated outBufs[i] */
-  size_t outBufsDataSizes[MTCODER__BLOCKS_MAX];
+  size_t outBufsDataSizes[MTCODER_BLOCKS_MAX];
   BoolInt mtCoder_WasConstructed;
   CMtCoder mtCoder;
-  Byte *outBufs[MTCODER__BLOCKS_MAX];
+  Byte *outBufs[MTCODER_BLOCKS_MAX];
 
   #endif
-
+};
-} CLzma2Enc;
 
 
 
@@ -396,30 +396,30 @@ CLzma2EncHandle Lzma2Enc_Create(ISzAllocPtr alloc, ISzAllocPtr allocBig)
   p->allocBig = allocBig;
   {
     unsigned i;
-    for (i = 0; i < MTCODER__THREADS_MAX; i++)
+    for (i = 0; i < MTCODER_THREADS_MAX; i++)
       p->coders[i].enc = NULL;
   }
   
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   p->mtCoder_WasConstructed = False;
   {
     unsigned i;
-    for (i = 0; i < MTCODER__BLOCKS_MAX; i++)
+    for (i = 0; i < MTCODER_BLOCKS_MAX; i++)
       p->outBufs[i] = NULL;
     p->outBufSize = 0;
   }
   #endif
 
-  return p;
+  return (CLzma2EncHandle)p;
 }
 
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
 static void Lzma2Enc_FreeOutBufs(CLzma2Enc *p)
 {
   unsigned i;
-  for (i = 0; i < MTCODER__BLOCKS_MAX; i++)
+  for (i = 0; i < MTCODER_BLOCKS_MAX; i++)
     if (p->outBufs[i])
     {
       ISzAlloc_Free(p->alloc, p->outBufs[i]);
@@ -430,12 +430,13 @@ static void Lzma2Enc_FreeOutBufs(CLzma2Enc *p)
 
 #endif
 
+// #define GET_CLzma2Enc_p  CLzma2Enc *p = (CLzma2Enc *)(void *)p;
 
-void Lzma2Enc_Destroy(CLzma2EncHandle pp)
+void Lzma2Enc_Destroy(CLzma2EncHandle p)
 {
-  CLzma2Enc *p = (CLzma2Enc *)pp;
+  // GET_CLzma2Enc_p
   unsigned i;
-  for (i = 0; i < MTCODER__THREADS_MAX; i++)
+  for (i = 0; i < MTCODER_THREADS_MAX; i++)
   {
     CLzma2EncInt *t = &p->coders[i];
     if (t->enc)
@@ -446,7 +447,7 @@ void Lzma2Enc_Destroy(CLzma2EncHandle pp)
   }
 
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   if (p->mtCoder_WasConstructed)
   {
     MtCoder_Destruct(&p->mtCoder);
@@ -458,13 +459,13 @@ void Lzma2Enc_Destroy(CLzma2EncHandle pp)
   ISzAlloc_Free(p->alloc, p->tempBufLzma);
   p->tempBufLzma = NULL;
 
-  ISzAlloc_Free(p->alloc, pp);
+  ISzAlloc_Free(p->alloc, p);
 }
 
 
-SRes Lzma2Enc_SetProps(CLzma2EncHandle pp, const CLzma2EncProps *props)
+SRes Lzma2Enc_SetProps(CLzma2EncHandle p, const CLzma2EncProps *props)
 {
-  CLzma2Enc *p = (CLzma2Enc *)pp;
+  // GET_CLzma2Enc_p
   CLzmaEncProps lzmaProps = props->lzmaProps;
   LzmaEncProps_Normalize(&lzmaProps);
   if (lzmaProps.lc + lzmaProps.lp > LZMA2_LCLP_MAX)
@@ -475,16 +476,16 @@ SRes Lzma2Enc_SetProps(CLzma2EncHandle pp, const CLzma2EncProps *props)
 }
 
 
-void Lzma2Enc_SetDataSize(CLzmaEncHandle pp, UInt64 expectedDataSiize)
+void Lzma2Enc_SetDataSize(CLzma2EncHandle p, UInt64 expectedDataSiize)
 {
-  CLzma2Enc *p = (CLzma2Enc *)pp;
+  // GET_CLzma2Enc_p
   p->expectedDataSize = expectedDataSiize;
 }
 
 
-Byte Lzma2Enc_WriteProperties(CLzma2EncHandle pp)
+Byte Lzma2Enc_WriteProperties(CLzma2EncHandle p)
 {
-  CLzma2Enc *p = (CLzma2Enc *)pp;
+  // GET_CLzma2Enc_p
   unsigned i;
   UInt32 dicSize = LzmaEncProps_GetDictSize(&p->props.lzmaProps);
   for (i = 0; i < 40; i++)
@@ -497,12 +498,12 @@ Byte Lzma2Enc_WriteProperties(CLzma2EncHandle pp)
 static SRes Lzma2Enc_EncodeMt1(
     CLzma2Enc *me,
     CLzma2EncInt *p,
-    ISeqOutStream *outStream,
+    ISeqOutStreamPtr outStream,
     Byte *outBuf, size_t *outBufSize,
-    ISeqInStream *inStream,
+    ISeqInStreamPtr inStream,
     const Byte *inData, size_t inDataSize,
     int finished,
-    ICompressProgress *progress)
+    ICompressProgressPtr progress)
 {
   UInt64 unpackTotal = 0;
   UInt64 packTotal = 0;
@@ -540,12 +541,12 @@ static SRes Lzma2Enc_EncodeMt1(
     }
   }
 
-  RINOK(Lzma2EncInt_InitStream(p, &me->props));
+  RINOK(Lzma2EncInt_InitStream(p, &me->props))
 
   for (;;)
   {
     SRes res = SZ_OK;
-    size_t inSizeCur = 0;
+    SizeT inSizeCur = 0;
 
     Lzma2EncInt_InitBlock(p);
     
@@ -559,7 +560,7 @@ static SRes Lzma2Enc_EncodeMt1(
       if (me->expectedDataSize != (UInt64)(Int64)-1
           && me->expectedDataSize >= unpackTotal)
         expected = me->expectedDataSize - unpackTotal;
-      if (me->props.blockSize != LZMA2_ENC_PROPS__BLOCK_SIZE__SOLID
+      if (me->props.blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID
           && expected > me->props.blockSize)
         expected = (size_t)me->props.blockSize;
 
@@ -569,14 +570,14 @@ static SRes Lzma2Enc_EncodeMt1(
           &limitedInStream.vt,
           LZMA2_KEEP_WINDOW_SIZE,
           me->alloc,
-          me->allocBig));
+          me->allocBig))
     }
     else
     {
-      inSizeCur = inDataSize - (size_t)unpackTotal;
+      inSizeCur = (SizeT)(inDataSize - (size_t)unpackTotal);
-      if (me->props.blockSize != LZMA2_ENC_PROPS__BLOCK_SIZE__SOLID
+      if (me->props.blockSize != LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID
           && inSizeCur > me->props.blockSize)
-        inSizeCur = (size_t)me->props.blockSize;
+        inSizeCur = (SizeT)(size_t)me->props.blockSize;
     
       // LzmaEnc_SetDataSize(p->enc, inSizeCur);
       
@@ -584,7 +585,7 @@ static SRes Lzma2Enc_EncodeMt1(
           inData + (size_t)unpackTotal, inSizeCur,
           LZMA2_KEEP_WINDOW_SIZE,
           me->alloc,
-          me->allocBig));
+          me->allocBig))
     }
 
     for (;;)
@@ -621,7 +622,7 @@ static SRes Lzma2Enc_EncodeMt1(
     
     unpackTotal += p->srcPos;
     
-    RINOK(res);
+    RINOK(res)
 
     if (p->srcPos != (inStream ? limitedInStream.processed : inSizeCur))
       return SZ_ERROR_FAIL;
@@ -652,12 +653,12 @@ static SRes Lzma2Enc_EncodeMt1(
 
 
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
-static SRes Lzma2Enc_MtCallback_Code(void *pp, unsigned coderIndex, unsigned outBufIndex,
+static SRes Lzma2Enc_MtCallback_Code(void *p, unsigned coderIndex, unsigned outBufIndex,
     const Byte *src, size_t srcSize, int finished)
 {
-  CLzma2Enc *me = (CLzma2Enc *)pp;
+  CLzma2Enc *me = (CLzma2Enc *)p;
   size_t destSize = me->outBufSize;
   SRes res;
   CMtProgressThunk progressThunk;
@@ -692,9 +693,9 @@ static SRes Lzma2Enc_MtCallback_Code(void *pp, unsigned coderIndex, unsigned out
 }
 
 
-static SRes Lzma2Enc_MtCallback_Write(void *pp, unsigned outBufIndex)
+static SRes Lzma2Enc_MtCallback_Write(void *p, unsigned outBufIndex)
 {
-  CLzma2Enc *me = (CLzma2Enc *)pp;
+  CLzma2Enc *me = (CLzma2Enc *)p;
   size_t size = me->outBufsDataSizes[outBufIndex];
   const Byte *data = me->outBufs[outBufIndex];
   
@@ -713,14 +714,14 @@ static SRes Lzma2Enc_MtCallback_Write(void *pp, unsigned outBufIndex)
 
 
 
-SRes Lzma2Enc_Encode2(CLzma2EncHandle pp,
+SRes Lzma2Enc_Encode2(CLzma2EncHandle p,
-    ISeqOutStream *outStream,
+    ISeqOutStreamPtr outStream,
     Byte *outBuf, size_t *outBufSize,
-    ISeqInStream *inStream,
+    ISeqInStreamPtr inStream,
     const Byte *inData, size_t inDataSize,
-    ICompressProgress *progress)
+    ICompressProgressPtr progress)
 {
-  CLzma2Enc *p = (CLzma2Enc *)pp;
+  // GET_CLzma2Enc_p
 
   if (inStream && inData)
     return SZ_ERROR_PARAM;
@@ -730,11 +731,11 @@ SRes Lzma2Enc_Encode2(CLzma2EncHandle pp,
 
   {
     unsigned i;
-    for (i = 0; i < MTCODER__THREADS_MAX; i++)
+    for (i = 0; i < MTCODER_THREADS_MAX; i++)
       p->coders[i].propsAreSet = False;
   }
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   
   if (p->props.numBlockThreads_Reduced > 1)
   {
@@ -772,7 +773,7 @@ SRes Lzma2Enc_Encode2(CLzma2EncHandle pp,
       return SZ_ERROR_PARAM; /* SZ_ERROR_MEM */
 
     {
-      size_t destBlockSize = p->mtCoder.blockSize + (p->mtCoder.blockSize >> 10) + 16;
+      const size_t destBlockSize = p->mtCoder.blockSize + (p->mtCoder.blockSize >> 10) + 16;
       if (destBlockSize < p->mtCoder.blockSize)
         return SZ_ERROR_PARAM;
       if (p->outBufSize != destBlockSize)
@@ -781,10 +782,11 @@ SRes Lzma2Enc_Encode2(CLzma2EncHandle pp,
     }
 
     p->mtCoder.numThreadsMax = (unsigned)p->props.numBlockThreads_Max;
+    p->mtCoder.numThreadGroups = p->props.numThreadGroups;
     p->mtCoder.expectedDataSize = p->expectedDataSize;
     
     {
-      SRes res = MtCoder_Code(&p->mtCoder);
+      const SRes res = MtCoder_Code(&p->mtCoder);
       if (!outStream)
         *outBufSize = (size_t)(p->outBuf - outBuf);
       return res;
@@ -801,3 +803,5 @@ SRes Lzma2Enc_Encode2(CLzma2EncHandle pp,
       True, /* finished */
       progress);
 }
+
+#undef PRF

C/Lzma2Enc.h

@@ -1,15 +1,15 @@
 /* Lzma2Enc.h -- LZMA2 Encoder
-2017-07-27 : Igor Pavlov : Public domain */
+2023-04-13 : Igor Pavlov : Public domain */
 
-#ifndef __LZMA2_ENC_H
+#ifndef ZIP7_INC_LZMA2_ENC_H
-#define __LZMA2_ENC_H
+#define ZIP7_INC_LZMA2_ENC_H
 
 #include "LzmaEnc.h"
 
 EXTERN_C_BEGIN
 
-#define LZMA2_ENC_PROPS__BLOCK_SIZE__AUTO 0
+#define LZMA2_ENC_PROPS_BLOCK_SIZE_AUTO   0
-#define LZMA2_ENC_PROPS__BLOCK_SIZE__SOLID ((UInt64)(Int64)-1)
+#define LZMA2_ENC_PROPS_BLOCK_SIZE_SOLID  ((UInt64)(Int64)-1)
 
 typedef struct
 {
@@ -18,6 +18,7 @@ typedef struct
   int numBlockThreads_Reduced;
   int numBlockThreads_Max;
   int numTotalThreads;
+  unsigned numThreadGroups; // 0 : no groups
 } CLzma2EncProps;
 
 void Lzma2EncProps_Init(CLzma2EncProps *p);
@@ -36,7 +37,9 @@ SRes:
   SZ_ERROR_THREAD - error in multithreading functions (only for Mt version)
 */
 
-typedef void * CLzma2EncHandle;
+typedef struct CLzma2Enc CLzma2Enc;
+typedef CLzma2Enc * CLzma2EncHandle;
+// Z7_DECLARE_HANDLE(CLzma2EncHandle)
 
 CLzma2EncHandle Lzma2Enc_Create(ISzAllocPtr alloc, ISzAllocPtr allocBig);
 void Lzma2Enc_Destroy(CLzma2EncHandle p);
@@ -44,11 +47,11 @@ SRes Lzma2Enc_SetProps(CLzma2EncHandle p, const CLzma2EncProps *props);
 void Lzma2Enc_SetDataSize(CLzma2EncHandle p, UInt64 expectedDataSiize);
 Byte Lzma2Enc_WriteProperties(CLzma2EncHandle p);
 SRes Lzma2Enc_Encode2(CLzma2EncHandle p,
-    ISeqOutStream *outStream,
+    ISeqOutStreamPtr outStream,
     Byte *outBuf, size_t *outBufSize,
-    ISeqInStream *inStream,
+    ISeqInStreamPtr inStream,
     const Byte *inData, size_t inDataSize,
-    ICompressProgress *progress);
+    ICompressProgressPtr progress);
 
 EXTERN_C_END
 

C/Lzma86.h

@@ -1,8 +1,8 @@
 /* Lzma86.h -- LZMA + x86 (BCJ) Filter
-2013-01-18 : Igor Pavlov : Public domain */
+2023-03-03 : Igor Pavlov : Public domain */
 
-#ifndef __LZMA86_H
+#ifndef ZIP7_INC_LZMA86_H
-#define __LZMA86_H
+#define ZIP7_INC_LZMA86_H
 
 #include "7zTypes.h"
 

C/Lzma86Dec.c

@@ -1,5 +1,5 @@
 /* Lzma86Dec.c -- LZMA + x86 (BCJ) Filter Decoder
-2016-05-16 : Igor Pavlov : Public domain */
+2023-03-03 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -46,9 +46,8 @@ SRes Lzma86_Decode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen)
     return res;
   if (useFilter == 1)
   {
-    UInt32 x86State;
+    UInt32 x86State = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL;
-    x86_Convert_Init(x86State);
+    z7_BranchConvSt_X86_Dec(dest, *destLen, 0, &x86State);
-    x86_Convert(dest, *destLen, 0, &x86State, 0);
   }
   return SZ_OK;
 }

C/Lzma86Enc.c

@@ -1,5 +1,5 @@
 /* Lzma86Enc.c -- LZMA + x86 (BCJ) Filter Encoder
-2018-07-04 : Igor Pavlov : Public domain */
+2023-03-03 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -46,9 +46,8 @@ int Lzma86_Encode(Byte *dest, size_t *destLen, const Byte *src, size_t srcLen,
       memcpy(filteredStream, src, srcLen);
     }
     {
-      UInt32 x86State;
+      UInt32 x86State = Z7_BRANCH_CONV_ST_X86_STATE_INIT_VAL;
-      x86_Convert_Init(x86State);
+      z7_BranchConvSt_X86_Enc(filteredStream, srcLen, 0, &x86State);
-      x86_Convert(filteredStream, srcLen, 0, &x86State, 1);
     }
   }
 

C/LzmaDec.c

@@ -1,5 +1,5 @@
 /* LzmaDec.c -- LZMA Decoder
-2021-04-01 : Igor Pavlov : Public domain */
+2023-04-07 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -8,15 +8,15 @@
 /* #include "CpuArch.h" */
 #include "LzmaDec.h"
 
-#define kNumTopBits 24
+// #define kNumTopBits 24
-#define kTopValue ((UInt32)1 << kNumTopBits)
+#define kTopValue ((UInt32)1 << 24)
 
 #define kNumBitModelTotalBits 11
 #define kBitModelTotal (1 << kNumBitModelTotalBits)
 
 #define RC_INIT_SIZE 5
 
-#ifndef _LZMA_DEC_OPT
+#ifndef Z7_LZMA_DEC_OPT
 
 #define kNumMoveBits 5
 #define NORMALIZE if (range < kTopValue) { range <<= 8; code = (code << 8) | (*buf++); }
@@ -25,14 +25,14 @@
 #define UPDATE_0(p) range = bound; *(p) = (CLzmaProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits));
 #define UPDATE_1(p) range -= bound; code -= bound; *(p) = (CLzmaProb)(ttt - (ttt >> kNumMoveBits));
 #define GET_BIT2(p, i, A0, A1) IF_BIT_0(p) \
-  { UPDATE_0(p); i = (i + i); A0; } else \
+  { UPDATE_0(p)  i = (i + i); A0; } else \
-  { UPDATE_1(p); i = (i + i) + 1; A1; }
+  { UPDATE_1(p)  i = (i + i) + 1; A1; }
 
 #define TREE_GET_BIT(probs, i) { GET_BIT2(probs + i, i, ;, ;); }
 
 #define REV_BIT(p, i, A0, A1) IF_BIT_0(p + i) \
-  { UPDATE_0(p + i); A0; } else \
+  { UPDATE_0(p + i)  A0; } else \
-  { UPDATE_1(p + i); A1; }
+  { UPDATE_1(p + i)  A1; }
 #define REV_BIT_VAR(  p, i, m) REV_BIT(p, i, i += m; m += m, m += m; i += m; )
 #define REV_BIT_CONST(p, i, m) REV_BIT(p, i, i += m;       , i += m * 2; )
 #define REV_BIT_LAST( p, i, m) REV_BIT(p, i, i -= m        , ; )
@@ -40,19 +40,19 @@
 #define TREE_DECODE(probs, limit, i) \
   { i = 1; do { TREE_GET_BIT(probs, i); } while (i < limit); i -= limit; }
 
-/* #define _LZMA_SIZE_OPT */
+/* #define Z7_LZMA_SIZE_OPT */
 
-#ifdef _LZMA_SIZE_OPT
+#ifdef Z7_LZMA_SIZE_OPT
 #define TREE_6_DECODE(probs, i) TREE_DECODE(probs, (1 << 6), i)
 #else
 #define TREE_6_DECODE(probs, i) \
   { i = 1; \
-  TREE_GET_BIT(probs, i); \
+  TREE_GET_BIT(probs, i) \
-  TREE_GET_BIT(probs, i); \
+  TREE_GET_BIT(probs, i) \
-  TREE_GET_BIT(probs, i); \
+  TREE_GET_BIT(probs, i) \
-  TREE_GET_BIT(probs, i); \
+  TREE_GET_BIT(probs, i) \
-  TREE_GET_BIT(probs, i); \
+  TREE_GET_BIT(probs, i) \
-  TREE_GET_BIT(probs, i); \
+  TREE_GET_BIT(probs, i) \
   i -= 0x40; }
 #endif
 
@@ -64,25 +64,25 @@
   probLit = prob + (offs + bit + symbol); \
   GET_BIT2(probLit, symbol, offs ^= bit; , ;)
 
-#endif // _LZMA_DEC_OPT
+#endif // Z7_LZMA_DEC_OPT
 
 
 #define NORMALIZE_CHECK if (range < kTopValue) { if (buf >= bufLimit) return DUMMY_INPUT_EOF; range <<= 8; code = (code << 8) | (*buf++); }
 
-#define IF_BIT_0_CHECK(p) ttt = *(p); NORMALIZE_CHECK; bound = (range >> kNumBitModelTotalBits) * (UInt32)ttt; if (code < bound)
+#define IF_BIT_0_CHECK(p) ttt = *(p); NORMALIZE_CHECK bound = (range >> kNumBitModelTotalBits) * (UInt32)ttt; if (code < bound)
 #define UPDATE_0_CHECK range = bound;
 #define UPDATE_1_CHECK range -= bound; code -= bound;
 #define GET_BIT2_CHECK(p, i, A0, A1) IF_BIT_0_CHECK(p) \
-  { UPDATE_0_CHECK; i = (i + i); A0; } else \
+  { UPDATE_0_CHECK  i = (i + i); A0; } else \
-  { UPDATE_1_CHECK; i = (i + i) + 1; A1; }
+  { UPDATE_1_CHECK  i = (i + i) + 1; A1; }
 #define GET_BIT_CHECK(p, i) GET_BIT2_CHECK(p, i, ; , ;)
 #define TREE_DECODE_CHECK(probs, limit, i) \
   { i = 1; do { GET_BIT_CHECK(probs + i, i) } while (i < limit); i -= limit; }
 
 
 #define REV_BIT_CHECK(p, i, m) IF_BIT_0_CHECK(p + i) \
-  { UPDATE_0_CHECK; i += m; m += m; } else \
+  { UPDATE_0_CHECK  i += m; m += m; } else \
-  { UPDATE_1_CHECK; m += m; i += m; }
+  { UPDATE_1_CHECK  m += m; i += m; }
 
 
 #define kNumPosBitsMax 4
@@ -224,14 +224,14 @@ Out:
 */
 
 
-#ifdef _LZMA_DEC_OPT
+#ifdef Z7_LZMA_DEC_OPT
 
-int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit);
+int Z7_FASTCALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit);
 
 #else
 
 static
-int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
+int Z7_FASTCALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
 {
   CLzmaProb *probs = GET_PROBS;
   unsigned state = (unsigned)p->state;
@@ -263,7 +263,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
     IF_BIT_0(prob)
     {
       unsigned symbol;
-      UPDATE_0(prob);
+      UPDATE_0(prob)
       prob = probs + Literal;
       if (processedPos != 0 || checkDicSize != 0)
         prob += (UInt32)3 * ((((processedPos << 8) + dic[(dicPos == 0 ? dicBufSize : dicPos) - 1]) & lpMask) << lc);
@@ -273,7 +273,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
       {
         state -= (state < 4) ? state : 3;
         symbol = 1;
-        #ifdef _LZMA_SIZE_OPT
+        #ifdef Z7_LZMA_SIZE_OPT
         do { NORMAL_LITER_DEC } while (symbol < 0x100);
         #else
         NORMAL_LITER_DEC
@@ -292,7 +292,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
         unsigned offs = 0x100;
         state -= (state < 10) ? 3 : 6;
         symbol = 1;
-        #ifdef _LZMA_SIZE_OPT
+        #ifdef Z7_LZMA_SIZE_OPT
         do
         {
           unsigned bit;
@@ -321,25 +321,25 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
     }
     
     {
-      UPDATE_1(prob);
+      UPDATE_1(prob)
       prob = probs + IsRep + state;
       IF_BIT_0(prob)
       {
-        UPDATE_0(prob);
+        UPDATE_0(prob)
         state += kNumStates;
         prob = probs + LenCoder;
       }
       else
       {
-        UPDATE_1(prob);
+        UPDATE_1(prob)
         prob = probs + IsRepG0 + state;
         IF_BIT_0(prob)
         {
-          UPDATE_0(prob);
+          UPDATE_0(prob)
           prob = probs + IsRep0Long + COMBINED_PS_STATE;
           IF_BIT_0(prob)
           {
-            UPDATE_0(prob);
+            UPDATE_0(prob)
   
             // that case was checked before with kBadRepCode
             // if (checkDicSize == 0 && processedPos == 0) { len = kMatchSpecLen_Error_Data + 1; break; }
@@ -353,30 +353,30 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
             state = state < kNumLitStates ? 9 : 11;
             continue;
           }
-          UPDATE_1(prob);
+          UPDATE_1(prob)
         }
         else
         {
           UInt32 distance;
-          UPDATE_1(prob);
+          UPDATE_1(prob)
           prob = probs + IsRepG1 + state;
           IF_BIT_0(prob)
           {
-            UPDATE_0(prob);
+            UPDATE_0(prob)
             distance = rep1;
           }
           else
           {
-            UPDATE_1(prob);
+            UPDATE_1(prob)
             prob = probs + IsRepG2 + state;
             IF_BIT_0(prob)
             {
-              UPDATE_0(prob);
+              UPDATE_0(prob)
               distance = rep2;
             }
             else
             {
-              UPDATE_1(prob);
+              UPDATE_1(prob)
               distance = rep3;
               rep3 = rep2;
             }
@@ -389,37 +389,37 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
         prob = probs + RepLenCoder;
       }
       
-      #ifdef _LZMA_SIZE_OPT
+      #ifdef Z7_LZMA_SIZE_OPT
       {
         unsigned lim, offset;
         CLzmaProb *probLen = prob + LenChoice;
         IF_BIT_0(probLen)
         {
-          UPDATE_0(probLen);
+          UPDATE_0(probLen)
           probLen = prob + LenLow + GET_LEN_STATE;
           offset = 0;
           lim = (1 << kLenNumLowBits);
         }
         else
         {
-          UPDATE_1(probLen);
+          UPDATE_1(probLen)
           probLen = prob + LenChoice2;
           IF_BIT_0(probLen)
           {
-            UPDATE_0(probLen);
+            UPDATE_0(probLen)
             probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits);
             offset = kLenNumLowSymbols;
             lim = (1 << kLenNumLowBits);
           }
           else
           {
-            UPDATE_1(probLen);
+            UPDATE_1(probLen)
             probLen = prob + LenHigh;
             offset = kLenNumLowSymbols * 2;
             lim = (1 << kLenNumHighBits);
           }
         }
-        TREE_DECODE(probLen, lim, len);
+        TREE_DECODE(probLen, lim, len)
         len += offset;
       }
       #else
@@ -427,32 +427,32 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
         CLzmaProb *probLen = prob + LenChoice;
         IF_BIT_0(probLen)
         {
-          UPDATE_0(probLen);
+          UPDATE_0(probLen)
           probLen = prob + LenLow + GET_LEN_STATE;
           len = 1;
-          TREE_GET_BIT(probLen, len);
+          TREE_GET_BIT(probLen, len)
-          TREE_GET_BIT(probLen, len);
+          TREE_GET_BIT(probLen, len)
-          TREE_GET_BIT(probLen, len);
+          TREE_GET_BIT(probLen, len)
           len -= 8;
         }
         else
         {
-          UPDATE_1(probLen);
+          UPDATE_1(probLen)
           probLen = prob + LenChoice2;
           IF_BIT_0(probLen)
           {
-            UPDATE_0(probLen);
+            UPDATE_0(probLen)
             probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits);
             len = 1;
-            TREE_GET_BIT(probLen, len);
+            TREE_GET_BIT(probLen, len)
-            TREE_GET_BIT(probLen, len);
+            TREE_GET_BIT(probLen, len)
-            TREE_GET_BIT(probLen, len);
+            TREE_GET_BIT(probLen, len)
           }
           else
           {
-            UPDATE_1(probLen);
+            UPDATE_1(probLen)
             probLen = prob + LenHigh;
-            TREE_DECODE(probLen, (1 << kLenNumHighBits), len);
+            TREE_DECODE(probLen, (1 << kLenNumHighBits), len)
             len += kLenNumLowSymbols * 2;
           }
         }
@@ -464,7 +464,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
         UInt32 distance;
         prob = probs + PosSlot +
             ((len < kNumLenToPosStates ? len : kNumLenToPosStates - 1) << kNumPosSlotBits);
-        TREE_6_DECODE(prob, distance);
+        TREE_6_DECODE(prob, distance)
         if (distance >= kStartPosModelIndex)
         {
           unsigned posSlot = (unsigned)distance;
@@ -479,7 +479,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
               distance++;
               do
               {
-                REV_BIT_VAR(prob, distance, m);
+                REV_BIT_VAR(prob, distance, m)
               }
               while (--numDirectBits);
               distance -= m;
@@ -514,10 +514,10 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
             distance <<= kNumAlignBits;
             {
               unsigned i = 1;
-              REV_BIT_CONST(prob, i, 1);
+              REV_BIT_CONST(prob, i, 1)
-              REV_BIT_CONST(prob, i, 2);
+              REV_BIT_CONST(prob, i, 2)
-              REV_BIT_CONST(prob, i, 4);
+              REV_BIT_CONST(prob, i, 4)
-              REV_BIT_LAST (prob, i, 8);
+              REV_BIT_LAST (prob, i, 8)
               distance |= i;
             }
             if (distance == (UInt32)0xFFFFFFFF)
@@ -592,7 +592,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
   }
   while (dicPos < limit && buf < bufLimit);
 
-  NORMALIZE;
+  NORMALIZE
   
   p->buf = buf;
   p->range = range;
@@ -613,7 +613,7 @@ int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit
 
 
 
-static void MY_FAST_CALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit)
+static void Z7_FASTCALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit)
 {
   unsigned len = (unsigned)p->remainLen;
   if (len == 0 /* || len >= kMatchSpecLenStart */)
@@ -683,7 +683,7 @@ and we support the following state of (p->checkDicSize):
     (p->checkDicSize == p->prop.dicSize)
 */
 
-static int MY_FAST_CALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
+static int Z7_FASTCALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
 {
   if (p->checkDicSize == 0)
   {
@@ -767,54 +767,54 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, const Byt
     else
     {
       unsigned len;
-      UPDATE_1_CHECK;
+      UPDATE_1_CHECK
 
       prob = probs + IsRep + state;
       IF_BIT_0_CHECK(prob)
       {
-        UPDATE_0_CHECK;
+        UPDATE_0_CHECK
         state = 0;
         prob = probs + LenCoder;
         res = DUMMY_MATCH;
       }
       else
       {
-        UPDATE_1_CHECK;
+        UPDATE_1_CHECK
         res = DUMMY_REP;
         prob = probs + IsRepG0 + state;
         IF_BIT_0_CHECK(prob)
         {
-          UPDATE_0_CHECK;
+          UPDATE_0_CHECK
           prob = probs + IsRep0Long + COMBINED_PS_STATE;
           IF_BIT_0_CHECK(prob)
           {
-            UPDATE_0_CHECK;
+            UPDATE_0_CHECK
             break;
           }
           else
           {
-            UPDATE_1_CHECK;
+            UPDATE_1_CHECK
           }
         }
         else
         {
-          UPDATE_1_CHECK;
+          UPDATE_1_CHECK
           prob = probs + IsRepG1 + state;
           IF_BIT_0_CHECK(prob)
           {
-            UPDATE_0_CHECK;
+            UPDATE_0_CHECK
           }
           else
           {
-            UPDATE_1_CHECK;
+            UPDATE_1_CHECK
             prob = probs + IsRepG2 + state;
             IF_BIT_0_CHECK(prob)
             {
-              UPDATE_0_CHECK;
+              UPDATE_0_CHECK
             }
             else
             {
-              UPDATE_1_CHECK;
+              UPDATE_1_CHECK
             }
           }
         }
@@ -826,31 +826,31 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, const Byt
         const CLzmaProb *probLen = prob + LenChoice;
         IF_BIT_0_CHECK(probLen)
         {
-          UPDATE_0_CHECK;
+          UPDATE_0_CHECK
           probLen = prob + LenLow + GET_LEN_STATE;
           offset = 0;
           limit = 1 << kLenNumLowBits;
         }
         else
         {
-          UPDATE_1_CHECK;
+          UPDATE_1_CHECK
           probLen = prob + LenChoice2;
           IF_BIT_0_CHECK(probLen)
           {
-            UPDATE_0_CHECK;
+            UPDATE_0_CHECK
             probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits);
             offset = kLenNumLowSymbols;
             limit = 1 << kLenNumLowBits;
           }
           else
           {
-            UPDATE_1_CHECK;
+            UPDATE_1_CHECK
             probLen = prob + LenHigh;
             offset = kLenNumLowSymbols * 2;
             limit = 1 << kLenNumHighBits;
           }
         }
-        TREE_DECODE_CHECK(probLen, limit, len);
+        TREE_DECODE_CHECK(probLen, limit, len)
         len += offset;
       }
 
@@ -860,7 +860,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, const Byt
         prob = probs + PosSlot +
             ((len < kNumLenToPosStates - 1 ? len : kNumLenToPosStates - 1) <<
             kNumPosSlotBits);
-        TREE_DECODE_CHECK(prob, 1 << kNumPosSlotBits, posSlot);
+        TREE_DECODE_CHECK(prob, 1 << kNumPosSlotBits, posSlot)
         if (posSlot >= kStartPosModelIndex)
         {
           unsigned numDirectBits = ((posSlot >> 1) - 1);
@@ -888,7 +888,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, const Byt
             unsigned m = 1;
             do
             {
-              REV_BIT_CHECK(prob, i, m);
+              REV_BIT_CHECK(prob, i, m)
             }
             while (--numDirectBits);
           }
@@ -897,7 +897,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, const Byt
     }
     break;
   }
-  NORMALIZE_CHECK;
+  NORMALIZE_CHECK
 
   *bufOut = buf;
   return res;
@@ -943,7 +943,7 @@ When the decoder lookahead, and the lookahead symbol is not end_marker, we have
 */
 
 
-#define RETURN__NOT_FINISHED__FOR_FINISH \
+#define RETURN_NOT_FINISHED_FOR_FINISH \
   *status = LZMA_STATUS_NOT_FINISHED; \
   return SZ_ERROR_DATA; // for strict mode
   // return SZ_OK; // for relaxed mode
@@ -1029,7 +1029,7 @@ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *sr
         }
         if (p->remainLen != 0)
         {
-          RETURN__NOT_FINISHED__FOR_FINISH;
+          RETURN_NOT_FINISHED_FOR_FINISH
         }
         checkEndMarkNow = 1;
       }
@@ -1072,7 +1072,7 @@ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *sr
             for (i = 0; i < (unsigned)dummyProcessed; i++)
               p->tempBuf[i] = src[i];
             // p->remainLen = kMatchSpecLen_Error_Data;
-            RETURN__NOT_FINISHED__FOR_FINISH;
+            RETURN_NOT_FINISHED_FOR_FINISH
           }
           
           bufLimit = src;
@@ -1150,7 +1150,7 @@ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *sr
             (*srcLen) += (unsigned)dummyProcessed - p->tempBufSize;
             p->tempBufSize = (unsigned)dummyProcessed;
             // p->remainLen = kMatchSpecLen_Error_Data;
-            RETURN__NOT_FINISHED__FOR_FINISH;
+            RETURN_NOT_FINISHED_FOR_FINISH
           }
         }
 
@@ -1299,8 +1299,8 @@ static SRes LzmaDec_AllocateProbs2(CLzmaDec *p, const CLzmaProps *propNew, ISzAl
 SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc)
 {
   CLzmaProps propNew;
-  RINOK(LzmaProps_Decode(&propNew, props, propsSize));
+  RINOK(LzmaProps_Decode(&propNew, props, propsSize))
-  RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc));
+  RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc))
   p->prop = propNew;
   return SZ_OK;
 }
@@ -1309,14 +1309,14 @@ SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAll
 {
   CLzmaProps propNew;
   SizeT dicBufSize;
-  RINOK(LzmaProps_Decode(&propNew, props, propsSize));
+  RINOK(LzmaProps_Decode(&propNew, props, propsSize))
-  RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc));
+  RINOK(LzmaDec_AllocateProbs2(p, &propNew, alloc))
 
   {
     UInt32 dictSize = propNew.dicSize;
     SizeT mask = ((UInt32)1 << 12) - 1;
          if (dictSize >= ((UInt32)1 << 30)) mask = ((UInt32)1 << 22) - 1;
-    else if (dictSize >= ((UInt32)1 << 22)) mask = ((UInt32)1 << 20) - 1;;
+    else if (dictSize >= ((UInt32)1 << 22)) mask = ((UInt32)1 << 20) - 1;
     dicBufSize = ((SizeT)dictSize + mask) & ~mask;
     if (dicBufSize < dictSize)
       dicBufSize = dictSize;
@@ -1348,8 +1348,8 @@ SRes LzmaDecode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen,
   *status = LZMA_STATUS_NOT_SPECIFIED;
   if (inSize < RC_INIT_SIZE)
     return SZ_ERROR_INPUT_EOF;
-  LzmaDec_Construct(&p);
+  LzmaDec_CONSTRUCT(&p)
-  RINOK(LzmaDec_AllocateProbs(&p, propData, propSize, alloc));
+  RINOK(LzmaDec_AllocateProbs(&p, propData, propSize, alloc))
   p.dic = dest;
   p.dicBufSize = outSize;
   LzmaDec_Init(&p);

C/LzmaDec.h

@@ -1,19 +1,19 @@
 /* LzmaDec.h -- LZMA Decoder
-2020-03-19 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
-#ifndef __LZMA_DEC_H
+#ifndef ZIP7_INC_LZMA_DEC_H
-#define __LZMA_DEC_H
+#define ZIP7_INC_LZMA_DEC_H
 
 #include "7zTypes.h"
 
 EXTERN_C_BEGIN
 
-/* #define _LZMA_PROB32 */
+/* #define Z7_LZMA_PROB32 */
-/* _LZMA_PROB32 can increase the speed on some CPUs,
+/* Z7_LZMA_PROB32 can increase the speed on some CPUs,
    but memory usage for CLzmaDec::probs will be doubled in that case */
 
 typedef
-#ifdef _LZMA_PROB32
+#ifdef Z7_LZMA_PROB32
   UInt32
 #else
   UInt16
@@ -25,7 +25,7 @@ typedef
 
 #define LZMA_PROPS_SIZE 5
 
-typedef struct _CLzmaProps
+typedef struct
 {
   Byte lc;
   Byte lp;
@@ -73,7 +73,8 @@ typedef struct
   Byte tempBuf[LZMA_REQUIRED_INPUT_MAX];
 } CLzmaDec;
 
-#define LzmaDec_Construct(p) { (p)->dic = NULL; (p)->probs = NULL; }
+#define LzmaDec_CONSTRUCT(p) { (p)->dic = NULL; (p)->probs = NULL; }
+#define LzmaDec_Construct(p) LzmaDec_CONSTRUCT(p)
 
 void LzmaDec_Init(CLzmaDec *p);
 

C/LzmaEnc.h

@@ -1,8 +1,8 @@
 /*  LzmaEnc.h -- LZMA Encoder
-2019-10-30 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __LZMA_ENC_H
+#ifndef ZIP7_INC_LZMA_ENC_H
-#define __LZMA_ENC_H
+#define ZIP7_INC_LZMA_ENC_H
 
 #include "7zTypes.h"
 
@@ -10,7 +10,7 @@ EXTERN_C_BEGIN
 
 #define LZMA_PROPS_SIZE 5
 
-typedef struct _CLzmaEncProps
+typedef struct
 {
   int level;       /* 0 <= level <= 9 */
   UInt32 dictSize; /* (1 << 12) <= dictSize <= (1 << 27) for 32-bit version
@@ -23,14 +23,19 @@ typedef struct _CLzmaEncProps
   int fb;          /* 5 <= fb <= 273, default = 32 */
   int btMode;      /* 0 - hashChain Mode, 1 - binTree mode - normal, default = 1 */
   int numHashBytes; /* 2, 3 or 4, default = 4 */
+  unsigned numHashOutBits;  /* default = ? */
   UInt32 mc;       /* 1 <= mc <= (1 << 30), default = 32 */
   unsigned writeEndMark;  /* 0 - do not write EOPM, 1 - write EOPM, default = 0 */
   int numThreads;  /* 1 or 2, default = 2 */
 
+  // int _pad;
+  Int32 affinityGroup;
+
   UInt64 reduceSize; /* estimated size of data that will be compressed. default = (UInt64)(Int64)-1.
                         Encoder uses this value to reduce dictionary size */
 
   UInt64 affinity;
+  UInt64 affinityInGroup;
 } CLzmaEncProps;
 
 void LzmaEncProps_Init(CLzmaEncProps *p);
@@ -51,7 +56,9 @@ SRes:
   SZ_ERROR_THREAD - error in multithreading functions (only for Mt version)
 */
 
-typedef void * CLzmaEncHandle;
+typedef struct CLzmaEnc CLzmaEnc;
+typedef CLzmaEnc * CLzmaEncHandle;
+// Z7_DECLARE_HANDLE(CLzmaEncHandle)
 
 CLzmaEncHandle LzmaEnc_Create(ISzAllocPtr alloc);
 void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAllocPtr alloc, ISzAllocPtr allocBig);
@@ -61,17 +68,17 @@ void LzmaEnc_SetDataSize(CLzmaEncHandle p, UInt64 expectedDataSiize);
 SRes LzmaEnc_WriteProperties(CLzmaEncHandle p, Byte *properties, SizeT *size);
 unsigned LzmaEnc_IsWriteEndMark(CLzmaEncHandle p);
 
-SRes LzmaEnc_Encode(CLzmaEncHandle p, ISeqOutStream *outStream, ISeqInStream *inStream,
+SRes LzmaEnc_Encode(CLzmaEncHandle p, ISeqOutStreamPtr outStream, ISeqInStreamPtr inStream,
-    ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
+    ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
 SRes LzmaEnc_MemEncode(CLzmaEncHandle p, Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen,
-    int writeEndMark, ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
+    int writeEndMark, ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
 
 
 /* ---------- One Call Interface ---------- */
 
 SRes LzmaEncode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen,
     const CLzmaEncProps *props, Byte *propsEncoded, SizeT *propsSize, int writeEndMark,
-    ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
+    ICompressProgressPtr progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
 
 EXTERN_C_END
 

C/LzmaLib.c

@@ -1,12 +1,14 @@
 /* LzmaLib.c -- LZMA library wrapper
-2015-06-13 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
+
+#include "Precomp.h"
 
 #include "Alloc.h"
 #include "LzmaDec.h"
 #include "LzmaEnc.h"
 #include "LzmaLib.h"
 
-MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen,
+Z7_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen,
   unsigned char *outProps, size_t *outPropsSize,
   int level, /* 0 <= level <= 9, default = 5 */
   unsigned dictSize, /* use (1 << N) or (3 << N). 4 KB < dictSize <= 128 MB */
@@ -32,7 +34,7 @@ MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char
 }
 
 
-MY_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t *srcLen,
+Z7_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t *srcLen,
   const unsigned char *props, size_t propsSize)
 {
   ELzmaStatus status;

C/LzmaLib.h

@@ -1,14 +1,14 @@
 /* LzmaLib.h -- LZMA library interface
-2021-04-03 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
-#ifndef __LZMA_LIB_H
+#ifndef ZIP7_INC_LZMA_LIB_H
-#define __LZMA_LIB_H
+#define ZIP7_INC_LZMA_LIB_H
 
 #include "7zTypes.h"
 
 EXTERN_C_BEGIN
 
-#define MY_STDAPI int MY_STD_CALL
+#define Z7_STDAPI int Z7_STDCALL
 
 #define LZMA_PROPS_SIZE 5
 
@@ -100,7 +100,7 @@ Returns:
   SZ_ERROR_THREAD     - errors in multithreading functions (only for Mt version)
 */
 
-MY_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen,
+Z7_STDAPI LzmaCompress(unsigned char *dest, size_t *destLen, const unsigned char *src, size_t srcLen,
   unsigned char *outProps, size_t *outPropsSize, /* *outPropsSize must be = 5 */
   int level,      /* 0 <= level <= 9, default = 5 */
   unsigned dictSize,  /* default = (1 << 24) */
@@ -130,7 +130,7 @@ Returns:
   SZ_ERROR_INPUT_EOF   - it needs more bytes in input buffer (src)
 */
 
-MY_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, SizeT *srcLen,
+Z7_STDAPI LzmaUncompress(unsigned char *dest, size_t *destLen, const unsigned char *src, SizeT *srcLen,
   const unsigned char *props, size_t propsSize);
 
 EXTERN_C_END

C/Md5.c

@@ -0,0 +1,206 @@
+/* Md5.c -- MD5 Hash
+: Igor Pavlov : Public domain
+This code is based on Colin Plumb's public domain md5.c code */
+
+#include "Precomp.h"
+
+#include <string.h>
+
+#include "Md5.h"
+#include "RotateDefs.h"
+#include "CpuArch.h"
+
+#define MD5_UPDATE_BLOCKS(p) Md5_UpdateBlocks
+
+Z7_NO_INLINE
+void Md5_Init(CMd5 *p)
+{
+  p->count = 0;
+  p->state[0] = 0x67452301;
+  p->state[1] = 0xefcdab89;
+  p->state[2] = 0x98badcfe;
+  p->state[3] = 0x10325476;
+}
+
+#if 0 && !defined(MY_CPU_LE_UNALIGN)
+// optional optimization for Big-endian processors or processors without unaligned access:
+// it is intended to reduce the number of complex LE32 memory reading from 64 to 16.
+// But some compilers (sparc, armt) are better without this optimization.
+#define Z7_MD5_USE_DATA32_ARRAY
+#endif
+
+#define LOAD_DATA(i)  GetUi32((const UInt32 *)(const void *)data + (i))
+
+#ifdef Z7_MD5_USE_DATA32_ARRAY
+#define D(i)  data32[i]
+#else
+#define D(i)  LOAD_DATA(i)
+#endif
+
+#define F1(x, y, z)   (z ^ (x & (y ^ z)))
+#define F2(x, y, z)   F1(z, x, y)
+#define F3(x, y, z)   (x ^ y ^ z)
+#define F4(x, y, z)   (y ^ (x | ~z))
+
+#define R1(i, f, start, step, w, x, y, z, s, k) \
+    w += D((start + step * (i)) % 16) + k; \
+    w += f(x, y, z); \
+    w = rotlFixed(w, s) + x; \
+
+#define R4(i4,  f, start, step, s0,s1,s2,s3, k0,k1,k2,k3) \
+    R1 (i4*4+0, f, start, step, a,b,c,d, s0, k0) \
+    R1 (i4*4+1, f, start, step, d,a,b,c, s1, k1) \
+    R1 (i4*4+2, f, start, step, c,d,a,b, s2, k2) \
+    R1 (i4*4+3, f, start, step, b,c,d,a, s3, k3) \
+
+#define R16(f, start, step, s0,s1,s2,s3, k00,k01,k02,k03, k10,k11,k12,k13, k20,k21,k22,k23, k30,k31,k32,k33)  \
+    R4 (0,  f, start, step, s0,s1,s2,s3, k00,k01,k02,k03) \
+    R4 (1,  f, start, step, s0,s1,s2,s3, k10,k11,k12,k13) \
+    R4 (2,  f, start, step, s0,s1,s2,s3, k20,k21,k22,k23) \
+    R4 (3,  f, start, step, s0,s1,s2,s3, k30,k31,k32,k33) \
+
+static
+Z7_NO_INLINE
+void Z7_FASTCALL Md5_UpdateBlocks(UInt32 state[4], const Byte *data, size_t numBlocks)
+{
+  UInt32 a, b, c, d;
+  // if (numBlocks == 0) return;
+  a = state[0];
+  b = state[1];
+  c = state[2];
+  d = state[3];
+  do
+  {
+#ifdef Z7_MD5_USE_DATA32_ARRAY
+    UInt32 data32[MD5_NUM_BLOCK_WORDS];
+    {
+#define LOAD_data32_x4(i) { \
+      data32[i    ] = LOAD_DATA(i    ); \
+      data32[i + 1] = LOAD_DATA(i + 1); \
+      data32[i + 2] = LOAD_DATA(i + 2); \
+      data32[i + 3] = LOAD_DATA(i + 3); }
+#if 1
+      LOAD_data32_x4 (0 * 4)
+      LOAD_data32_x4 (1 * 4)
+      LOAD_data32_x4 (2 * 4)
+      LOAD_data32_x4 (3 * 4)
+#else
+      unsigned i;
+      for (i = 0; i < MD5_NUM_BLOCK_WORDS; i += 4)
+      {
+        LOAD_data32_x4(i)
+      }
+#endif
+    }
+#endif
+
+    R16 (F1, 0, 1,  7,12,17,22, 0xd76aa478, 0xe8c7b756, 0x242070db, 0xc1bdceee,
+                                0xf57c0faf, 0x4787c62a, 0xa8304613, 0xfd469501,
+                                0x698098d8, 0x8b44f7af, 0xffff5bb1, 0x895cd7be,
+                                0x6b901122, 0xfd987193, 0xa679438e, 0x49b40821)
+    R16 (F2, 1, 5,  5, 9,14,20, 0xf61e2562, 0xc040b340, 0x265e5a51, 0xe9b6c7aa,
+                                0xd62f105d, 0x02441453, 0xd8a1e681, 0xe7d3fbc8,
+                                0x21e1cde6, 0xc33707d6, 0xf4d50d87, 0x455a14ed,
+                                0xa9e3e905, 0xfcefa3f8, 0x676f02d9, 0x8d2a4c8a)
+    R16 (F3, 5, 3,  4,11,16,23, 0xfffa3942, 0x8771f681, 0x6d9d6122, 0xfde5380c,
+                                0xa4beea44, 0x4bdecfa9, 0xf6bb4b60, 0xbebfbc70,
+                                0x289b7ec6, 0xeaa127fa, 0xd4ef3085, 0x04881d05,
+                                0xd9d4d039, 0xe6db99e5, 0x1fa27cf8, 0xc4ac5665)
+    R16 (F4, 0, 7,  6,10,15,21, 0xf4292244, 0x432aff97, 0xab9423a7, 0xfc93a039,
+                                0x655b59c3, 0x8f0ccc92, 0xffeff47d, 0x85845dd1,
+                                0x6fa87e4f, 0xfe2ce6e0, 0xa3014314, 0x4e0811a1,
+                                0xf7537e82, 0xbd3af235, 0x2ad7d2bb, 0xeb86d391)
+
+    a += state[0];
+    b += state[1];
+    c += state[2];
+    d += state[3];
+    
+    state[0] = a;
+    state[1] = b;
+    state[2] = c;
+    state[3] = d;
+    
+    data += MD5_BLOCK_SIZE;
+  }
+  while (--numBlocks);
+}
+
+
+#define Md5_UpdateBlock(p) MD5_UPDATE_BLOCKS(p)(p->state, p->buffer, 1)
+
+void Md5_Update(CMd5 *p, const Byte *data, size_t size)
+{
+  if (size == 0)
+    return;
+  {
+    const unsigned pos = (unsigned)p->count & (MD5_BLOCK_SIZE - 1);
+    const unsigned num = MD5_BLOCK_SIZE - pos;
+    p->count += size;
+    if (num > size)
+    {
+      memcpy(p->buffer + pos, data, size);
+      return;
+    }
+    if (pos != 0)
+    {
+      size -= num;
+      memcpy(p->buffer + pos, data, num);
+      data += num;
+      Md5_UpdateBlock(p);
+    }
+  }
+  {
+    const size_t numBlocks = size >> 6;
+    if (numBlocks)
+    MD5_UPDATE_BLOCKS(p)(p->state, data, numBlocks);
+    size &= MD5_BLOCK_SIZE - 1;
+    if (size == 0)
+      return;
+    data += (numBlocks << 6);
+    memcpy(p->buffer, data, size);
+  }
+}
+
+
+void Md5_Final(CMd5 *p, Byte *digest)
+{
+  unsigned pos = (unsigned)p->count & (MD5_BLOCK_SIZE - 1);
+  p->buffer[pos++] = 0x80;
+  if (pos > (MD5_BLOCK_SIZE - 4 * 2))
+  {
+    while (pos != MD5_BLOCK_SIZE) { p->buffer[pos++] = 0; }
+    // memset(&p->buf.buffer[pos], 0, MD5_BLOCK_SIZE - pos);
+    Md5_UpdateBlock(p);
+    pos = 0;
+  }
+  memset(&p->buffer[pos], 0, (MD5_BLOCK_SIZE - 4 * 2) - pos);
+  {
+    const UInt64 numBits = p->count << 3;
+#if defined(MY_CPU_LE_UNALIGN)
+    SetUi64 (p->buffer + MD5_BLOCK_SIZE - 4 * 2, numBits)
+#else
+    SetUi32a(p->buffer + MD5_BLOCK_SIZE - 4 * 2, (UInt32)(numBits))
+    SetUi32a(p->buffer + MD5_BLOCK_SIZE - 4 * 1, (UInt32)(numBits >> 32))
+#endif
+  }
+  Md5_UpdateBlock(p);
+
+  SetUi32(digest,      p->state[0])
+  SetUi32(digest + 4,  p->state[1])
+  SetUi32(digest + 8,  p->state[2])
+  SetUi32(digest + 12, p->state[3])
+  
+  Md5_Init(p);
+}
+
+#undef R1
+#undef R4
+#undef R16
+#undef D
+#undef LOAD_DATA
+#undef LOAD_data32_x4
+#undef F1
+#undef F2
+#undef F3
+#undef F4

C/Md5.h

@@ -0,0 +1,34 @@
+/* Md5.h -- MD5 Hash
+: Igor Pavlov : Public domain */
+
+#ifndef ZIP7_INC_MD5_H
+#define ZIP7_INC_MD5_H
+
+#include "7zTypes.h"
+
+EXTERN_C_BEGIN
+
+#define MD5_NUM_BLOCK_WORDS  16
+#define MD5_NUM_DIGEST_WORDS  4
+
+#define MD5_BLOCK_SIZE   (MD5_NUM_BLOCK_WORDS * 4)
+#define MD5_DIGEST_SIZE  (MD5_NUM_DIGEST_WORDS * 4)
+
+typedef struct
+{
+  UInt64 count;
+  UInt64 _pad_1;
+  // we want 16-bytes alignment here
+  UInt32 state[MD5_NUM_DIGEST_WORDS];
+  UInt64 _pad_2[4];
+  // we want 64-bytes alignment here
+  Byte buffer[MD5_BLOCK_SIZE];
+} CMd5;
+
+void Md5_Init(CMd5 *p);
+void Md5_Update(CMd5 *p, const Byte *data, size_t size);
+void Md5_Final(CMd5 *p, Byte *digest);
+
+EXTERN_C_END
+
+#endif

C/MtCoder.c

@@ -1,28 +1,28 @@
 /* MtCoder.c -- Multi-thread Coder
-2021-12-21 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include "MtCoder.h"
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
-static SRes MtProgressThunk_Progress(const ICompressProgress *pp, UInt64 inSize, UInt64 outSize)
+static SRes MtProgressThunk_Progress(ICompressProgressPtr pp, UInt64 inSize, UInt64 outSize)
 {
-  CMtProgressThunk *thunk = CONTAINER_FROM_VTBL(pp, CMtProgressThunk, vt);
+  Z7_CONTAINER_FROM_VTBL_TO_DECL_VAR_pp_vt_p(CMtProgressThunk)
   UInt64 inSize2 = 0;
   UInt64 outSize2 = 0;
   if (inSize != (UInt64)(Int64)-1)
   {
-    inSize2 = inSize - thunk->inSize;
+    inSize2 = inSize - p->inSize;
-    thunk->inSize = inSize;
+    p->inSize = inSize;
   }
   if (outSize != (UInt64)(Int64)-1)
   {
-    outSize2 = outSize - thunk->outSize;
+    outSize2 = outSize - p->outSize;
-    thunk->outSize = outSize;
+    p->outSize = outSize;
   }
-  return MtProgress_ProgressAdd(thunk->mtProgress, inSize2, outSize2);
+  return MtProgress_ProgressAdd(p->mtProgress, inSize2, outSize2);
 }
 
 
@@ -36,25 +36,31 @@ void MtProgressThunk_CreateVTable(CMtProgressThunk *p)
 #define RINOK_THREAD(x) { if ((x) != 0) return SZ_ERROR_THREAD; }
 
 
-static WRes ArEvent_OptCreate_And_Reset(CEvent *p)
-{
-  if (Event_IsCreated(p))
-    return Event_Reset(p);
-  return AutoResetEvent_CreateNotSignaled(p);
-}
-
-
 static THREAD_FUNC_DECL ThreadFunc(void *pp);
 
 
-static SRes MtCoderThread_CreateAndStart(CMtCoderThread *t)
+static SRes MtCoderThread_CreateAndStart(CMtCoderThread *t
+#ifdef _WIN32
+    , CMtCoder * const mtc
+#endif
+    )
 {
-  WRes wres = ArEvent_OptCreate_And_Reset(&t->startEvent);
+  WRes wres = AutoResetEvent_OptCreate_And_Reset(&t->startEvent);
+  // printf("\n====== MtCoderThread_CreateAndStart : \n");
   if (wres == 0)
   {
     t->stop = False;
     if (!Thread_WasCreated(&t->thread))
-      wres = Thread_Create(&t->thread, ThreadFunc, t);
+    {
+#ifdef _WIN32
+      if (mtc->numThreadGroups)
+        wres = Thread_Create_With_Group(&t->thread, ThreadFunc, t,
+            ThreadNextGroup_GetNext(&mtc->nextGroup), // group
+            0); // affinityMask
+      else
+#endif
+        wres = Thread_Create(&t->thread, ThreadFunc, t);
+    }
     if (wres == 0)
       wres = Event_Set(&t->startEvent);
   }
@@ -64,6 +70,7 @@ static SRes MtCoderThread_CreateAndStart(CMtCoderThread *t)
 }
 
 
+Z7_FORCE_INLINE
 static void MtCoderThread_Destruct(CMtCoderThread *t)
 {
   if (Thread_WasCreated(&t->thread))
@@ -84,24 +91,6 @@ static void MtCoderThread_Destruct(CMtCoderThread *t)
 
 
 
-static SRes FullRead(ISeqInStream *stream, Byte *data, size_t *processedSize)
-{
-  size_t size = *processedSize;
-  *processedSize = 0;
-  while (size != 0)
-  {
-    size_t cur = size;
-    SRes res = ISeqInStream_Read(stream, data, &cur);
-    *processedSize += cur;
-    data += cur;
-    size -= cur;
-    RINOK(res);
-    if (cur == 0)
-      return SZ_OK;
-  }
-  return SZ_OK;
-}
-
 
 /*
   ThreadFunc2() returns:
@@ -111,7 +100,7 @@ static SRes FullRead(ISeqInStream *stream, Byte *data, size_t *processedSize)
 
 static SRes ThreadFunc2(CMtCoderThread *t)
 {
-  CMtCoder *mtc = t->mtCoder;
+  CMtCoder * const mtc = t->mtCoder;
 
   for (;;)
   {
@@ -152,7 +141,7 @@ static SRes ThreadFunc2(CMtCoderThread *t)
         }
         if (res == SZ_OK)
         {
-          res = FullRead(mtc->inStream, t->inBuf, &size);
+          res = SeqInStream_ReadMax(mtc->inStream, t->inBuf, &size);
           readProcessed = mtc->readProcessed + size;
           mtc->readProcessed = readProcessed;
         }
@@ -211,7 +200,11 @@ static SRes ThreadFunc2(CMtCoderThread *t)
       if (mtc->numStartedThreads < mtc->numStartedThreadsLimit
           && mtc->expectedDataSize != readProcessed)
       {
-        res = MtCoderThread_CreateAndStart(&mtc->threads[mtc->numStartedThreads]);
+        res = MtCoderThread_CreateAndStart(&mtc->threads[mtc->numStartedThreads]
+#ifdef _WIN32
+            , mtc
+#endif
+            );
         if (res == SZ_OK)
           mtc->numStartedThreads++;
         else
@@ -247,13 +240,13 @@ static SRes ThreadFunc2(CMtCoderThread *t)
     }
 
     {
-      CMtCoderBlock *block = &mtc->blocks[bi];
+      CMtCoderBlock * const block = &mtc->blocks[bi];
       block->res = res;
       block->bufIndex = bufIndex;
       block->finished = finished;
     }
     
-    #ifdef MTCODER__USE_WRITE_THREAD
+    #ifdef MTCODER_USE_WRITE_THREAD
       RINOK_THREAD(Event_Set(&mtc->writeEvents[bi]))
     #else
     {
@@ -337,7 +330,7 @@ static SRes ThreadFunc2(CMtCoderThread *t)
 
 static THREAD_FUNC_DECL ThreadFunc(void *pp)
 {
-  CMtCoderThread *t = (CMtCoderThread *)pp;
+  CMtCoderThread * const t = (CMtCoderThread *)pp;
   for (;;)
   {
     if (Event_Wait(&t->startEvent) != 0)
@@ -345,16 +338,16 @@ static THREAD_FUNC_DECL ThreadFunc(void *pp)
     if (t->stop)
       return 0;
     {
-      SRes res = ThreadFunc2(t);
+      const SRes res = ThreadFunc2(t);
       CMtCoder *mtc = t->mtCoder;
       if (res != SZ_OK)
       {
         MtProgress_SetError(&mtc->mtProgress, res);
       }
       
-      #ifndef MTCODER__USE_WRITE_THREAD
+      #ifndef MTCODER_USE_WRITE_THREAD
       {
-        unsigned numFinished = (unsigned)InterlockedIncrement(&mtc->numFinishedThreads);
+        const unsigned numFinished = (unsigned)InterlockedIncrement(&mtc->numFinishedThreads);
         if (numFinished == mtc->numStartedThreads)
           if (Event_Set(&mtc->finishedEvent) != 0)
             return (THREAD_FUNC_RET_TYPE)SZ_ERROR_THREAD;
@@ -372,6 +365,7 @@ void MtCoder_Construct(CMtCoder *p)
   
   p->blockSize = 0;
   p->numThreadsMax = 0;
+  p->numThreadGroups = 0;
   p->expectedDataSize = (UInt64)(Int64)-1;
 
   p->inStream = NULL;
@@ -389,7 +383,7 @@ void MtCoder_Construct(CMtCoder *p)
   Event_Construct(&p->readEvent);
   Semaphore_Construct(&p->blocksSemaphore);
 
-  for (i = 0; i < MTCODER__THREADS_MAX; i++)
+  for (i = 0; i < MTCODER_THREADS_MAX; i++)
   {
     CMtCoderThread *t = &p->threads[i];
     t->mtCoder = p;
@@ -397,11 +391,11 @@ void MtCoder_Construct(CMtCoder *p)
     t->inBuf = NULL;
     t->stop = False;
     Event_Construct(&t->startEvent);
-    Thread_Construct(&t->thread);
+    Thread_CONSTRUCT(&t->thread)
   }
 
-  #ifdef MTCODER__USE_WRITE_THREAD
+  #ifdef MTCODER_USE_WRITE_THREAD
-    for (i = 0; i < MTCODER__BLOCKS_MAX; i++)
+    for (i = 0; i < MTCODER_BLOCKS_MAX; i++)
       Event_Construct(&p->writeEvents[i]);
   #else
     Event_Construct(&p->finishedEvent);
@@ -424,14 +418,14 @@ static void MtCoder_Free(CMtCoder *p)
     Event_Set(&p->readEvent);
   */
 
-  for (i = 0; i < MTCODER__THREADS_MAX; i++)
+  for (i = 0; i < MTCODER_THREADS_MAX; i++)
     MtCoderThread_Destruct(&p->threads[i]);
 
   Event_Close(&p->readEvent);
   Semaphore_Close(&p->blocksSemaphore);
 
-  #ifdef MTCODER__USE_WRITE_THREAD
+  #ifdef MTCODER_USE_WRITE_THREAD
-    for (i = 0; i < MTCODER__BLOCKS_MAX; i++)
+    for (i = 0; i < MTCODER_BLOCKS_MAX; i++)
       Event_Close(&p->writeEvents[i]);
   #else
     Event_Close(&p->finishedEvent);
@@ -455,20 +449,22 @@ SRes MtCoder_Code(CMtCoder *p)
   unsigned i;
   SRes res = SZ_OK;
 
-  if (numThreads > MTCODER__THREADS_MAX)
+  // printf("\n====== MtCoder_Code : \n");
-    numThreads = MTCODER__THREADS_MAX;
+
-  numBlocksMax = MTCODER__GET_NUM_BLOCKS_FROM_THREADS(numThreads);
+  if (numThreads > MTCODER_THREADS_MAX)
+      numThreads = MTCODER_THREADS_MAX;
+  numBlocksMax = MTCODER_GET_NUM_BLOCKS_FROM_THREADS(numThreads);
   
   if (p->blockSize < ((UInt32)1 << 26)) numBlocksMax++;
   if (p->blockSize < ((UInt32)1 << 24)) numBlocksMax++;
   if (p->blockSize < ((UInt32)1 << 22)) numBlocksMax++;
 
-  if (numBlocksMax > MTCODER__BLOCKS_MAX)
+  if (numBlocksMax > MTCODER_BLOCKS_MAX)
-    numBlocksMax = MTCODER__BLOCKS_MAX;
+      numBlocksMax = MTCODER_BLOCKS_MAX;
 
   if (p->blockSize != p->allocatedBufsSize)
   {
-    for (i = 0; i < MTCODER__THREADS_MAX; i++)
+    for (i = 0; i < MTCODER_THREADS_MAX; i++)
     {
       CMtCoderThread *t = &p->threads[i];
       if (t->inBuf)
@@ -484,23 +480,23 @@ SRes MtCoder_Code(CMtCoder *p)
 
   MtProgress_Init(&p->mtProgress, p->progress);
 
-  #ifdef MTCODER__USE_WRITE_THREAD
+  #ifdef MTCODER_USE_WRITE_THREAD
     for (i = 0; i < numBlocksMax; i++)
     {
-      RINOK_THREAD(ArEvent_OptCreate_And_Reset(&p->writeEvents[i]));
+      RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->writeEvents[i]))
     }
   #else
-    RINOK_THREAD(ArEvent_OptCreate_And_Reset(&p->finishedEvent));
+    RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->finishedEvent))
   #endif
 
   {
-    RINOK_THREAD(ArEvent_OptCreate_And_Reset(&p->readEvent));
+    RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->readEvent))
-    RINOK_THREAD(Semaphore_OptCreateInit(&p->blocksSemaphore, numBlocksMax, numBlocksMax));
+    RINOK_THREAD(Semaphore_OptCreateInit(&p->blocksSemaphore, (UInt32)numBlocksMax, (UInt32)numBlocksMax))
   }
 
-  for (i = 0; i < MTCODER__BLOCKS_MAX - 1; i++)
+  for (i = 0; i < MTCODER_BLOCKS_MAX - 1; i++)
     p->freeBlockList[i] = i + 1;
-  p->freeBlockList[MTCODER__BLOCKS_MAX - 1] = (unsigned)(int)-1;
+  p->freeBlockList[MTCODER_BLOCKS_MAX - 1] = (unsigned)(int)-1;
   p->freeBlockHead = 0;
 
   p->readProcessed = 0;
@@ -508,26 +504,37 @@ SRes MtCoder_Code(CMtCoder *p)
   p->numBlocksMax = numBlocksMax;
   p->stopReading = False;
 
-  #ifndef MTCODER__USE_WRITE_THREAD
+  #ifndef MTCODER_USE_WRITE_THREAD
     p->writeIndex = 0;
     p->writeRes = SZ_OK;
-    for (i = 0; i < MTCODER__BLOCKS_MAX; i++)
+    for (i = 0; i < MTCODER_BLOCKS_MAX; i++)
       p->ReadyBlocks[i] = False;
     p->numFinishedThreads = 0;
   #endif
 
   p->numStartedThreadsLimit = numThreads;
   p->numStartedThreads = 0;
+  ThreadNextGroup_Init(&p->nextGroup, p->numThreadGroups, 0); // startGroup
 
   // for (i = 0; i < numThreads; i++)
   {
+    // here we create new thread for first block.
+    // And each new thread will create another new thread after block reading
+    // until numStartedThreadsLimit is reached.
     CMtCoderThread *nextThread = &p->threads[p->numStartedThreads++];
-    RINOK(MtCoderThread_CreateAndStart(nextThread));
+    {
+      const SRes res2 = MtCoderThread_CreateAndStart(nextThread
+#ifdef _WIN32
+            , p
+#endif
+            );
+      RINOK(res2)
+    }
   }
 
   RINOK_THREAD(Event_Set(&p->readEvent))
 
-  #ifdef MTCODER__USE_WRITE_THREAD
+  #ifdef MTCODER_USE_WRITE_THREAD
   {
     unsigned bi = 0;
 
@@ -539,9 +546,9 @@ SRes MtCoder_Code(CMtCoder *p)
       RINOK_THREAD(Event_Wait(&p->writeEvents[bi]))
 
       {
-        const CMtCoderBlock *block = &p->blocks[bi];
+        const CMtCoderBlock * const block = &p->blocks[bi];
-        unsigned bufIndex = block->bufIndex;
+        const unsigned bufIndex = block->bufIndex;
-        BoolInt finished = block->finished;
+        const BoolInt finished = block->finished;
         if (res == SZ_OK && block->res != SZ_OK)
           res = block->res;
 
@@ -571,7 +578,7 @@ SRes MtCoder_Code(CMtCoder *p)
   }
   #else
   {
-    WRes wres = Event_Wait(&p->finishedEvent);
+    const WRes wres = Event_Wait(&p->finishedEvent);
     res = MY_SRes_HRESULT_FROM_WRes(wres);
   }
   #endif
@@ -582,7 +589,7 @@ SRes MtCoder_Code(CMtCoder *p)
   if (res == SZ_OK)
     res = p->mtProgress.res;
 
-  #ifndef MTCODER__USE_WRITE_THREAD
+  #ifndef MTCODER_USE_WRITE_THREAD
     if (res == SZ_OK)
       res = p->writeRes;
   #endif
@@ -593,3 +600,5 @@ SRes MtCoder_Code(CMtCoder *p)
 }
 
 #endif
+
+#undef RINOK_THREAD

C/MtCoder.h

@@ -1,30 +1,30 @@
 /* MtCoder.h -- Multi-thread Coder
-2018-07-04 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __MT_CODER_H
+#ifndef ZIP7_INC_MT_CODER_H
-#define __MT_CODER_H
+#define ZIP7_INC_MT_CODER_H
 
 #include "MtDec.h"
 
 EXTERN_C_BEGIN
 
 /*
-  if (    defined MTCODER__USE_WRITE_THREAD) : main thread writes all data blocks to output stream
+  if (    defined MTCODER_USE_WRITE_THREAD) : main thread writes all data blocks to output stream
-  if (not defined MTCODER__USE_WRITE_THREAD) : any coder thread can write data blocks to output stream
+  if (not defined MTCODER_USE_WRITE_THREAD) : any coder thread can write data blocks to output stream
 */
-/* #define MTCODER__USE_WRITE_THREAD */
+/* #define MTCODER_USE_WRITE_THREAD */
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
-  #define MTCODER__GET_NUM_BLOCKS_FROM_THREADS(numThreads) ((numThreads) + (numThreads) / 8 + 1)
+  #define MTCODER_GET_NUM_BLOCKS_FROM_THREADS(numThreads) ((numThreads) + (numThreads) / 8 + 1)
-  #define MTCODER__THREADS_MAX 64
+  #define MTCODER_THREADS_MAX 256
-  #define MTCODER__BLOCKS_MAX (MTCODER__GET_NUM_BLOCKS_FROM_THREADS(MTCODER__THREADS_MAX) + 3)
+  #define MTCODER_BLOCKS_MAX (MTCODER_GET_NUM_BLOCKS_FROM_THREADS(MTCODER_THREADS_MAX) + 3)
 #else
-  #define MTCODER__THREADS_MAX 1
+  #define MTCODER_THREADS_MAX 1
-  #define MTCODER__BLOCKS_MAX 1
+  #define MTCODER_BLOCKS_MAX 1
 #endif
 
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
 
 typedef struct
@@ -37,15 +37,15 @@ typedef struct
 
 void MtProgressThunk_CreateVTable(CMtProgressThunk *p);
     
-#define MtProgressThunk_Init(p) { (p)->inSize = 0; (p)->outSize = 0; }
+#define MtProgressThunk_INIT(p) { (p)->inSize = 0; (p)->outSize = 0; }
 
 
-struct _CMtCoder;
+struct CMtCoder_;
 
 
 typedef struct
 {
-  struct _CMtCoder *mtCoder;
+  struct CMtCoder_ *mtCoder;
   unsigned index;
   int stop;
   Byte *inBuf;
@@ -71,19 +71,20 @@ typedef struct
 } CMtCoderBlock;
 
 
-typedef struct _CMtCoder
+typedef struct CMtCoder_
 {
   /* input variables */
   
   size_t blockSize;        /* size of input block */
   unsigned numThreadsMax;
+  unsigned numThreadGroups;
   UInt64 expectedDataSize;
 
-  ISeqInStream *inStream;
+  ISeqInStreamPtr inStream;
   const Byte *inData;
   size_t inDataSize;
 
-  ICompressProgress *progress;
+  ICompressProgressPtr progress;
   ISzAllocPtr allocBig;
 
   IMtCoderCallback2 *mtCallback;
@@ -100,13 +101,13 @@ typedef struct _CMtCoder
   BoolInt stopReading;
   SRes readRes;
 
-  #ifdef MTCODER__USE_WRITE_THREAD
+  #ifdef MTCODER_USE_WRITE_THREAD
-    CAutoResetEvent writeEvents[MTCODER__BLOCKS_MAX];
+    CAutoResetEvent writeEvents[MTCODER_BLOCKS_MAX];
   #else
     CAutoResetEvent finishedEvent;
     SRes writeRes;
     unsigned writeIndex;
-    Byte ReadyBlocks[MTCODER__BLOCKS_MAX];
+    Byte ReadyBlocks[MTCODER_BLOCKS_MAX];
     LONG numFinishedThreads;
   #endif
 
@@ -120,11 +121,13 @@ typedef struct _CMtCoder
   CCriticalSection cs;
 
   unsigned freeBlockHead;
-  unsigned freeBlockList[MTCODER__BLOCKS_MAX];
+  unsigned freeBlockList[MTCODER_BLOCKS_MAX];
 
   CMtProgress mtProgress;
-  CMtCoderBlock blocks[MTCODER__BLOCKS_MAX];
+  CMtCoderBlock blocks[MTCODER_BLOCKS_MAX];
-  CMtCoderThread threads[MTCODER__THREADS_MAX];
+  CMtCoderThread threads[MTCODER_THREADS_MAX];
+
+  CThreadNextGroup nextGroup;
 } CMtCoder;
 
 

C/MtDec.c

@@ -1,5 +1,5 @@
 /* MtDec.c -- Multi-thread Decoder
-2021-12-21 : Igor Pavlov : Public domain */
+2024-02-20 : Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -14,7 +14,7 @@
 
 #include "MtDec.h"
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
 #ifdef SHOW_DEBUG_INFO
 #define PRF(x) x
@@ -24,7 +24,7 @@
 
 #define PRF_STR_INT(s, d) PRF(printf("\n" s " %d\n", (unsigned)d))
 
-void MtProgress_Init(CMtProgress *p, ICompressProgress *progress)
+void MtProgress_Init(CMtProgress *p, ICompressProgressPtr progress)
 {
   p->progress = progress;
   p->res = SZ_OK;
@@ -81,36 +81,28 @@ void MtProgress_SetError(CMtProgress *p, SRes res)
 #define RINOK_THREAD(x) RINOK_WRes(x)
 
 
-static WRes ArEvent_OptCreate_And_Reset(CEvent *p)
+struct CMtDecBufLink_
 {
-  if (Event_IsCreated(p))
+  struct CMtDecBufLink_ *next;
-    return Event_Reset(p);
-  return AutoResetEvent_CreateNotSignaled(p);
-}
-
-
-struct __CMtDecBufLink
-{
-  struct __CMtDecBufLink *next;
   void *pad[3];
 };
 
-typedef struct __CMtDecBufLink CMtDecBufLink;
+typedef struct CMtDecBufLink_ CMtDecBufLink;
 
 #define MTDEC__LINK_DATA_OFFSET sizeof(CMtDecBufLink)
 #define MTDEC__DATA_PTR_FROM_LINK(link) ((Byte *)(link) + MTDEC__LINK_DATA_OFFSET)
 
 
 
-static THREAD_FUNC_DECL ThreadFunc(void *pp);
+static THREAD_FUNC_DECL MtDec_ThreadFunc(void *pp);
 
 
 static WRes MtDecThread_CreateEvents(CMtDecThread *t)
 {
-  WRes wres = ArEvent_OptCreate_And_Reset(&t->canWrite);
+  WRes wres = AutoResetEvent_OptCreate_And_Reset(&t->canWrite);
   if (wres == 0)
   {
-    wres = ArEvent_OptCreate_And_Reset(&t->canRead);
+    wres = AutoResetEvent_OptCreate_And_Reset(&t->canRead);
     if (wres == 0)
       return SZ_OK;
   }
@@ -126,7 +118,7 @@ static SRes MtDecThread_CreateAndStart(CMtDecThread *t)
   {
     if (Thread_WasCreated(&t->thread))
       return SZ_OK;
-    wres = Thread_Create(&t->thread, ThreadFunc, t);
+    wres = Thread_Create(&t->thread, MtDec_ThreadFunc, t);
     if (wres == 0)
       return SZ_OK;
   }
@@ -167,7 +159,7 @@ static void MtDecThread_CloseThread(CMtDecThread *t)
 static void MtDec_CloseThreads(CMtDec *p)
 {
   unsigned i;
-  for (i = 0; i < MTDEC__THREADS_MAX; i++)
+  for (i = 0; i < MTDEC_THREADS_MAX; i++)
     MtDecThread_CloseThread(&p->threads[i]);
 }
 
@@ -179,25 +171,6 @@ static void MtDecThread_Destruct(CMtDecThread *t)
 
 
 
-static SRes FullRead(ISeqInStream *stream, Byte *data, size_t *processedSize)
-{
-  size_t size = *processedSize;
-  *processedSize = 0;
-  while (size != 0)
-  {
-    size_t cur = size;
-    SRes res = ISeqInStream_Read(stream, data, &cur);
-    *processedSize += cur;
-    data += cur;
-    size -= cur;
-    RINOK(res);
-    if (cur == 0)
-      return SZ_OK;
-  }
-  return SZ_OK;
-}
-
-
 static SRes MtDec_GetError_Spec(CMtDec *p, UInt64 interruptIndex, BoolInt *wasInterrupted)
 {
   SRes res;
@@ -253,7 +226,7 @@ Byte *MtDec_GetCrossBuff(CMtDec *p)
 
 
 /*
-  ThreadFunc2() returns:
+  MtDec_ThreadFunc2() returns:
   0      - in all normal cases (even for stream error or memory allocation error)
   (!= 0) - WRes error return by system threading function
 */
@@ -261,11 +234,11 @@ Byte *MtDec_GetCrossBuff(CMtDec *p)
 // #define MTDEC_ProgessStep (1 << 22)
 #define MTDEC_ProgessStep (1 << 0)
 
-static WRes ThreadFunc2(CMtDecThread *t)
+static WRes MtDec_ThreadFunc2(CMtDecThread *t)
 {
   CMtDec *p = t->mtDec;
 
-  PRF_STR_INT("ThreadFunc2", t->index);
+  PRF_STR_INT("MtDec_ThreadFunc2", t->index)
 
   // SetThreadAffinityMask(GetCurrentThread(), 1 << t->index);
 
@@ -295,13 +268,13 @@ static WRes ThreadFunc2(CMtDecThread *t)
     // CMtDecCallbackInfo parse;
     CMtDecThread *nextThread;
 
-    PRF_STR_INT("=============== Event_Wait(&t->canRead)", t->index);
+    PRF_STR_INT("=============== Event_Wait(&t->canRead)", t->index)
 
-    RINOK_THREAD(Event_Wait(&t->canRead));
+    RINOK_THREAD(Event_Wait(&t->canRead))
     if (p->exitThread)
       return 0;
 
-    PRF_STR_INT("after Event_Wait(&t->canRead)", t->index);
+    PRF_STR_INT("after Event_Wait(&t->canRead)", t->index)
 
     // if (t->index == 3) return 19; // for test
 
@@ -373,7 +346,7 @@ static WRes ThreadFunc2(CMtDecThread *t)
             {
               size = p->inBufSize;
               
-              res = FullRead(p->inStream, data, &size);
+              res = SeqInStream_ReadMax(p->inStream, data, &size);
               
               // size = 10; // test
 
@@ -615,7 +588,7 @@ static WRes ThreadFunc2(CMtDecThread *t)
     // if ( !finish ) we must call Event_Set(&nextThread->canWrite) in any case
     // if (  finish ) we switch to single-thread mode and there are 2 ways at the end of current iteration (current block):
     //   - if (needContinue) after Write(&needContinue), we restore decoding with new iteration
-    //   - otherwise we stop decoding and exit from ThreadFunc2()
+    //   - otherwise we stop decoding and exit from MtDec_ThreadFunc2()
 
     // Don't change (finish) variable in the further code
 
@@ -688,7 +661,7 @@ static WRes ThreadFunc2(CMtDecThread *t)
 
     // ---------- WRITE ----------
    
-    RINOK_THREAD(Event_Wait(&t->canWrite));
+    RINOK_THREAD(Event_Wait(&t->canWrite))
 
   {
     BoolInt isErrorMode = False;
@@ -801,14 +774,14 @@ static WRes ThreadFunc2(CMtDecThread *t)
 
     if (!finish)
     {
-      RINOK_THREAD(Event_Set(&nextThread->canWrite));
+      RINOK_THREAD(Event_Set(&nextThread->canWrite))
     }
     else
     {
       if (needContinue)
       {
         // we restore decoding with new iteration
-        RINOK_THREAD(Event_Set(&p->threads[0].canWrite));
+        RINOK_THREAD(Event_Set(&p->threads[0].canWrite))
       }
       else
       {
@@ -817,7 +790,7 @@ static WRes ThreadFunc2(CMtDecThread *t)
           return SZ_OK;
         p->exitThread = True;
       }
-      RINOK_THREAD(Event_Set(&p->threads[0].canRead));
+      RINOK_THREAD(Event_Set(&p->threads[0].canRead))
     }
   }
   }
@@ -836,7 +809,17 @@ static WRes ThreadFunc2(CMtDecThread *t)
 #endif
 
 
-static THREAD_FUNC_DECL ThreadFunc1(void *pp)
+typedef
+  #ifdef _WIN32
+    UINT_PTR
+  #elif 1
+    uintptr_t
+  #else
+    ptrdiff_t
+  #endif
+    MY_uintptr_t;
+
+static THREAD_FUNC_DECL MtDec_ThreadFunc1(void *pp)
 {
   WRes res;
 
@@ -845,10 +828,10 @@ static THREAD_FUNC_DECL ThreadFunc1(void *pp)
 
   // fprintf(stdout, "\n%d = %p\n", t->index, &t);
 
-  res = ThreadFunc2(t);
+  res = MtDec_ThreadFunc2(t);
   p = t->mtDec;
   if (res == 0)
-    return (THREAD_FUNC_RET_TYPE)(UINT_PTR)p->exitThreadWRes;
+    return (THREAD_FUNC_RET_TYPE)(MY_uintptr_t)p->exitThreadWRes;
   {
     // it's unexpected situation for some threading function error
     if (p->exitThreadWRes == 0)
@@ -859,17 +842,17 @@ static THREAD_FUNC_DECL ThreadFunc1(void *pp)
     Event_Set(&p->threads[0].canWrite);
     MtProgress_SetError(&p->mtProgress, MY_SRes_HRESULT_FROM_WRes(res));
   }
-  return (THREAD_FUNC_RET_TYPE)(UINT_PTR)res;
+  return (THREAD_FUNC_RET_TYPE)(MY_uintptr_t)res;
 }
 
-static MY_NO_INLINE THREAD_FUNC_DECL ThreadFunc(void *pp)
+static Z7_NO_INLINE THREAD_FUNC_DECL MtDec_ThreadFunc(void *pp)
 {
   #ifdef USE_ALLOCA
   CMtDecThread *t = (CMtDecThread *)pp;
   // fprintf(stderr, "\n%d = %p - before", t->index, &t);
   t->allocaPtr = alloca(t->index * 128);
   #endif
-  return ThreadFunc1(pp);
+  return MtDec_ThreadFunc1(pp);
 }
 
 
@@ -883,7 +866,7 @@ int MtDec_PrepareRead(CMtDec *p)
     
   {
     unsigned i;
-    for (i = 0; i < MTDEC__THREADS_MAX; i++)
+    for (i = 0; i < MTDEC_THREADS_MAX; i++)
       if (i > p->numStartedThreads
           || p->numFilledThreads <=
             (i >= p->filledThreadStart ?
@@ -987,7 +970,7 @@ void MtDec_Construct(CMtDec *p)
 
   p->allocatedBufsSize = 0;
 
-  for (i = 0; i < MTDEC__THREADS_MAX; i++)
+  for (i = 0; i < MTDEC_THREADS_MAX; i++)
   {
     CMtDecThread *t = &p->threads[i];
     t->mtDec = p;
@@ -995,7 +978,7 @@ void MtDec_Construct(CMtDec *p)
     t->inBuf = NULL;
     Event_Construct(&t->canRead);
     Event_Construct(&t->canWrite);
-    Thread_Construct(&t->thread);
+    Thread_CONSTRUCT(&t->thread)
   }
 
   // Event_Construct(&p->finishedEvent);
@@ -1010,7 +993,7 @@ static void MtDec_Free(CMtDec *p)
 
   p->exitThread = True;
 
-  for (i = 0; i < MTDEC__THREADS_MAX; i++)
+  for (i = 0; i < MTDEC_THREADS_MAX; i++)
     MtDecThread_Destruct(&p->threads[i]);
 
   // Event_Close(&p->finishedEvent);
@@ -1061,15 +1044,15 @@ SRes MtDec_Code(CMtDec *p)
 
   {
     unsigned numThreads = p->numThreadsMax;
-    if (numThreads > MTDEC__THREADS_MAX)
+    if (numThreads > MTDEC_THREADS_MAX)
-      numThreads = MTDEC__THREADS_MAX;
+      numThreads = MTDEC_THREADS_MAX;
     p->numStartedThreads_Limit = numThreads;
     p->numStartedThreads = 0;
   }
 
   if (p->inBufSize != p->allocatedBufsSize)
   {
-    for (i = 0; i < MTDEC__THREADS_MAX; i++)
+    for (i = 0; i < MTDEC_THREADS_MAX; i++)
     {
       CMtDecThread *t = &p->threads[i];
       if (t->inBuf)
@@ -1086,7 +1069,7 @@ SRes MtDec_Code(CMtDec *p)
 
   MtProgress_Init(&p->mtProgress, p->progress);
 
-  // RINOK_THREAD(ArEvent_OptCreate_And_Reset(&p->finishedEvent));
+  // RINOK_THREAD(AutoResetEvent_OptCreate_And_Reset(&p->finishedEvent))
   p->exitThread = False;
   p->exitThreadWRes = 0;
 
@@ -1098,8 +1081,8 @@ SRes MtDec_Code(CMtDec *p)
     wres = MtDecThread_CreateEvents(nextThread);
     if (wres == 0) { wres = Event_Set(&nextThread->canWrite);
     if (wres == 0) { wres = Event_Set(&nextThread->canRead);
-    if (wres == 0) { THREAD_FUNC_RET_TYPE res = ThreadFunc(nextThread);
+    if (wres == 0) { THREAD_FUNC_RET_TYPE res = MtDec_ThreadFunc(nextThread);
-    wres = (WRes)(UINT_PTR)res;
+    wres = (WRes)(MY_uintptr_t)res;
     if (wres != 0)
     {
       p->needContinue = False;
@@ -1137,3 +1120,5 @@ SRes MtDec_Code(CMtDec *p)
 }
 
 #endif
+
+#undef PRF

C/MtDec.h

@@ -1,46 +1,46 @@
 /* MtDec.h -- Multi-thread Decoder
-2020-03-05 : Igor Pavlov : Public domain */
+2023-04-02 : Igor Pavlov : Public domain */
 
-#ifndef __MT_DEC_H
+#ifndef ZIP7_INC_MT_DEC_H
-#define __MT_DEC_H
+#define ZIP7_INC_MT_DEC_H
 
 #include "7zTypes.h"
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 #include "Threads.h"
 #endif
 
 EXTERN_C_BEGIN
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
 
-#ifndef _7ZIP_ST
+#ifndef Z7_ST
-  #define MTDEC__THREADS_MAX 32
+  #define MTDEC_THREADS_MAX 32
 #else
-  #define MTDEC__THREADS_MAX 1
+  #define MTDEC_THREADS_MAX 1
 #endif
 
 
 typedef struct
 {
-  ICompressProgress *progress;
+  ICompressProgressPtr progress;
   SRes res;
   UInt64 totalInSize;
   UInt64 totalOutSize;
   CCriticalSection cs;
 } CMtProgress;
 
-void MtProgress_Init(CMtProgress *p, ICompressProgress *progress);
+void MtProgress_Init(CMtProgress *p, ICompressProgressPtr progress);
 SRes MtProgress_Progress_ST(CMtProgress *p);
 SRes MtProgress_ProgressAdd(CMtProgress *p, UInt64 inSize, UInt64 outSize);
 SRes MtProgress_GetError(CMtProgress *p);
 void MtProgress_SetError(CMtProgress *p, SRes res);
 
-struct _CMtDec;
+struct CMtDec;
 
 typedef struct
 {
-  struct _CMtDec *mtDec;
+  struct CMtDec_ *mtDec;
   unsigned index;
   void *inBuf;
 
@@ -117,7 +117,7 @@ typedef struct
 
 
 
-typedef struct _CMtDec
+typedef struct CMtDec_
 {
   /* input variables */
   
@@ -126,11 +126,11 @@ typedef struct _CMtDec
   // size_t inBlockMax;
   unsigned numThreadsMax_2;
 
-  ISeqInStream *inStream;
+  ISeqInStreamPtr inStream;
   // const Byte *inData;
   // size_t inDataSize;
 
-  ICompressProgress *progress;
+  ICompressProgressPtr progress;
   ISzAllocPtr alloc;
 
   IMtDecCallback2 *mtCallback;
@@ -171,11 +171,11 @@ typedef struct _CMtDec
   unsigned filledThreadStart;
   unsigned numFilledThreads;
 
-  #ifndef _7ZIP_ST
+  #ifndef Z7_ST
   BoolInt needInterrupt;
   UInt64 interruptIndex;
   CMtProgress mtProgress;
-  CMtDecThread threads[MTDEC__THREADS_MAX];
+  CMtDecThread threads[MTDEC_THREADS_MAX];
   #endif
 } CMtDec;
 

C/Ppmd.h

@@ -1,9 +1,9 @@
 /* Ppmd.h -- PPMD codec common code
-2021-04-13 : Igor Pavlov : Public domain
+2023-03-05 : Igor Pavlov : Public domain
 This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */
 
-#ifndef __PPMD_H
+#ifndef ZIP7_INC_PPMD_H
-#define __PPMD_H
+#define ZIP7_INC_PPMD_H
 
 #include "CpuArch.h"
 
@@ -48,8 +48,10 @@ typedef struct
   Byte Count;  /* Count to next change of Shift */
 } CPpmd_See;
 
-#define Ppmd_See_Update(p)  if ((p)->Shift < PPMD_PERIOD_BITS && --(p)->Count == 0) \
+#define Ppmd_See_UPDATE(p) \
-    { (p)->Summ = (UInt16)((p)->Summ << 1); (p)->Count = (Byte)(3 << (p)->Shift++); }
+  { if ((p)->Shift < PPMD_PERIOD_BITS && --(p)->Count == 0) \
+    { (p)->Summ = (UInt16)((p)->Summ << 1); \
+      (p)->Count = (Byte)(3 << (p)->Shift++); }}
 
 
 typedef struct

C/Ppmd7.c

@@ -1,5 +1,5 @@
 /* Ppmd7.c -- PPMdH codec
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */
 
 #include "Precomp.h"
@@ -14,7 +14,7 @@ This code is based on PPMd var.H (2001): Dmitry Shkarin : Public domain */
 MY_ALIGN(16)
 static const Byte PPMD7_kExpEscape[16] = { 25, 14, 9, 7, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2 };
 MY_ALIGN(16)
-static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051};
+static const UInt16 PPMD7_kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051};
 
 #define MAX_FREQ 124
 #define UNIT_SIZE 12
@@ -33,7 +33,7 @@ static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x
 #define ONE_STATE(ctx) Ppmd7Context_OneState(ctx)
 #define SUFFIX(ctx) CTX((ctx)->Suffix)
 
-typedef CPpmd7_Context * CTX_PTR;
+typedef CPpmd7_Context * PPMD7_CTX_PTR;
 
 struct CPpmd7_Node_;
 
@@ -107,14 +107,14 @@ BoolInt Ppmd7_Alloc(CPpmd7 *p, UInt32 size, ISzAllocPtr alloc)
 // ---------- Internal Memory Allocator ----------
 
 /* We can use CPpmd7_Node in list of free units (as in Ppmd8)
-   But we still need one additional list walk pass in GlueFreeBlocks().
+   But we still need one additional list walk pass in Ppmd7_GlueFreeBlocks().
-   So we use simple CPpmd_Void_Ref instead of CPpmd7_Node in InsertNode() / RemoveNode()
+   So we use simple CPpmd_Void_Ref instead of CPpmd7_Node in Ppmd7_InsertNode() / Ppmd7_RemoveNode()
 */
 
 #define EMPTY_NODE 0
 
 
-static void InsertNode(CPpmd7 *p, void *node, unsigned indx)
+static void Ppmd7_InsertNode(CPpmd7 *p, void *node, unsigned indx)
 {
   *((CPpmd_Void_Ref *)node) = p->FreeList[indx];
   // ((CPpmd7_Node *)node)->Next = (CPpmd7_Node_Ref)p->FreeList[indx];
@@ -124,7 +124,7 @@ static void InsertNode(CPpmd7 *p, void *node, unsigned indx)
 }
 
 
-static void *RemoveNode(CPpmd7 *p, unsigned indx)
+static void *Ppmd7_RemoveNode(CPpmd7 *p, unsigned indx)
 {
   CPpmd_Void_Ref *node = (CPpmd_Void_Ref *)Ppmd7_GetPtr(p, p->FreeList[indx]);
   p->FreeList[indx] = *node;
@@ -134,32 +134,32 @@ static void *RemoveNode(CPpmd7 *p, unsigned indx)
 }
 
 
-static void SplitBlock(CPpmd7 *p, void *ptr, unsigned oldIndx, unsigned newIndx)
+static void Ppmd7_SplitBlock(CPpmd7 *p, void *ptr, unsigned oldIndx, unsigned newIndx)
 {
   unsigned i, nu = I2U(oldIndx) - I2U(newIndx);
   ptr = (Byte *)ptr + U2B(I2U(newIndx));
   if (I2U(i = U2I(nu)) != nu)
   {
     unsigned k = I2U(--i);
-    InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1);
+    Ppmd7_InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1);
   }
-  InsertNode(p, ptr, i);
+  Ppmd7_InsertNode(p, ptr, i);
 }
 
 
 /* we use CPpmd7_Node_Union union to solve XLC -O2 strict pointer aliasing problem */
 
-typedef union _CPpmd7_Node_Union
+typedef union
 {
   CPpmd7_Node     Node;
   CPpmd7_Node_Ref NextRef;
 } CPpmd7_Node_Union;
 
-/* Original PPmdH (Ppmd7) code uses doubly linked list in GlueFreeBlocks()
+/* Original PPmdH (Ppmd7) code uses doubly linked list in Ppmd7_GlueFreeBlocks()
    we use single linked list similar to Ppmd8 code */
 
 
-static void GlueFreeBlocks(CPpmd7 *p)
+static void Ppmd7_GlueFreeBlocks(CPpmd7 *p)
 {
   /*
   we use first UInt16 field of 12-bytes UNITs as record type stamp
@@ -239,27 +239,27 @@ static void GlueFreeBlocks(CPpmd7 *p)
     if (nu == 0)
       continue;
     for (; nu > 128; nu -= 128, node += 128)
-      InsertNode(p, node, PPMD_NUM_INDEXES - 1);
+      Ppmd7_InsertNode(p, node, PPMD_NUM_INDEXES - 1);
     if (I2U(i = U2I(nu)) != nu)
     {
       unsigned k = I2U(--i);
-      InsertNode(p, node + k, (unsigned)nu - k - 1);
+      Ppmd7_InsertNode(p, node + k, (unsigned)nu - k - 1);
     }
-    InsertNode(p, node, i);
+    Ppmd7_InsertNode(p, node, i);
   }
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static void *AllocUnitsRare(CPpmd7 *p, unsigned indx)
+static void *Ppmd7_AllocUnitsRare(CPpmd7 *p, unsigned indx)
 {
   unsigned i;
   
   if (p->GlueCount == 0)
   {
-    GlueFreeBlocks(p);
+    Ppmd7_GlueFreeBlocks(p);
     if (p->FreeList[indx] != 0)
-      return RemoveNode(p, indx);
+      return Ppmd7_RemoveNode(p, indx);
   }
   
   i = indx;
@@ -277,17 +277,17 @@ static void *AllocUnitsRare(CPpmd7 *p, unsigned indx)
   while (p->FreeList[i] == 0);
 
   {
-    void *block = RemoveNode(p, i);
+    void *block = Ppmd7_RemoveNode(p, i);
-    SplitBlock(p, block, i, indx);
+    Ppmd7_SplitBlock(p, block, i, indx);
     return block;
   }
 }
 
 
-static void *AllocUnits(CPpmd7 *p, unsigned indx)
+static void *Ppmd7_AllocUnits(CPpmd7 *p, unsigned indx)
 {
   if (p->FreeList[indx] != 0)
-    return RemoveNode(p, indx);
+    return Ppmd7_RemoveNode(p, indx);
   {
     UInt32 numBytes = U2B(I2U(indx));
     Byte *lo = p->LoUnit;
@@ -297,13 +297,22 @@ static void *AllocUnits(CPpmd7 *p, unsigned indx)
       return lo;
     }
   }
-  return AllocUnitsRare(p, indx);
+  return Ppmd7_AllocUnitsRare(p, indx);
 }
 
 
-#define MyMem12Cpy(dest, src, num) \
+#define MEM_12_CPY(dest, src, num) \
-  { UInt32 *d = (UInt32 *)dest; const UInt32 *z = (const UInt32 *)src; UInt32 n = num; \
+  { UInt32 *d = (UInt32 *)(dest); \
-    do { d[0] = z[0]; d[1] = z[1]; d[2] = z[2]; z += 3; d += 3; } while (--n); }
+    const UInt32 *z = (const UInt32 *)(src); \
+    unsigned n = (num); \
+    do { \
+      d[0] = z[0]; \
+      d[1] = z[1]; \
+      d[2] = z[2]; \
+      z += 3; \
+      d += 3; \
+    } while (--n); \
+  }
 
 
 /*
@@ -315,12 +324,12 @@ static void *ShrinkUnits(CPpmd7 *p, void *oldPtr, unsigned oldNU, unsigned newNU
     return oldPtr;
   if (p->FreeList[i1] != 0)
   {
-    void *ptr = RemoveNode(p, i1);
+    void *ptr = Ppmd7_RemoveNode(p, i1);
-    MyMem12Cpy(ptr, oldPtr, newNU);
+    MEM_12_CPY(ptr, oldPtr, newNU)
-    InsertNode(p, oldPtr, i0);
+    Ppmd7_InsertNode(p, oldPtr, i0);
     return ptr;
   }
-  SplitBlock(p, oldPtr, i0, i1);
+  Ppmd7_SplitBlock(p, oldPtr, i0, i1);
   return oldPtr;
 }
 */
@@ -329,14 +338,14 @@ static void *ShrinkUnits(CPpmd7 *p, void *oldPtr, unsigned oldNU, unsigned newNU
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
 static void SetSuccessor(CPpmd_State *p, CPpmd_Void_Ref v)
 {
-  Ppmd_SET_SUCCESSOR(p, v);
+  Ppmd_SET_SUCCESSOR(p, v)
 }
 
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static
-void RestartModel(CPpmd7 *p)
+void Ppmd7_RestartModel(CPpmd7 *p)
 {
   unsigned i, k;
 
@@ -352,8 +361,8 @@ void RestartModel(CPpmd7 *p)
   p->PrevSuccess = 0;
 
   {
-    CPpmd7_Context *mc = (CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */
+    CPpmd7_Context *mc = (PPMD7_CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */
-    CPpmd_State *s = (CPpmd_State *)p->LoUnit; /* AllocUnits(p, PPMD_NUM_INDEXES - 1); */
+    CPpmd_State *s = (CPpmd_State *)p->LoUnit; /* Ppmd7_AllocUnits(p, PPMD_NUM_INDEXES - 1); */
     
     p->LoUnit += U2B(256 / 2);
     p->MaxContext = p->MinContext = mc;
@@ -391,7 +400,7 @@ void RestartModel(CPpmd7 *p)
     {
       unsigned m;
       UInt16 *dest = p->BinSumm[i] + k;
-      UInt16 val = (UInt16)(PPMD_BIN_SCALE - kInitBinEsc[k] / (i + 2));
+      const UInt16 val = (UInt16)(PPMD_BIN_SCALE - PPMD7_kInitBinEsc[k] / (i + 2));
       for (m = 0; m < 64; m += 8)
         dest[m] = val;
     }
@@ -423,13 +432,13 @@ void Ppmd7_Init(CPpmd7 *p, unsigned maxOrder)
 {
   p->MaxOrder = maxOrder;
 
-  RestartModel(p);
+  Ppmd7_RestartModel(p);
 }
 
 
 
 /*
-  CreateSuccessors()
+  Ppmd7_CreateSuccessors()
   It's called when (FoundState->Successor) is RAW-Successor,
   that is the link to position in Raw text.
   So we create Context records and write the links to
@@ -445,10 +454,10 @@ void Ppmd7_Init(CPpmd7 *p, unsigned maxOrder)
   also it can return pointer to real context of same order,
 */
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static CTX_PTR CreateSuccessors(CPpmd7 *p)
+static PPMD7_CTX_PTR Ppmd7_CreateSuccessors(CPpmd7 *p)
 {
-  CTX_PTR c = p->MinContext;
+  PPMD7_CTX_PTR c = p->MinContext;
   CPpmd_Byte_Ref upBranch = (CPpmd_Byte_Ref)SUCCESSOR(p->FoundState);
   Byte newSym, newFreq;
   unsigned numPs = 0;
@@ -522,15 +531,15 @@ static CTX_PTR CreateSuccessors(CPpmd7 *p)
 
   do
   {
-    CTX_PTR c1;
+    PPMD7_CTX_PTR c1;
     /* = AllocContext(p); */
     if (p->HiUnit != p->LoUnit)
-      c1 = (CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE);
+      c1 = (PPMD7_CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE);
     else if (p->FreeList[0] != 0)
-      c1 = (CTX_PTR)RemoveNode(p, 0);
+      c1 = (PPMD7_CTX_PTR)Ppmd7_RemoveNode(p, 0);
     else
     {
-      c1 = (CTX_PTR)AllocUnitsRare(p, 0);
+      c1 = (PPMD7_CTX_PTR)Ppmd7_AllocUnitsRare(p, 0);
       if (!c1)
         return NULL;
     }
@@ -550,16 +559,16 @@ static CTX_PTR CreateSuccessors(CPpmd7 *p)
 
 
 
-#define SwapStates(s) \
+#define SWAP_STATES(s) \
   { CPpmd_State tmp = s[0]; s[0] = s[-1]; s[-1] = tmp; }
 
 
 void Ppmd7_UpdateModel(CPpmd7 *p);
-MY_NO_INLINE
+Z7_NO_INLINE
 void Ppmd7_UpdateModel(CPpmd7 *p)
 {
   CPpmd_Void_Ref maxSuccessor, minSuccessor;
-  CTX_PTR c, mc;
+  PPMD7_CTX_PTR c, mc;
   unsigned s0, ns;
 
 
@@ -592,7 +601,7 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
         
         if (s[0].Freq >= s[-1].Freq)
         {
-          SwapStates(s);
+          SWAP_STATES(s)
           s--;
         }
       }
@@ -610,10 +619,10 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
   {
     /* MAX ORDER context */
     /* (FoundState->Successor) is RAW-Successor. */
-    p->MaxContext = p->MinContext = CreateSuccessors(p);
+    p->MaxContext = p->MinContext = Ppmd7_CreateSuccessors(p);
     if (!p->MinContext)
     {
-      RestartModel(p);
+      Ppmd7_RestartModel(p);
       return;
     }
     SetSuccessor(p->FoundState, REF(p->MinContext));
@@ -629,7 +638,7 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
     p->Text = text;
     if (text >= p->UnitsStart)
     {
-      RestartModel(p);
+      Ppmd7_RestartModel(p);
       return;
     }
     maxSuccessor = REF(text);
@@ -645,10 +654,10 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
     if (minSuccessor <= maxSuccessor)
     {
       // minSuccessor is RAW-Successor. So we will create real contexts records:
-      CTX_PTR cs = CreateSuccessors(p);
+      PPMD7_CTX_PTR cs = Ppmd7_CreateSuccessors(p);
       if (!cs)
       {
-        RestartModel(p);
+        Ppmd7_RestartModel(p);
         return;
       }
       minSuccessor = REF(cs);
@@ -711,27 +720,27 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
       if ((ns1 & 1) == 0)
       {
         /* Expand for one UNIT */
-        unsigned oldNU = ns1 >> 1;
+        const unsigned oldNU = ns1 >> 1;
-        unsigned i = U2I(oldNU);
+        const unsigned i = U2I(oldNU);
         if (i != U2I((size_t)oldNU + 1))
         {
-          void *ptr = AllocUnits(p, i + 1);
+          void *ptr = Ppmd7_AllocUnits(p, i + 1);
           void *oldPtr;
           if (!ptr)
           {
-            RestartModel(p);
+            Ppmd7_RestartModel(p);
             return;
           }
           oldPtr = STATS(c);
-          MyMem12Cpy(ptr, oldPtr, oldNU);
+          MEM_12_CPY(ptr, oldPtr, oldNU)
-          InsertNode(p, oldPtr, i);
+          Ppmd7_InsertNode(p, oldPtr, i);
           c->Union4.Stats = STATS_REF(ptr);
         }
       }
       sum = c->Union2.SummFreq;
       /* max increase of Escape_Freq is 3 here.
          total increase of Union2.SummFreq for all symbols is less than 256 here */
-      sum += (UInt32)(2 * ns1 < ns) + 2 * ((unsigned)(4 * ns1 <= ns) & (sum <= 8 * ns1));
+      sum += (UInt32)(unsigned)((2 * ns1 < ns) + 2 * ((unsigned)(4 * ns1 <= ns) & (sum <= 8 * ns1)));
       /* original PPMdH uses 16-bit variable for (sum) here.
          But (sum < 0x9000). So we don't truncate (sum) to 16-bit */
       // sum = (UInt16)sum;
@@ -739,10 +748,10 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
     else
     {
       // instead of One-symbol context we create 2-symbol context
-      CPpmd_State *s = (CPpmd_State*)AllocUnits(p, 0);
+      CPpmd_State *s = (CPpmd_State*)Ppmd7_AllocUnits(p, 0);
       if (!s)
       {
-        RestartModel(p);
+        Ppmd7_RestartModel(p);
         return;
       }
       {
@@ -761,7 +770,7 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
         // (max(s->freq) == 120), when we convert from 1-symbol into 2-symbol context
         s->Freq = (Byte)freq;
         // max(InitEsc = PPMD7_kExpEscape[*]) is 25. So the max(escapeFreq) is 26 here
-        sum = freq + p->InitEsc + (ns > 3);
+        sum = (UInt32)(freq + p->InitEsc + (ns > 3));
       }
     }
     
@@ -795,8 +804,8 @@ void Ppmd7_UpdateModel(CPpmd7 *p)
   
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static void Rescale(CPpmd7 *p)
+static void Ppmd7_Rescale(CPpmd7 *p)
 {
   unsigned i, adder, sumFreq, escFreq;
   CPpmd_State *stats = STATS(p->MinContext);
@@ -885,7 +894,7 @@ static void Rescale(CPpmd7 *p)
       *s = *stats;
       s->Freq = (Byte)freq; // (freq <= 260 / 4)
       p->FoundState = s;
-      InsertNode(p, stats, U2I(n0));
+      Ppmd7_InsertNode(p, stats, U2I(n0));
       return;
     }
     
@@ -899,13 +908,13 @@ static void Rescale(CPpmd7 *p)
       {
         if (p->FreeList[i1] != 0)
         {
-          void *ptr = RemoveNode(p, i1);
+          void *ptr = Ppmd7_RemoveNode(p, i1);
           p->MinContext->Union4.Stats = STATS_REF(ptr);
-          MyMem12Cpy(ptr, (const void *)stats, n1);
+          MEM_12_CPY(ptr, (const void *)stats, n1)
-          InsertNode(p, stats, i0);
+          Ppmd7_InsertNode(p, stats, i0);
         }
         else
-          SplitBlock(p, stats, i0, i1);
+          Ppmd7_SplitBlock(p, stats, i0, i1);
       }
     }
   }
@@ -933,10 +942,10 @@ CPpmd_See *Ppmd7_MakeEscFreq(CPpmd7 *p, unsigned numMasked, UInt32 *escFreq)
         p->HiBitsFlag;
     {
       // if (see->Summ) field is larger than 16-bit, we need only low 16 bits of Summ
-      unsigned summ = (UInt16)see->Summ; // & 0xFFFF
+      const unsigned summ = (UInt16)see->Summ; // & 0xFFFF
-      unsigned r = (summ >> see->Shift);
+      const unsigned r = (summ >> see->Shift);
       see->Summ = (UInt16)(summ - r);
-      *escFreq = r + (r == 0);
+      *escFreq = (UInt32)(r + (r == 0));
     }
   }
   else
@@ -948,9 +957,9 @@ CPpmd_See *Ppmd7_MakeEscFreq(CPpmd7 *p, unsigned numMasked, UInt32 *escFreq)
 }
 
 
-static void NextContext(CPpmd7 *p)
+static void Ppmd7_NextContext(CPpmd7 *p)
 {
-  CTX_PTR c = CTX(SUCCESSOR(p->FoundState));
+  PPMD7_CTX_PTR c = CTX(SUCCESSOR(p->FoundState));
   if (p->OrderFall == 0 && (const Byte *)c > p->Text)
     p->MaxContext = p->MinContext = c;
   else
@@ -967,12 +976,12 @@ void Ppmd7_Update1(CPpmd7 *p)
   s->Freq = (Byte)freq;
   if (freq > s[-1].Freq)
   {
-    SwapStates(s);
+    SWAP_STATES(s)
     p->FoundState = --s;
     if (freq > MAX_FREQ)
-      Rescale(p);
+      Ppmd7_Rescale(p);
   }
-  NextContext(p);
+  Ppmd7_NextContext(p);
 }
 
 
@@ -981,15 +990,15 @@ void Ppmd7_Update1_0(CPpmd7 *p)
   CPpmd_State *s = p->FoundState;
   CPpmd7_Context *mc = p->MinContext;
   unsigned freq = s->Freq;
-  unsigned summFreq = mc->Union2.SummFreq;
+  const unsigned summFreq = mc->Union2.SummFreq;
   p->PrevSuccess = (2 * freq > summFreq);
-  p->RunLength += (int)p->PrevSuccess;
+  p->RunLength += (Int32)p->PrevSuccess;
   mc->Union2.SummFreq = (UInt16)(summFreq + 4);
   freq += 4;
   s->Freq = (Byte)freq;
   if (freq > MAX_FREQ)
-    Rescale(p);
+    Ppmd7_Rescale(p);
-  NextContext(p);
+  Ppmd7_NextContext(p);
 }
 
 
@@ -1000,7 +1009,7 @@ void Ppmd7_UpdateBin(CPpmd7 *p)
   p->FoundState->Freq = (Byte)(freq + (freq < 128));
   p->PrevSuccess = 1;
   p->RunLength++;
-  NextContext(p);
+  Ppmd7_NextContext(p);
 }
 */
 
@@ -1013,7 +1022,7 @@ void Ppmd7_Update2(CPpmd7 *p)
   p->MinContext->Union2.SummFreq = (UInt16)(p->MinContext->Union2.SummFreq + 4);
   s->Freq = (Byte)freq;
   if (freq > MAX_FREQ)
-    Rescale(p);
+    Ppmd7_Rescale(p);
   Ppmd7_UpdateModel(p);
 }
 
@@ -1042,8 +1051,8 @@ Last UNIT of array at offset (Size - 12) is root order-0 CPpmd7_Context record.
 The code can free UNITs memory blocks that were allocated to store CPpmd_State vectors.
 The code doesn't free UNITs allocated for CPpmd7_Context records.
 
-The code calls RestartModel(), when there is no free memory for allocation.
+The code calls Ppmd7_RestartModel(), when there is no free memory for allocation.
-And RestartModel() changes the state to orignal start state, with full free block.
+And Ppmd7_RestartModel() changes the state to orignal start state, with full free block.
 
 
 The code allocates UNITs with the following order:
@@ -1051,14 +1060,14 @@ The code allocates UNITs with the following order:
 Allocation of 1 UNIT for Context record
   - from free space (HiUnit) down to (LoUnit)
   - from FreeList[0]
-  - AllocUnitsRare()
+  - Ppmd7_AllocUnitsRare()
 
-AllocUnits() for CPpmd_State vectors:
+Ppmd7_AllocUnits() for CPpmd_State vectors:
   - from FreeList[i]
   - from free space (LoUnit) up to (HiUnit)
-  - AllocUnitsRare()
+  - Ppmd7_AllocUnitsRare()
 
-AllocUnitsRare()
+Ppmd7_AllocUnitsRare()
   - if (GlueCount == 0)
        {  Glue lists, GlueCount = 255, allocate from FreeList[i]] }
   - loop for all higher sized FreeList[...] lists
@@ -1093,8 +1102,8 @@ The PPMd code tries to fulfill the condition:
 We have (Sum(Stats[].Freq) <= 256 * 124), because of (MAX_FREQ = 124)
 So (4 = 128 - 124) is average reserve for Escape_Freq for each symbol.
 If (CPpmd_State::Freq) is not aligned for 4, the reserve can be 5, 6 or 7.
-SummFreq and Escape_Freq can be changed in Rescale() and *Update*() functions.
+SummFreq and Escape_Freq can be changed in Ppmd7_Rescale() and *Update*() functions.
-Rescale() can remove symbols only from max-order contexts. So Escape_Freq can increase after multiple calls of Rescale() for
+Ppmd7_Rescale() can remove symbols only from max-order contexts. So Escape_Freq can increase after multiple calls of Ppmd7_Rescale() for
 max-order context.
 
 When the PPMd code still break (Total <= RC::Range) condition in range coder,
@@ -1102,3 +1111,21 @@ we have two ways to resolve that problem:
   1) we can report error, if we want to keep compatibility with original PPMd code that has no fix for such cases.
   2) we can reduce (Total) value to (RC::Range) by reducing (Escape_Freq) part of (Total) value.
 */
+
+#undef MAX_FREQ
+#undef UNIT_SIZE
+#undef U2B
+#undef U2I
+#undef I2U
+#undef I2U_UInt16
+#undef REF
+#undef STATS_REF
+#undef CTX
+#undef STATS
+#undef ONE_STATE
+#undef SUFFIX
+#undef NODE
+#undef EMPTY_NODE
+#undef MEM_12_CPY
+#undef SUCCESSOR
+#undef SWAP_STATES

C/Ppmd7.h

@@ -1,11 +1,11 @@
 /* Ppmd7.h -- Ppmd7 (PPMdH) compression codec
-2021-04-13 : Igor Pavlov : Public domain
+2023-04-02 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.H (2001): Dmitry Shkarin : Public domain */
  
 
-#ifndef __PPMD7_H
+#ifndef ZIP7_INC_PPMD7_H
-#define __PPMD7_H
+#define ZIP7_INC_PPMD7_H
 
 #include "Ppmd.h"
 
@@ -55,7 +55,7 @@ typedef struct
   UInt32 Range;
   UInt32 Code;
   UInt32 Low;
-  IByteIn *Stream;
+  IByteInPtr Stream;
 } CPpmd7_RangeDec;
 
 
@@ -66,7 +66,7 @@ typedef struct
   // Byte _dummy_[3];
   UInt64 Low;
   UInt64 CacheSize;
-  IByteOut *Stream;
+  IByteOutPtr Stream;
 } CPpmd7z_RangeEnc;
 
 

C/Ppmd7Dec.c

@@ -1,5 +1,5 @@
 /* Ppmd7Dec.c -- Ppmd7z (PPMdH with 7z Range Coder) Decoder
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.H (2001): Dmitry Shkarin : Public domain */
 
@@ -8,7 +8,7 @@ This code is based on:
 
 #include "Ppmd7.h"
 
-#define kTopValue (1 << 24)
+#define kTopValue ((UInt32)1 << 24)
 
 
 #define READ_BYTE(p) IByteIn_Read((p)->Stream)
@@ -37,9 +37,9 @@ BoolInt Ppmd7z_RangeDec_Init(CPpmd7_RangeDec *p)
 
 #define R (&p->rc.dec)
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
-// MY_NO_INLINE
+// Z7_NO_INLINE
-static void RangeDec_Decode(CPpmd7 *p, UInt32 start, UInt32 size)
+static void Ppmd7z_RD_Decode(CPpmd7 *p, UInt32 start, UInt32 size)
 {
 
   
@@ -48,18 +48,18 @@ static void RangeDec_Decode(CPpmd7 *p, UInt32 start, UInt32 size)
   RC_NORM_LOCAL(R)
 }
 
-#define RC_Decode(start, size) RangeDec_Decode(p, start, size);
+#define RC_Decode(start, size)  Ppmd7z_RD_Decode(p, start, size);
-#define RC_DecodeFinal(start, size) RC_Decode(start, size) RC_NORM_REMOTE(R)
+#define RC_DecodeFinal(start, size)  RC_Decode(start, size)  RC_NORM_REMOTE(R)
-#define RC_GetThreshold(total) (R->Code / (R->Range /= (total)))
+#define RC_GetThreshold(total)  (R->Code / (R->Range /= (total)))
 
 
 #define CTX(ref) ((CPpmd7_Context *)Ppmd7_GetContext(p, ref))
-typedef CPpmd7_Context * CTX_PTR;
+// typedef CPpmd7_Context * CTX_PTR;
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
 void Ppmd7_UpdateModel(CPpmd7 *p);
 
-#define MASK(sym) ((unsigned char *)charMask)[sym]
+#define MASK(sym)  ((Byte *)charMask)[sym]
-// MY_FORCE_INLINE
+// Z7_FORCE_INLINE
 // static
 int Ppmd7z_DecodeSymbol(CPpmd7 *p)
 {
@@ -70,7 +70,7 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
     CPpmd_State *s = Ppmd7_GetStats(p, p->MinContext);
     unsigned i;
     UInt32 count, hiCnt;
-    UInt32 summFreq = p->MinContext->Union2.SummFreq;
+    const UInt32 summFreq = p->MinContext->Union2.SummFreq;
 
     
     
@@ -81,7 +81,7 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
     if ((Int32)(count -= s->Freq) < 0)
     {
       Byte sym;
-      RC_DecodeFinal(0, s->Freq);
+      RC_DecodeFinal(0, s->Freq)
       p->FoundState = s;
       sym = s->Symbol;
       Ppmd7_Update1_0(p);
@@ -96,7 +96,7 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
       if ((Int32)(count -= (++s)->Freq) < 0)
       {
         Byte sym;
-        RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq);
+        RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq)
         p->FoundState = s;
         sym = s->Symbol;
         Ppmd7_Update1(p);
@@ -109,10 +109,10 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
       return PPMD7_SYM_ERROR;
     
     hiCnt -= count;
-    RC_Decode(hiCnt, summFreq - hiCnt);
+    RC_Decode(hiCnt, summFreq - hiCnt)
 
     p->HiBitsFlag = PPMD7_HiBitsFlag_3(p->FoundState->Symbol);
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     // i = p->MinContext->NumStats - 1;
     // do { MASK((--s)->Symbol) = 0; } while (--i);
     {
@@ -120,8 +120,8 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -152,7 +152,7 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
       // Ppmd7_UpdateBin(p);
       {
         unsigned freq = s->Freq;
-        CTX_PTR c = CTX(SUCCESSOR(s));
+        CPpmd7_Context *c = CTX(SUCCESSOR(s));
         sym = s->Symbol;
         p->FoundState = s;
         p->PrevSuccess = 1;
@@ -176,7 +176,7 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
     R->Range -= size0;
     RC_NORM_LOCAL(R)
     
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     MASK(Ppmd7Context_OneState(p->MinContext)->Symbol) = 0;
     p->PrevSuccess = 0;
   }
@@ -209,17 +209,17 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
       unsigned num2 = num / 2;
       
       num &= 1;
-      hiCnt = (s->Freq & (unsigned)(MASK(s->Symbol))) & (0 - (UInt32)num);
+      hiCnt = (s->Freq & (UInt32)(MASK(s->Symbol))) & (0 - (UInt32)num);
       s += num;
       p->MinContext = mc;
 
       do
       {
-        unsigned sym0 = s[0].Symbol;
+        const unsigned sym0 = s[0].Symbol;
-        unsigned sym1 = s[1].Symbol;
+        const unsigned sym1 = s[1].Symbol;
         s += 2;
-        hiCnt += (s[-2].Freq & (unsigned)(MASK(sym0)));
+        hiCnt += (s[-2].Freq & (UInt32)(MASK(sym0)));
-        hiCnt += (s[-1].Freq & (unsigned)(MASK(sym1)));
+        hiCnt += (s[-1].Freq & (UInt32)(MASK(sym1)));
       }
       while (--num2);
     }
@@ -238,20 +238,20 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
 
       s = Ppmd7_GetStats(p, p->MinContext);
       hiCnt = count;
-      // count -= s->Freq & (unsigned)(MASK(s->Symbol));
+      // count -= s->Freq & (UInt32)(MASK(s->Symbol));
       // if ((Int32)count >= 0)
       {
         for (;;)
         {
-          count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
+          count -= s->Freq & (UInt32)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
-          // count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
+          // count -= s->Freq & (UInt32)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
-        };
+        }
       }
       s--;
-      RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq);
+      RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq)
 
       // new (see->Summ) value can overflow over 16-bits in some rare cases
-      Ppmd_See_Update(see);
+      Ppmd_See_UPDATE(see)
       p->FoundState = s;
       sym = s->Symbol;
       Ppmd7_Update2(p);
@@ -261,7 +261,7 @@ int Ppmd7z_DecodeSymbol(CPpmd7 *p)
     if (count >= freqSum)
       return PPMD7_SYM_ERROR;
     
-    RC_Decode(hiCnt, freqSum - hiCnt);
+    RC_Decode(hiCnt, freqSum - hiCnt)
 
     // We increase (see->Summ) for sum of Freqs of all non_Masked symbols.
     // new (see->Summ) value can overflow over 16-bits in some rare cases
@@ -295,3 +295,18 @@ Byte *Ppmd7z_DecodeSymbols(CPpmd7 *p, Byte *buf, const Byte *lim)
   return buf;
 }
 */
+
+#undef kTopValue
+#undef READ_BYTE
+#undef RC_NORM_BASE
+#undef RC_NORM_1
+#undef RC_NORM
+#undef RC_NORM_LOCAL
+#undef RC_NORM_REMOTE
+#undef R
+#undef RC_Decode
+#undef RC_DecodeFinal
+#undef RC_GetThreshold
+#undef CTX
+#undef SUCCESSOR
+#undef MASK

C/Ppmd7Enc.c

@@ -1,5 +1,5 @@
 /* Ppmd7Enc.c -- Ppmd7z (PPMdH with 7z Range Coder) Encoder
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.H (2001): Dmitry Shkarin : Public domain */
 
@@ -8,7 +8,7 @@ This code is based on:
 
 #include "Ppmd7.h"
 
-#define kTopValue (1 << 24)
+#define kTopValue ((UInt32)1 << 24)
 
 #define R (&p->rc.enc)
 
@@ -20,8 +20,8 @@ void Ppmd7z_Init_RangeEnc(CPpmd7 *p)
   R->CacheSize = 1;
 }
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static void RangeEnc_ShiftLow(CPpmd7 *p)
+static void Ppmd7z_RangeEnc_ShiftLow(CPpmd7 *p)
 {
   if ((UInt32)R->Low < (UInt32)0xFF000000 || (unsigned)(R->Low >> 32) != 0)
   {
@@ -38,53 +38,53 @@ static void RangeEnc_ShiftLow(CPpmd7 *p)
   R->Low = (UInt32)((UInt32)R->Low << 8);
 }
 
-#define RC_NORM_BASE(p) if (R->Range < kTopValue) { R->Range <<= 8; RangeEnc_ShiftLow(p);
+#define RC_NORM_BASE(p) if (R->Range < kTopValue) { R->Range <<= 8;  Ppmd7z_RangeEnc_ShiftLow(p);
-#define RC_NORM_1(p) RC_NORM_BASE(p) }
+#define RC_NORM_1(p)    RC_NORM_BASE(p) }
-#define RC_NORM(p) RC_NORM_BASE(p) RC_NORM_BASE(p) }}
+#define RC_NORM(p)      RC_NORM_BASE(p)  RC_NORM_BASE(p) }}
 
 // we must use only one type of Normalization from two: LOCAL or REMOTE
 #define RC_NORM_LOCAL(p)    // RC_NORM(p)
 #define RC_NORM_REMOTE(p)   RC_NORM(p)
 
 /*
-#define RangeEnc_Encode(p, start, _size_) \
+#define Ppmd7z_RangeEnc_Encode(p, start, _size_) \
   { UInt32 size = _size_; \
     R->Low += start * R->Range; \
     R->Range *= size; \
     RC_NORM_LOCAL(p); }
 */
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
-// MY_NO_INLINE
+// Z7_NO_INLINE
-static void RangeEnc_Encode(CPpmd7 *p, UInt32 start, UInt32 size)
+static void Ppmd7z_RangeEnc_Encode(CPpmd7 *p, UInt32 start, UInt32 size)
 {
   R->Low += start * R->Range;
   R->Range *= size;
-  RC_NORM_LOCAL(p);
+  RC_NORM_LOCAL(p)
 }
 
 void Ppmd7z_Flush_RangeEnc(CPpmd7 *p)
 {
   unsigned i;
   for (i = 0; i < 5; i++)
-    RangeEnc_ShiftLow(p);
+    Ppmd7z_RangeEnc_ShiftLow(p);
 }
 
 
 
-#define RC_Encode(start, size) RangeEnc_Encode(p, start, size);
+#define RC_Encode(start, size)  Ppmd7z_RangeEnc_Encode(p, start, size);
-#define RC_EncodeFinal(start, size) RC_Encode(start, size); RC_NORM_REMOTE(p);
+#define RC_EncodeFinal(start, size)  RC_Encode(start, size) RC_NORM_REMOTE(p)
 
 #define CTX(ref) ((CPpmd7_Context *)Ppmd7_GetContext(p, ref))
 #define SUFFIX(ctx) CTX((ctx)->Suffix)
-typedef CPpmd7_Context * CTX_PTR;
+// typedef CPpmd7_Context * CTX_PTR;
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
 
 void Ppmd7_UpdateModel(CPpmd7 *p);
 
-#define MASK(sym) ((unsigned char *)charMask)[sym]
+#define MASK(sym)  ((Byte *)charMask)[sym]
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
 static
 void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
 {
@@ -104,7 +104,7 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
     if (s->Symbol == symbol)
     {
       // R->Range /= p->MinContext->Union2.SummFreq;
-      RC_EncodeFinal(0, s->Freq);
+      RC_EncodeFinal(0, s->Freq)
       p->FoundState = s;
       Ppmd7_Update1_0(p);
       return;
@@ -117,7 +117,7 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
       if ((++s)->Symbol == symbol)
       {
         // R->Range /= p->MinContext->Union2.SummFreq;
-        RC_EncodeFinal(sum, s->Freq);
+        RC_EncodeFinal(sum, s->Freq)
         p->FoundState = s;
         Ppmd7_Update1(p);
         return;
@@ -127,10 +127,10 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
     while (--i);
 
     // R->Range /= p->MinContext->Union2.SummFreq;
-    RC_Encode(sum, p->MinContext->Union2.SummFreq - sum);
+    RC_Encode(sum, p->MinContext->Union2.SummFreq - sum)
     
     p->HiBitsFlag = PPMD7_HiBitsFlag_3(p->FoundState->Symbol);
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     // MASK(s->Symbol) = 0;
     // i = p->MinContext->NumStats - 1;
     // do { MASK((--s)->Symbol) = 0; } while (--i);
@@ -139,8 +139,8 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -153,20 +153,20 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
     UInt16 *prob = Ppmd7_GetBinSumm(p);
     CPpmd_State *s = Ppmd7Context_OneState(p->MinContext);
     UInt32 pr = *prob;
-    UInt32 bound = (R->Range >> 14) * pr;
+    const UInt32 bound = (R->Range >> 14) * pr;
     pr = PPMD_UPDATE_PROB_1(pr);
     if (s->Symbol == symbol)
     {
       *prob = (UInt16)(pr + (1 << PPMD_INT_BITS));
       // RangeEnc_EncodeBit_0(p, bound);
       R->Range = bound;
-      RC_NORM_1(p);
+      RC_NORM_1(p)
       
       // p->FoundState = s;
       // Ppmd7_UpdateBin(p);
       {
-        unsigned freq = s->Freq;
+        const unsigned freq = s->Freq;
-        CTX_PTR c = CTX(SUCCESSOR(s));
+        CPpmd7_Context *c = CTX(SUCCESSOR(s));
         p->FoundState = s;
         p->PrevSuccess = 1;
         p->RunLength++;
@@ -187,7 +187,7 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
     R->Range -= bound;
     RC_NORM_LOCAL(p)
     
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     MASK(s->Symbol) = 0;
     p->PrevSuccess = 0;
   }
@@ -248,14 +248,14 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
 
     do
     {
-      unsigned cur = s->Symbol;
+      const unsigned cur = s->Symbol;
       if ((int)cur == symbol)
       {
-        UInt32 low = sum;
+        const UInt32 low = sum;
-        UInt32 freq = s->Freq;
+        const UInt32 freq = s->Freq;
         unsigned num2;
 
-        Ppmd_See_Update(see);
+        Ppmd_See_UPDATE(see)
         p->FoundState = s;
         sum += escFreq;
 
@@ -265,21 +265,20 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
         if (num2 != 0)
         {
           s += i;
-          for (;;)
+          do
           {
-            unsigned sym0 = s[0].Symbol;
+            const unsigned sym0 = s[0].Symbol;
-            unsigned sym1 = s[1].Symbol;
+            const unsigned sym1 = s[1].Symbol;
             s += 2;
             sum += (s[-2].Freq & (unsigned)(MASK(sym0)));
             sum += (s[-1].Freq & (unsigned)(MASK(sym1)));
-            if (--num2 == 0)
-              break;
           }
+          while (--num2);
         }
 
         
         R->Range /= sum;
-        RC_EncodeFinal(low, freq);
+        RC_EncodeFinal(low, freq)
         Ppmd7_Update2(p);
         return;
       }
@@ -289,21 +288,21 @@ void Ppmd7z_EncodeSymbol(CPpmd7 *p, int symbol)
     while (--i);
     
     {
-      UInt32 total = sum + escFreq;
+      const UInt32 total = sum + escFreq;
       see->Summ = (UInt16)(see->Summ + total);
 
       R->Range /= total;
-      RC_Encode(sum, escFreq);
+      RC_Encode(sum, escFreq)
     }
 
     {
-      CPpmd_State *s2 = Ppmd7_GetStats(p, p->MinContext);
+      const CPpmd_State *s2 = Ppmd7_GetStats(p, p->MinContext);
       s--;
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -321,3 +320,18 @@ void Ppmd7z_EncodeSymbols(CPpmd7 *p, const Byte *buf, const Byte *lim)
     Ppmd7z_EncodeSymbol(p, *buf);
   }
 }
+
+#undef kTopValue
+#undef WRITE_BYTE
+#undef RC_NORM_BASE
+#undef RC_NORM_1
+#undef RC_NORM
+#undef RC_NORM_LOCAL
+#undef RC_NORM_REMOTE
+#undef R
+#undef RC_Encode
+#undef RC_EncodeFinal
+#undef SUFFIX
+#undef CTX
+#undef SUCCESSOR
+#undef MASK

C/Ppmd7aDec.c

@@ -1,5 +1,5 @@
 /* Ppmd7aDec.c -- PPMd7a (PPMdH) Decoder
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.H (2001): Dmitry Shkarin : Public domain
   Carryless rangecoder (1999): Dmitry Subbotin : Public domain */
@@ -8,8 +8,8 @@ This code is based on:
 
 #include "Ppmd7.h"
 
-#define kTop (1 << 24)
+#define kTop ((UInt32)1 << 24)
-#define kBot (1 << 15)
+#define kBot ((UInt32)1 << 15)
 
 #define READ_BYTE(p) IByteIn_Read((p)->Stream)
 
@@ -37,9 +37,9 @@ BoolInt Ppmd7a_RangeDec_Init(CPpmd7_RangeDec *p)
 
 #define R (&p->rc.dec)
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
-// MY_NO_INLINE
+// Z7_NO_INLINE
-static void RangeDec_Decode(CPpmd7 *p, UInt32 start, UInt32 size)
+static void Ppmd7a_RD_Decode(CPpmd7 *p, UInt32 start, UInt32 size)
 {
   start *= R->Range;
   R->Low += start;
@@ -48,9 +48,9 @@ static void RangeDec_Decode(CPpmd7 *p, UInt32 start, UInt32 size)
   RC_NORM_LOCAL(R)
 }
 
-#define RC_Decode(start, size) RangeDec_Decode(p, start, size);
+#define RC_Decode(start, size)  Ppmd7a_RD_Decode(p, start, size);
-#define RC_DecodeFinal(start, size) RC_Decode(start, size) RC_NORM_REMOTE(R)
+#define RC_DecodeFinal(start, size)  RC_Decode(start, size)  RC_NORM_REMOTE(R)
-#define RC_GetThreshold(total) (R->Code / (R->Range /= (total)))
+#define RC_GetThreshold(total)  (R->Code / (R->Range /= (total)))
 
 
 #define CTX(ref) ((CPpmd7_Context *)Ppmd7_GetContext(p, ref))
@@ -58,7 +58,7 @@ typedef CPpmd7_Context * CTX_PTR;
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
 void Ppmd7_UpdateModel(CPpmd7 *p);
 
-#define MASK(sym) ((unsigned char *)charMask)[sym]
+#define MASK(sym)  ((Byte *)charMask)[sym]
 
 
 int Ppmd7a_DecodeSymbol(CPpmd7 *p)
@@ -70,7 +70,7 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
     CPpmd_State *s = Ppmd7_GetStats(p, p->MinContext);
     unsigned i;
     UInt32 count, hiCnt;
-    UInt32 summFreq = p->MinContext->Union2.SummFreq;
+    const UInt32 summFreq = p->MinContext->Union2.SummFreq;
 
     if (summFreq > R->Range)
       return PPMD7_SYM_ERROR;
@@ -81,7 +81,7 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
     if ((Int32)(count -= s->Freq) < 0)
     {
       Byte sym;
-      RC_DecodeFinal(0, s->Freq);
+      RC_DecodeFinal(0, s->Freq)
       p->FoundState = s;
       sym = s->Symbol;
       Ppmd7_Update1_0(p);
@@ -96,7 +96,7 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
       if ((Int32)(count -= (++s)->Freq) < 0)
       {
         Byte sym;
-        RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq);
+        RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq)
         p->FoundState = s;
         sym = s->Symbol;
         Ppmd7_Update1(p);
@@ -109,10 +109,10 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
       return PPMD7_SYM_ERROR;
     
     hiCnt -= count;
-    RC_Decode(hiCnt, summFreq - hiCnt);
+    RC_Decode(hiCnt, summFreq - hiCnt)
 
     p->HiBitsFlag = PPMD7_HiBitsFlag_3(p->FoundState->Symbol);
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     // i = p->MinContext->NumStats - 1;
     // do { MASK((--s)->Symbol) = 0; } while (--i);
     {
@@ -120,8 +120,8 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -176,7 +176,7 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
     R->Range = (R->Range & ~((UInt32)PPMD_BIN_SCALE - 1)) - size0;
     RC_NORM_LOCAL(R)
     
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     MASK(Ppmd7Context_OneState(p->MinContext)->Symbol) = 0;
     p->PrevSuccess = 0;
   }
@@ -209,17 +209,17 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
       unsigned num2 = num / 2;
       
       num &= 1;
-      hiCnt = (s->Freq & (unsigned)(MASK(s->Symbol))) & (0 - (UInt32)num);
+      hiCnt = (s->Freq & (UInt32)(MASK(s->Symbol))) & (0 - (UInt32)num);
       s += num;
       p->MinContext = mc;
 
       do
       {
-        unsigned sym0 = s[0].Symbol;
+        const unsigned sym0 = s[0].Symbol;
-        unsigned sym1 = s[1].Symbol;
+        const unsigned sym1 = s[1].Symbol;
         s += 2;
-        hiCnt += (s[-2].Freq & (unsigned)(MASK(sym0)));
+        hiCnt += (s[-2].Freq & (UInt32)(MASK(sym0)));
-        hiCnt += (s[-1].Freq & (unsigned)(MASK(sym1)));
+        hiCnt += (s[-1].Freq & (UInt32)(MASK(sym1)));
       }
       while (--num2);
     }
@@ -238,20 +238,20 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
 
       s = Ppmd7_GetStats(p, p->MinContext);
       hiCnt = count;
-      // count -= s->Freq & (unsigned)(MASK(s->Symbol));
+      // count -= s->Freq & (UInt32)(MASK(s->Symbol));
       // if ((Int32)count >= 0)
       {
         for (;;)
         {
-          count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
+          count -= s->Freq & (UInt32)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
-          // count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
+          // count -= s->Freq & (UInt32)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
-        };
+        }
       }
       s--;
-      RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq);
+      RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq)
 
       // new (see->Summ) value can overflow over 16-bits in some rare cases
-      Ppmd_See_Update(see);
+      Ppmd_See_UPDATE(see)
       p->FoundState = s;
       sym = s->Symbol;
       Ppmd7_Update2(p);
@@ -261,7 +261,7 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
     if (count >= freqSum)
       return PPMD7_SYM_ERROR;
     
-    RC_Decode(hiCnt, freqSum - hiCnt);
+    RC_Decode(hiCnt, freqSum - hiCnt)
 
     // We increase (see->Summ) for sum of Freqs of all non_Masked symbols.
     // new (see->Summ) value can overflow over 16-bits in some rare cases
@@ -277,3 +277,19 @@ int Ppmd7a_DecodeSymbol(CPpmd7 *p)
     while (s != s2);
   }
 }
+
+#undef kTop
+#undef kBot
+#undef READ_BYTE
+#undef RC_NORM_BASE
+#undef RC_NORM_1
+#undef RC_NORM
+#undef RC_NORM_LOCAL
+#undef RC_NORM_REMOTE
+#undef R
+#undef RC_Decode
+#undef RC_DecodeFinal
+#undef RC_GetThreshold
+#undef CTX
+#undef SUCCESSOR
+#undef MASK

C/Ppmd8.c

@@ -1,5 +1,5 @@
 /* Ppmd8.c -- PPMdI codec
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on PPMd var.I (2002): Dmitry Shkarin : Public domain */
 
 #include "Precomp.h"
@@ -14,7 +14,7 @@ This code is based on PPMd var.I (2002): Dmitry Shkarin : Public domain */
 MY_ALIGN(16)
 static const Byte PPMD8_kExpEscape[16] = { 25, 14, 9, 7, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2 };
 MY_ALIGN(16)
-static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051};
+static const UInt16 PPMD8_kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051};
 
 #define MAX_FREQ 124
 #define UNIT_SIZE 12
@@ -33,7 +33,7 @@ static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x
 #define ONE_STATE(ctx) Ppmd8Context_OneState(ctx)
 #define SUFFIX(ctx) CTX((ctx)->Suffix)
 
-typedef CPpmd8_Context * CTX_PTR;
+typedef CPpmd8_Context * PPMD8_CTX_PTR;
 
 struct CPpmd8_Node_;
 
@@ -114,7 +114,7 @@ BoolInt Ppmd8_Alloc(CPpmd8 *p, UInt32 size, ISzAllocPtr alloc)
 #define EMPTY_NODE 0xFFFFFFFF
 
 
-static void InsertNode(CPpmd8 *p, void *node, unsigned indx)
+static void Ppmd8_InsertNode(CPpmd8 *p, void *node, unsigned indx)
 {
   ((CPpmd8_Node *)node)->Stamp = EMPTY_NODE;
   ((CPpmd8_Node *)node)->Next = (CPpmd8_Node_Ref)p->FreeList[indx];
@@ -124,7 +124,7 @@ static void InsertNode(CPpmd8 *p, void *node, unsigned indx)
 }
 
 
-static void *RemoveNode(CPpmd8 *p, unsigned indx)
+static void *Ppmd8_RemoveNode(CPpmd8 *p, unsigned indx)
 {
   CPpmd8_Node *node = NODE((CPpmd8_Node_Ref)p->FreeList[indx]);
   p->FreeList[indx] = node->Next;
@@ -134,16 +134,16 @@ static void *RemoveNode(CPpmd8 *p, unsigned indx)
 }
 
 
-static void SplitBlock(CPpmd8 *p, void *ptr, unsigned oldIndx, unsigned newIndx)
+static void Ppmd8_SplitBlock(CPpmd8 *p, void *ptr, unsigned oldIndx, unsigned newIndx)
 {
   unsigned i, nu = I2U(oldIndx) - I2U(newIndx);
   ptr = (Byte *)ptr + U2B(I2U(newIndx));
   if (I2U(i = U2I(nu)) != nu)
   {
     unsigned k = I2U(--i);
-    InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1);
+    Ppmd8_InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1);
   }
-  InsertNode(p, ptr, i);
+  Ppmd8_InsertNode(p, ptr, i);
 }
 
 
@@ -159,7 +159,7 @@ static void SplitBlock(CPpmd8 *p, void *ptr, unsigned oldIndx, unsigned newIndx)
 
 
 
-static void GlueFreeBlocks(CPpmd8 *p)
+static void Ppmd8_GlueFreeBlocks(CPpmd8 *p)
 {
   /*
   we use first UInt32 field of 12-bytes UNITs as record type stamp
@@ -239,27 +239,27 @@ static void GlueFreeBlocks(CPpmd8 *p)
     if (nu == 0)
       continue;
     for (; nu > 128; nu -= 128, node += 128)
-      InsertNode(p, node, PPMD_NUM_INDEXES - 1);
+      Ppmd8_InsertNode(p, node, PPMD_NUM_INDEXES - 1);
     if (I2U(i = U2I(nu)) != nu)
     {
       unsigned k = I2U(--i);
-      InsertNode(p, node + k, (unsigned)nu - k - 1);
+      Ppmd8_InsertNode(p, node + k, (unsigned)nu - k - 1);
     }
-    InsertNode(p, node, i);
+    Ppmd8_InsertNode(p, node, i);
   }
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static void *AllocUnitsRare(CPpmd8 *p, unsigned indx)
+static void *Ppmd8_AllocUnitsRare(CPpmd8 *p, unsigned indx)
 {
   unsigned i;
   
   if (p->GlueCount == 0)
   {
-    GlueFreeBlocks(p);
+    Ppmd8_GlueFreeBlocks(p);
     if (p->FreeList[indx] != 0)
-      return RemoveNode(p, indx);
+      return Ppmd8_RemoveNode(p, indx);
   }
   
   i = indx;
@@ -277,17 +277,17 @@ static void *AllocUnitsRare(CPpmd8 *p, unsigned indx)
   while (p->FreeList[i] == 0);
   
   {
-    void *block = RemoveNode(p, i);
+    void *block = Ppmd8_RemoveNode(p, i);
-    SplitBlock(p, block, i, indx);
+    Ppmd8_SplitBlock(p, block, i, indx);
     return block;
   }
 }
 
 
-static void *AllocUnits(CPpmd8 *p, unsigned indx)
+static void *Ppmd8_AllocUnits(CPpmd8 *p, unsigned indx)
 {
   if (p->FreeList[indx] != 0)
-    return RemoveNode(p, indx);
+    return Ppmd8_RemoveNode(p, indx);
   {
     UInt32 numBytes = U2B(I2U(indx));
     Byte *lo = p->LoUnit;
@@ -297,13 +297,22 @@ static void *AllocUnits(CPpmd8 *p, unsigned indx)
       return lo;
     }
   }
-  return AllocUnitsRare(p, indx);
+  return Ppmd8_AllocUnitsRare(p, indx);
 }
 
 
-#define MyMem12Cpy(dest, src, num) \
+#define MEM_12_CPY(dest, src, num) \
-  { UInt32 *d = (UInt32 *)dest; const UInt32 *z = (const UInt32 *)src; UInt32 n = num; \
+  { UInt32 *d = (UInt32 *)(dest); \
-    do { d[0] = z[0]; d[1] = z[1]; d[2] = z[2]; z += 3; d += 3; } while (--n); }
+    const UInt32 *z = (const UInt32 *)(src); \
+    unsigned n = (num); \
+    do { \
+      d[0] = z[0]; \
+      d[1] = z[1]; \
+      d[2] = z[2]; \
+      z += 3; \
+      d += 3; \
+    } while (--n); \
+  }
 
 
 
@@ -315,26 +324,26 @@ static void *ShrinkUnits(CPpmd8 *p, void *oldPtr, unsigned oldNU, unsigned newNU
     return oldPtr;
   if (p->FreeList[i1] != 0)
   {
-    void *ptr = RemoveNode(p, i1);
+    void *ptr = Ppmd8_RemoveNode(p, i1);
-    MyMem12Cpy(ptr, oldPtr, newNU);
+    MEM_12_CPY(ptr, oldPtr, newNU)
-    InsertNode(p, oldPtr, i0);
+    Ppmd8_InsertNode(p, oldPtr, i0);
     return ptr;
   }
-  SplitBlock(p, oldPtr, i0, i1);
+  Ppmd8_SplitBlock(p, oldPtr, i0, i1);
   return oldPtr;
 }
 
 
 static void FreeUnits(CPpmd8 *p, void *ptr, unsigned nu)
 {
-  InsertNode(p, ptr, U2I(nu));
+  Ppmd8_InsertNode(p, ptr, U2I(nu));
 }
 
 
 static void SpecialFreeUnit(CPpmd8 *p, void *ptr)
 {
   if ((Byte *)ptr != p->UnitsStart)
-    InsertNode(p, ptr, 0);
+    Ppmd8_InsertNode(p, ptr, 0);
   else
   {
     #ifdef PPMD8_FREEZE_SUPPORT
@@ -352,10 +361,10 @@ static void *MoveUnitsUp(CPpmd8 *p, void *oldPtr, unsigned nu)
   void *ptr;
   if ((Byte *)oldPtr > p->UnitsStart + (1 << 14) || REF(oldPtr) > p->FreeList[indx])
     return oldPtr;
-  ptr = RemoveNode(p, indx);
+  ptr = Ppmd8_RemoveNode(p, indx);
-  MyMem12Cpy(ptr, oldPtr, nu);
+  MEM_12_CPY(ptr, oldPtr, nu)
   if ((Byte *)oldPtr != p->UnitsStart)
-    InsertNode(p, oldPtr, indx);
+    Ppmd8_InsertNode(p, oldPtr, indx);
   else
     p->UnitsStart += U2B(I2U(indx));
   return ptr;
@@ -411,22 +420,22 @@ static void ExpandTextArea(CPpmd8 *p)
 
 
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
-static void SetSuccessor(CPpmd_State *p, CPpmd_Void_Ref v)
+static void Ppmd8State_SetSuccessor(CPpmd_State *p, CPpmd_Void_Ref v)
 {
-  Ppmd_SET_SUCCESSOR(p, v);
+  Ppmd_SET_SUCCESSOR(p, v)
 }
 
 #define RESET_TEXT(offs) { p->Text = p->Base + p->AlignOffset + (offs); }
 
-MY_NO_INLINE
+Z7_NO_INLINE
 static
-void RestartModel(CPpmd8 *p)
+void Ppmd8_RestartModel(CPpmd8 *p)
 {
   unsigned i, k, m;
 
   memset(p->FreeList, 0, sizeof(p->FreeList));
   memset(p->Stamps, 0, sizeof(p->Stamps));
-  RESET_TEXT(0);
+  RESET_TEXT(0)
   p->HiUnit = p->Text + p->Size;
   p->LoUnit = p->UnitsStart = p->HiUnit - p->Size / 8 / UNIT_SIZE * 7 * UNIT_SIZE;
   p->GlueCount = 0;
@@ -436,8 +445,8 @@ void RestartModel(CPpmd8 *p)
   p->PrevSuccess = 0;
 
   {
-    CPpmd8_Context *mc = (CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */
+    CPpmd8_Context *mc = (PPMD8_CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */
-    CPpmd_State *s = (CPpmd_State *)p->LoUnit; /* AllocUnits(p, PPMD_NUM_INDEXES - 1); */
+    CPpmd_State *s = (CPpmd_State *)p->LoUnit; /* Ppmd8_AllocUnits(p, PPMD_NUM_INDEXES - 1); */
     
     p->LoUnit += U2B(256 / 2);
     p->MaxContext = p->MinContext = mc;
@@ -452,7 +461,7 @@ void RestartModel(CPpmd8 *p)
     {
       s->Symbol = (Byte)i;
       s->Freq = 1;
-      SetSuccessor(s, 0);
+      Ppmd8State_SetSuccessor(s, 0);
     }
   }
 
@@ -475,7 +484,7 @@ void RestartModel(CPpmd8 *p)
     {
       unsigned r;
       UInt16 *dest = p->BinSumm[m] + k;
-      UInt16 val = (UInt16)(PPMD_BIN_SCALE - kInitBinEsc[k] / (i + 1));
+      const UInt16 val = (UInt16)(PPMD_BIN_SCALE - PPMD8_kInitBinEsc[k] / (i + 1));
       for (r = 0; r < 64; r += 8)
         dest[r] = val;
     }
@@ -507,7 +516,7 @@ void Ppmd8_Init(CPpmd8 *p, unsigned maxOrder, unsigned restoreMethod)
 {
   p->MaxOrder = maxOrder;
   p->RestoreMethod = restoreMethod;
-  RestartModel(p);
+  Ppmd8_RestartModel(p);
 }
 
 
@@ -531,7 +540,7 @@ Refresh() is called when we remove some symbols (successors) in context.
 It increases Escape_Freq for sum of all removed symbols.
 */
 
-static void Refresh(CPpmd8 *p, CTX_PTR ctx, unsigned oldNU, unsigned scale)
+static void Refresh(CPpmd8 *p, PPMD8_CTX_PTR ctx, unsigned oldNU, unsigned scale)
 {
   unsigned i = ctx->NumStats, escFreq, sumFreq, flags;
   CPpmd_State *s = (CPpmd_State *)ShrinkUnits(p, STATS(ctx), oldNU, (i + 2) >> 1);
@@ -581,7 +590,7 @@ static void Refresh(CPpmd8 *p, CTX_PTR ctx, unsigned oldNU, unsigned scale)
 }
 
 
-static void SwapStates(CPpmd_State *t1, CPpmd_State *t2)
+static void SWAP_STATES(CPpmd_State *t1, CPpmd_State *t2)
 {
   CPpmd_State tmp = *t1;
   *t1 = *t2;
@@ -597,7 +606,7 @@ CutOff() reduces contexts:
   if the (Union4.Stats) is close to (UnitsStart), it moves it up.
 */
 
-static CPpmd_Void_Ref CutOff(CPpmd8 *p, CTX_PTR ctx, unsigned order)
+static CPpmd_Void_Ref CutOff(CPpmd8 *p, PPMD8_CTX_PTR ctx, unsigned order)
 {
   int ns = ctx->NumStats;
   unsigned nu;
@@ -613,7 +622,7 @@ static CPpmd_Void_Ref CutOff(CPpmd8 *p, CTX_PTR ctx, unsigned order)
         successor = CutOff(p, CTX(successor), order + 1);
       else
         successor = 0;
-      SetSuccessor(s, successor);
+      Ppmd8State_SetSuccessor(s, successor);
       if (successor || order <= 9) /* O_BOUND */
         return REF(ctx);
     }
@@ -630,11 +639,11 @@ static CPpmd_Void_Ref CutOff(CPpmd8 *p, CTX_PTR ctx, unsigned order)
     if ((UInt32)((Byte *)stats - p->UnitsStart) <= (1 << 14)
         && (CPpmd_Void_Ref)ctx->Union4.Stats <= p->FreeList[indx])
     {
-      void *ptr = RemoveNode(p, indx);
+      void *ptr = Ppmd8_RemoveNode(p, indx);
       ctx->Union4.Stats = STATS_REF(ptr);
-      MyMem12Cpy(ptr, (const void *)stats, nu);
+      MEM_12_CPY(ptr, (const void *)stats, nu)
       if ((Byte *)stats != p->UnitsStart)
-        InsertNode(p, stats, indx);
+        Ppmd8_InsertNode(p, stats, indx);
       else
         p->UnitsStart += U2B(I2U(indx));
       stats = ptr;
@@ -656,16 +665,16 @@ static CPpmd_Void_Ref CutOff(CPpmd8 *p, CTX_PTR ctx, unsigned order)
         }
         else
         {
-          SwapStates(s, s2);
+          SWAP_STATES(s, s2);
-          SetSuccessor(s2, 0);
+          Ppmd8State_SetSuccessor(s2, 0);
         }
       }
       else
       {
         if (order < p->MaxOrder)
-          SetSuccessor(s, CutOff(p, CTX(successor), order + 1));
+          Ppmd8State_SetSuccessor(s, CutOff(p, CTX(successor), order + 1));
         else
-          SetSuccessor(s, 0);
+          Ppmd8State_SetSuccessor(s, 0);
       }
     }
     while (--s >= stats);
@@ -711,7 +720,7 @@ RemoveBinContexts()
   removes Bin Context without Successor, if suffix of that context is also binary.
 */
 
-static CPpmd_Void_Ref RemoveBinContexts(CPpmd8 *p, CTX_PTR ctx, unsigned order)
+static CPpmd_Void_Ref RemoveBinContexts(CPpmd8 *p, PPMD8_CTX_PTR ctx, unsigned order)
 {
   if (!ctx->NumStats)
   {
@@ -721,7 +730,7 @@ static CPpmd_Void_Ref RemoveBinContexts(CPpmd8 *p, CTX_PTR ctx, unsigned order)
       successor = RemoveBinContexts(p, CTX(successor), order + 1);
     else
       successor = 0;
-    SetSuccessor(s, successor);
+    Ppmd8State_SetSuccessor(s, successor);
     /* Suffix context can be removed already, since different (high-order)
        Successors may refer to same context. So we check Flags == 0xFF (Stamp == EMPTY_NODE) */
     if (!successor && (!SUFFIX(ctx)->NumStats || SUFFIX(ctx)->Flags == 0xFF))
@@ -737,9 +746,9 @@ static CPpmd_Void_Ref RemoveBinContexts(CPpmd8 *p, CTX_PTR ctx, unsigned order)
     {
       CPpmd_Void_Ref successor = SUCCESSOR(s);
       if ((Byte *)Ppmd8_GetPtr(p, successor) >= p->UnitsStart && order < p->MaxOrder)
-        SetSuccessor(s, RemoveBinContexts(p, CTX(successor), order + 1));
+        Ppmd8State_SetSuccessor(s, RemoveBinContexts(p, CTX(successor), order + 1));
       else
-        SetSuccessor(s, 0);
+        Ppmd8State_SetSuccessor(s, 0);
     }
     while (--s >= STATS(ctx));
   }
@@ -767,15 +776,15 @@ static UInt32 GetUsedMemory(const CPpmd8 *p)
 #endif
 
 
-static void RestoreModel(CPpmd8 *p, CTX_PTR ctxError
+static void RestoreModel(CPpmd8 *p, PPMD8_CTX_PTR ctxError
     #ifdef PPMD8_FREEZE_SUPPORT
-    , CTX_PTR fSuccessor
+    , PPMD8_CTX_PTR fSuccessor
     #endif
     )
 {
-  CTX_PTR c;
+  PPMD8_CTX_PTR c;
   CPpmd_State *s;
-  RESET_TEXT(0);
+  RESET_TEXT(0)
 
   // we go here in cases of error of allocation for context (c1)
   // Order(MinContext) < Order(ctxError) <= Order(MaxContext)
@@ -831,7 +840,7 @@ static void RestoreModel(CPpmd8 *p, CTX_PTR ctxError
   else
   #endif
   if (p->RestoreMethod == PPMD8_RESTORE_METHOD_RESTART || GetUsedMemory(p) < (p->Size >> 1))
-    RestartModel(p);
+    Ppmd8_RestartModel(p);
   else
   {
     while (p->MaxContext->Suffix)
@@ -850,8 +859,8 @@ static void RestoreModel(CPpmd8 *p, CTX_PTR ctxError
 
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static CTX_PTR CreateSuccessors(CPpmd8 *p, BoolInt skip, CPpmd_State *s1, CTX_PTR c)
+static PPMD8_CTX_PTR Ppmd8_CreateSuccessors(CPpmd8 *p, BoolInt skip, CPpmd_State *s1, PPMD8_CTX_PTR c)
 {
 
   CPpmd_Byte_Ref upBranch = (CPpmd_Byte_Ref)SUCCESSOR(p->FoundState);
@@ -927,15 +936,15 @@ static CTX_PTR CreateSuccessors(CPpmd8 *p, BoolInt skip, CPpmd_State *s1, CTX_PT
 
   do
   {
-    CTX_PTR c1;
+    PPMD8_CTX_PTR c1;
     /* = AllocContext(p); */
     if (p->HiUnit != p->LoUnit)
-      c1 = (CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE);
+      c1 = (PPMD8_CTX_PTR)(void *)(p->HiUnit -= UNIT_SIZE);
     else if (p->FreeList[0] != 0)
-      c1 = (CTX_PTR)RemoveNode(p, 0);
+      c1 = (PPMD8_CTX_PTR)Ppmd8_RemoveNode(p, 0);
     else
     {
-      c1 = (CTX_PTR)AllocUnitsRare(p, 0);
+      c1 = (PPMD8_CTX_PTR)Ppmd8_AllocUnitsRare(p, 0);
       if (!c1)
         return NULL;
     }
@@ -943,9 +952,9 @@ static CTX_PTR CreateSuccessors(CPpmd8 *p, BoolInt skip, CPpmd_State *s1, CTX_PT
     c1->NumStats = 0;
     c1->Union2.State2.Symbol = newSym;
     c1->Union2.State2.Freq = newFreq;
-    SetSuccessor(ONE_STATE(c1), upBranch);
+    Ppmd8State_SetSuccessor(ONE_STATE(c1), upBranch);
     c1->Suffix = REF(c);
-    SetSuccessor(ps[--numPs], REF(c1));
+    Ppmd8State_SetSuccessor(ps[--numPs], REF(c1));
     c = c1;
   }
   while (numPs != 0);
@@ -954,10 +963,10 @@ static CTX_PTR CreateSuccessors(CPpmd8 *p, BoolInt skip, CPpmd_State *s1, CTX_PT
 }
 
 
-static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
+static PPMD8_CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, PPMD8_CTX_PTR c)
 {
   CPpmd_State *s = NULL;
-  CTX_PTR c1 = c;
+  PPMD8_CTX_PTR c1 = c;
   CPpmd_Void_Ref upBranch = REF(p->Text);
   
   #ifdef PPMD8_FREEZE_SUPPORT
@@ -967,7 +976,7 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
   ps[numPs++] = p->FoundState;
   #endif
 
-  SetSuccessor(p->FoundState, upBranch);
+  Ppmd8State_SetSuccessor(p->FoundState, upBranch);
   p->OrderFall++;
 
   for (;;)
@@ -985,8 +994,8 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
         #ifdef PPMD8_FREEZE_SUPPORT
         if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE)
         {
-          do { SetSuccessor(ps[--numPs], REF(c)); } while (numPs);
+          do { Ppmd8State_SetSuccessor(ps[--numPs], REF(c)); } while (numPs);
-          RESET_TEXT(1);
+          RESET_TEXT(1)
           p->OrderFall = 1;
         }
         #endif
@@ -1014,7 +1023,7 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
     #ifdef PPMD8_FREEZE_SUPPORT
     ps[numPs++] = s;
     #endif
-    SetSuccessor(s, upBranch);
+    Ppmd8State_SetSuccessor(s, upBranch);
     p->OrderFall++;
   }
   
@@ -1022,8 +1031,8 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
   if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE)
   {
     c = CTX(SUCCESSOR(s));
-    do { SetSuccessor(ps[--numPs], REF(c)); } while (numPs);
+    do { Ppmd8State_SetSuccessor(ps[--numPs], REF(c)); } while (numPs);
-    RESET_TEXT(1);
+    RESET_TEXT(1)
     p->OrderFall = 1;
     return c;
   }
@@ -1031,15 +1040,15 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
   #endif
   if (SUCCESSOR(s) <= upBranch)
   {
-    CTX_PTR successor;
+    PPMD8_CTX_PTR successor;
     CPpmd_State *s2 = p->FoundState;
     p->FoundState = s;
 
-    successor = CreateSuccessors(p, False, NULL, c);
+    successor = Ppmd8_CreateSuccessors(p, False, NULL, c);
     if (!successor)
-      SetSuccessor(s, 0);
+      Ppmd8State_SetSuccessor(s, 0);
     else
-      SetSuccessor(s, REF(successor));
+      Ppmd8State_SetSuccessor(s, REF(successor));
     p->FoundState = s2;
   }
   
@@ -1047,7 +1056,7 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
     CPpmd_Void_Ref successor = SUCCESSOR(s);
     if (p->OrderFall == 1 && c1 == p->MaxContext)
     {
-      SetSuccessor(p->FoundState, successor);
+      Ppmd8State_SetSuccessor(p->FoundState, successor);
       p->Text--;
     }
     if (successor == 0)
@@ -1059,11 +1068,11 @@ static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c)
 
 
 void Ppmd8_UpdateModel(CPpmd8 *p);
-MY_NO_INLINE
+Z7_NO_INLINE
 void Ppmd8_UpdateModel(CPpmd8 *p)
 {
   CPpmd_Void_Ref maxSuccessor, minSuccessor = SUCCESSOR(p->FoundState);
-  CTX_PTR c;
+  PPMD8_CTX_PTR c;
   unsigned s0, ns, fFreq = p->FoundState->Freq;
   Byte flag, fSymbol = p->FoundState->Symbol;
   {
@@ -1096,7 +1105,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
         
         if (s[0].Freq >= s[-1].Freq)
         {
-          SwapStates(&s[0], &s[-1]);
+          SWAP_STATES(&s[0], &s[-1]);
           s--;
         }
       }
@@ -1112,14 +1121,14 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
   c = p->MaxContext;
   if (p->OrderFall == 0 && minSuccessor)
   {
-    CTX_PTR cs = CreateSuccessors(p, True, s, p->MinContext);
+    PPMD8_CTX_PTR cs = Ppmd8_CreateSuccessors(p, True, s, p->MinContext);
     if (!cs)
     {
-      SetSuccessor(p->FoundState, 0);
+      Ppmd8State_SetSuccessor(p->FoundState, 0);
       RESTORE_MODEL(c, CTX(minSuccessor));
       return;
     }
-    SetSuccessor(p->FoundState, REF(cs));
+    Ppmd8State_SetSuccessor(p->FoundState, REF(cs));
     p->MinContext = p->MaxContext = cs;
     return;
   }
@@ -1141,7 +1150,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
 
   if (!minSuccessor)
   {
-    CTX_PTR cs = ReduceOrder(p, s, p->MinContext);
+    PPMD8_CTX_PTR cs = ReduceOrder(p, s, p->MinContext);
     if (!cs)
     {
       RESTORE_MODEL(c, NULL);
@@ -1151,7 +1160,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
   }
   else if ((Byte *)Ppmd8_GetPtr(p, minSuccessor) < p->UnitsStart)
   {
-    CTX_PTR cs = CreateSuccessors(p, False, s, p->MinContext);
+    PPMD8_CTX_PTR cs = Ppmd8_CreateSuccessors(p, False, s, p->MinContext);
     if (!cs)
     {
       RESTORE_MODEL(c, NULL);
@@ -1169,7 +1178,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
   else if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE)
   {
     maxSuccessor = minSuccessor;
-    RESET_TEXT(0);
+    RESET_TEXT(0)
     p->OrderFall = 0;
   }
   #endif
@@ -1215,11 +1224,11 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
       if ((ns1 & 1) != 0)
       {
         /* Expand for one UNIT */
-        unsigned oldNU = (ns1 + 1) >> 1;
+        const unsigned oldNU = (ns1 + 1) >> 1;
-        unsigned i = U2I(oldNU);
+        const unsigned i = U2I(oldNU);
         if (i != U2I((size_t)oldNU + 1))
         {
-          void *ptr = AllocUnits(p, i + 1);
+          void *ptr = Ppmd8_AllocUnits(p, i + 1);
           void *oldPtr;
           if (!ptr)
           {
@@ -1227,15 +1236,15 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
             return;
           }
           oldPtr = STATS(c);
-          MyMem12Cpy(ptr, oldPtr, oldNU);
+          MEM_12_CPY(ptr, oldPtr, oldNU)
-          InsertNode(p, oldPtr, i);
+          Ppmd8_InsertNode(p, oldPtr, i);
           c->Union4.Stats = STATS_REF(ptr);
         }
       }
       sum = c->Union2.SummFreq;
       /* max increase of Escape_Freq is 1 here.
          an average increase is 1/3 per symbol */
-      sum += (3 * ns1 + 1 < ns);
+      sum += (UInt32)(unsigned)(3 * ns1 + 1 < ns);
       /* original PPMdH uses 16-bit variable for (sum) here.
          But (sum < ???). Do we need to truncate (sum) to 16-bit */
       // sum = (UInt16)sum;
@@ -1243,7 +1252,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
     else
     {
       
-      CPpmd_State *s = (CPpmd_State*)AllocUnits(p, 0);
+      CPpmd_State *s = (CPpmd_State*)Ppmd8_AllocUnits(p, 0);
       if (!s)
       {
         RESTORE_MODEL(c, CTX(minSuccessor));
@@ -1255,7 +1264,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
         s->Symbol = c->Union2.State2.Symbol;
         s->Successor_0 = c->Union4.State4.Successor_0;
         s->Successor_1 = c->Union4.State4.Successor_1;
-        // SetSuccessor(s, c->Union4.Stats);  // call it only for debug purposes to check the order of
+        // Ppmd8State_SetSuccessor(s, c->Union4.Stats);  // call it only for debug purposes to check the order of
                                               // (Successor_0 and Successor_1) in LE/BE.
         c->Union4.Stats = REF(s);
         if (freq < MAX_FREQ / 4 - 1)
@@ -1265,7 +1274,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
 
         s->Freq = (Byte)freq;
 
-        sum = freq + p->InitEsc + (ns > 2);   // Ppmd8 (> 2)
+        sum = (UInt32)(freq + p->InitEsc + (ns > 2));   // Ppmd8 (> 2)
       }
     }
 
@@ -1275,7 +1284,7 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
       UInt32 sf = (UInt32)s0 + sum;
       s->Symbol = fSymbol;
       c->NumStats = (Byte)(ns1 + 1);
-      SetSuccessor(s, maxSuccessor);
+      Ppmd8State_SetSuccessor(s, maxSuccessor);
       c->Flags |= flag;
       if (cf < 6 * sf)
       {
@@ -1299,8 +1308,8 @@ void Ppmd8_UpdateModel(CPpmd8 *p)
   
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
-static void Rescale(CPpmd8 *p)
+static void Ppmd8_Rescale(CPpmd8 *p)
 {
   unsigned i, adder, sumFreq, escFreq;
   CPpmd_State *stats = STATS(p->MinContext);
@@ -1389,7 +1398,7 @@ static void Rescale(CPpmd8 *p)
       *s = *stats;
       s->Freq = (Byte)freq;
       p->FoundState = s;
-      InsertNode(p, stats, U2I(n0));
+      Ppmd8_InsertNode(p, stats, U2I(n0));
       return;
     }
 
@@ -1437,10 +1446,10 @@ CPpmd_See *Ppmd8_MakeEscFreq(CPpmd8 *p, unsigned numMasked1, UInt32 *escFreq)
 
     {
       // if (see->Summ) field is larger than 16-bit, we need only low 16 bits of Summ
-      unsigned summ = (UInt16)see->Summ; // & 0xFFFF
+      const unsigned summ = (UInt16)see->Summ; // & 0xFFFF
-      unsigned r = (summ >> see->Shift);
+      const unsigned r = (summ >> see->Shift);
       see->Summ = (UInt16)(summ - r);
-      *escFreq = r + (r == 0);
+      *escFreq = (UInt32)(r + (r == 0));
     }
   }
   else
@@ -1452,9 +1461,9 @@ CPpmd_See *Ppmd8_MakeEscFreq(CPpmd8 *p, unsigned numMasked1, UInt32 *escFreq)
 }
 
  
-static void NextContext(CPpmd8 *p)
+static void Ppmd8_NextContext(CPpmd8 *p)
 {
-  CTX_PTR c = CTX(SUCCESSOR(p->FoundState));
+  PPMD8_CTX_PTR c = CTX(SUCCESSOR(p->FoundState));
   if (p->OrderFall == 0 && (const Byte *)c >= p->UnitsStart)
     p->MaxContext = p->MinContext = c;
   else
@@ -1471,12 +1480,12 @@ void Ppmd8_Update1(CPpmd8 *p)
   s->Freq = (Byte)freq;
   if (freq > s[-1].Freq)
   {
-    SwapStates(s, &s[-1]);
+    SWAP_STATES(s, &s[-1]);
     p->FoundState = --s;
     if (freq > MAX_FREQ)
-      Rescale(p);
+      Ppmd8_Rescale(p);
   }
-  NextContext(p);
+  Ppmd8_NextContext(p);
 }
 
 
@@ -1485,15 +1494,15 @@ void Ppmd8_Update1_0(CPpmd8 *p)
   CPpmd_State *s = p->FoundState;
   CPpmd8_Context *mc = p->MinContext;
   unsigned freq = s->Freq;
-  unsigned summFreq = mc->Union2.SummFreq;
+  const unsigned summFreq = mc->Union2.SummFreq;
   p->PrevSuccess = (2 * freq >= summFreq); // Ppmd8 (>=)
-  p->RunLength += (int)p->PrevSuccess;
+  p->RunLength += (Int32)p->PrevSuccess;
   mc->Union2.SummFreq = (UInt16)(summFreq + 4);
   freq += 4;
   s->Freq = (Byte)freq;
   if (freq > MAX_FREQ)
-    Rescale(p);
+    Ppmd8_Rescale(p);
-  NextContext(p);
+  Ppmd8_NextContext(p);
 }
 
 
@@ -1504,7 +1513,7 @@ void Ppmd8_UpdateBin(CPpmd8 *p)
   p->FoundState->Freq = (Byte)(freq + (freq < 196)); // Ppmd8 (196)
   p->PrevSuccess = 1;
   p->RunLength++;
-  NextContext(p);
+  Ppmd8_NextContext(p);
 }
 */
 
@@ -1517,7 +1526,7 @@ void Ppmd8_Update2(CPpmd8 *p)
   p->MinContext->Union2.SummFreq = (UInt16)(p->MinContext->Union2.SummFreq + 4);
   s->Freq = (Byte)freq;
   if (freq > MAX_FREQ)
-    Rescale(p);
+    Ppmd8_Rescale(p);
   Ppmd8_UpdateModel(p);
 }
 
@@ -1526,7 +1535,7 @@ void Ppmd8_Update2(CPpmd8 *p)
   GlueCount, and Glue method
   BinSum
   See / EscFreq
-  CreateSuccessors updates more suffix contexts
+  Ppmd8_CreateSuccessors updates more suffix contexts
   Ppmd8_UpdateModel consts.
   PrevSuccess Update
 
@@ -1535,3 +1544,31 @@ Flags:
   (1 << 3) - there is symbol in Stats with (sym >= 0x40) in
   (1 << 4) - main symbol of context is (sym >= 0x40)
 */
+
+#undef RESET_TEXT
+#undef FLAG_RESCALED
+#undef FLAG_PREV_HIGH
+#undef HiBits_Prepare
+#undef HiBits_Convert_3
+#undef HiBits_Convert_4
+#undef PPMD8_HiBitsFlag_3
+#undef PPMD8_HiBitsFlag_4
+#undef RESTORE_MODEL
+
+#undef MAX_FREQ
+#undef UNIT_SIZE
+#undef U2B
+#undef U2I
+#undef I2U
+
+#undef REF
+#undef STATS_REF
+#undef CTX
+#undef STATS
+#undef ONE_STATE
+#undef SUFFIX
+#undef NODE
+#undef EMPTY_NODE
+#undef MEM_12_CPY
+#undef SUCCESSOR
+#undef SWAP_STATES

C/Ppmd8.h

@@ -1,11 +1,11 @@
 /* Ppmd8.h -- Ppmd8 (PPMdI) compression codec
-2021-04-13 : Igor Pavlov : Public domain
+2023-04-02 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.I (2002): Dmitry Shkarin : Public domain
   Carryless rangecoder (1999): Dmitry Subbotin : Public domain */
 
-#ifndef __PPMD8_H
+#ifndef ZIP7_INC_PPMD8_H
-#define __PPMD8_H
+#define ZIP7_INC_PPMD8_H
 
 #include "Ppmd.h"
 
@@ -87,8 +87,8 @@ typedef struct
   UInt32 Low;
   union
   {
-    IByteIn *In;
+    IByteInPtr In;
-    IByteOut *Out;
+    IByteOutPtr Out;
   } Stream;
 
   Byte Indx2Units[PPMD_NUM_INDEXES + 2]; // +2 for alignment

C/Ppmd8Dec.c

@@ -1,5 +1,5 @@
 /* Ppmd8Dec.c -- Ppmd8 (PPMdI) Decoder
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.I (2002): Dmitry Shkarin : Public domain
   Carryless rangecoder (1999): Dmitry Subbotin : Public domain */
@@ -8,8 +8,8 @@ This code is based on:
 
 #include "Ppmd8.h"
 
-#define kTop (1 << 24)
+#define kTop ((UInt32)1 << 24)
-#define kBot (1 << 15)
+#define kBot ((UInt32)1 << 15)
 
 #define READ_BYTE(p) IByteIn_Read((p)->Stream.In)
 
@@ -37,9 +37,9 @@ BoolInt Ppmd8_Init_RangeDec(CPpmd8 *p)
 
 #define R p
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
-// MY_NO_INLINE
+// Z7_NO_INLINE
-static void RangeDec_Decode(CPpmd8 *p, UInt32 start, UInt32 size)
+static void Ppmd8_RD_Decode(CPpmd8 *p, UInt32 start, UInt32 size)
 {
   start *= R->Range;
   R->Low += start;
@@ -48,17 +48,17 @@ static void RangeDec_Decode(CPpmd8 *p, UInt32 start, UInt32 size)
   RC_NORM_LOCAL(R)
 }
 
-#define RC_Decode(start, size) RangeDec_Decode(p, start, size);
+#define RC_Decode(start, size)  Ppmd8_RD_Decode(p, start, size);
-#define RC_DecodeFinal(start, size) RC_Decode(start, size) RC_NORM_REMOTE(R)
+#define RC_DecodeFinal(start, size)  RC_Decode(start, size)  RC_NORM_REMOTE(R)
-#define RC_GetThreshold(total) (R->Code / (R->Range /= (total)))
+#define RC_GetThreshold(total)  (R->Code / (R->Range /= (total)))
 
 
 #define CTX(ref) ((CPpmd8_Context *)Ppmd8_GetContext(p, ref))
-typedef CPpmd8_Context * CTX_PTR;
+// typedef CPpmd8_Context * CTX_PTR;
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
 void Ppmd8_UpdateModel(CPpmd8 *p);
 
-#define MASK(sym) ((unsigned char *)charMask)[sym]
+#define MASK(sym)  ((Byte *)charMask)[sym]
 
 
 int Ppmd8_DecodeSymbol(CPpmd8 *p)
@@ -81,7 +81,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
     if ((Int32)(count -= s->Freq) < 0)
     {
       Byte sym;
-      RC_DecodeFinal(0, s->Freq);
+      RC_DecodeFinal(0, s->Freq)
       p->FoundState = s;
       sym = s->Symbol;
       Ppmd8_Update1_0(p);
@@ -96,7 +96,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
       if ((Int32)(count -= (++s)->Freq) < 0)
       {
         Byte sym;
-        RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq);
+        RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq)
         p->FoundState = s;
         sym = s->Symbol;
         Ppmd8_Update1(p);
@@ -109,10 +109,10 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
       return PPMD8_SYM_ERROR;
 
     hiCnt -= count;
-    RC_Decode(hiCnt, summFreq - hiCnt);
+    RC_Decode(hiCnt, summFreq - hiCnt)
     
     
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     // i = p->MinContext->NumStats - 1;
     // do { MASK((--s)->Symbol) = 0; } while (--i);
     {
@@ -120,8 +120,8 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -152,7 +152,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
       // Ppmd8_UpdateBin(p);
       {
         unsigned freq = s->Freq;
-        CTX_PTR c = CTX(SUCCESSOR(s));
+        CPpmd8_Context *c = CTX(SUCCESSOR(s));
         sym = s->Symbol;
         p->FoundState = s;
         p->PrevSuccess = 1;
@@ -176,7 +176,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
     R->Range = (R->Range & ~((UInt32)PPMD_BIN_SCALE - 1)) - size0;
     RC_NORM_LOCAL(R)
     
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     MASK(Ppmd8Context_OneState(p->MinContext)->Symbol) = 0;
     p->PrevSuccess = 0;
   }
@@ -209,17 +209,17 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
       unsigned num2 = num / 2;
 
       num &= 1;
-      hiCnt = (s->Freq & (unsigned)(MASK(s->Symbol))) & (0 - (UInt32)num);
+      hiCnt = (s->Freq & (UInt32)(MASK(s->Symbol))) & (0 - (UInt32)num);
       s += num;
       p->MinContext = mc;
 
       do
       {
-        unsigned sym0 = s[0].Symbol;
+        const unsigned sym0 = s[0].Symbol;
-        unsigned sym1 = s[1].Symbol;
+        const unsigned sym1 = s[1].Symbol;
         s += 2;
-        hiCnt += (s[-2].Freq & (unsigned)(MASK(sym0)));
+        hiCnt += (s[-2].Freq & (UInt32)(MASK(sym0)));
-        hiCnt += (s[-1].Freq & (unsigned)(MASK(sym1)));
+        hiCnt += (s[-1].Freq & (UInt32)(MASK(sym1)));
       }
       while (--num2);
     }
@@ -227,7 +227,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
     see = Ppmd8_MakeEscFreq(p, numMasked, &freqSum);
     freqSum += hiCnt;
     freqSum2 = freqSum;
-    PPMD8_CORRECT_SUM_RANGE(R, freqSum2);
+    PPMD8_CORRECT_SUM_RANGE(R, freqSum2)
 
 
     count = RC_GetThreshold(freqSum2);
@@ -235,7 +235,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
     if (count < hiCnt)
     {
       Byte sym;
-      // Ppmd_See_Update(see); // new (see->Summ) value can overflow over 16-bits in some rare cases
+      // Ppmd_See_UPDATE(see) // new (see->Summ) value can overflow over 16-bits in some rare cases
       s = Ppmd8_GetStats(p, p->MinContext);
       hiCnt = count;
 
@@ -243,15 +243,15 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
       {
         for (;;)
         {
-          count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
+          count -= s->Freq & (UInt32)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
-          // count -= s->Freq & (unsigned)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
+          // count -= s->Freq & (UInt32)(MASK((s)->Symbol)); s++; if ((Int32)count < 0) break;
         }
       }
       s--;
-      RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq);
+      RC_DecodeFinal((hiCnt - count) - s->Freq, s->Freq)
 
       // new (see->Summ) value can overflow over 16-bits in some rare cases
-      Ppmd_See_Update(see);
+      Ppmd_See_UPDATE(see)
       p->FoundState = s;
       sym = s->Symbol;
       Ppmd8_Update2(p);
@@ -261,7 +261,7 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
     if (count >= freqSum2)
       return PPMD8_SYM_ERROR;
     
-    RC_Decode(hiCnt, freqSum2 - hiCnt);
+    RC_Decode(hiCnt, freqSum2 - hiCnt)
     
     // We increase (see->Summ) for sum of Freqs of all non_Masked symbols.
     // new (see->Summ) value can overflow over 16-bits in some rare cases
@@ -277,3 +277,19 @@ int Ppmd8_DecodeSymbol(CPpmd8 *p)
     while (s != s2);
   }
 }
+
+#undef kTop
+#undef kBot
+#undef READ_BYTE
+#undef RC_NORM_BASE
+#undef RC_NORM_1
+#undef RC_NORM
+#undef RC_NORM_LOCAL
+#undef RC_NORM_REMOTE
+#undef R
+#undef RC_Decode
+#undef RC_DecodeFinal
+#undef RC_GetThreshold
+#undef CTX
+#undef SUCCESSOR
+#undef MASK

C/Ppmd8Enc.c

@@ -1,5 +1,5 @@
 /* Ppmd8Enc.c -- Ppmd8 (PPMdI) Encoder
-2021-04-13 : Igor Pavlov : Public domain
+2023-09-07 : Igor Pavlov : Public domain
 This code is based on:
   PPMd var.I (2002): Dmitry Shkarin : Public domain
   Carryless rangecoder (1999): Dmitry Subbotin : Public domain */
@@ -8,8 +8,8 @@ This code is based on:
 
 #include "Ppmd8.h"
 
-#define kTop (1 << 24)
+#define kTop ((UInt32)1 << 24)
-#define kBot (1 << 15)
+#define kBot ((UInt32)1 << 15)
 
 #define WRITE_BYTE(p) IByteOut_Write(p->Stream.Out, (Byte)(p->Low >> 24))
 
@@ -54,13 +54,13 @@ void Ppmd8_Flush_RangeEnc(CPpmd8 *p)
 
 
 
-MY_FORCE_INLINE
+Z7_FORCE_INLINE
-// MY_NO_INLINE
+// Z7_NO_INLINE
-static void RangeEnc_Encode(CPpmd8 *p, UInt32 start, UInt32 size, UInt32 total)
+static void Ppmd8_RangeEnc_Encode(CPpmd8 *p, UInt32 start, UInt32 size, UInt32 total)
 {
   R->Low += start * (R->Range /= total);
   R->Range *= size;
-  RC_NORM_LOCAL(R);
+  RC_NORM_LOCAL(R)
 }
 
 
@@ -72,19 +72,19 @@ static void RangeEnc_Encode(CPpmd8 *p, UInt32 start, UInt32 size, UInt32 total)
 
 
 
-#define RC_Encode(start, size, total) RangeEnc_Encode(p, start, size, total);
+#define RC_Encode(start, size, total)  Ppmd8_RangeEnc_Encode(p, start, size, total);
-#define RC_EncodeFinal(start, size, total) RC_Encode(start, size, total); RC_NORM_REMOTE(p);
+#define RC_EncodeFinal(start, size, total)  RC_Encode(start, size, total)  RC_NORM_REMOTE(p)
 
 #define CTX(ref) ((CPpmd8_Context *)Ppmd8_GetContext(p, ref))
 
-typedef CPpmd8_Context * CTX_PTR;
+// typedef CPpmd8_Context * CTX_PTR;
 #define SUCCESSOR(p) Ppmd_GET_SUCCESSOR(p)
 
 void Ppmd8_UpdateModel(CPpmd8 *p);
 
-#define MASK(sym) ((unsigned char *)charMask)[sym]
+#define MASK(sym)  ((Byte *)charMask)[sym]
 
-// MY_FORCE_INLINE
+// Z7_FORCE_INLINE
 // static
 void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
 {
@@ -104,7 +104,7 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
     if (s->Symbol == symbol)
     {
 
-      RC_EncodeFinal(0, s->Freq, summFreq);
+      RC_EncodeFinal(0, s->Freq, summFreq)
       p->FoundState = s;
       Ppmd8_Update1_0(p);
       return;
@@ -117,7 +117,7 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
       if ((++s)->Symbol == symbol)
       {
 
-        RC_EncodeFinal(sum, s->Freq, summFreq);
+        RC_EncodeFinal(sum, s->Freq, summFreq)
         p->FoundState = s;
         Ppmd8_Update1(p);
         return;
@@ -127,10 +127,10 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
     while (--i);
     
     
-    RC_Encode(sum, summFreq - sum, summFreq);
+    RC_Encode(sum, summFreq - sum, summFreq)
     
     
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     // MASK(s->Symbol) = 0;
     // i = p->MinContext->NumStats;
     // do { MASK((--s)->Symbol) = 0; } while (--i);
@@ -139,8 +139,8 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -153,20 +153,20 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
     UInt16 *prob = Ppmd8_GetBinSumm(p);
     CPpmd_State *s = Ppmd8Context_OneState(p->MinContext);
     UInt32 pr = *prob;
-    UInt32 bound = (R->Range >> 14) * pr;
+    const UInt32 bound = (R->Range >> 14) * pr;
     pr = PPMD_UPDATE_PROB_1(pr);
     if (s->Symbol == symbol)
     {
       *prob = (UInt16)(pr + (1 << PPMD_INT_BITS));
       // RangeEnc_EncodeBit_0(p, bound);
       R->Range = bound;
-      RC_NORM(R);
+      RC_NORM(R)
 
       // p->FoundState = s;
       // Ppmd8_UpdateBin(p);
       {
-        unsigned freq = s->Freq;
+        const unsigned freq = s->Freq;
-        CTX_PTR c = CTX(SUCCESSOR(s));
+        CPpmd8_Context *c = CTX(SUCCESSOR(s));
         p->FoundState = s;
         p->PrevSuccess = 1;
         p->RunLength++;
@@ -187,7 +187,7 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
     R->Range = (R->Range & ~((UInt32)PPMD_BIN_SCALE - 1)) - bound;
     RC_NORM_LOCAL(R)
 
-    PPMD_SetAllBitsIn256Bytes(charMask);
+    PPMD_SetAllBitsIn256Bytes(charMask)
     MASK(s->Symbol) = 0;
     p->PrevSuccess = 0;
   }
@@ -248,14 +248,14 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
 
     do
     {
-      unsigned cur = s->Symbol;
+      const unsigned cur = s->Symbol;
       if ((int)cur == symbol)
       {
-        UInt32 low = sum;
+        const UInt32 low = sum;
-        UInt32 freq = s->Freq;
+        const UInt32 freq = s->Freq;
         unsigned num2;
 
-        Ppmd_See_Update(see);
+        Ppmd_See_UPDATE(see)
         p->FoundState = s;
         sum += escFreq;
 
@@ -265,21 +265,20 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
         if (num2 != 0)
         {
           s += i;
-          for (;;)
+          do
           {
-            unsigned sym0 = s[0].Symbol;
+            const unsigned sym0 = s[0].Symbol;
-            unsigned sym1 = s[1].Symbol;
+            const unsigned sym1 = s[1].Symbol;
             s += 2;
             sum += (s[-2].Freq & (unsigned)(MASK(sym0)));
             sum += (s[-1].Freq & (unsigned)(MASK(sym1)));
-            if (--num2 == 0)
-              break;
           }
+          while (--num2);
         }
 
-        PPMD8_CORRECT_SUM_RANGE(p, sum);
+        PPMD8_CORRECT_SUM_RANGE(p, sum)
 
-        RC_EncodeFinal(low, freq, sum);
+        RC_EncodeFinal(low, freq, sum)
         Ppmd8_Update2(p);
         return;
       }
@@ -291,19 +290,19 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
     {
       UInt32 total = sum + escFreq;
       see->Summ = (UInt16)(see->Summ + total);
-      PPMD8_CORRECT_SUM_RANGE(p, total);
+      PPMD8_CORRECT_SUM_RANGE(p, total)
 
-      RC_Encode(sum, total - sum, total);
+      RC_Encode(sum, total - sum, total)
     }
 
     {
-      CPpmd_State *s2 = Ppmd8_GetStats(p, p->MinContext);
+      const CPpmd_State *s2 = Ppmd8_GetStats(p, p->MinContext);
       s--;
       MASK(s->Symbol) = 0;
       do
       {
-        unsigned sym0 = s2[0].Symbol;
+        const unsigned sym0 = s2[0].Symbol;
-        unsigned sym1 = s2[1].Symbol;
+        const unsigned sym1 = s2[1].Symbol;
         s2 += 2;
         MASK(sym0) = 0;
         MASK(sym1) = 0;
@@ -312,3 +311,27 @@ void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol)
     }
   }
 }
+
+
+
+
+
+
+
+
+
+#undef kTop
+#undef kBot
+#undef WRITE_BYTE
+#undef RC_NORM_BASE
+#undef RC_NORM_1
+#undef RC_NORM
+#undef RC_NORM_LOCAL
+#undef RC_NORM_REMOTE
+#undef R
+#undef RC_Encode
+#undef RC_EncodeFinal
+
+#undef CTX
+#undef SUCCESSOR
+#undef MASK

C/Precomp.h

@@ -1,10 +1,127 @@
-/* Precomp.h -- StdAfx
+/* Precomp.h -- precompilation file
-2013-11-12 : Igor Pavlov : Public domain */
+2024-01-25 : Igor Pavlov : Public domain */
 
-#ifndef __7Z_PRECOMP_H
+#ifndef ZIP7_INC_PRECOMP_H
-#define __7Z_PRECOMP_H
+#define ZIP7_INC_PRECOMP_H
+
+/*
+  this file must be included before another *.h files and before <windows.h>.
+  this file is included from the following files:
+    C\*.c
+    C\Util\*\Precomp.h   <-  C\Util\*\*.c
+    CPP\Common\Common.h  <-  *\StdAfx.h    <-  *\*.cpp
+
+  this file can set the following macros:
+    Z7_LARGE_PAGES 1
+    Z7_LONG_PATH 1
+    Z7_WIN32_WINNT_MIN  0x0500 (or higher) : we require at least win2000+ for 7-Zip
+    _WIN32_WINNT        0x0500 (or higher)
+    WINVER  _WIN32_WINNT
+    UNICODE 1
+    _UNICODE 1
+*/
 
 #include "Compiler.h"
-/* #include "7zTypes.h" */
+
+#ifdef _MSC_VER
+// #pragma warning(disable : 4206) // nonstandard extension used : translation unit is empty
+#if _MSC_VER >= 1912
+// #pragma warning(disable : 5039) // pointer or reference to potentially throwing function passed to 'extern "C"' function under - EHc.Undefined behavior may occur if this function throws an exception.
+#endif
+#endif
+
+/*
+// for debug:
+#define UNICODE 1
+#define _UNICODE 1
+#define  _WIN32_WINNT  0x0500  // win2000
+#ifndef WINVER
+  #define WINVER  _WIN32_WINNT
+#endif
+*/
+
+#ifdef _WIN32
+/*
+  this "Precomp.h" file must be included before <windows.h>,
+  if we want to define _WIN32_WINNT before <windows.h>.
+*/
+
+#ifndef Z7_LARGE_PAGES
+#ifndef Z7_NO_LARGE_PAGES
+#define Z7_LARGE_PAGES 1
+#endif
+#endif
+
+#ifndef Z7_LONG_PATH
+#ifndef Z7_NO_LONG_PATH
+#define Z7_LONG_PATH 1
+#endif
+#endif
+
+#ifndef Z7_DEVICE_FILE
+#ifndef Z7_NO_DEVICE_FILE
+// #define Z7_DEVICE_FILE 1
+#endif
+#endif
+
+// we don't change macros if included after <windows.h>
+#ifndef _WINDOWS_
+
+#ifndef Z7_WIN32_WINNT_MIN
+  #if defined(_M_ARM64) || defined(__aarch64__)
+    // #define Z7_WIN32_WINNT_MIN  0x0a00  // win10
+    #define Z7_WIN32_WINNT_MIN  0x0600  // vista
+  #elif defined(_M_ARM) && defined(_M_ARMT) && defined(_M_ARM_NT)
+    // #define Z7_WIN32_WINNT_MIN  0x0602  // win8
+    #define Z7_WIN32_WINNT_MIN  0x0600  // vista
+  #elif defined(_M_X64) || defined(_M_AMD64) || defined(__x86_64__) || defined(_M_IA64)
+    #define Z7_WIN32_WINNT_MIN  0x0503  // win2003
+  // #elif defined(_M_IX86) || defined(__i386__)
+  //   #define Z7_WIN32_WINNT_MIN  0x0500  // win2000
+  #else // x86 and another(old) systems
+    #define Z7_WIN32_WINNT_MIN  0x0500  // win2000
+    // #define Z7_WIN32_WINNT_MIN  0x0502  // win2003 // for debug
+  #endif
+#endif // Z7_WIN32_WINNT_MIN
+
+
+#ifndef Z7_DO_NOT_DEFINE_WIN32_WINNT
+#ifdef _WIN32_WINNT
+  // #error Stop_Compiling_Bad_WIN32_WINNT
+#else
+  #ifndef Z7_NO_DEFINE_WIN32_WINNT
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #define _WIN32_WINNT  Z7_WIN32_WINNT_MIN
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+  #endif
+#endif // _WIN32_WINNT
+
+#ifndef WINVER
+  #define WINVER  _WIN32_WINNT
+#endif
+#endif // Z7_DO_NOT_DEFINE_WIN32_WINNT
+
+
+#ifndef _MBCS
+#ifndef Z7_NO_UNICODE
+// UNICODE and _UNICODE are used by <windows.h> and by 7-zip code.
+
+#ifndef UNICODE
+#define UNICODE 1
+#endif
+
+#ifndef _UNICODE
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+#define _UNICODE 1
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+#endif
+
+#endif // Z7_NO_UNICODE
+#endif // _MBCS
+#endif // _WINDOWS_
+
+// #include "7zWindows.h"
+
+#endif // _WIN32
 
 #endif

C/RotateDefs.h

@@ -1,14 +1,14 @@
 /* RotateDefs.h -- Rotate functions
-2015-03-25 : Igor Pavlov : Public domain */
+2023-06-18 : Igor Pavlov : Public domain */
 
-#ifndef __ROTATE_DEFS_H
+#ifndef ZIP7_INC_ROTATE_DEFS_H
-#define __ROTATE_DEFS_H
+#define ZIP7_INC_ROTATE_DEFS_H
 
 #ifdef _MSC_VER
 
 #include <stdlib.h>
 
-/* don't use _rotl with MINGW. It can insert slow call to function. */
+/* don't use _rotl with old MINGW. It can insert slow call to function. */
  
 /* #if (_MSC_VER >= 1200) */
 #pragma intrinsic(_rotl)
@@ -18,12 +18,32 @@
 #define rotlFixed(x, n) _rotl((x), (n))
 #define rotrFixed(x, n) _rotr((x), (n))
 
+#if (_MSC_VER >= 1300)
+#define Z7_ROTL64(x, n) _rotl64((x), (n))
+#define Z7_ROTR64(x, n) _rotr64((x), (n))
+#else
+#define Z7_ROTL64(x, n) (((x) << (n)) | ((x) >> (64 - (n))))
+#define Z7_ROTR64(x, n) (((x) >> (n)) | ((x) << (64 - (n))))
+#endif
+
 #else
 
 /* new compilers can translate these macros to fast commands. */
 
+#if defined(__clang__) && (__clang_major__ >= 4) \
+  || defined(__GNUC__) && (__GNUC__ >= 5)
+/* GCC 4.9.0 and clang 3.5 can recognize more correct version: */
+#define rotlFixed(x, n) (((x) << (n)) | ((x) >> (-(n) & 31)))
+#define rotrFixed(x, n) (((x) >> (n)) | ((x) << (-(n) & 31)))
+#define Z7_ROTL64(x, n) (((x) << (n)) | ((x) >> (-(n) & 63)))
+#define Z7_ROTR64(x, n) (((x) >> (n)) | ((x) << (-(n) & 63)))
+#else
+/* for old GCC / clang: */
 #define rotlFixed(x, n) (((x) << (n)) | ((x) >> (32 - (n))))
 #define rotrFixed(x, n) (((x) >> (n)) | ((x) << (32 - (n))))
+#define Z7_ROTL64(x, n) (((x) << (n)) | ((x) >> (64 - (n))))
+#define Z7_ROTR64(x, n) (((x) >> (n)) | ((x) << (64 - (n))))
+#endif
 
 #endif
 

C/Sha1.c

@@ -1,64 +1,60 @@
 /* Sha1.c -- SHA-1 Hash
-2021-07-13 : Igor Pavlov : Public domain
+: Igor Pavlov : Public domain
 This code is based on public domain code of Steve Reid from Wei Dai's Crypto++ library. */
 
 #include "Precomp.h"
 
 #include <string.h>
 
-#include "CpuArch.h"
-#include "RotateDefs.h"
 #include "Sha1.h"
-
+#include "RotateDefs.h"
-#if defined(_MSC_VER) && (_MSC_VER < 1900)
+#include "CpuArch.h"
-// #define USE_MY_MM
-#endif
 
 #ifdef MY_CPU_X86_OR_AMD64
-  #ifdef _MSC_VER
+  #if   defined(Z7_LLVM_CLANG_VERSION)  && (Z7_LLVM_CLANG_VERSION  >= 30800) \
-    #if _MSC_VER >= 1200
+     || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 50100) \
-      #define _SHA_SUPPORTED
+     || defined(Z7_GCC_VERSION)         && (Z7_GCC_VERSION         >= 40900) \
-    #endif
+     || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1600) \
-  #elif defined(__clang__)
+     || defined(_MSC_VER) && (_MSC_VER >= 1200)
-    #if (__clang_major__ >= 8) // fix that check
+      #define Z7_COMPILER_SHA1_SUPPORTED
-      #define _SHA_SUPPORTED
-    #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 8) // fix that check
-      #define _SHA_SUPPORTED
-    #endif
-  #elif defined(__INTEL_COMPILER)
-    #if (__INTEL_COMPILER >= 1800) // fix that check
-      #define _SHA_SUPPORTED
-    #endif
   #endif
-#elif defined(MY_CPU_ARM_OR_ARM64)
+#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE) \
-  #ifdef _MSC_VER
+   && (!defined(Z7_MSC_VER_ORIGINAL) || (_MSC_VER >= 1929) && (_MSC_FULL_VER >= 192930037))
-    #if _MSC_VER >= 1910 && _MSC_VER >= 1929 && _MSC_FULL_VER >= 192930037
+  #if   defined(__ARM_FEATURE_SHA2) \
-      #define _SHA_SUPPORTED
+     || defined(__ARM_FEATURE_CRYPTO)
+    #define Z7_COMPILER_SHA1_SUPPORTED
+  #else
+    #if  defined(MY_CPU_ARM64) \
+      || defined(__ARM_ARCH) && (__ARM_ARCH >= 4) \
+      || defined(Z7_MSC_VER_ORIGINAL)
+    #if  defined(__ARM_FP) && \
+          (   defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30800) \
+           || defined(__GNUC__) && (__GNUC__ >= 6) \
+          ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1910)
+    #if  defined(MY_CPU_ARM64) \
+      || !defined(Z7_CLANG_VERSION) \
+      || defined(__ARM_NEON) && \
+          (Z7_CLANG_VERSION < 170000 || \
+           Z7_CLANG_VERSION > 170001)
+      #define Z7_COMPILER_SHA1_SUPPORTED
     #endif
-  #elif defined(__clang__)
-    #if (__clang_major__ >= 8) // fix that check
-      #define _SHA_SUPPORTED
     #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 6) // fix that check
-      #define _SHA_SUPPORTED
     #endif
   #endif
 #endif
 
-void MY_FAST_CALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks);
 
-#ifdef _SHA_SUPPORTED
+#ifdef Z7_COMPILER_SHA1_SUPPORTED
-  void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks);
+  void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks);
 
-  static SHA1_FUNC_UPDATE_BLOCKS g_FUNC_UPDATE_BLOCKS = Sha1_UpdateBlocks;
+  static SHA1_FUNC_UPDATE_BLOCKS g_SHA1_FUNC_UPDATE_BLOCKS = Sha1_UpdateBlocks;
-  static SHA1_FUNC_UPDATE_BLOCKS g_FUNC_UPDATE_BLOCKS_HW;
+  static SHA1_FUNC_UPDATE_BLOCKS g_SHA1_FUNC_UPDATE_BLOCKS_HW;
 
-  #define UPDATE_BLOCKS(p) p->func_UpdateBlocks
+  #define SHA1_UPDATE_BLOCKS(p) p->v.vars.func_UpdateBlocks
 #else
-  #define UPDATE_BLOCKS(p) Sha1_UpdateBlocks
+  #define SHA1_UPDATE_BLOCKS(p) Sha1_UpdateBlocks
 #endif
 
 
@@ -66,16 +62,16 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 {
   SHA1_FUNC_UPDATE_BLOCKS func = Sha1_UpdateBlocks;
   
-  #ifdef _SHA_SUPPORTED
+  #ifdef Z7_COMPILER_SHA1_SUPPORTED
     if (algo != SHA1_ALGO_SW)
     {
       if (algo == SHA1_ALGO_DEFAULT)
-        func = g_FUNC_UPDATE_BLOCKS;
+        func = g_SHA1_FUNC_UPDATE_BLOCKS;
       else
       {
         if (algo != SHA1_ALGO_HW)
           return False;
-        func = g_FUNC_UPDATE_BLOCKS_HW;
+        func = g_SHA1_FUNC_UPDATE_BLOCKS_HW;
         if (!func)
           return False;
       }
@@ -85,27 +81,28 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
       return False;
   #endif
 
-  p->func_UpdateBlocks = func;
+  p->v.vars.func_UpdateBlocks = func;
   return True;
 }
 
 
 /* define it for speed optimization */
-// #define _SHA1_UNROLL
+// #define Z7_SHA1_UNROLL
 
 // allowed unroll steps: (1, 2, 4, 5, 20)
 
-#ifdef _SHA1_UNROLL
+#undef Z7_SHA1_BIG_W
+#ifdef Z7_SHA1_UNROLL
   #define STEP_PRE  20
   #define STEP_MAIN 20
 #else
-  #define _SHA1_BIG_W
+  #define Z7_SHA1_BIG_W
   #define STEP_PRE  5
   #define STEP_MAIN 5
 #endif
 
 
-#ifdef _SHA1_BIG_W
+#ifdef Z7_SHA1_BIG_W
   #define kNumW 80
   #define w(i) W[i]
 #else
@@ -150,11 +147,11 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 */
 
 #define M5(i, fx, wx0, wx1) \
-    T5 ( a,b,c,d,e, fx, wx0((i)  ) ); \
+    T5 ( a,b,c,d,e, fx, wx0((i)  ) ) \
-    T5 ( e,a,b,c,d, fx, wx1((i)+1) ); \
+    T5 ( e,a,b,c,d, fx, wx1((i)+1) ) \
-    T5 ( d,e,a,b,c, fx, wx1((i)+2) ); \
+    T5 ( d,e,a,b,c, fx, wx1((i)+2) ) \
-    T5 ( c,d,e,a,b, fx, wx1((i)+3) ); \
+    T5 ( c,d,e,a,b, fx, wx1((i)+3) ) \
-    T5 ( b,c,d,e,a, fx, wx1((i)+4) ); \
+    T5 ( b,c,d,e,a, fx, wx1((i)+4) ) \
 
 #define R5(i, fx, wx) \
     M5 ( i, fx, wx, wx) \
@@ -163,17 +160,17 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 #if STEP_PRE > 5
 
   #define R20_START \
-    R5 (  0, f0, w0); \
+    R5 (  0, f0, w0) \
-    R5 (  5, f0, w0); \
+    R5 (  5, f0, w0) \
-    R5 ( 10, f0, w0); \
+    R5 ( 10, f0, w0) \
-    M5 ( 15, f0, w0, w1); \
+    M5 ( 15, f0, w0, w1) \
   
   #elif STEP_PRE == 5
   
   #define R20_START \
     { size_t i; for (i = 0; i < 15; i += STEP_PRE) \
-      { R5(i, f0, w0); } } \
+      { R5(i, f0, w0) } } \
-    M5 ( 15, f0, w0, w1); \
+    M5 ( 15, f0, w0, w1) \
 
 #else
 
@@ -187,8 +184,8 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 
   #define R20_START \
     { size_t i; for (i = 0; i < 16; i += STEP_PRE) \
-      { R_PRE(i, f0, w0); } } \
+      { R_PRE(i, f0, w0) } } \
-    R4 ( 16, f0, w1); \
+    R4 ( 16, f0, w1) \
 
 #endif
 
@@ -197,10 +194,10 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 #if STEP_MAIN > 5
 
   #define R20(ii, fx) \
-    R5 ( (ii)     , fx, w1); \
+    R5 ( (ii)     , fx, w1) \
-    R5 ( (ii) + 5 , fx, w1); \
+    R5 ( (ii) + 5 , fx, w1) \
-    R5 ( (ii) + 10, fx, w1); \
+    R5 ( (ii) + 10, fx, w1) \
-    R5 ( (ii) + 15, fx, w1); \
+    R5 ( (ii) + 15, fx, w1) \
 
 #else
 
@@ -216,7 +213,7 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 
   #define R20(ii, fx)  \
     { size_t i; for (i = (ii); i < (ii) + 20; i += STEP_MAIN) \
-      { R_MAIN(i, fx, w1); } } \
+      { R_MAIN(i, fx, w1) } } \
 
 #endif
 
@@ -224,7 +221,7 @@ BoolInt Sha1_SetFunction(CSha1 *p, unsigned algo)
 
 void Sha1_InitState(CSha1 *p)
 {
-  p->count = 0;
+  p->v.vars.count = 0;
   p->state[0] = 0x67452301;
   p->state[1] = 0xEFCDAB89;
   p->state[2] = 0x98BADCFE;
@@ -234,9 +231,9 @@ void Sha1_InitState(CSha1 *p)
 
 void Sha1_Init(CSha1 *p)
 {
-  p->func_UpdateBlocks =
+  p->v.vars.func_UpdateBlocks =
-  #ifdef _SHA_SUPPORTED
+  #ifdef Z7_COMPILER_SHA1_SUPPORTED
-      g_FUNC_UPDATE_BLOCKS;
+      g_SHA1_FUNC_UPDATE_BLOCKS;
   #else
       NULL;
   #endif
@@ -244,12 +241,12 @@ void Sha1_Init(CSha1 *p)
 }
 
 
-MY_NO_INLINE
+Z7_NO_INLINE
-void MY_FAST_CALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks)
 {
   UInt32 a, b, c, d, e;
   UInt32 W[kNumW];
-  // if (numBlocks != 0x1264378347) return;
+
   if (numBlocks == 0)
     return;
 
@@ -266,9 +263,9 @@ void MY_FAST_CALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t nu
   #endif
 
   R20_START
-  R20(20, f1);
+  R20(20, f1)
-  R20(40, f2);
+  R20(40, f2)
-  R20(60, f3);
+  R20(60, f3)
 
   a += state[0];
   b += state[1];
@@ -282,32 +279,27 @@ void MY_FAST_CALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t nu
   state[3] = d;
   state[4] = e;
 
-  data += 64;
+  data += SHA1_BLOCK_SIZE;
   }
   while (--numBlocks);
 }
 
 
-#define Sha1_UpdateBlock(p) UPDATE_BLOCKS(p)(p->state, p->buffer, 1)
+#define Sha1_UpdateBlock(p) SHA1_UPDATE_BLOCKS(p)(p->state, p->buffer, 1)
 
 void Sha1_Update(CSha1 *p, const Byte *data, size_t size)
 {
   if (size == 0)
     return;
-
   {
-    unsigned pos = (unsigned)p->count & 0x3F;
+    const unsigned pos = (unsigned)p->v.vars.count & (SHA1_BLOCK_SIZE - 1);
-    unsigned num;
+    const unsigned num = SHA1_BLOCK_SIZE - pos;
-    
+    p->v.vars.count += size;
-    p->count += size;
-    
-    num = 64 - pos;
     if (num > size)
     {
       memcpy(p->buffer + pos, data, size);
       return;
     }
-    
     if (pos != 0)
     {
       size -= num;
@@ -317,9 +309,10 @@ void Sha1_Update(CSha1 *p, const Byte *data, size_t size)
     }
   }
   {
-    size_t numBlocks = size >> 6;
+    const size_t numBlocks = size >> 6;
-    UPDATE_BLOCKS(p)(p->state, data, numBlocks);
+    // if (numBlocks)
-    size &= 0x3F;
+    SHA1_UPDATE_BLOCKS(p)(p->state, data, numBlocks);
+    size &= SHA1_BLOCK_SIZE - 1;
     if (size == 0)
       return;
     data += (numBlocks << 6);
@@ -330,52 +323,28 @@ void Sha1_Update(CSha1 *p, const Byte *data, size_t size)
 
 void Sha1_Final(CSha1 *p, Byte *digest)
 {
-  unsigned pos = (unsigned)p->count & 0x3F;
+  unsigned pos = (unsigned)p->v.vars.count & (SHA1_BLOCK_SIZE - 1);
-  
-
   p->buffer[pos++] = 0x80;
-  
+  if (pos > (SHA1_BLOCK_SIZE - 4 * 2))
-  if (pos > (64 - 8))
   {
-    while (pos != 64) { p->buffer[pos++] = 0; }
+    while (pos != SHA1_BLOCK_SIZE) { p->buffer[pos++] = 0; }
-    // memset(&p->buf.buffer[pos], 0, 64 - pos);
+    // memset(&p->buf.buffer[pos], 0, SHA1_BLOCK_SIZE - pos);
     Sha1_UpdateBlock(p);
     pos = 0;
   }
-
+  memset(&p->buffer[pos], 0, (SHA1_BLOCK_SIZE - 4 * 2) - pos);
-  /*
-  if (pos & 3)
   {
-    p->buffer[pos] = 0;
+    const UInt64 numBits = p->v.vars.count << 3;
-    p->buffer[pos + 1] = 0;
+    SetBe32(p->buffer + SHA1_BLOCK_SIZE - 4 * 2, (UInt32)(numBits >> 32))
-    p->buffer[pos + 2] = 0;
+    SetBe32(p->buffer + SHA1_BLOCK_SIZE - 4 * 1, (UInt32)(numBits))
-    pos += 3;
-    pos &= ~3;
   }
-  {
-    for (; pos < 64 - 8; pos += 4)
-      *(UInt32 *)(&p->buffer[pos]) = 0;
-  }
-  */
-
-  memset(&p->buffer[pos], 0, (64 - 8) - pos);
-
-  {
-    UInt64 numBits = (p->count << 3);
-    SetBe32(p->buffer + 64 - 8, (UInt32)(numBits >> 32));
-    SetBe32(p->buffer + 64 - 4, (UInt32)(numBits));
-  }
-  
   Sha1_UpdateBlock(p);
 
-  SetBe32(digest,      p->state[0]);
+  SetBe32(digest,      p->state[0])
-  SetBe32(digest + 4,  p->state[1]);
+  SetBe32(digest + 4,  p->state[1])
-  SetBe32(digest + 8,  p->state[2]);
+  SetBe32(digest + 8,  p->state[2])
-  SetBe32(digest + 12, p->state[3]);
+  SetBe32(digest + 12, p->state[3])
-  SetBe32(digest + 16, p->state[4]);
+  SetBe32(digest + 16, p->state[4])
-  
-
-
   
   Sha1_InitState(p);
 }
@@ -383,11 +352,11 @@ void Sha1_Final(CSha1 *p, Byte *digest)
 
 void Sha1_PrepareBlock(const CSha1 *p, Byte *block, unsigned size)
 {
-  const UInt64 numBits = (p->count + size) << 3;
+  const UInt64 numBits = (p->v.vars.count + size) << 3;
-  SetBe32(&((UInt32 *)(void *)block)[SHA1_NUM_BLOCK_WORDS - 2], (UInt32)(numBits >> 32));
+  SetBe32(&((UInt32 *)(void *)block)[SHA1_NUM_BLOCK_WORDS - 2], (UInt32)(numBits >> 32))
-  SetBe32(&((UInt32 *)(void *)block)[SHA1_NUM_BLOCK_WORDS - 1], (UInt32)(numBits));
+  SetBe32(&((UInt32 *)(void *)block)[SHA1_NUM_BLOCK_WORDS - 1], (UInt32)(numBits))
   // SetBe32((UInt32 *)(block + size), 0x80000000);
-  SetUi32((UInt32 *)(void *)(block + size), 0x80);
+  SetUi32((UInt32 *)(void *)(block + size), 0x80)
   size += 4;
   while (size != (SHA1_NUM_BLOCK_WORDS - 2) * 4)
   {
@@ -407,67 +376,66 @@ void Sha1_GetBlockDigest(const CSha1 *p, const Byte *data, Byte *destDigest)
   st[3] = p->state[3];
   st[4] = p->state[4];
   
-  UPDATE_BLOCKS(p)(st, data, 1);
+  SHA1_UPDATE_BLOCKS(p)(st, data, 1);
 
-  SetBe32(destDigest + 0    , st[0]);
+  SetBe32(destDigest + 0    , st[0])
-  SetBe32(destDigest + 1 * 4, st[1]);
+  SetBe32(destDigest + 1 * 4, st[1])
-  SetBe32(destDigest + 2 * 4, st[2]);
+  SetBe32(destDigest + 2 * 4, st[2])
-  SetBe32(destDigest + 3 * 4, st[3]);
+  SetBe32(destDigest + 3 * 4, st[3])
-  SetBe32(destDigest + 4 * 4, st[4]);
+  SetBe32(destDigest + 4 * 4, st[4])
 }
 
 
-void Sha1Prepare()
+void Sha1Prepare(void)
 {
-  #ifdef _SHA_SUPPORTED
+#ifdef Z7_COMPILER_SHA1_SUPPORTED
   SHA1_FUNC_UPDATE_BLOCKS f, f_hw;
   f = Sha1_UpdateBlocks;
   f_hw = NULL;
-  #ifdef MY_CPU_X86_OR_AMD64
+#ifdef MY_CPU_X86_OR_AMD64
-  #ifndef USE_MY_MM
   if (CPU_IsSupported_SHA()
       && CPU_IsSupported_SSSE3()
-      // && CPU_IsSupported_SSE41()
       )
-  #endif
+#else
-  #else
   if (CPU_IsSupported_SHA1())
-  #endif
+#endif
   {
     // printf("\n========== HW SHA1 ======== \n");
-    #if defined(MY_CPU_ARM_OR_ARM64) && defined(_MSC_VER)
+#if 1 && defined(MY_CPU_ARM_OR_ARM64) && defined(Z7_MSC_VER_ORIGINAL) && (_MSC_FULL_VER < 192930037)
     /* there was bug in MSVC compiler for ARM64 -O2 before version VS2019 16.10 (19.29.30037).
-       It generated incorrect SHA-1 code.
+       It generated incorrect SHA-1 code. */
-       21.03 : we test sha1-hardware code at runtime initialization */
+      #pragma message("== SHA1 code can work incorrectly with this compiler")
-
+      #error Stop_Compiling_MSC_Compiler_BUG_SHA1
-      #pragma message("== SHA1 code: MSC compiler : failure-check code was inserted")
+#endif
-
-      UInt32 state[5] = { 0, 1, 2, 3, 4 } ;
-      Byte data[64];
-      unsigned i;
-      for (i = 0; i < sizeof(data); i += 2)
-      {
-        data[i    ] = (Byte)(i);
-        data[i + 1] = (Byte)(i + 1);
-      }
-
-      Sha1_UpdateBlocks_HW(state, data, sizeof(data) / 64);
-    
-      if (   state[0] != 0x9acd7297
-          || state[1] != 0x4624d898
-          || state[2] != 0x0bf079f0
-          || state[3] != 0x031e61b3
-          || state[4] != 0x8323fe20)
-      {
-        // printf("\n========== SHA-1 hardware version failure ======== \n");
-      }
-      else
-    #endif
       {
         f = f_hw = Sha1_UpdateBlocks_HW;
       }
   }
-  g_FUNC_UPDATE_BLOCKS    = f;
+  g_SHA1_FUNC_UPDATE_BLOCKS    = f;
-  g_FUNC_UPDATE_BLOCKS_HW = f_hw;
+  g_SHA1_FUNC_UPDATE_BLOCKS_HW = f_hw;
-  #endif
+#endif
 }
+
+#undef kNumW
+#undef w
+#undef w0
+#undef w1
+#undef f0
+#undef f1
+#undef f2
+#undef f3
+#undef T1
+#undef T5
+#undef M5
+#undef R1
+#undef R2
+#undef R4
+#undef R5
+#undef R20_START
+#undef R_PRE
+#undef R_MAIN
+#undef STEP_PRE
+#undef STEP_MAIN
+#undef Z7_SHA1_BIG_W
+#undef Z7_SHA1_UNROLL
+#undef Z7_COMPILER_SHA1_SUPPORTED

C/Sha1.h

@@ -1,8 +1,8 @@
 /* Sha1.h -- SHA-1 Hash
-2021-02-08 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __7Z_SHA1_H
+#ifndef ZIP7_INC_SHA1_H
-#define __7Z_SHA1_H
+#define ZIP7_INC_SHA1_H
 
 #include "7zTypes.h"
 
@@ -14,7 +14,10 @@ EXTERN_C_BEGIN
 #define SHA1_BLOCK_SIZE   (SHA1_NUM_BLOCK_WORDS * 4)
 #define SHA1_DIGEST_SIZE  (SHA1_NUM_DIGEST_WORDS * 4)
 
-typedef void (MY_FAST_CALL *SHA1_FUNC_UPDATE_BLOCKS)(UInt32 state[5], const Byte *data, size_t numBlocks);
+
+
+
+typedef void (Z7_FASTCALL *SHA1_FUNC_UPDATE_BLOCKS)(UInt32 state[5], const Byte *data, size_t numBlocks);
 
 /*
   if (the system supports different SHA1 code implementations)
@@ -32,11 +35,18 @@ typedef void (MY_FAST_CALL *SHA1_FUNC_UPDATE_BLOCKS)(UInt32 state[5], const Byte
 
 typedef struct
 {
-  SHA1_FUNC_UPDATE_BLOCKS func_UpdateBlocks;
+  union
-  UInt64 count;
+  {
-  UInt64 __pad_2[2];
+    struct
+    {
+      SHA1_FUNC_UPDATE_BLOCKS func_UpdateBlocks;
+      UInt64 count;
+    } vars;
+    UInt64 _pad_64bit[4];
+    void *_pad_align_ptr[2];
+  } v;
   UInt32 state[SHA1_NUM_DIGEST_WORDS];
-  UInt32 __pad_3[3];
+  UInt32 _pad_3[3];
   Byte buffer[SHA1_BLOCK_SIZE];
 } CSha1;
 
@@ -62,7 +72,7 @@ void Sha1_Final(CSha1 *p, Byte *digest);
 void Sha1_PrepareBlock(const CSha1 *p, Byte *block, unsigned size);
 void Sha1_GetBlockDigest(const CSha1 *p, const Byte *data, Byte *destDigest);
 
-// void MY_FAST_CALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks);
+// void Z7_FASTCALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks);
 
 /*
 call Sha1Prepare() once at program start.

C/Sha1Opt.c

@@ -1,71 +1,53 @@
 /* Sha1Opt.c -- SHA-1 optimized code for SHA-1 hardware instructions
-2021-04-01 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
-
+#include "Compiler.h"
-#if defined(_MSC_VER)
-#if (_MSC_VER < 1900) && (_MSC_VER >= 1200)
-// #define USE_MY_MM
-#endif
-#endif
-
 #include "CpuArch.h"
 
+// #define Z7_USE_HW_SHA_STUB // for debug
 #ifdef MY_CPU_X86_OR_AMD64
-  #if defined(__clang__)
+  #if defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1600) // fix that check
-    #if (__clang_major__ >= 8) // fix that check
       #define USE_HW_SHA
-      #ifndef __SHA__
+  #elif defined(Z7_LLVM_CLANG_VERSION)  && (Z7_LLVM_CLANG_VERSION  >= 30800) \
+     || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 50100) \
+     || defined(Z7_GCC_VERSION)         && (Z7_GCC_VERSION         >= 40900)
+      #define USE_HW_SHA
+      #if !defined(__INTEL_COMPILER)
+      // icc defines __GNUC__, but icc doesn't support __attribute__(__target__)
+      #if !defined(__SHA__) || !defined(__SSSE3__)
         #define ATTRIB_SHA __attribute__((__target__("sha,ssse3")))
-        #if defined(_MSC_VER)
-          // SSSE3: for clang-cl:
-          #include <tmmintrin.h>
-          #define __SHA__
-        #endif
       #endif
-      #pragma clang diagnostic ignored "-Wvector-conversion"
-    #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 8) // fix that check
-      #define USE_HW_SHA
-      #ifndef __SHA__
-        #define ATTRIB_SHA __attribute__((__target__("sha,ssse3")))
-        // #pragma GCC target("sha,ssse3")
       #endif
-    #endif
-  #elif defined(__INTEL_COMPILER)
-    #if (__INTEL_COMPILER >= 1800) // fix that check
-      #define USE_HW_SHA
-    #endif
   #elif defined(_MSC_VER)
-    #ifdef USE_MY_MM
+    #if (_MSC_VER >= 1900)
-      #define USE_VER_MIN 1300
-    #else
-      #define USE_VER_MIN 1910
-    #endif
-    #if _MSC_VER >= USE_VER_MIN
       #define USE_HW_SHA
+    #else
+      #define Z7_USE_HW_SHA_STUB
     #endif
   #endif
 // #endif // MY_CPU_X86_OR_AMD64
+#ifndef USE_HW_SHA
+  // #define Z7_USE_HW_SHA_STUB // for debug
+#endif
 
 #ifdef USE_HW_SHA
 
 // #pragma message("Sha1 HW")
-// #include <wmmintrin.h>
 
-#if !defined(_MSC_VER) || (_MSC_VER >= 1900)
+
+
+
+// sse/sse2/ssse3:
+#include <tmmintrin.h>
+// sha*:
 #include <immintrin.h>
+
+#if defined (__clang__) && defined(_MSC_VER)
+  #if !defined(__SHA__)
+    #include <shaintrin.h>
+  #endif
 #else
-#include <emmintrin.h>
-
-#if defined(_MSC_VER) && (_MSC_VER >= 1600)
-// #include <intrin.h>
-#endif
-
-#ifdef USE_MY_MM
-#include "My_mm.h"
-#endif
 
 #endif
 
@@ -87,84 +69,69 @@ SHA:
   _mm_sha1*
 */
 
-#define ADD_EPI32(dest, src)      dest = _mm_add_epi32(dest, src);
 #define XOR_SI128(dest, src)      dest = _mm_xor_si128(dest, src);
 #define SHUFFLE_EPI8(dest, mask)  dest = _mm_shuffle_epi8(dest, mask);
 #define SHUFFLE_EPI32(dest, mask) dest = _mm_shuffle_epi32(dest, mask);
-
+#ifdef __clang__
-#define SHA1_RND4(abcd, e0, f)  abcd = _mm_sha1rnds4_epu32(abcd, e0, f);
+#define SHA1_RNDS4_RET_TYPE_CAST (__m128i)
-#define SHA1_NEXTE(e, m)        e = _mm_sha1nexte_epu32(e, m);
+#else
-
+#define SHA1_RNDS4_RET_TYPE_CAST
-
+#endif
-
+#define SHA1_RND4(abcd, e0, f)    abcd = SHA1_RNDS4_RET_TYPE_CAST _mm_sha1rnds4_epu32(abcd, e0, f);
-
+#define SHA1_NEXTE(e, m)          e = _mm_sha1nexte_epu32(e, m);
-
+#define ADD_EPI32(dest, src)      dest = _mm_add_epi32(dest, src);
-#define SHA1_MSG1(dest, src) dest = _mm_sha1msg1_epu32(dest, src);
+#define SHA1_MSG1(dest, src)      dest = _mm_sha1msg1_epu32(dest, src);
-#define SHA1_MSG2(dest, src) dest = _mm_sha1msg2_epu32(dest, src);
+#define SHA1_MSG2(dest, src)      dest = _mm_sha1msg2_epu32(dest, src);
-
 
 #define LOAD_SHUFFLE(m, k) \
     m = _mm_loadu_si128((const __m128i *)(const void *)(data + (k) * 16)); \
-    SHUFFLE_EPI8(m, mask); \
+    SHUFFLE_EPI8(m, mask) \
-
-#define SM1(m0, m1, m2, m3) \
-    SHA1_MSG1(m0, m1); \
-
-#define SM2(m0, m1, m2, m3) \
-    XOR_SI128(m3, m1); \
-    SHA1_MSG2(m3, m2); \
-
-#define SM3(m0, m1, m2, m3) \
-    XOR_SI128(m3, m1); \
-    SM1(m0, m1, m2, m3) \
-    SHA1_MSG2(m3, m2); \
 
 #define NNN(m0, m1, m2, m3)
 
+#define SM1(m0, m1, m2, m3) \
+    SHA1_MSG1(m0, m1) \
 
+#define SM2(m0, m1, m2, m3) \
+    XOR_SI128(m3, m1) \
+    SHA1_MSG2(m3, m2) \
 
+#define SM3(m0, m1, m2, m3) \
+    XOR_SI128(m3, m1) \
+    SM1(m0, m1, m2, m3) \
+    SHA1_MSG2(m3, m2) \
 
-
+#define R4(k, m0, m1, m2, m3, e0, e1, OP) \
-
-
-
-
-
-
-
-
-
-
-
-
-#define R4(k, e0, e1, m0, m1, m2, m3, OP) \
     e1 = abcd; \
-    SHA1_RND4(abcd, e0, (k) / 5); \
+    SHA1_RND4(abcd, e0, (k) / 5) \
-    SHA1_NEXTE(e1, m1); \
+    SHA1_NEXTE(e1, m1) \
-    OP(m0, m1, m2, m3); \
+    OP(m0, m1, m2, m3) \
+
+
 
 #define R16(k, mx, OP0, OP1, OP2, OP3) \
-    R4 ( (k)*4+0, e0,e1, m0,m1,m2,m3, OP0 ) \
+    R4 ( (k)*4+0, m0,m1,m2,m3, e0,e1, OP0 ) \
-    R4 ( (k)*4+1, e1,e0, m1,m2,m3,m0, OP1 ) \
+    R4 ( (k)*4+1, m1,m2,m3,m0, e1,e0, OP1 ) \
-    R4 ( (k)*4+2, e0,e1, m2,m3,m0,m1, OP2 ) \
+    R4 ( (k)*4+2, m2,m3,m0,m1, e0,e1, OP2 ) \
-    R4 ( (k)*4+3, e1,e0, m3,mx,m1,m2, OP3 ) \
+    R4 ( (k)*4+3, m3,mx,m1,m2, e1,e0, OP3 ) \
 
 #define PREPARE_STATE \
-    SHUFFLE_EPI32 (abcd, 0x1B); \
+    SHUFFLE_EPI32 (abcd, 0x1B) \
-    SHUFFLE_EPI32 (e0,   0x1B); \
+    SHUFFLE_EPI32 (e0,   0x1B) \
 
 
 
 
 
-void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks);
 #ifdef ATTRIB_SHA
 ATTRIB_SHA
 #endif
-void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks)
 {
   const __m128i mask = _mm_set_epi32(0x00010203, 0x04050607, 0x08090a0b, 0x0c0d0e0f);
 
+  
   __m128i abcd, e0;
 
   if (numBlocks == 0)
@@ -190,15 +157,15 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t
     LOAD_SHUFFLE (m2, 2)
     LOAD_SHUFFLE (m3, 3)
 
-    ADD_EPI32(e0, m0);
+    ADD_EPI32(e0, m0)
     
-    R16 ( 0, m0, SM1, SM3, SM3, SM3 );
+    R16 ( 0, m0, SM1, SM3, SM3, SM3 )
-    R16 ( 1, m0, SM3, SM3, SM3, SM3 );
+    R16 ( 1, m0, SM3, SM3, SM3, SM3 )
-    R16 ( 2, m0, SM3, SM3, SM3, SM3 );
+    R16 ( 2, m0, SM3, SM3, SM3, SM3 )
-    R16 ( 3, m0, SM3, SM3, SM3, SM3 );
+    R16 ( 3, m0, SM3, SM3, SM3, SM3 )
-    R16 ( 4, e2, SM2, NNN, NNN, NNN );
+    R16 ( 4, e2, SM2, NNN, NNN, NNN )
     
-    ADD_EPI32(abcd, abcd_save);
+    ADD_EPI32(abcd, abcd_save)
     
     data += 64;
   }
@@ -207,78 +174,155 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t
   PREPARE_STATE
 
   _mm_storeu_si128((__m128i *) (void *) state, abcd);
-  *(state+4) = (UInt32)_mm_cvtsi128_si32(e0);
+  *(state + 4) = (UInt32)_mm_cvtsi128_si32(e0);
 }
 
 #endif // USE_HW_SHA
 
-#elif defined(MY_CPU_ARM_OR_ARM64)
+#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE) \
-
+   && (!defined(Z7_MSC_VER_ORIGINAL) || (_MSC_VER >= 1929) && (_MSC_FULL_VER >= 192930037))
-  #if defined(__clang__)
+  #if   defined(__ARM_FEATURE_SHA2) \
-    #if (__clang_major__ >= 8) // fix that check
+     || defined(__ARM_FEATURE_CRYPTO)
+    #define USE_HW_SHA
+  #else
+    #if  defined(MY_CPU_ARM64) \
+      || defined(__ARM_ARCH) && (__ARM_ARCH >= 4) \
+      || defined(Z7_MSC_VER_ORIGINAL)
+    #if  defined(__ARM_FP) && \
+          (   defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30800) \
+           || defined(__GNUC__) && (__GNUC__ >= 6) \
+          ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1910)
+    #if  defined(MY_CPU_ARM64) \
+      || !defined(Z7_CLANG_VERSION) \
+      || defined(__ARM_NEON) && \
+          (Z7_CLANG_VERSION < 170000 || \
+           Z7_CLANG_VERSION > 170001)
       #define USE_HW_SHA
     #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 6) // fix that check
-      #define USE_HW_SHA
     #endif
-  #elif defined(_MSC_VER)
-    #if _MSC_VER >= 1910
-      #define USE_HW_SHA
     #endif
   #endif
 
 #ifdef USE_HW_SHA
 
 // #pragma message("=== Sha1 HW === ")
+// __ARM_FEATURE_CRYPTO macro is deprecated in favor of the finer grained feature macro __ARM_FEATURE_SHA2
 
 #if defined(__clang__) || defined(__GNUC__)
+#if !defined(__ARM_FEATURE_SHA2) && \
+    !defined(__ARM_FEATURE_CRYPTO)
   #ifdef MY_CPU_ARM64
+#if defined(__clang__)
+    #define ATTRIB_SHA __attribute__((__target__("crypto")))
+#else
     #define ATTRIB_SHA __attribute__((__target__("+crypto")))
+#endif
   #else
+#if defined(__clang__) && (__clang_major__ >= 1)
+    #define ATTRIB_SHA __attribute__((__target__("armv8-a,sha2")))
+#else
     #define ATTRIB_SHA __attribute__((__target__("fpu=crypto-neon-fp-armv8")))
+#endif
   #endif
+#endif
 #else
   // _MSC_VER
   // for arm32
   #define _ARM_USE_NEW_NEON_INTRINSICS
 #endif
 
-#if defined(_MSC_VER) && defined(MY_CPU_ARM64)
+#if defined(Z7_MSC_VER_ORIGINAL) && defined(MY_CPU_ARM64)
 #include <arm64_neon.h>
 #else
-#include <arm_neon.h>
+
+#if defined(__clang__) && __clang_major__ < 16
+#if !defined(__ARM_FEATURE_SHA2) && \
+    !defined(__ARM_FEATURE_CRYPTO)
+//     #pragma message("=== we set __ARM_FEATURE_CRYPTO 1 === ")
+    Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #define Z7_ARM_FEATURE_CRYPTO_WAS_SET 1
+// #if defined(__clang__) && __clang_major__ < 13
+    #define __ARM_FEATURE_CRYPTO 1
+// #else
+    #define __ARM_FEATURE_SHA2 1
+// #endif
+    Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
 #endif
+#endif // clang
+
+#if defined(__clang__)
+
+#if defined(__ARM_ARCH) && __ARM_ARCH < 8
+    Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+//    #pragma message("#define __ARM_ARCH 8")
+    #undef  __ARM_ARCH
+    #define __ARM_ARCH 8
+    Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+#endif
+
+#endif // clang
+
+#include <arm_neon.h>
+
+#if defined(Z7_ARM_FEATURE_CRYPTO_WAS_SET) && \
+    defined(__ARM_FEATURE_CRYPTO) && \
+    defined(__ARM_FEATURE_SHA2)
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #undef __ARM_FEATURE_CRYPTO
+    #undef __ARM_FEATURE_SHA2
+    #undef Z7_ARM_FEATURE_CRYPTO_WAS_SET
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+//    #pragma message("=== we undefine __ARM_FEATURE_CRYPTO === ")
+#endif
+
+#endif // Z7_MSC_VER_ORIGINAL
 
 typedef uint32x4_t v128;
 // typedef __n128 v128; // MSVC
-
+// the bug in clang 3.8.1:
-#ifdef MY_CPU_BE
+// __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1);
-  #define MY_rev32_for_LE(x)
+#if defined(__clang__) && (__clang_major__ <= 9)
-#else
+#pragma GCC diagnostic ignored "-Wvector-conversion"
-  #define MY_rev32_for_LE(x) x = vreinterpretq_u32_u8(vrev32q_u8(vreinterpretq_u8_u32(x)))
 #endif
 
-#define LOAD_128(_p)      (*(const v128 *)(const void *)(_p))
+#ifdef MY_CPU_BE
-#define STORE_128(_p, _v) *(v128 *)(void *)(_p) = (_v)
+  #define MY_rev32_for_LE(x) x
+#else
+  #define MY_rev32_for_LE(x) vrev32q_u8(x)
+#endif
+
+#define LOAD_128_32(_p)       vld1q_u32(_p)
+#define LOAD_128_8(_p)        vld1q_u8 (_p)
+#define STORE_128_32(_p, _v)  vst1q_u32(_p, _v)
 
 #define LOAD_SHUFFLE(m, k) \
-    m = LOAD_128((data + (k) * 16)); \
+    m = vreinterpretq_u32_u8( \
-    MY_rev32_for_LE(m); \
+        MY_rev32_for_LE( \
+        LOAD_128_8(data + (k) * 16))); \
 
-#define SU0(dest, src2, src3) dest = vsha1su0q_u32(dest, src2, src3);
+#define N0(dest, src2, src3)
-#define SU1(dest, src)        dest = vsha1su1q_u32(dest, src);
+#define N1(dest, src)
-#define C(e)                  abcd = vsha1cq_u32(abcd, e, t);
+#define U0(dest, src2, src3)  dest = vsha1su0q_u32(dest, src2, src3);
-#define P(e)                  abcd = vsha1pq_u32(abcd, e, t);
+#define U1(dest, src)         dest = vsha1su1q_u32(dest, src);
-#define M(e)                  abcd = vsha1mq_u32(abcd, e, t);
+#define C(e)                  abcd = vsha1cq_u32(abcd, e, t)
+#define P(e)                  abcd = vsha1pq_u32(abcd, e, t)
+#define M(e)                  abcd = vsha1mq_u32(abcd, e, t)
 #define H(e)                  e = vsha1h_u32(vgetq_lane_u32(abcd, 0))
 #define T(m, c)               t = vaddq_u32(m, c)
 
-void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
+#define R16(d0,d1,d2,d3, f0,z0, f1,z1, f2,z2, f3,z3, w0,w1,w2,w3) \
+    T(m0, d0);  f0(m3, m0, m1)  z0(m2, m1)  H(e1);  w0(e0); \
+    T(m1, d1);  f1(m0, m1, m2)  z1(m3, m2)  H(e0);  w1(e1); \
+    T(m2, d2);  f2(m1, m2, m3)  z2(m0, m3)  H(e1);  w2(e0); \
+    T(m3, d3);  f3(m2, m3, m0)  z3(m1, m0)  H(e0);  w3(e1); \
+
+
+void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
 #ifdef ATTRIB_SHA
 ATTRIB_SHA
 #endif
-void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
 {
   v128 abcd;
   v128 c0, c1, c2, c3;
@@ -292,7 +336,7 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t
   c2 = vdupq_n_u32(0x8f1bbcdc);
   c3 = vdupq_n_u32(0xca62c1d6);
 
-  abcd = LOAD_128(&state[0]);
+  abcd = LOAD_128_32(&state[0]);
   e0 = state[4];
   
   do
@@ -310,26 +354,11 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t
     LOAD_SHUFFLE (m2, 2)
     LOAD_SHUFFLE (m3, 3)
                      
-    T(m0, c0);                                  H(e1); C(e0);
+    R16 ( c0,c0,c0,c0, N0,N1, U0,N1, U0,U1, U0,U1, C,C,C,C )
-    T(m1, c0);  SU0(m0, m1, m2);                H(e0); C(e1);
+    R16 ( c0,c1,c1,c1, U0,U1, U0,U1, U0,U1, U0,U1, C,P,P,P )
-    T(m2, c0);  SU0(m1, m2, m3);  SU1(m0, m3);  H(e1); C(e0);
+    R16 ( c1,c1,c2,c2, U0,U1, U0,U1, U0,U1, U0,U1, P,P,M,M )
-    T(m3, c0);  SU0(m2, m3, m0);  SU1(m1, m0);  H(e0); C(e1);
+    R16 ( c2,c2,c2,c3, U0,U1, U0,U1, U0,U1, U0,U1, M,M,M,P )
-    T(m0, c0);  SU0(m3, m0, m1);  SU1(m2, m1);  H(e1); C(e0);
+    R16 ( c3,c3,c3,c3, U0,U1, N0,U1, N0,N1, N0,N1, P,P,P,P )
-    T(m1, c1);  SU0(m0, m1, m2);  SU1(m3, m2);  H(e0); P(e1);
-    T(m2, c1);  SU0(m1, m2, m3);  SU1(m0, m3);  H(e1); P(e0);
-    T(m3, c1);  SU0(m2, m3, m0);  SU1(m1, m0);  H(e0); P(e1);
-    T(m0, c1);  SU0(m3, m0, m1);  SU1(m2, m1);  H(e1); P(e0);
-    T(m1, c1);  SU0(m0, m1, m2);  SU1(m3, m2);  H(e0); P(e1);
-    T(m2, c2);  SU0(m1, m2, m3);  SU1(m0, m3);  H(e1); M(e0);
-    T(m3, c2);  SU0(m2, m3, m0);  SU1(m1, m0);  H(e0); M(e1);
-    T(m0, c2);  SU0(m3, m0, m1);  SU1(m2, m1);  H(e1); M(e0);
-    T(m1, c2);  SU0(m0, m1, m2);  SU1(m3, m2);  H(e0); M(e1);
-    T(m2, c2);  SU0(m1, m2, m3);  SU1(m0, m3);  H(e1); M(e0);
-    T(m3, c3);  SU0(m2, m3, m0);  SU1(m1, m0);  H(e0); P(e1);
-    T(m0, c3);  SU0(m3, m0, m1);  SU1(m2, m1);  H(e1); P(e0);
-    T(m1, c3);                    SU1(m3, m2);  H(e0); P(e1);
-    T(m2, c3);                                  H(e1); P(e0);
-    T(m3, c3);                                  H(e0); P(e1);
                                                                                                                      
     abcd = vaddq_u32(abcd, abcd_save);
     e0 += e0_save;
@@ -338,7 +367,7 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t
   }
   while (--numBlocks);
 
-  STORE_128(&state[0], abcd);
+  STORE_128_32(&state[0], abcd);
   state[4] = e0;
 }
 
@@ -346,19 +375,16 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t
 
 #endif // MY_CPU_ARM_OR_ARM64
 
-
+#if !defined(USE_HW_SHA) && defined(Z7_USE_HW_SHA_STUB)
-#ifndef USE_HW_SHA
-
 // #error Stop_Compiling_UNSUPPORTED_SHA
 // #include <stdlib.h>
-
 // #include "Sha1.h"
-void MY_FAST_CALL Sha1_UpdateBlocks(UInt32 state[5], const Byte *data, size_t numBlocks);
+// #if defined(_MSC_VER)
-
 #pragma message("Sha1   HW-SW stub was used")
-
+// #endif
-void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha1_UpdateBlocks   (UInt32 state[5], const Byte *data, size_t numBlocks);
-void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t numBlocks)
 {
   Sha1_UpdateBlocks(state, data, numBlocks);
   /*
@@ -369,5 +395,30 @@ void MY_FAST_CALL Sha1_UpdateBlocks_HW(UInt32 state[5], const Byte *data, size_t
   return;
   */
 }
-
 #endif
+
+#undef U0
+#undef U1
+#undef N0
+#undef N1
+#undef C
+#undef P
+#undef M
+#undef H
+#undef T
+#undef MY_rev32_for_LE
+#undef NNN
+#undef LOAD_128
+#undef STORE_128
+#undef LOAD_SHUFFLE
+#undef SM1
+#undef SM2
+#undef SM3
+#undef NNN
+#undef R4
+#undef R16
+#undef PREPARE_STATE
+#undef USE_HW_SHA
+#undef ATTRIB_SHA
+#undef USE_VER_MIN
+#undef Z7_USE_HW_SHA_STUB

C/Sha256.c

@@ -1,64 +1,60 @@
 /* Sha256.c -- SHA-256 Hash
-2021-04-01 : Igor Pavlov : Public domain
+: Igor Pavlov : Public domain
 This code is based on public domain code from Wei Dai's Crypto++ library. */
 
 #include "Precomp.h"
 
 #include <string.h>
 
-#include "CpuArch.h"
-#include "RotateDefs.h"
 #include "Sha256.h"
-
+#include "RotateDefs.h"
-#if defined(_MSC_VER) && (_MSC_VER < 1900)
+#include "CpuArch.h"
-// #define USE_MY_MM
-#endif
 
 #ifdef MY_CPU_X86_OR_AMD64
-  #ifdef _MSC_VER
+  #if   defined(Z7_LLVM_CLANG_VERSION)  && (Z7_LLVM_CLANG_VERSION  >= 30800) \
-    #if _MSC_VER >= 1200
+     || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 50100) \
-      #define _SHA_SUPPORTED
+     || defined(Z7_GCC_VERSION)         && (Z7_GCC_VERSION         >= 40900) \
-    #endif
+     || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1600) \
-  #elif defined(__clang__)
+     || defined(_MSC_VER) && (_MSC_VER >= 1200)
-    #if (__clang_major__ >= 8) // fix that check
+      #define Z7_COMPILER_SHA256_SUPPORTED
-      #define _SHA_SUPPORTED
-    #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 8) // fix that check
-      #define _SHA_SUPPORTED
-    #endif
-  #elif defined(__INTEL_COMPILER)
-    #if (__INTEL_COMPILER >= 1800) // fix that check
-      #define _SHA_SUPPORTED
-    #endif
   #endif
-#elif defined(MY_CPU_ARM_OR_ARM64)
+#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE)
-  #ifdef _MSC_VER
+
-    #if _MSC_VER >= 1910
+  #if   defined(__ARM_FEATURE_SHA2) \
-      #define _SHA_SUPPORTED
+     || defined(__ARM_FEATURE_CRYPTO)
+    #define Z7_COMPILER_SHA256_SUPPORTED
+  #else
+    #if  defined(MY_CPU_ARM64) \
+      || defined(__ARM_ARCH) && (__ARM_ARCH >= 4) \
+      || defined(Z7_MSC_VER_ORIGINAL)
+    #if  defined(__ARM_FP) && \
+          (   defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30800) \
+           || defined(__GNUC__) && (__GNUC__ >= 6) \
+          ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1910)
+    #if  defined(MY_CPU_ARM64) \
+      || !defined(Z7_CLANG_VERSION) \
+      || defined(__ARM_NEON) && \
+          (Z7_CLANG_VERSION < 170000 || \
+           Z7_CLANG_VERSION > 170001)
+      #define Z7_COMPILER_SHA256_SUPPORTED
     #endif
-  #elif defined(__clang__)
-    #if (__clang_major__ >= 8) // fix that check
-      #define _SHA_SUPPORTED
     #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 6) // fix that check
-      #define _SHA_SUPPORTED
     #endif
   #endif
 #endif
 
-void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks);
 
-#ifdef _SHA_SUPPORTED
+#ifdef Z7_COMPILER_SHA256_SUPPORTED
-  void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
+  void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
 
-  static SHA256_FUNC_UPDATE_BLOCKS g_FUNC_UPDATE_BLOCKS = Sha256_UpdateBlocks;
+  static SHA256_FUNC_UPDATE_BLOCKS g_SHA256_FUNC_UPDATE_BLOCKS = Sha256_UpdateBlocks;
-  static SHA256_FUNC_UPDATE_BLOCKS g_FUNC_UPDATE_BLOCKS_HW;
+  static SHA256_FUNC_UPDATE_BLOCKS g_SHA256_FUNC_UPDATE_BLOCKS_HW;
 
-  #define UPDATE_BLOCKS(p) p->func_UpdateBlocks
+  #define SHA256_UPDATE_BLOCKS(p) p->v.vars.func_UpdateBlocks
 #else
-  #define UPDATE_BLOCKS(p) Sha256_UpdateBlocks
+  #define SHA256_UPDATE_BLOCKS(p) Sha256_UpdateBlocks
 #endif
 
 
@@ -66,16 +62,16 @@ BoolInt Sha256_SetFunction(CSha256 *p, unsigned algo)
 {
   SHA256_FUNC_UPDATE_BLOCKS func = Sha256_UpdateBlocks;
   
-  #ifdef _SHA_SUPPORTED
+  #ifdef Z7_COMPILER_SHA256_SUPPORTED
     if (algo != SHA256_ALGO_SW)
     {
       if (algo == SHA256_ALGO_DEFAULT)
-        func = g_FUNC_UPDATE_BLOCKS;
+        func = g_SHA256_FUNC_UPDATE_BLOCKS;
       else
       {
         if (algo != SHA256_ALGO_HW)
           return False;
-        func = g_FUNC_UPDATE_BLOCKS_HW;
+        func = g_SHA256_FUNC_UPDATE_BLOCKS_HW;
         if (!func)
           return False;
       }
@@ -85,24 +81,25 @@ BoolInt Sha256_SetFunction(CSha256 *p, unsigned algo)
       return False;
   #endif
 
-  p->func_UpdateBlocks = func;
+  p->v.vars.func_UpdateBlocks = func;
   return True;
 }
 
 
 /* define it for speed optimization */
 
-#ifdef _SFX
+#ifdef Z7_SFX
   #define STEP_PRE 1
   #define STEP_MAIN 1
 #else
   #define STEP_PRE 2
   #define STEP_MAIN 4
-  // #define _SHA256_UNROLL
+  // #define Z7_SHA256_UNROLL
 #endif
 
+#undef Z7_SHA256_BIG_W
 #if STEP_MAIN != 16
-  #define _SHA256_BIG_W
+  #define Z7_SHA256_BIG_W
 #endif
 
 
@@ -110,7 +107,7 @@ BoolInt Sha256_SetFunction(CSha256 *p, unsigned algo)
 
 void Sha256_InitState(CSha256 *p)
 {
-  p->count = 0;
+  p->v.vars.count = 0;
   p->state[0] = 0x6a09e667;
   p->state[1] = 0xbb67ae85;
   p->state[2] = 0x3c6ef372;
@@ -121,21 +118,28 @@ void Sha256_InitState(CSha256 *p)
   p->state[7] = 0x5be0cd19;
 }
 
+
+
+
+
+
+
+
 void Sha256_Init(CSha256 *p)
 {
-  p->func_UpdateBlocks =
+  p->v.vars.func_UpdateBlocks =
-  #ifdef _SHA_SUPPORTED
+  #ifdef Z7_COMPILER_SHA256_SUPPORTED
-      g_FUNC_UPDATE_BLOCKS;
+      g_SHA256_FUNC_UPDATE_BLOCKS;
   #else
       NULL;
   #endif
   Sha256_InitState(p);
 }
 
-#define S0(x) (rotrFixed(x, 2) ^ rotrFixed(x,13) ^ rotrFixed(x, 22))
+#define S0(x) (rotrFixed(x, 2) ^ rotrFixed(x,13) ^ rotrFixed(x,22))
-#define S1(x) (rotrFixed(x, 6) ^ rotrFixed(x,11) ^ rotrFixed(x, 25))
+#define S1(x) (rotrFixed(x, 6) ^ rotrFixed(x,11) ^ rotrFixed(x,25))
 #define s0(x) (rotrFixed(x, 7) ^ rotrFixed(x,18) ^ (x >> 3))
-#define s1(x) (rotrFixed(x,17) ^ rotrFixed(x,19) ^ (x >> 10))
+#define s1(x) (rotrFixed(x,17) ^ rotrFixed(x,19) ^ (x >>10))
 
 #define Ch(x,y,z) (z^(x&(y^z)))
 #define Maj(x,y,z) ((x&y)|(z&(x|y)))
@@ -145,7 +149,7 @@ void Sha256_Init(CSha256 *p)
 
 #define blk2_main(j, i)  s1(w(j, (i)-2)) + w(j, (i)-7) + s0(w(j, (i)-15))
 
-#ifdef _SHA256_BIG_W
+#ifdef Z7_SHA256_BIG_W
     // we use +i instead of +(i) to change the order to solve CLANG compiler warning for signed/unsigned.
     #define w(j, i)     W[(size_t)(j) + i]
     #define blk2(j, i)  (w(j, i) = w(j, (i)-16) + blk2_main(j, i))
@@ -176,7 +180,7 @@ void Sha256_Init(CSha256 *p)
 #define R1_PRE(i)  T1( W_PRE, i)
 #define R1_MAIN(i) T1( W_MAIN, i)
 
-#if (!defined(_SHA256_UNROLL) || STEP_MAIN < 8) && (STEP_MAIN >= 4)
+#if (!defined(Z7_SHA256_UNROLL) || STEP_MAIN < 8) && (STEP_MAIN >= 4)
 #define R2_MAIN(i) \
     R1_MAIN(i) \
     R1_MAIN(i + 1) \
@@ -185,7 +189,7 @@ void Sha256_Init(CSha256 *p)
 
 
 
-#if defined(_SHA256_UNROLL) && STEP_MAIN >= 8
+#if defined(Z7_SHA256_UNROLL) && STEP_MAIN >= 8
 
 #define T4( a,b,c,d,e,f,g,h, wx, i) \
     h += S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \
@@ -223,14 +227,10 @@ void Sha256_Init(CSha256 *p)
 
 #endif
 
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
 
-// static
+extern
-extern MY_ALIGN(64)
+MY_ALIGN(64) const UInt32 SHA256_K_ARRAY[64];
-const UInt32 SHA256_K_ARRAY[64];
+MY_ALIGN(64) const UInt32 SHA256_K_ARRAY[64] = {
-
-MY_ALIGN(64)
-const UInt32 SHA256_K_ARRAY[64] = {
   0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5,
   0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
   0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3,
@@ -249,26 +249,28 @@ const UInt32 SHA256_K_ARRAY[64] = {
   0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
 };
 
+
+
+
+
 #define K SHA256_K_ARRAY
 
-
+Z7_NO_INLINE
-MY_NO_INLINE
+void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks)
-void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks)
 {
   UInt32 W
-  #ifdef _SHA256_BIG_W
+#ifdef Z7_SHA256_BIG_W
       [64];
-  #else
+#else
       [16];
-  #endif
+#endif
-
   unsigned j;
-
   UInt32 a,b,c,d,e,f,g,h;
-
+#if !defined(Z7_SHA256_UNROLL) || (STEP_MAIN <= 4) || (STEP_PRE <= 4)
-  #if !defined(_SHA256_UNROLL) || (STEP_MAIN <= 4) || (STEP_PRE <= 4)
   UInt32 tmp;
-  #endif
+#endif
+  
+  if (numBlocks == 0) return;
 
   a = state[0];
   b = state[1];
@@ -279,7 +281,7 @@ void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t
   g = state[6];
   h = state[7];
 
-  while (numBlocks)
+  do
   {
 
   for (j = 0; j < 16; j += STEP_PRE)
@@ -297,12 +299,12 @@ void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t
 
     #else
 
-      R1_PRE(0);
+      R1_PRE(0)
       #if STEP_PRE >= 2
-      R1_PRE(1);
+      R1_PRE(1)
       #if STEP_PRE >= 4
-      R1_PRE(2);
+      R1_PRE(2)
-      R1_PRE(3);
+      R1_PRE(3)
       #endif
       #endif
     
@@ -311,32 +313,32 @@ void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t
 
   for (j = 16; j < 64; j += STEP_MAIN)
   {
-    #if defined(_SHA256_UNROLL) && STEP_MAIN >= 8
+    #if defined(Z7_SHA256_UNROLL) && STEP_MAIN >= 8
 
       #if STEP_MAIN < 8
-      R4_MAIN(0);
+      R4_MAIN(0)
       #else
-      R8_MAIN(0);
+      R8_MAIN(0)
       #if STEP_MAIN == 16
-      R8_MAIN(8);
+      R8_MAIN(8)
       #endif
       #endif
 
     #else
       
-      R1_MAIN(0);
+      R1_MAIN(0)
       #if STEP_MAIN >= 2
-      R1_MAIN(1);
+      R1_MAIN(1)
       #if STEP_MAIN >= 4
-      R2_MAIN(2);
+      R2_MAIN(2)
       #if STEP_MAIN >= 8
-      R2_MAIN(4);
+      R2_MAIN(4)
-      R2_MAIN(6);
+      R2_MAIN(6)
       #if STEP_MAIN >= 16
-      R2_MAIN(8);
+      R2_MAIN(8)
-      R2_MAIN(10);
+      R2_MAIN(10)
-      R2_MAIN(12);
+      R2_MAIN(12)
-      R2_MAIN(14);
+      R2_MAIN(14)
       #endif
       #endif
       #endif
@@ -353,40 +355,27 @@ void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t
   g += state[6]; state[6] = g;
   h += state[7]; state[7] = h;
 
-  data += 64;
+  data += SHA256_BLOCK_SIZE;
-  numBlocks--;
   }
-
+  while (--numBlocks);
-  /* Wipe variables */
-  /* memset(W, 0, sizeof(W)); */
 }
 
-#undef S0
-#undef S1
-#undef s0
-#undef s1
-#undef K
 
-#define Sha256_UpdateBlock(p) UPDATE_BLOCKS(p)(p->state, p->buffer, 1)
+#define Sha256_UpdateBlock(p) SHA256_UPDATE_BLOCKS(p)(p->state, p->buffer, 1)
 
 void Sha256_Update(CSha256 *p, const Byte *data, size_t size)
 {
   if (size == 0)
     return;
-
   {
-    unsigned pos = (unsigned)p->count & 0x3F;
+    const unsigned pos = (unsigned)p->v.vars.count & (SHA256_BLOCK_SIZE - 1);
-    unsigned num;
+    const unsigned num = SHA256_BLOCK_SIZE - pos;
-    
+    p->v.vars.count += size;
-    p->count += size;
-    
-    num = 64 - pos;
     if (num > size)
     {
       memcpy(p->buffer + pos, data, size);
       return;
     }
-    
     if (pos != 0)
     {
       size -= num;
@@ -396,9 +385,10 @@ void Sha256_Update(CSha256 *p, const Byte *data, size_t size)
     }
   }
   {
-    size_t numBlocks = size >> 6;
+    const size_t numBlocks = size >> 6;
-    UPDATE_BLOCKS(p)(p->state, data, numBlocks);
+    // if (numBlocks)
-    size &= 0x3F;
+    SHA256_UPDATE_BLOCKS(p)(p->state, data, numBlocks);
+    size &= SHA256_BLOCK_SIZE - 1;
     if (size == 0)
       return;
     data += (numBlocks << 6);
@@ -409,78 +399,94 @@ void Sha256_Update(CSha256 *p, const Byte *data, size_t size)
 
 void Sha256_Final(CSha256 *p, Byte *digest)
 {
-  unsigned pos = (unsigned)p->count & 0x3F;
+  unsigned pos = (unsigned)p->v.vars.count & (SHA256_BLOCK_SIZE - 1);
-  unsigned i;
-  
   p->buffer[pos++] = 0x80;
-  
+  if (pos > (SHA256_BLOCK_SIZE - 4 * 2))
-  if (pos > (64 - 8))
   {
-    while (pos != 64) { p->buffer[pos++] = 0; }
+    while (pos != SHA256_BLOCK_SIZE) { p->buffer[pos++] = 0; }
-    // memset(&p->buf.buffer[pos], 0, 64 - pos);
+    // memset(&p->buf.buffer[pos], 0, SHA256_BLOCK_SIZE - pos);
     Sha256_UpdateBlock(p);
     pos = 0;
   }
-
+  memset(&p->buffer[pos], 0, (SHA256_BLOCK_SIZE - 4 * 2) - pos);
-  /*
-  if (pos & 3)
   {
-    p->buffer[pos] = 0;
+    const UInt64 numBits = p->v.vars.count << 3;
-    p->buffer[pos + 1] = 0;
+    SetBe32(p->buffer + SHA256_BLOCK_SIZE - 4 * 2, (UInt32)(numBits >> 32))
-    p->buffer[pos + 2] = 0;
+    SetBe32(p->buffer + SHA256_BLOCK_SIZE - 4 * 1, (UInt32)(numBits))
-    pos += 3;
-    pos &= ~3;
   }
-  {
-    for (; pos < 64 - 8; pos += 4)
-      *(UInt32 *)(&p->buffer[pos]) = 0;
-  }
-  */
-
-  memset(&p->buffer[pos], 0, (64 - 8) - pos);
-
-  {
-    UInt64 numBits = (p->count << 3);
-    SetBe32(p->buffer + 64 - 8, (UInt32)(numBits >> 32));
-    SetBe32(p->buffer + 64 - 4, (UInt32)(numBits));
-  }
-  
   Sha256_UpdateBlock(p);
-
+#if 1 && defined(MY_CPU_BE)
-  for (i = 0; i < 8; i += 2)
+  memcpy(digest, p->state, SHA256_DIGEST_SIZE);
+#else
   {
-    UInt32 v0 = p->state[i];
+    unsigned i;
-    UInt32 v1 = p->state[(size_t)i + 1];
+    for (i = 0; i < 8; i += 2)
-    SetBe32(digest    , v0);
+    {
-    SetBe32(digest + 4, v1);
+      const UInt32 v0 = p->state[i];
-    digest += 8;
+      const UInt32 v1 = p->state[(size_t)i + 1];
+      SetBe32(digest    , v0)
+      SetBe32(digest + 4, v1)
+      digest += 4 * 2;
+    }
   }
 
+
+
+
+#endif
   Sha256_InitState(p);
 }
 
 
-void Sha256Prepare()
+void Sha256Prepare(void)
 {
-  #ifdef _SHA_SUPPORTED
+#ifdef Z7_COMPILER_SHA256_SUPPORTED
   SHA256_FUNC_UPDATE_BLOCKS f, f_hw;
   f = Sha256_UpdateBlocks;
   f_hw = NULL;
-  #ifdef MY_CPU_X86_OR_AMD64
+#ifdef MY_CPU_X86_OR_AMD64
-  #ifndef USE_MY_MM
   if (CPU_IsSupported_SHA()
       && CPU_IsSupported_SSSE3()
-      // && CPU_IsSupported_SSE41()
       )
-  #endif
+#else
-  #else
   if (CPU_IsSupported_SHA2())
-  #endif
+#endif
   {
     // printf("\n========== HW SHA256 ======== \n");
     f = f_hw = Sha256_UpdateBlocks_HW;
   }
-  g_FUNC_UPDATE_BLOCKS    = f;
+  g_SHA256_FUNC_UPDATE_BLOCKS    = f;
-  g_FUNC_UPDATE_BLOCKS_HW = f_hw;
+  g_SHA256_FUNC_UPDATE_BLOCKS_HW = f_hw;
-  #endif
+#endif
 }
+
+#undef U64C
+#undef K
+#undef S0
+#undef S1
+#undef s0
+#undef s1
+#undef Ch
+#undef Maj
+#undef W_MAIN
+#undef W_PRE
+#undef w
+#undef blk2_main
+#undef blk2
+#undef T1
+#undef T4
+#undef T8
+#undef R1_PRE
+#undef R1_MAIN
+#undef R2_MAIN
+#undef R4
+#undef R4_PRE
+#undef R4_MAIN
+#undef R8
+#undef R8_PRE
+#undef R8_MAIN
+#undef STEP_PRE
+#undef STEP_MAIN
+#undef Z7_SHA256_BIG_W
+#undef Z7_SHA256_UNROLL
+#undef Z7_COMPILER_SHA256_SUPPORTED

C/Sha256.h

@@ -1,8 +1,8 @@
 /* Sha256.h -- SHA-256 Hash
-2021-01-01 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
-#ifndef __7Z_SHA256_H
+#ifndef ZIP7_INC_SHA256_H
-#define __7Z_SHA256_H
+#define ZIP7_INC_SHA256_H
 
 #include "7zTypes.h"
 
@@ -14,7 +14,10 @@ EXTERN_C_BEGIN
 #define SHA256_BLOCK_SIZE   (SHA256_NUM_BLOCK_WORDS * 4)
 #define SHA256_DIGEST_SIZE  (SHA256_NUM_DIGEST_WORDS * 4)
 
-typedef void (MY_FAST_CALL *SHA256_FUNC_UPDATE_BLOCKS)(UInt32 state[8], const Byte *data, size_t numBlocks);
+
+
+
+typedef void (Z7_FASTCALL *SHA256_FUNC_UPDATE_BLOCKS)(UInt32 state[8], const Byte *data, size_t numBlocks);
 
 /*
   if (the system supports different SHA256 code implementations)
@@ -32,9 +35,16 @@ typedef void (MY_FAST_CALL *SHA256_FUNC_UPDATE_BLOCKS)(UInt32 state[8], const By
 
 typedef struct
 {
-  SHA256_FUNC_UPDATE_BLOCKS func_UpdateBlocks;
+  union
-  UInt64 count;
+  {
-  UInt64 __pad_2[2];
+    struct
+    {
+      SHA256_FUNC_UPDATE_BLOCKS func_UpdateBlocks;
+      UInt64 count;
+    } vars;
+    UInt64 _pad_64bit[4];
+    void *_pad_align_ptr[2];
+  } v;
   UInt32 state[SHA256_NUM_DIGEST_WORDS];
 
   Byte buffer[SHA256_BLOCK_SIZE];
@@ -62,7 +72,7 @@ void Sha256_Final(CSha256 *p, Byte *digest);
 
 
 
-// void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks);
+// void Z7_FASTCALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks);
 
 /*
 call Sha256Prepare() once at program start.

C/Sha256Opt.c

@@ -1,71 +1,53 @@
 /* Sha256Opt.c -- SHA-256 optimized code for SHA-256 hardware instructions
-2021-04-01 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
-
+#include "Compiler.h"
-#if defined(_MSC_VER)
-#if (_MSC_VER < 1900) && (_MSC_VER >= 1200)
-// #define USE_MY_MM
-#endif
-#endif
-
 #include "CpuArch.h"
 
+// #define Z7_USE_HW_SHA_STUB // for debug
 #ifdef MY_CPU_X86_OR_AMD64
-  #if defined(__clang__)
+  #if defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1600) // fix that check
-    #if (__clang_major__ >= 8) // fix that check
       #define USE_HW_SHA
-      #ifndef __SHA__
+  #elif defined(Z7_LLVM_CLANG_VERSION)  && (Z7_LLVM_CLANG_VERSION  >= 30800) \
+     || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 50100) \
+     || defined(Z7_GCC_VERSION)         && (Z7_GCC_VERSION         >= 40900)
+      #define USE_HW_SHA
+      #if !defined(__INTEL_COMPILER)
+      // icc defines __GNUC__, but icc doesn't support __attribute__(__target__)
+      #if !defined(__SHA__) || !defined(__SSSE3__)
         #define ATTRIB_SHA __attribute__((__target__("sha,ssse3")))
-        #if defined(_MSC_VER)
-          // SSSE3: for clang-cl:
-          #include <tmmintrin.h>
-          #define __SHA__
-        #endif
       #endif
-
-    #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 8) // fix that check
-      #define USE_HW_SHA
-      #ifndef __SHA__
-        #define ATTRIB_SHA __attribute__((__target__("sha,ssse3")))
-        // #pragma GCC target("sha,ssse3")
       #endif
-    #endif
-  #elif defined(__INTEL_COMPILER)
-    #if (__INTEL_COMPILER >= 1800) // fix that check
-      #define USE_HW_SHA
-    #endif
   #elif defined(_MSC_VER)
-    #ifdef USE_MY_MM
+    #if (_MSC_VER >= 1900)
-      #define USE_VER_MIN 1300
-    #else
-      #define USE_VER_MIN 1910
-    #endif
-    #if _MSC_VER >= USE_VER_MIN
       #define USE_HW_SHA
+    #else
+      #define Z7_USE_HW_SHA_STUB
     #endif
   #endif
 // #endif // MY_CPU_X86_OR_AMD64
+#ifndef USE_HW_SHA
+  // #define Z7_USE_HW_SHA_STUB // for debug
+#endif
 
 #ifdef USE_HW_SHA
 
 // #pragma message("Sha256 HW")
-// #include <wmmintrin.h>
 
-#if !defined(_MSC_VER) || (_MSC_VER >= 1900)
+
+
+
+// sse/sse2/ssse3:
+#include <tmmintrin.h>
+// sha*:
 #include <immintrin.h>
+
+#if defined (__clang__) && defined(_MSC_VER)
+  #if !defined(__SHA__)
+    #include <shaintrin.h>
+  #endif
 #else
-#include <emmintrin.h>
-
-#if defined(_MSC_VER) && (_MSC_VER >= 1600)
-// #include <intrin.h>
-#endif
-
-#ifdef USE_MY_MM
-#include "My_mm.h"
-#endif
 
 #endif
 
@@ -94,60 +76,44 @@ SHA:
 extern
 MY_ALIGN(64)
 const UInt32 SHA256_K_ARRAY[64];
-
 #define K SHA256_K_ARRAY
 
 
-#define ADD_EPI32(dest, src) dest = _mm_add_epi32(dest, src);
+#define ADD_EPI32(dest, src)      dest = _mm_add_epi32(dest, src);
-#define SHA256_MSG1(dest, src) dest = _mm_sha256msg1_epu32(dest, src);
+#define SHA256_MSG1(dest, src)    dest = _mm_sha256msg1_epu32(dest, src);
-#define SHA25G_MSG2(dest, src) dest = _mm_sha256msg2_epu32(dest, src);
+#define SHA256_MSG2(dest, src)    dest = _mm_sha256msg2_epu32(dest, src);
-
 
 #define LOAD_SHUFFLE(m, k) \
     m = _mm_loadu_si128((const __m128i *)(const void *)(data + (k) * 16)); \
     m = _mm_shuffle_epi8(m, mask); \
 
-#define SM1(g0, g1, g2, g3) \
+#define NNN(m0, m1, m2, m3)
-    SHA256_MSG1(g3, g0); \
 
-#define SM2(g0, g1, g2, g3) \
+#define SM1(m1, m2, m3, m0) \
-    tmp = _mm_alignr_epi8(g1, g0, 4); \
+    SHA256_MSG1(m0, m1); \
-    ADD_EPI32(g2, tmp); \
-    SHA25G_MSG2(g2, g1); \
-
-// #define LS0(k, g0, g1, g2, g3) LOAD_SHUFFLE(g0, k)
-// #define LS1(k, g0, g1, g2, g3) LOAD_SHUFFLE(g1, k+1)
-
-
-#define NNN(g0, g1, g2, g3)
 
+#define SM2(m2, m3, m0, m1) \
+    ADD_EPI32(m0, _mm_alignr_epi8(m3, m2, 4)) \
+    SHA256_MSG2(m0, m3); \
 
 #define RND2(t0, t1) \
     t0 = _mm_sha256rnds2_epu32(t0, t1, msg);
 
-#define RND2_0(m, k) \
+
-    msg = _mm_add_epi32(m, *(const __m128i *) (const void *) &K[(k) * 4]); \
+
+#define R4(k, m0, m1, m2, m3, OP0, OP1) \
+    msg = _mm_add_epi32(m0, *(const __m128i *) (const void *) &K[(k) * 4]); \
     RND2(state0, state1); \
     msg = _mm_shuffle_epi32(msg, 0x0E); \
-
+    OP0(m0, m1, m2, m3) \
-
-#define RND2_1 \
     RND2(state1, state0); \
-
+    OP1(m0, m1, m2, m3) \
-
-// We use scheme with 3 rounds ahead for SHA256_MSG1 / 2 rounds ahead for SHA256_MSG2
-
-#define R4(k, g0, g1, g2, g3, OP0, OP1) \
-    RND2_0(g0, k); \
-    OP0(g0, g1, g2, g3); \
-    RND2_1; \
-    OP1(g0, g1, g2, g3); \
 
 #define R16(k, OP0, OP1, OP2, OP3, OP4, OP5, OP6, OP7) \
-    R4 ( (k)*4+0, m0, m1, m2, m3, OP0, OP1 ) \
+    R4 ( (k)*4+0, m0,m1,m2,m3, OP0, OP1 ) \
-    R4 ( (k)*4+1, m1, m2, m3, m0, OP2, OP3 ) \
+    R4 ( (k)*4+1, m1,m2,m3,m0, OP2, OP3 ) \
-    R4 ( (k)*4+2, m2, m3, m0, m1, OP4, OP5 ) \
+    R4 ( (k)*4+2, m2,m3,m0,m1, OP4, OP5 ) \
-    R4 ( (k)*4+3, m3, m0, m1, m2, OP6, OP7 ) \
+    R4 ( (k)*4+3, m3,m0,m1,m2, OP6, OP7 ) \
 
 #define PREPARE_STATE \
     tmp    = _mm_shuffle_epi32(state0, 0x1B); /* abcd */ \
@@ -157,15 +123,16 @@ const UInt32 SHA256_K_ARRAY[64];
     state1 = _mm_unpackhi_epi64(state1, tmp); /* abef */ \
 
 
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
 #ifdef ATTRIB_SHA
 ATTRIB_SHA
 #endif
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
 {
   const __m128i mask = _mm_set_epi32(0x0c0d0e0f, 0x08090a0b, 0x04050607, 0x00010203);
-  __m128i tmp;
+   
-  __m128i state0, state1;
+  
+  __m128i tmp, state0, state1;
 
   if (numBlocks == 0)
     return;
@@ -192,13 +159,13 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
 
 
 
-    R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 );
+    R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 )
-    R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 );
+    R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
-    R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 );
+    R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
-    R16 ( 3, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN );
+    R16 ( 3, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN )
     
-    ADD_EPI32(state0, state0_save);
+    ADD_EPI32(state0, state0_save)
-    ADD_EPI32(state1, state1_save);
+    ADD_EPI32(state1, state1_save)
     
     data += 64;
   }
@@ -212,19 +179,28 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
 
 #endif // USE_HW_SHA
 
-#elif defined(MY_CPU_ARM_OR_ARM64)
+#elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE)
   
-  #if defined(__clang__)
+  #if   defined(__ARM_FEATURE_SHA2) \
-    #if (__clang_major__ >= 8) // fix that check
+     || defined(__ARM_FEATURE_CRYPTO)
+    #define USE_HW_SHA
+  #else
+    #if  defined(MY_CPU_ARM64) \
+      || defined(__ARM_ARCH) && (__ARM_ARCH >= 4) \
+      || defined(Z7_MSC_VER_ORIGINAL)
+    #if  defined(__ARM_FP) && \
+          (   defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 30800) \
+           || defined(__GNUC__) && (__GNUC__ >= 6) \
+          ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1910)
+    #if  defined(MY_CPU_ARM64) \
+      || !defined(Z7_CLANG_VERSION) \
+      || defined(__ARM_NEON) && \
+          (Z7_CLANG_VERSION < 170000 || \
+           Z7_CLANG_VERSION > 170001)
       #define USE_HW_SHA
     #endif
-  #elif defined(__GNUC__)
-    #if (__GNUC__ >= 6) // fix that check
-      #define USE_HW_SHA
     #endif
-  #elif defined(_MSC_VER)
-    #if _MSC_VER >= 1910
-      #define USE_HW_SHA
     #endif
   #endif
 
@@ -232,63 +208,144 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
 
 // #pragma message("=== Sha256 HW === ")
 
+
 #if defined(__clang__) || defined(__GNUC__)
+#if !defined(__ARM_FEATURE_SHA2) && \
+    !defined(__ARM_FEATURE_CRYPTO)
   #ifdef MY_CPU_ARM64
+#if defined(__clang__)
+    #define ATTRIB_SHA __attribute__((__target__("crypto")))
+#else
     #define ATTRIB_SHA __attribute__((__target__("+crypto")))
+#endif
   #else
+#if defined(__clang__) && (__clang_major__ >= 1)
+    #define ATTRIB_SHA __attribute__((__target__("armv8-a,sha2")))
+#else
     #define ATTRIB_SHA __attribute__((__target__("fpu=crypto-neon-fp-armv8")))
+#endif
   #endif
+#endif
 #else
   // _MSC_VER
   // for arm32
   #define _ARM_USE_NEW_NEON_INTRINSICS
 #endif
 
-#if defined(_MSC_VER) && defined(MY_CPU_ARM64)
+#if defined(Z7_MSC_VER_ORIGINAL) && defined(MY_CPU_ARM64)
 #include <arm64_neon.h>
 #else
-#include <arm_neon.h>
+
+#if defined(__clang__) && __clang_major__ < 16
+#if !defined(__ARM_FEATURE_SHA2) && \
+    !defined(__ARM_FEATURE_CRYPTO)
+//     #pragma message("=== we set __ARM_FEATURE_CRYPTO 1 === ")
+    Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #define Z7_ARM_FEATURE_CRYPTO_WAS_SET 1
+// #if defined(__clang__) && __clang_major__ < 13
+    #define __ARM_FEATURE_CRYPTO 1
+// #else
+    #define __ARM_FEATURE_SHA2 1
+// #endif
+    Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
 #endif
+#endif // clang
+
+#if defined(__clang__)
+
+#if defined(__ARM_ARCH) && __ARM_ARCH < 8
+    Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+//    #pragma message("#define __ARM_ARCH 8")
+    #undef  __ARM_ARCH
+    #define __ARM_ARCH 8
+    Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+#endif
+
+#endif // clang
+
+#include <arm_neon.h>
+
+#if defined(Z7_ARM_FEATURE_CRYPTO_WAS_SET) && \
+    defined(__ARM_FEATURE_CRYPTO) && \
+    defined(__ARM_FEATURE_SHA2)
+Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #undef __ARM_FEATURE_CRYPTO
+    #undef __ARM_FEATURE_SHA2
+    #undef Z7_ARM_FEATURE_CRYPTO_WAS_SET
+Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+//    #pragma message("=== we undefine __ARM_FEATURE_CRYPTO === ")
+#endif
+
+#endif // Z7_MSC_VER_ORIGINAL
 
 typedef uint32x4_t v128;
 // typedef __n128 v128; // MSVC
 
 #ifdef MY_CPU_BE
-  #define MY_rev32_for_LE(x)
+  #define MY_rev32_for_LE(x) x
 #else
-  #define MY_rev32_for_LE(x) x = vreinterpretq_u32_u8(vrev32q_u8(vreinterpretq_u8_u32(x)))
+  #define MY_rev32_for_LE(x) vrev32q_u8(x)
 #endif
 
-#define LOAD_128(_p)      (*(const v128 *)(const void *)(_p))
+#if 1 // 0 for debug
-#define STORE_128(_p, _v) *(v128 *)(void *)(_p) = (_v)
+// for arm32: it works slower by some reason than direct code
+/*
+for arm32 it generates:
+MSVC-2022, GCC-9:
+    vld1.32 {d18,d19}, [r10]
+    vst1.32 {d4,d5}, [r3]
+    vld1.8  {d20-d21}, [r4]
+there is no align hint (like [r10:128]).  So instruction allows unaligned access
+*/
+#define LOAD_128_32(_p)       vld1q_u32(_p)
+#define LOAD_128_8(_p)        vld1q_u8 (_p)
+#define STORE_128_32(_p, _v)  vst1q_u32(_p, _v)
+#else
+/*
+for arm32:
+MSVC-2022:
+    vldm r10,{d18,d19}
+    vstm r3,{d4,d5}
+    does it require strict alignment?
+GCC-9:
+    vld1.64 {d30-d31}, [r0:64]
+    vldr  d28, [r0, #16]
+    vldr  d29, [r0, #24]
+    vst1.64 {d30-d31}, [r0:64]
+    vstr  d28, [r0, #16]
+    vstr  d29, [r0, #24]
+there is hint [r0:64], so does it requires 64-bit alignment.
+*/
+#define LOAD_128_32(_p)       (*(const v128 *)(const void *)(_p))
+#define LOAD_128_8(_p)        vreinterpretq_u8_u32(*(const v128 *)(const void *)(_p))
+#define STORE_128_32(_p, _v)  *(v128 *)(void *)(_p) = (_v)
+#endif
 
 #define LOAD_SHUFFLE(m, k) \
-    m = LOAD_128((data + (k) * 16)); \
+    m = vreinterpretq_u32_u8( \
-    MY_rev32_for_LE(m); \
+        MY_rev32_for_LE( \
+        LOAD_128_8(data + (k) * 16))); \
 
 // K array must be aligned for 16-bytes at least.
 extern
 MY_ALIGN(64)
 const UInt32 SHA256_K_ARRAY[64];
-
 #define K SHA256_K_ARRAY
 
-
 #define SHA256_SU0(dest, src)        dest = vsha256su0q_u32(dest, src);
-#define SHA25G_SU1(dest, src2, src3) dest = vsha256su1q_u32(dest, src2, src3);
+#define SHA256_SU1(dest, src2, src3) dest = vsha256su1q_u32(dest, src2, src3);
 
-#define SM1(g0, g1, g2, g3)  SHA256_SU0(g3, g0)
+#define SM1(m0, m1, m2, m3)  SHA256_SU0(m3, m0)
-#define SM2(g0, g1, g2, g3)  SHA25G_SU1(g2, g0, g1)
+#define SM2(m0, m1, m2, m3)  SHA256_SU1(m2, m0, m1)
-#define NNN(g0, g1, g2, g3)
+#define NNN(m0, m1, m2, m3)
 
-
+#define R4(k, m0, m1, m2, m3, OP0, OP1) \
-#define R4(k, g0, g1, g2, g3, OP0, OP1) \
+    msg = vaddq_u32(m0, *(const v128 *) (const void *) &K[(k) * 4]); \
-    msg = vaddq_u32(g0, *(const v128 *) (const void *) &K[(k) * 4]); \
     tmp = state0; \
     state0 = vsha256hq_u32( state0, state1, msg ); \
     state1 = vsha256h2q_u32( state1, tmp, msg ); \
-    OP0(g0, g1, g2, g3); \
+    OP0(m0, m1, m2, m3); \
-    OP1(g0, g1, g2, g3); \
+    OP1(m0, m1, m2, m3); \
 
 
 #define R16(k, OP0, OP1, OP2, OP3, OP4, OP5, OP6, OP7) \
@@ -298,19 +355,19 @@ const UInt32 SHA256_K_ARRAY[64];
     R4 ( (k)*4+3, m3, m0, m1, m2, OP6, OP7 ) \
 
 
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
 #ifdef ATTRIB_SHA
 ATTRIB_SHA
 #endif
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
 {
   v128 state0, state1;
 
   if (numBlocks == 0)
     return;
 
-  state0 = LOAD_128(&state[0]);
+  state0 = LOAD_128_32(&state[0]);
-  state1 = LOAD_128(&state[4]);
+  state1 = LOAD_128_32(&state[4]);
   
   do
   {
@@ -326,10 +383,10 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
     LOAD_SHUFFLE (m2, 2)
     LOAD_SHUFFLE (m3, 3)
 
-    R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 );
+    R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 )
-    R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 );
+    R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
-    R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 );
+    R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
-    R16 ( 3, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN );
+    R16 ( 3, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN )
     
     state0 = vaddq_u32(state0, state0_save);
     state1 = vaddq_u32(state1, state1_save);
@@ -338,8 +395,8 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
   }
   while (--numBlocks);
 
-  STORE_128(&state[0], state0);
+  STORE_128_32(&state[0], state0);
-  STORE_128(&state[4], state1);
+  STORE_128_32(&state[4], state1);
 }
 
 #endif // USE_HW_SHA
@@ -347,18 +404,19 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
 #endif // MY_CPU_ARM_OR_ARM64
 
 
-#ifndef USE_HW_SHA
+#if !defined(USE_HW_SHA) && defined(Z7_USE_HW_SHA_STUB)
-
 // #error Stop_Compiling_UNSUPPORTED_SHA
 // #include <stdlib.h>
-
+// We can compile this file with another C compiler,
+// or we can compile asm version.
+// So we can generate real code instead of this stub function.
 // #include "Sha256.h"
-void MY_FAST_CALL Sha256_UpdateBlocks(UInt32 state[8], const Byte *data, size_t numBlocks);
+// #if defined(_MSC_VER)
-
 #pragma message("Sha256 HW-SW stub was used")
-
+// #endif
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha256_UpdateBlocks   (UInt32 state[8], const Byte *data, size_t numBlocks);
-void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
+void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size_t numBlocks)
 {
   Sha256_UpdateBlocks(state, data, numBlocks);
   /*
@@ -369,5 +427,25 @@ void MY_FAST_CALL Sha256_UpdateBlocks_HW(UInt32 state[8], const Byte *data, size
   return;
   */
 }
-
 #endif
+
+
+#undef K
+#undef RND2
+#undef MY_rev32_for_LE
+
+#undef NNN
+#undef LOAD_128
+#undef STORE_128
+#undef LOAD_SHUFFLE
+#undef SM1
+#undef SM2
+
+
+#undef R4
+#undef R16
+#undef PREPARE_STATE
+#undef USE_HW_SHA
+#undef ATTRIB_SHA
+#undef USE_VER_MIN
+#undef Z7_USE_HW_SHA_STUB

C/Sha3.c

@@ -0,0 +1,359 @@
+/* Sha3.c -- SHA-3 Hash
+: Igor Pavlov : Public domain
+This code is based on public domain code from Wei Dai's Crypto++ library. */
+
+#include "Precomp.h"
+
+#include <string.h>
+
+#include "Sha3.h"
+#include "RotateDefs.h"
+#include "CpuArch.h"
+
+#define U64C(x) UINT64_CONST(x)
+
+static
+MY_ALIGN(64)
+const UInt64 SHA3_K_ARRAY[24] =
+{
+  U64C(0x0000000000000001), U64C(0x0000000000008082),
+  U64C(0x800000000000808a), U64C(0x8000000080008000),
+  U64C(0x000000000000808b), U64C(0x0000000080000001),
+  U64C(0x8000000080008081), U64C(0x8000000000008009),
+  U64C(0x000000000000008a), U64C(0x0000000000000088),
+  U64C(0x0000000080008009), U64C(0x000000008000000a),
+  U64C(0x000000008000808b), U64C(0x800000000000008b),
+  U64C(0x8000000000008089), U64C(0x8000000000008003),
+  U64C(0x8000000000008002), U64C(0x8000000000000080),
+  U64C(0x000000000000800a), U64C(0x800000008000000a),
+  U64C(0x8000000080008081), U64C(0x8000000000008080),
+  U64C(0x0000000080000001), U64C(0x8000000080008008)
+};
+
+void Sha3_Init(CSha3 *p)
+{
+  p->count = 0;
+  memset(p->state, 0, sizeof(p->state));
+}
+
+#define GET_state(i, a)   UInt64 a = state[i];
+#define SET_state(i, a)   state[i] = a;
+
+#define LS_5(M, i, a0,a1,a2,a3,a4) \
+        M ((i) * 5    , a0) \
+        M ((i) * 5 + 1, a1) \
+        M ((i) * 5 + 2, a2) \
+        M ((i) * 5 + 3, a3) \
+        M ((i) * 5 + 4, a4) \
+
+#define LS_25(M) \
+        LS_5 (M, 0, a50, a51, a52, a53, a54) \
+        LS_5 (M, 1, a60, a61, a62, a63, a64) \
+        LS_5 (M, 2, a70, a71, a72, a73, a74) \
+        LS_5 (M, 3, a80, a81, a82, a83, a84) \
+        LS_5 (M, 4, a90, a91, a92, a93, a94) \
+
+
+#define XOR_1(i, a0) \
+        a0 ^= GetUi64(data + (i) * 8); \
+
+#define XOR_4(i, a0,a1,a2,a3) \
+        XOR_1 ((i)    , a0); \
+        XOR_1 ((i) + 1, a1); \
+        XOR_1 ((i) + 2, a2); \
+        XOR_1 ((i) + 3, a3); \
+
+#define D(d,b1,b2) \
+        d = b1 ^ Z7_ROTL64(b2, 1);
+
+#define D5 \
+        D (d0, c4, c1) \
+        D (d1, c0, c2) \
+        D (d2, c1, c3) \
+        D (d3, c2, c4) \
+        D (d4, c3, c0) \
+
+#define C0(c,a,d) \
+        c = a ^ d; \
+
+#define C(c,a,d,k) \
+        c = a ^ d; \
+        c = Z7_ROTL64(c, k); \
+
+#define E4(e1,e2,e3,e4) \
+        e1 = c1 ^ (~c2 & c3); \
+        e2 = c2 ^ (~c3 & c4); \
+        e3 = c3 ^ (~c4 & c0); \
+        e4 = c4 ^ (~c0 & c1); \
+
+#define CK(   v0,w0,    \
+              v1,w1,k1, \
+              v2,w2,k2, \
+              v3,w3,k3, \
+              v4,w4,k4, e0,e1,e2,e3,e4, keccak_c) \
+        C0(c0,v0,w0)    \
+        C (c1,v1,w1,k1) \
+        C (c2,v2,w2,k2) \
+        C (c3,v3,w3,k3) \
+        C (c4,v4,w4,k4) \
+        e0 = c0 ^ (~c1 & c2) ^ keccak_c; \
+        E4(e1,e2,e3,e4) \
+
+#define CE(   v0,w0,k0, \
+              v1,w1,k1, \
+              v2,w2,k2, \
+              v3,w3,k3, \
+              v4,w4,k4, e0,e1,e2,e3,e4) \
+        C (c0,v0,w0,k0) \
+        C (c1,v1,w1,k1) \
+        C (c2,v2,w2,k2) \
+        C (c3,v3,w3,k3) \
+        C (c4,v4,w4,k4) \
+        e0 = c0 ^ (~c1 & c2); \
+        E4(e1,e2,e3,e4) \
+
+// numBlocks != 0
+static
+Z7_NO_INLINE
+void Z7_FASTCALL Sha3_UpdateBlocks(UInt64 state[SHA3_NUM_STATE_WORDS],
+    const Byte *data, size_t numBlocks, size_t blockSize)
+{
+  LS_25 (GET_state)
+
+  do
+  {
+    unsigned round;
+                              XOR_4 ( 0, a50, a51, a52, a53)
+                              XOR_4 ( 4, a54, a60, a61, a62)
+                              XOR_1 ( 8, a63)
+    if (blockSize > 8 *  9) { XOR_4 ( 9, a64, a70, a71, a72)  // sha3-384
+    if (blockSize > 8 * 13) { XOR_4 (13, a73, a74, a80, a81)  // sha3-256
+    if (blockSize > 8 * 17) { XOR_1 (17, a82)                 // sha3-224
+    if (blockSize > 8 * 18) { XOR_1 (18, a83)                 // shake128
+                              XOR_1 (19, a84)
+                              XOR_1 (20, a90) }}}}
+    data += blockSize;
+
+    for (round = 0; round < 24; round += 2)
+    {
+      UInt64 c0, c1, c2, c3, c4;
+      UInt64 d0, d1, d2, d3, d4;
+      UInt64 e50, e51, e52, e53, e54;
+      UInt64 e60, e61, e62, e63, e64;
+      UInt64 e70, e71, e72, e73, e74;
+      UInt64 e80, e81, e82, e83, e84;
+      UInt64 e90, e91, e92, e93, e94;
+
+      c0 = a50^a60^a70^a80^a90;
+      c1 = a51^a61^a71^a81^a91;
+      c2 = a52^a62^a72^a82^a92;
+      c3 = a53^a63^a73^a83^a93;
+      c4 = a54^a64^a74^a84^a94;
+      D5
+      CK( a50, d0,
+          a61, d1, 44,
+          a72, d2, 43,
+          a83, d3, 21,
+          a94, d4, 14, e50, e51, e52, e53, e54, SHA3_K_ARRAY[round])
+      CE( a53, d3, 28,
+          a64, d4, 20,
+          a70, d0,  3,
+          a81, d1, 45,
+          a92, d2, 61, e60, e61, e62, e63, e64)
+      CE( a51, d1,  1,
+          a62, d2,  6,
+          a73, d3, 25,
+          a84, d4,  8,
+          a90, d0, 18, e70, e71, e72, e73, e74)
+      CE( a54, d4, 27,
+          a60, d0, 36,
+          a71, d1, 10,
+          a82, d2, 15,
+          a93, d3, 56, e80, e81, e82, e83, e84)
+      CE( a52, d2, 62,
+          a63, d3, 55,
+          a74, d4, 39,
+          a80, d0, 41,
+          a91, d1,  2, e90, e91, e92, e93, e94)
+      
+      // ---------- ROUND + 1 ----------
+
+      c0 = e50^e60^e70^e80^e90;
+      c1 = e51^e61^e71^e81^e91;
+      c2 = e52^e62^e72^e82^e92;
+      c3 = e53^e63^e73^e83^e93;
+      c4 = e54^e64^e74^e84^e94;
+      D5
+      CK( e50, d0,
+          e61, d1, 44,
+          e72, d2, 43,
+          e83, d3, 21,
+          e94, d4, 14, a50, a51, a52, a53, a54, SHA3_K_ARRAY[(size_t)round + 1])
+      CE( e53, d3, 28,
+          e64, d4, 20,
+          e70, d0,  3,
+          e81, d1, 45,
+          e92, d2, 61, a60, a61, a62, a63, a64)
+      CE( e51, d1,  1,
+          e62, d2,  6,
+          e73, d3, 25,
+          e84, d4,  8,
+          e90, d0, 18, a70, a71, a72, a73, a74)
+      CE (e54, d4, 27,
+          e60, d0, 36,
+          e71, d1, 10,
+          e82, d2, 15,
+          e93, d3, 56, a80, a81, a82, a83, a84)
+      CE (e52, d2, 62,
+          e63, d3, 55,
+          e74, d4, 39,
+          e80, d0, 41,
+          e91, d1,  2, a90, a91, a92, a93, a94)
+    }
+  }
+  while (--numBlocks);
+
+  LS_25 (SET_state)
+}
+
+
+#define Sha3_UpdateBlock(p) \
+        Sha3_UpdateBlocks(p->state, p->buffer, 1, p->blockSize)
+
+void Sha3_Update(CSha3 *p, const Byte *data, size_t size)
+{
+/*
+  for (;;)
+  {
+    if (size == 0)
+      return;
+    unsigned cur = p->blockSize - p->count;
+    if (cur > size)
+      cur = (unsigned)size;
+    size -= cur;
+    unsigned pos = p->count;
+    p->count = pos + cur;
+    while (pos & 7)
+    {
+      if (cur == 0)
+        return;
+      Byte *pb = &(((Byte *)p->state)[pos]);
+      *pb = (Byte)(*pb ^ *data++);
+      cur--;
+      pos++;
+    }
+    if (cur >= 8)
+    {
+      do
+      {
+        *(UInt64 *)(void *)&(((Byte *)p->state)[pos]) ^= GetUi64(data);
+        data += 8;
+        pos += 8;
+        cur -= 8;
+      }
+      while (cur >= 8);
+    }
+    if (pos != p->blockSize)
+    {
+      if (cur)
+      {
+        Byte *pb = &(((Byte *)p->state)[pos]);
+        do
+        {
+          *pb = (Byte)(*pb ^ *data++);
+          pb++;
+        }
+        while (--cur);
+      }
+      return;
+    }
+    Sha3_UpdateBlock(p->state);
+    p->count = 0;
+  }
+*/
+  if (size == 0)
+    return;
+  {
+    const unsigned pos = p->count;
+    const unsigned num = p->blockSize - pos;
+    if (num > size)
+    {
+      p->count = pos + (unsigned)size;
+      memcpy(p->buffer + pos, data, size);
+      return;
+    }
+    if (pos != 0)
+    {
+      size -= num;
+      memcpy(p->buffer + pos, data, num);
+      data += num;
+      Sha3_UpdateBlock(p);
+    }
+  }
+  if (size >= p->blockSize)
+  {
+    const size_t numBlocks = size / p->blockSize;
+    const Byte *dataOld = data;
+    data += numBlocks * p->blockSize;
+    size = (size_t)(dataOld + size - data);
+    Sha3_UpdateBlocks(p->state, dataOld, numBlocks, p->blockSize);
+  }
+  p->count = (unsigned)size;
+  if (size)
+    memcpy(p->buffer, data, size);
+}
+
+
+// we support only (digestSize % 4 == 0) cases
+void Sha3_Final(CSha3 *p, Byte *digest, unsigned digestSize, unsigned shake)
+{
+  memset(p->buffer + p->count, 0, p->blockSize - p->count);
+  // we write bits markers from low to higher in current byte:
+  //   - if sha-3 : 2 bits : 0,1
+  //   - if shake : 4 bits : 1111
+  // then we write bit 1 to same byte.
+  // And we write bit 1 to highest bit of last byte of block.
+  p->buffer[p->count] = (Byte)(shake ? 0x1f : 0x06);
+  // we need xor operation (^= 0x80) here because we must write 0x80 bit
+  // to same byte as (0x1f : 0x06), if (p->count == p->blockSize - 1) !!!
+  p->buffer[p->blockSize - 1] ^= 0x80;
+/*
+  ((Byte *)p->state)[p->count] ^= (Byte)(shake ? 0x1f : 0x06);
+  ((Byte *)p->state)[p->blockSize - 1] ^= 0x80;
+*/
+  Sha3_UpdateBlock(p);
+#if 1 && defined(MY_CPU_LE)
+  memcpy(digest, p->state, digestSize);
+#else
+  {
+    const unsigned numWords = digestSize >> 3;
+    unsigned i;
+    for (i = 0; i < numWords; i++)
+    {
+      const UInt64 v = p->state[i];
+      SetUi64(digest, v)
+      digest += 8;
+    }
+    if (digestSize & 4) // for SHA3-224
+    {
+      const UInt32 v = (UInt32)p->state[numWords];
+      SetUi32(digest, v)
+    }
+  }
+#endif
+  Sha3_Init(p);
+}
+
+#undef GET_state
+#undef SET_state
+#undef LS_5
+#undef LS_25
+#undef XOR_1
+#undef XOR_4
+#undef D
+#undef D5
+#undef C0
+#undef C
+#undef E4
+#undef CK
+#undef CE

C/Sha3.h

@@ -0,0 +1,36 @@
+/* Sha3.h -- SHA-3 Hash
+: Igor Pavlov : Public domain */
+
+#ifndef ZIP7_INC_MD5_H
+#define ZIP7_INC_MD5_H
+
+#include "7zTypes.h"
+
+EXTERN_C_BEGIN
+
+#define SHA3_NUM_STATE_WORDS  25
+
+#define SHA3_BLOCK_SIZE_FROM_DIGEST_SIZE(digestSize) \
+    (SHA3_NUM_STATE_WORDS * 8 - (digestSize) * 2)
+
+typedef struct
+{
+  UInt32 count;     // < blockSize
+  UInt32 blockSize; // <= SHA3_NUM_STATE_WORDS * 8
+  UInt64 _pad1[3];
+  // we want 32-bytes alignment here
+  UInt64 state[SHA3_NUM_STATE_WORDS];
+  UInt64 _pad2[3];
+  // we want 64-bytes alignment here
+  Byte buffer[SHA3_NUM_STATE_WORDS * 8]; // last bytes will be unused with predefined blockSize values
+} CSha3;
+
+#define Sha3_SET_blockSize(p, blockSize) { (p)->blockSize = (blockSize); }
+
+void Sha3_Init(CSha3 *p);
+void Sha3_Update(CSha3 *p, const Byte *data, size_t size);
+void Sha3_Final(CSha3 *p, Byte *digest, unsigned digestSize, unsigned shake);
+
+EXTERN_C_END
+
+#endif

C/Sha512.c

@@ -0,0 +1,711 @@
+/* Sha512.c -- SHA-512 Hash
+: Igor Pavlov : Public domain
+This code is based on public domain code from Wei Dai's Crypto++ library. */
+
+#include "Precomp.h"
+
+#include <string.h>
+
+#include "Sha512.h"
+#include "RotateDefs.h"
+#include "CpuArch.h"
+
+#ifdef MY_CPU_X86_OR_AMD64
+  #if   defined(Z7_LLVM_CLANG_VERSION)  && (Z7_LLVM_CLANG_VERSION  >= 170001) \
+     || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 170001) \
+     || defined(Z7_GCC_VERSION)         && (Z7_GCC_VERSION         >= 140000) \
+     || defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 2400) && (__INTEL_COMPILER <= 9900) \
+     || defined(_MSC_VER) && (_MSC_VER >= 1940)
+      #define Z7_COMPILER_SHA512_SUPPORTED
+  #endif
+#elif defined(MY_CPU_ARM64) && defined(MY_CPU_LE)
+  #if defined(__ARM_FEATURE_SHA512)
+    #define Z7_COMPILER_SHA512_SUPPORTED
+  #else
+    #if (defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 130000) \
+           || defined(__GNUC__) && (__GNUC__ >= 9) \
+        ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1940) // fix it
+      #define Z7_COMPILER_SHA512_SUPPORTED
+    #endif
+  #endif
+#endif
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+void Z7_FASTCALL Sha512_UpdateBlocks(UInt64 state[8], const Byte *data, size_t numBlocks);
+
+#ifdef Z7_COMPILER_SHA512_SUPPORTED
+  void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks);
+
+  static SHA512_FUNC_UPDATE_BLOCKS g_SHA512_FUNC_UPDATE_BLOCKS = Sha512_UpdateBlocks;
+  static SHA512_FUNC_UPDATE_BLOCKS g_SHA512_FUNC_UPDATE_BLOCKS_HW;
+
+  #define SHA512_UPDATE_BLOCKS(p) p->v.vars.func_UpdateBlocks
+#else
+  #define SHA512_UPDATE_BLOCKS(p) Sha512_UpdateBlocks
+#endif
+
+
+BoolInt Sha512_SetFunction(CSha512 *p, unsigned algo)
+{
+  SHA512_FUNC_UPDATE_BLOCKS func = Sha512_UpdateBlocks;
+  
+  #ifdef Z7_COMPILER_SHA512_SUPPORTED
+    if (algo != SHA512_ALGO_SW)
+    {
+      if (algo == SHA512_ALGO_DEFAULT)
+        func = g_SHA512_FUNC_UPDATE_BLOCKS;
+      else
+      {
+        if (algo != SHA512_ALGO_HW)
+          return False;
+        func = g_SHA512_FUNC_UPDATE_BLOCKS_HW;
+        if (!func)
+          return False;
+      }
+    }
+  #else
+    if (algo > 1)
+      return False;
+  #endif
+
+  p->v.vars.func_UpdateBlocks = func;
+  return True;
+}
+
+
+/* define it for speed optimization */
+
+#if 0 // 1 for size optimization
+  #define STEP_PRE 1
+  #define STEP_MAIN 1
+#else
+  #define STEP_PRE 2
+  #define STEP_MAIN 4
+  // #define Z7_SHA512_UNROLL
+#endif
+
+#undef Z7_SHA512_BIG_W
+#if STEP_MAIN != 16
+  #define Z7_SHA512_BIG_W
+#endif
+
+
+#define U64C(x) UINT64_CONST(x)
+
+static MY_ALIGN(64) const UInt64 SHA512_INIT_ARRAYS[4][8] = {
+{ U64C(0x8c3d37c819544da2), U64C(0x73e1996689dcd4d6), U64C(0x1dfab7ae32ff9c82), U64C(0x679dd514582f9fcf),
+  U64C(0x0f6d2b697bd44da8), U64C(0x77e36f7304c48942), U64C(0x3f9d85a86a1d36c8), U64C(0x1112e6ad91d692a1)
+},
+{ U64C(0x22312194fc2bf72c), U64C(0x9f555fa3c84c64c2), U64C(0x2393b86b6f53b151), U64C(0x963877195940eabd),
+  U64C(0x96283ee2a88effe3), U64C(0xbe5e1e2553863992), U64C(0x2b0199fc2c85b8aa), U64C(0x0eb72ddc81c52ca2)
+},
+{ U64C(0xcbbb9d5dc1059ed8), U64C(0x629a292a367cd507), U64C(0x9159015a3070dd17), U64C(0x152fecd8f70e5939),
+  U64C(0x67332667ffc00b31), U64C(0x8eb44a8768581511), U64C(0xdb0c2e0d64f98fa7), U64C(0x47b5481dbefa4fa4)
+},
+{ U64C(0x6a09e667f3bcc908), U64C(0xbb67ae8584caa73b), U64C(0x3c6ef372fe94f82b), U64C(0xa54ff53a5f1d36f1),
+  U64C(0x510e527fade682d1), U64C(0x9b05688c2b3e6c1f), U64C(0x1f83d9abfb41bd6b), U64C(0x5be0cd19137e2179)
+}};
+
+void Sha512_InitState(CSha512 *p, unsigned digestSize)
+{
+  p->v.vars.count = 0;
+  memcpy(p->state, SHA512_INIT_ARRAYS[(size_t)(digestSize >> 4) - 1], sizeof(p->state));
+}
+
+void Sha512_Init(CSha512 *p, unsigned digestSize)
+{
+  p->v.vars.func_UpdateBlocks =
+  #ifdef Z7_COMPILER_SHA512_SUPPORTED
+      g_SHA512_FUNC_UPDATE_BLOCKS;
+  #else
+      NULL;
+  #endif
+  Sha512_InitState(p, digestSize);
+}
+
+#define S0(x) (Z7_ROTR64(x,28) ^ Z7_ROTR64(x,34) ^ Z7_ROTR64(x,39))
+#define S1(x) (Z7_ROTR64(x,14) ^ Z7_ROTR64(x,18) ^ Z7_ROTR64(x,41))
+#define s0(x) (Z7_ROTR64(x, 1) ^ Z7_ROTR64(x, 8) ^ (x >> 7))
+#define s1(x) (Z7_ROTR64(x,19) ^ Z7_ROTR64(x,61) ^ (x >> 6))
+
+#define Ch(x,y,z) (z^(x&(y^z)))
+#define Maj(x,y,z) ((x&y)|(z&(x|y)))
+
+
+#define W_PRE(i) (W[(i) + (size_t)(j)] = GetBe64(data + ((size_t)(j) + i) * 8))
+
+#define blk2_main(j, i)  s1(w(j, (i)-2)) + w(j, (i)-7) + s0(w(j, (i)-15))
+
+#ifdef Z7_SHA512_BIG_W
+    // we use +i instead of +(i) to change the order to solve CLANG compiler warning for signed/unsigned.
+    #define w(j, i)     W[(size_t)(j) + i]
+    #define blk2(j, i)  (w(j, i) = w(j, (i)-16) + blk2_main(j, i))
+#else
+    #if STEP_MAIN == 16
+        #define w(j, i)  W[(i) & 15]
+    #else
+        #define w(j, i)  W[((size_t)(j) + (i)) & 15]
+    #endif
+    #define blk2(j, i)  (w(j, i) += blk2_main(j, i))
+#endif
+
+#define W_MAIN(i)  blk2(j, i)
+
+
+#define T1(wx, i) \
+    tmp = h + S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \
+    h = g; \
+    g = f; \
+    f = e; \
+    e = d + tmp; \
+    tmp += S0(a) + Maj(a, b, c); \
+    d = c; \
+    c = b; \
+    b = a; \
+    a = tmp; \
+
+#define R1_PRE(i)  T1( W_PRE, i)
+#define R1_MAIN(i) T1( W_MAIN, i)
+
+#if (!defined(Z7_SHA512_UNROLL) || STEP_MAIN < 8) && (STEP_MAIN >= 4)
+#define R2_MAIN(i) \
+    R1_MAIN(i) \
+    R1_MAIN(i + 1) \
+
+#endif
+
+
+
+#if defined(Z7_SHA512_UNROLL) && STEP_MAIN >= 8
+
+#define T4( a,b,c,d,e,f,g,h, wx, i) \
+    h += S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \
+    tmp = h; \
+    h += d; \
+    d = tmp + S0(a) + Maj(a, b, c); \
+
+#define R4( wx, i) \
+    T4 ( a,b,c,d,e,f,g,h, wx, (i  )); \
+    T4 ( d,a,b,c,h,e,f,g, wx, (i+1)); \
+    T4 ( c,d,a,b,g,h,e,f, wx, (i+2)); \
+    T4 ( b,c,d,a,f,g,h,e, wx, (i+3)); \
+
+#define R4_PRE(i)  R4( W_PRE, i)
+#define R4_MAIN(i) R4( W_MAIN, i)
+
+
+#define T8( a,b,c,d,e,f,g,h, wx, i) \
+    h += S1(e) + Ch(e,f,g) + K[(i)+(size_t)(j)] + wx(i); \
+    d += h; \
+    h += S0(a) + Maj(a, b, c); \
+
+#define R8( wx, i) \
+    T8 ( a,b,c,d,e,f,g,h, wx, i  ); \
+    T8 ( h,a,b,c,d,e,f,g, wx, i+1); \
+    T8 ( g,h,a,b,c,d,e,f, wx, i+2); \
+    T8 ( f,g,h,a,b,c,d,e, wx, i+3); \
+    T8 ( e,f,g,h,a,b,c,d, wx, i+4); \
+    T8 ( d,e,f,g,h,a,b,c, wx, i+5); \
+    T8 ( c,d,e,f,g,h,a,b, wx, i+6); \
+    T8 ( b,c,d,e,f,g,h,a, wx, i+7); \
+
+#define R8_PRE(i)  R8( W_PRE, i)
+#define R8_MAIN(i) R8( W_MAIN, i)
+
+#endif
+
+
+extern
+MY_ALIGN(64) const UInt64 SHA512_K_ARRAY[80];
+MY_ALIGN(64) const UInt64 SHA512_K_ARRAY[80] = {
+  U64C(0x428a2f98d728ae22), U64C(0x7137449123ef65cd), U64C(0xb5c0fbcfec4d3b2f), U64C(0xe9b5dba58189dbbc),
+  U64C(0x3956c25bf348b538), U64C(0x59f111f1b605d019), U64C(0x923f82a4af194f9b), U64C(0xab1c5ed5da6d8118),
+  U64C(0xd807aa98a3030242), U64C(0x12835b0145706fbe), U64C(0x243185be4ee4b28c), U64C(0x550c7dc3d5ffb4e2),
+  U64C(0x72be5d74f27b896f), U64C(0x80deb1fe3b1696b1), U64C(0x9bdc06a725c71235), U64C(0xc19bf174cf692694),
+  U64C(0xe49b69c19ef14ad2), U64C(0xefbe4786384f25e3), U64C(0x0fc19dc68b8cd5b5), U64C(0x240ca1cc77ac9c65),
+  U64C(0x2de92c6f592b0275), U64C(0x4a7484aa6ea6e483), U64C(0x5cb0a9dcbd41fbd4), U64C(0x76f988da831153b5),
+  U64C(0x983e5152ee66dfab), U64C(0xa831c66d2db43210), U64C(0xb00327c898fb213f), U64C(0xbf597fc7beef0ee4),
+  U64C(0xc6e00bf33da88fc2), U64C(0xd5a79147930aa725), U64C(0x06ca6351e003826f), U64C(0x142929670a0e6e70),
+  U64C(0x27b70a8546d22ffc), U64C(0x2e1b21385c26c926), U64C(0x4d2c6dfc5ac42aed), U64C(0x53380d139d95b3df),
+  U64C(0x650a73548baf63de), U64C(0x766a0abb3c77b2a8), U64C(0x81c2c92e47edaee6), U64C(0x92722c851482353b),
+  U64C(0xa2bfe8a14cf10364), U64C(0xa81a664bbc423001), U64C(0xc24b8b70d0f89791), U64C(0xc76c51a30654be30),
+  U64C(0xd192e819d6ef5218), U64C(0xd69906245565a910), U64C(0xf40e35855771202a), U64C(0x106aa07032bbd1b8),
+  U64C(0x19a4c116b8d2d0c8), U64C(0x1e376c085141ab53), U64C(0x2748774cdf8eeb99), U64C(0x34b0bcb5e19b48a8),
+  U64C(0x391c0cb3c5c95a63), U64C(0x4ed8aa4ae3418acb), U64C(0x5b9cca4f7763e373), U64C(0x682e6ff3d6b2b8a3),
+  U64C(0x748f82ee5defb2fc), U64C(0x78a5636f43172f60), U64C(0x84c87814a1f0ab72), U64C(0x8cc702081a6439ec),
+  U64C(0x90befffa23631e28), U64C(0xa4506cebde82bde9), U64C(0xbef9a3f7b2c67915), U64C(0xc67178f2e372532b),
+  U64C(0xca273eceea26619c), U64C(0xd186b8c721c0c207), U64C(0xeada7dd6cde0eb1e), U64C(0xf57d4f7fee6ed178),
+  U64C(0x06f067aa72176fba), U64C(0x0a637dc5a2c898a6), U64C(0x113f9804bef90dae), U64C(0x1b710b35131c471b),
+  U64C(0x28db77f523047d84), U64C(0x32caab7b40c72493), U64C(0x3c9ebe0a15c9bebc), U64C(0x431d67c49c100d4c),
+  U64C(0x4cc5d4becb3e42b6), U64C(0x597f299cfc657e2a), U64C(0x5fcb6fab3ad6faec), U64C(0x6c44198c4a475817)
+};
+
+#define K SHA512_K_ARRAY
+
+Z7_NO_INLINE
+void Z7_FASTCALL Sha512_UpdateBlocks(UInt64 state[8], const Byte *data, size_t numBlocks)
+{
+  UInt64 W
+#ifdef Z7_SHA512_BIG_W
+      [80];
+#else
+      [16];
+#endif
+  unsigned j;
+  UInt64 a,b,c,d,e,f,g,h;
+#if !defined(Z7_SHA512_UNROLL) || (STEP_MAIN <= 4) || (STEP_PRE <= 4)
+  UInt64 tmp;
+#endif
+
+  if (numBlocks == 0) return;
+  
+  a = state[0];
+  b = state[1];
+  c = state[2];
+  d = state[3];
+  e = state[4];
+  f = state[5];
+  g = state[6];
+  h = state[7];
+
+  do
+  {
+
+  for (j = 0; j < 16; j += STEP_PRE)
+  {
+    #if STEP_PRE > 4
+
+      #if STEP_PRE < 8
+      R4_PRE(0);
+      #else
+      R8_PRE(0);
+      #if STEP_PRE == 16
+      R8_PRE(8);
+      #endif
+      #endif
+
+    #else
+
+      R1_PRE(0)
+      #if STEP_PRE >= 2
+      R1_PRE(1)
+      #if STEP_PRE >= 4
+      R1_PRE(2)
+      R1_PRE(3)
+      #endif
+      #endif
+    
+    #endif
+  }
+
+  for (j = 16; j < 80; j += STEP_MAIN)
+  {
+    #if defined(Z7_SHA512_UNROLL) && STEP_MAIN >= 8
+
+      #if STEP_MAIN < 8
+      R4_MAIN(0)
+      #else
+      R8_MAIN(0)
+      #if STEP_MAIN == 16
+      R8_MAIN(8)
+      #endif
+      #endif
+
+    #else
+      
+      R1_MAIN(0)
+      #if STEP_MAIN >= 2
+      R1_MAIN(1)
+      #if STEP_MAIN >= 4
+      R2_MAIN(2)
+      #if STEP_MAIN >= 8
+      R2_MAIN(4)
+      R2_MAIN(6)
+      #if STEP_MAIN >= 16
+      R2_MAIN(8)
+      R2_MAIN(10)
+      R2_MAIN(12)
+      R2_MAIN(14)
+      #endif
+      #endif
+      #endif
+      #endif
+    #endif
+  }
+
+  a += state[0]; state[0] = a;
+  b += state[1]; state[1] = b;
+  c += state[2]; state[2] = c;
+  d += state[3]; state[3] = d;
+  e += state[4]; state[4] = e;
+  f += state[5]; state[5] = f;
+  g += state[6]; state[6] = g;
+  h += state[7]; state[7] = h;
+
+  data += SHA512_BLOCK_SIZE;
+  }
+  while (--numBlocks);
+}
+
+
+#define Sha512_UpdateBlock(p) SHA512_UPDATE_BLOCKS(p)(p->state, p->buffer, 1)
+
+void Sha512_Update(CSha512 *p, const Byte *data, size_t size)
+{
+  if (size == 0)
+    return;
+  {
+    const unsigned pos = (unsigned)p->v.vars.count & (SHA512_BLOCK_SIZE - 1);
+    const unsigned num = SHA512_BLOCK_SIZE - pos;
+    p->v.vars.count += size;
+    if (num > size)
+    {
+      memcpy(p->buffer + pos, data, size);
+      return;
+    }
+    if (pos != 0)
+    {
+      size -= num;
+      memcpy(p->buffer + pos, data, num);
+      data += num;
+      Sha512_UpdateBlock(p);
+    }
+  }
+  {
+    const size_t numBlocks = size >> 7;
+    // if (numBlocks)
+    SHA512_UPDATE_BLOCKS(p)(p->state, data, numBlocks);
+    size &= SHA512_BLOCK_SIZE - 1;
+    if (size == 0)
+      return;
+    data += (numBlocks << 7);
+    memcpy(p->buffer, data, size);
+  }
+}
+
+
+void Sha512_Final(CSha512 *p, Byte *digest, unsigned digestSize)
+{
+  unsigned pos = (unsigned)p->v.vars.count & (SHA512_BLOCK_SIZE - 1);
+  p->buffer[pos++] = 0x80;
+  if (pos > (SHA512_BLOCK_SIZE - 8 * 2))
+  {
+    while (pos != SHA512_BLOCK_SIZE) { p->buffer[pos++] = 0; }
+    // memset(&p->buf.buffer[pos], 0, SHA512_BLOCK_SIZE - pos);
+    Sha512_UpdateBlock(p);
+    pos = 0;
+  }
+  memset(&p->buffer[pos], 0, (SHA512_BLOCK_SIZE - 8 * 2) - pos);
+  {
+    const UInt64 numBits = p->v.vars.count << 3;
+    SetBe64(p->buffer + SHA512_BLOCK_SIZE - 8 * 2, 0) // = (p->v.vars.count >> (64 - 3)); (high 64-bits)
+    SetBe64(p->buffer + SHA512_BLOCK_SIZE - 8 * 1, numBits)
+  }
+  Sha512_UpdateBlock(p);
+#if 1 && defined(MY_CPU_BE)
+  memcpy(digest, p->state, digestSize);
+#else
+  {
+    const unsigned numWords = digestSize >> 3;
+    unsigned i;
+    for (i = 0; i < numWords; i++)
+    {
+      const UInt64 v = p->state[i];
+      SetBe64(digest, v)
+      digest += 8;
+    }
+    if (digestSize & 4) // digestSize == SHA512_224_DIGEST_SIZE
+    {
+      const UInt32 v = (UInt32)((p->state[numWords]) >> 32);
+      SetBe32(digest, v)
+    }
+  }
+#endif
+  Sha512_InitState(p, digestSize);
+}
+
+
+
+// #define Z7_SHA512_PROBE_DEBUG // for debug
+
+#if defined(Z7_SHA512_PROBE_DEBUG) || defined(Z7_COMPILER_SHA512_SUPPORTED)
+
+#if defined(Z7_SHA512_PROBE_DEBUG) \
+    || defined(_WIN32) && defined(MY_CPU_ARM64)
+#ifndef Z7_SHA512_USE_PROBE
+#define Z7_SHA512_USE_PROBE
+#endif
+#endif
+
+#ifdef Z7_SHA512_USE_PROBE
+
+#ifdef Z7_SHA512_PROBE_DEBUG
+#include <stdio.h>
+#define PRF(x) x
+#else
+#define PRF(x)
+#endif
+
+#if 0 || !defined(_MSC_VER) // 1 || : for debug LONGJMP mode
+// MINGW doesn't support __try. So we use signal() / longjmp().
+// Note: signal() / longjmp() probably is not thread-safe.
+// So we must call Sha512Prepare() from main thread at program start.
+#ifndef Z7_SHA512_USE_LONGJMP
+#define Z7_SHA512_USE_LONGJMP
+#endif
+#endif
+
+#ifdef Z7_SHA512_USE_LONGJMP
+#include <signal.h>
+#include <setjmp.h>
+static jmp_buf g_Sha512_jmp_buf;
+// static int g_Sha512_Unsupported;
+
+#if defined(__GNUC__) && (__GNUC__ >= 8) \
+    || defined(__clang__) && (__clang_major__ >= 3)
+  __attribute__((noreturn))
+#endif
+static void Z7_CDECL Sha512_signal_Handler(int v)
+{
+  PRF(printf("======== Sha512_signal_Handler = %x\n", (unsigned)v);)
+  // g_Sha512_Unsupported = 1;
+  longjmp(g_Sha512_jmp_buf, 1);
+}
+#endif // Z7_SHA512_USE_LONGJMP
+
+
+#if defined(_WIN32)
+#include "7zWindows.h"
+#endif
+
+#if defined(MY_CPU_ARM64)
+// #define Z7_SHA512_USE_SIMPLIFIED_PROBE // for debug
+#endif
+
+#ifdef Z7_SHA512_USE_SIMPLIFIED_PROBE
+#include <arm_neon.h>
+#if defined(__clang__)
+  __attribute__((__target__("sha3")))
+#elif !defined(_MSC_VER)
+  __attribute__((__target__("arch=armv8.2-a+sha3")))
+#endif
+#endif
+static BoolInt CPU_IsSupported_SHA512_Probe(void)
+{
+  PRF(printf("\n== CPU_IsSupported_SHA512_Probe\n");)
+#if defined(_WIN32) && defined(MY_CPU_ARM64)
+  // we have no SHA512 flag for IsProcessorFeaturePresent() still.
+  if (!CPU_IsSupported_CRYPTO())
+    return False;
+  PRF(printf("==== Registry check\n");)
+  {
+    // we can't read ID_AA64ISAR0_EL1 register from application.
+    // but ID_AA64ISAR0_EL1 register is mapped to "CP 4030" registry value.
+    HKEY key = NULL;
+    LONG res = RegOpenKeyEx(HKEY_LOCAL_MACHINE,
+        TEXT("HARDWARE\\DESCRIPTION\\System\\CentralProcessor\\0"),
+        0, KEY_READ, &key);
+    if (res != ERROR_SUCCESS)
+      return False;
+    {
+      DWORD type = 0;
+      DWORD count = sizeof(UInt64);
+      UInt64 val = 0;
+      res = RegQueryValueEx(key, TEXT("CP 4030"), NULL,
+          &type, (LPBYTE)&val, &count);
+      RegCloseKey(key);
+      if (res != ERROR_SUCCESS
+          || type != REG_QWORD
+          || count != sizeof(UInt64)
+          || ((unsigned)(val >> 12) & 0xf) != 2)
+        return False;
+      // we parse SHA2 field of ID_AA64ISAR0_EL1 register:
+      //   0 : No SHA2 instructions implemented
+      //   1 : SHA256 implemented
+      //   2 : SHA256 and SHA512 implemented
+    }
+  }
+#endif // defined(_WIN32) && defined(MY_CPU_ARM64)
+
+
+#if 1  // 0 for debug to disable SHA512 PROBE code
+
+/*
+----- SHA512 PROBE -----
+
+We suppose that "CP 4030" registry reading is enough.
+But we use additional SHA512 PROBE code, because
+we can catch exception here, and we don't catch exceptions,
+if we call Sha512 functions from main code.
+
+NOTE: arm64 PROBE code doesn't work, if we call it via Wine in linux-arm64.
+The program just stops.
+Also x64 version of PROBE code doesn't work, if we run it via Intel SDE emulator
+without SHA512 support (-skl switch),
+The program stops, and we have message from SDE:
+  TID 0 SDE-ERROR: Executed instruction not valid for specified chip (SKYLAKE): vsha512msg1
+But we still want to catch that exception instead of process stopping.
+Does this PROBE code work in native Windows-arm64 (with/without sha512 hw instructions)?
+Are there any ways to fix the problems with arm64-wine and x64-SDE cases?
+*/
+
+  PRF(printf("==== CPU_IsSupported_SHA512 PROBE\n");)
+  {
+    BoolInt isSupported = False;
+#ifdef Z7_SHA512_USE_LONGJMP
+    void (Z7_CDECL *signal_prev)(int);
+    /*
+    if (g_Sha512_Unsupported)
+    {
+      PRF(printf("==== g_Sha512_Unsupported\n");)
+      return False;
+    }
+    */
+    printf("====== signal(SIGILL)\n");
+    signal_prev = signal(SIGILL, Sha512_signal_Handler);
+    if (signal_prev == SIG_ERR)
+    {
+      PRF(printf("====== signal fail\n");)
+      return False;
+    }
+    // PRF(printf("==== signal_prev = %p\n", (void *)signal_prev);)
+    // docs: Before the specified function is executed,
+    // the value of func is set to SIG_DFL.
+    // So we can exit if (setjmp(g_Sha512_jmp_buf) != 0).
+    PRF(printf("====== setjmp\n");)
+    if (!setjmp(g_Sha512_jmp_buf))
+#else //  Z7_SHA512_USE_LONGJMP
+
+#ifdef _MSC_VER
+#ifdef __clang_major__
+  #pragma GCC diagnostic ignored "-Wlanguage-extension-token"
+#endif
+    __try
+#endif
+#endif //  Z7_SHA512_USE_LONGJMP
+
+    {
+#if defined(Z7_COMPILER_SHA512_SUPPORTED)
+#ifdef Z7_SHA512_USE_SIMPLIFIED_PROBE
+      // simplified sha512 check for arm64:
+      const uint64x2_t a = vdupq_n_u64(1);
+      const uint64x2_t b = vsha512hq_u64(a, a, a);
+      PRF(printf("======== vsha512hq_u64 probe\n");)
+      if ((UInt32)vgetq_lane_u64(b, 0) == 0x11800002)
+#else
+      MY_ALIGN(16)
+      UInt64 temp[SHA512_NUM_DIGEST_WORDS + SHA512_NUM_BLOCK_WORDS];
+      memset(temp, 0x5a, sizeof(temp));
+      PRF(printf("======== Sha512_UpdateBlocks_HW\n");)
+      Sha512_UpdateBlocks_HW(temp,
+          (const Byte *)(const void *)(temp + SHA512_NUM_DIGEST_WORDS), 1);
+      // PRF(printf("======== t = %x\n", (UInt32)temp[0]);)
+      if ((UInt32)temp[0] == 0xa33cfdf7)
+#endif
+      {
+        PRF(printf("======== PROBE SHA512: SHA512 is supported\n");)
+        isSupported = True;
+      }
+#else // Z7_COMPILER_SHA512_SUPPORTED
+      // for debug : we generate bad instrction or raise exception.
+      // __except() doesn't catch raise() calls.
+#ifdef Z7_SHA512_USE_LONGJMP
+      PRF(printf("====== raise(SIGILL)\n");)
+      raise(SIGILL);
+#else
+#if defined(_MSC_VER) && defined(MY_CPU_X86)
+      __asm  ud2
+#endif
+#endif // Z7_SHA512_USE_LONGJMP
+#endif // Z7_COMPILER_SHA512_SUPPORTED
+    }
+
+#ifdef Z7_SHA512_USE_LONGJMP
+    PRF(printf("====== restore signal SIGILL\n");)
+    signal(SIGILL, signal_prev);
+#elif _MSC_VER
+    __except (EXCEPTION_EXECUTE_HANDLER)
+    {
+      PRF(printf("==== CPU_IsSupported_SHA512 __except(EXCEPTION_EXECUTE_HANDLER)\n");)
+    }
+#endif
+    PRF(printf("== return (sha512 supported) = %d\n", isSupported);)
+    return isSupported;
+  }
+#else
+  // without SHA512 PROBE code
+  return True;
+#endif
+}
+
+#endif // Z7_SHA512_USE_PROBE
+#endif // defined(Z7_SHA512_PROBE_DEBUG) || defined(Z7_COMPILER_SHA512_SUPPORTED)
+
+
+void Sha512Prepare(void)
+{
+#ifdef Z7_COMPILER_SHA512_SUPPORTED
+  SHA512_FUNC_UPDATE_BLOCKS f, f_hw;
+  f = Sha512_UpdateBlocks;
+  f_hw = NULL;
+#ifdef Z7_SHA512_USE_PROBE
+  if (CPU_IsSupported_SHA512_Probe())
+#elif defined(MY_CPU_X86_OR_AMD64)
+  if (CPU_IsSupported_SHA512() && CPU_IsSupported_AVX2())
+#else
+  if (CPU_IsSupported_SHA512())
+#endif
+  {
+    // printf("\n========== HW SHA512 ======== \n");
+    f = f_hw = Sha512_UpdateBlocks_HW;
+  }
+  g_SHA512_FUNC_UPDATE_BLOCKS    = f;
+  g_SHA512_FUNC_UPDATE_BLOCKS_HW = f_hw;
+#elif defined(Z7_SHA512_PROBE_DEBUG)
+  CPU_IsSupported_SHA512_Probe(); // for debug
+#endif
+}
+
+
+#undef K
+#undef S0
+#undef S1
+#undef s0
+#undef s1
+#undef Ch
+#undef Maj
+#undef W_MAIN
+#undef W_PRE
+#undef w
+#undef blk2_main
+#undef blk2
+#undef T1
+#undef T4
+#undef T8
+#undef R1_PRE
+#undef R1_MAIN
+#undef R2_MAIN
+#undef R4
+#undef R4_PRE
+#undef R4_MAIN
+#undef R8
+#undef R8_PRE
+#undef R8_MAIN
+#undef STEP_PRE
+#undef STEP_MAIN
+#undef Z7_SHA512_BIG_W
+#undef Z7_SHA512_UNROLL
+#undef Z7_COMPILER_SHA512_SUPPORTED

C/Sha512.h

@@ -0,0 +1,86 @@
+/* Sha512.h -- SHA-512 Hash
+: Igor Pavlov : Public domain */
+
+#ifndef ZIP7_INC_SHA512_H
+#define ZIP7_INC_SHA512_H
+
+#include "7zTypes.h"
+
+EXTERN_C_BEGIN
+
+#define SHA512_NUM_BLOCK_WORDS  16
+#define SHA512_NUM_DIGEST_WORDS  8
+
+#define SHA512_BLOCK_SIZE   (SHA512_NUM_BLOCK_WORDS * 8)
+#define SHA512_DIGEST_SIZE  (SHA512_NUM_DIGEST_WORDS * 8)
+#define SHA512_224_DIGEST_SIZE  (224 / 8)
+#define SHA512_256_DIGEST_SIZE  (256 / 8)
+#define SHA512_384_DIGEST_SIZE  (384 / 8)
+
+typedef void (Z7_FASTCALL *SHA512_FUNC_UPDATE_BLOCKS)(UInt64 state[8], const Byte *data, size_t numBlocks);
+
+/*
+  if (the system supports different SHA512 code implementations)
+  {
+    (CSha512::func_UpdateBlocks) will be used
+    (CSha512::func_UpdateBlocks) can be set by
+       Sha512_Init()        - to default (fastest)
+       Sha512_SetFunction() - to any algo
+  }
+  else
+  {
+    (CSha512::func_UpdateBlocks) is ignored.
+  }
+*/
+
+typedef struct
+{
+  union
+  {
+    struct
+    {
+      SHA512_FUNC_UPDATE_BLOCKS func_UpdateBlocks;
+      UInt64 count;
+    } vars;
+    UInt64 _pad_64bit[8];
+    void *_pad_align_ptr[2];
+  } v;
+  UInt64 state[SHA512_NUM_DIGEST_WORDS];
+  
+  Byte buffer[SHA512_BLOCK_SIZE];
+} CSha512;
+
+
+#define SHA512_ALGO_DEFAULT 0
+#define SHA512_ALGO_SW      1
+#define SHA512_ALGO_HW      2
+
+/*
+Sha512_SetFunction()
+return:
+  0 - (algo) value is not supported, and func_UpdateBlocks was not changed
+  1 - func_UpdateBlocks was set according (algo) value.
+*/
+
+BoolInt Sha512_SetFunction(CSha512 *p, unsigned algo);
+// we support only these (digestSize) values: 224/8, 256/8, 384/8, 512/8
+void Sha512_InitState(CSha512 *p, unsigned digestSize);
+void Sha512_Init(CSha512 *p, unsigned digestSize);
+void Sha512_Update(CSha512 *p, const Byte *data, size_t size);
+void Sha512_Final(CSha512 *p, Byte *digest, unsigned digestSize);
+
+
+
+
+// void Z7_FASTCALL Sha512_UpdateBlocks(UInt64 state[8], const Byte *data, size_t numBlocks);
+
+/*
+call Sha512Prepare() once at program start.
+It prepares all supported implementations, and detects the fastest implementation.
+*/
+
+void Sha512Prepare(void);
+
+EXTERN_C_END
+
+#endif

C/Sha512Opt.c

@@ -0,0 +1,395 @@
+/* Sha512Opt.c -- SHA-512 optimized code for SHA-512 hardware instructions
+: Igor Pavlov : Public domain */
+
+#include "Precomp.h"
+#include "Compiler.h"
+#include "CpuArch.h"
+
+// #define Z7_USE_HW_SHA_STUB // for debug
+#ifdef MY_CPU_X86_OR_AMD64
+  #if defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 2400) && (__INTEL_COMPILER <= 9900) // fix it
+      #define USE_HW_SHA
+  #elif defined(Z7_LLVM_CLANG_VERSION)  && (Z7_LLVM_CLANG_VERSION  >= 170001) \
+     || defined(Z7_APPLE_CLANG_VERSION) && (Z7_APPLE_CLANG_VERSION >= 170001) \
+     || defined(Z7_GCC_VERSION)         && (Z7_GCC_VERSION         >= 140000)
+      #define USE_HW_SHA
+      #if !defined(__INTEL_COMPILER)
+      // icc defines __GNUC__, but icc doesn't support __attribute__(__target__)
+      #if !defined(__SHA512__) || !defined(__AVX2__)
+        #define ATTRIB_SHA512 __attribute__((__target__("sha512,avx2")))
+      #endif
+      #endif
+  #elif defined(Z7_MSC_VER_ORIGINAL)
+    #if (_MSC_VER >= 1940)
+      #define USE_HW_SHA
+    #else
+      // #define Z7_USE_HW_SHA_STUB
+    #endif
+  #endif
+// #endif // MY_CPU_X86_OR_AMD64
+#ifndef USE_HW_SHA
+  // #define Z7_USE_HW_SHA_STUB // for debug
+#endif
+
+#ifdef USE_HW_SHA
+
+// #pragma message("Sha512 HW")
+
+#include <immintrin.h>
+
+#if defined (__clang__) && defined(_MSC_VER)
+  #if !defined(__AVX__)
+    #include <avxintrin.h>
+  #endif
+  #if !defined(__AVX2__)
+    #include <avx2intrin.h>
+  #endif
+  #if !defined(__SHA512__)
+    #include <sha512intrin.h>
+  #endif
+#else
+
+#endif
+
+/*
+SHA512 uses:
+AVX:
+  _mm256_loadu_si256  (vmovdqu)
+  _mm256_storeu_si256
+  _mm256_set_epi32    (unused)
+AVX2:
+  _mm256_add_epi64     : vpaddq
+  _mm256_shuffle_epi8  : vpshufb
+  _mm256_shuffle_epi32 : pshufd
+  _mm256_blend_epi32   : vpblendd
+  _mm256_permute4x64_epi64 : vpermq     : 3c
+  _mm256_permute2x128_si256: vperm2i128 : 3c
+  _mm256_extracti128_si256 : vextracti128  : 3c
+SHA512:
+  _mm256_sha512*
+*/
+
+// K array must be aligned for 32-bytes at least.
+// The compiler can look align attribute and selects
+//  vmovdqu - for code without align attribute
+//  vmovdqa - for code with    align attribute
+extern
+MY_ALIGN(64)
+const UInt64 SHA512_K_ARRAY[80];
+#define K SHA512_K_ARRAY
+
+
+#define ADD_EPI64(dest, src)      dest = _mm256_add_epi64(dest, src);
+#define SHA512_MSG1(dest, src)    dest = _mm256_sha512msg1_epi64(dest, _mm256_extracti128_si256(src, 0));
+#define SHA512_MSG2(dest, src)    dest = _mm256_sha512msg2_epi64(dest, src);
+
+#define LOAD_SHUFFLE(m, k) \
+    m = _mm256_loadu_si256((const __m256i *)(const void *)(data + (k) * 32)); \
+    m = _mm256_shuffle_epi8(m, mask); \
+
+#define NNN(m0, m1, m2, m3)
+
+#define SM1(m1, m2, m3, m0) \
+    SHA512_MSG1(m0, m1); \
+            
+#define SM2(m2, m3, m0, m1) \
+    ADD_EPI64(m0, _mm256_permute4x64_epi64(_mm256_blend_epi32(m2, m3, 3), 0x39)); \
+    SHA512_MSG2(m0, m3); \
+
+#define RND2(t0, t1, lane) \
+    t0 = _mm256_sha512rnds2_epi64(t0, t1, _mm256_extracti128_si256(msg, lane));
+
+
+
+#define R4(k, m0, m1, m2, m3, OP0, OP1) \
+    msg = _mm256_add_epi64(m0, *(const __m256i *) (const void *) &K[(k) * 4]); \
+    RND2(state0, state1, 0);  OP0(m0, m1, m2, m3) \
+    RND2(state1, state0, 1);  OP1(m0, m1, m2, m3) \
+
+
+
+
+#define R16(k, OP0, OP1, OP2, OP3, OP4, OP5, OP6, OP7) \
+    R4 ( (k)*4+0, m0,m1,m2,m3, OP0, OP1 ) \
+    R4 ( (k)*4+1, m1,m2,m3,m0, OP2, OP3 ) \
+    R4 ( (k)*4+2, m2,m3,m0,m1, OP4, OP5 ) \
+    R4 ( (k)*4+3, m3,m0,m1,m2, OP6, OP7 ) \
+
+#define PREPARE_STATE \
+    state0 = _mm256_shuffle_epi32(state0, 0x4e);              /* cdab */ \
+    state1 = _mm256_shuffle_epi32(state1, 0x4e);              /* ghef */ \
+    tmp = state0; \
+    state0 = _mm256_permute2x128_si256(state0, state1, 0x13); /* cdgh */ \
+    state1 = _mm256_permute2x128_si256(tmp,    state1, 2);    /* abef */ \
+
+
+void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks);
+#ifdef ATTRIB_SHA512
+ATTRIB_SHA512
+#endif
+void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks)
+{
+  const __m256i mask = _mm256_set_epi32(
+      0x08090a0b,0x0c0d0e0f, 0x00010203,0x04050607,
+      0x08090a0b,0x0c0d0e0f, 0x00010203,0x04050607);
+  __m256i tmp, state0, state1;
+
+  if (numBlocks == 0)
+    return;
+
+  state0 = _mm256_loadu_si256((const __m256i *) (const void *) &state[0]);
+  state1 = _mm256_loadu_si256((const __m256i *) (const void *) &state[4]);
+  
+  PREPARE_STATE
+
+  do
+  {
+    __m256i state0_save, state1_save;
+    __m256i m0, m1, m2, m3;
+    __m256i msg;
+    // #define msg tmp
+
+    state0_save = state0;
+    state1_save = state1;
+    
+    LOAD_SHUFFLE (m0, 0)
+    LOAD_SHUFFLE (m1, 1)
+    LOAD_SHUFFLE (m2, 2)
+    LOAD_SHUFFLE (m3, 3)
+
+
+
+    R16 ( 0, NNN, NNN, SM1, NNN, SM1, SM2, SM1, SM2 )
+    R16 ( 1, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
+    R16 ( 2, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
+    R16 ( 3, SM1, SM2, SM1, SM2, SM1, SM2, SM1, SM2 )
+    R16 ( 4, SM1, SM2, NNN, SM2, NNN, NNN, NNN, NNN )
+    ADD_EPI64(state0, state0_save)
+    ADD_EPI64(state1, state1_save)
+    
+    data += 128;
+  }
+  while (--numBlocks);
+
+  PREPARE_STATE
+
+  _mm256_storeu_si256((__m256i *) (void *) &state[0], state0);
+  _mm256_storeu_si256((__m256i *) (void *) &state[4], state1);
+}
+
+#endif // USE_HW_SHA
+
+// gcc 8.5 also supports sha512, but we need also support in assembler that is called by gcc
+#elif defined(MY_CPU_ARM64) && defined(MY_CPU_LE)
+  
+  #if defined(__ARM_FEATURE_SHA512)
+    #define USE_HW_SHA
+  #else
+    #if (defined(Z7_CLANG_VERSION) && (Z7_CLANG_VERSION >= 130000) \
+           || defined(__GNUC__) && (__GNUC__ >= 9) \
+          ) \
+      || defined(Z7_MSC_VER_ORIGINAL) && (_MSC_VER >= 1940) // fix it
+      #define USE_HW_SHA
+    #endif
+  #endif
+
+#ifdef USE_HW_SHA
+
+// #pragma message("=== Sha512 HW === ")
+
+
+#if defined(__clang__) || defined(__GNUC__)
+#if !defined(__ARM_FEATURE_SHA512)
+// #pragma message("=== we define SHA3 ATTRIB_SHA512 === ")
+#if defined(__clang__)
+    #define ATTRIB_SHA512 __attribute__((__target__("sha3"))) // "armv8.2-a,sha3"
+#else
+    #define ATTRIB_SHA512 __attribute__((__target__("arch=armv8.2-a+sha3")))
+#endif
+#endif
+#endif
+
+
+#if defined(Z7_MSC_VER_ORIGINAL)
+#include <arm64_neon.h>
+#else
+
+#if defined(__clang__) && __clang_major__ < 16
+#if !defined(__ARM_FEATURE_SHA512)
+// #pragma message("=== we set __ARM_FEATURE_SHA512 1 === ")
+    Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #define Z7_ARM_FEATURE_SHA512_WAS_SET 1
+    #define __ARM_FEATURE_SHA512 1
+    Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+#endif
+#endif // clang
+
+#include <arm_neon.h>
+
+#if defined(Z7_ARM_FEATURE_SHA512_WAS_SET) && \
+    defined(__ARM_FEATURE_SHA512)
+    Z7_DIAGNOSTIC_IGNORE_BEGIN_RESERVED_MACRO_IDENTIFIER
+    #undef __ARM_FEATURE_SHA512
+    #undef Z7_ARM_FEATURE_SHA512_WAS_SET
+    Z7_DIAGNOSTIC_IGNORE_END_RESERVED_MACRO_IDENTIFIER
+// #pragma message("=== we undefine __ARM_FEATURE_CRYPTO === ")
+#endif
+
+#endif // Z7_MSC_VER_ORIGINAL
+
+typedef uint64x2_t v128_64;
+// typedef __n128 v128_64; // MSVC
+
+#ifdef MY_CPU_BE
+  #define MY_rev64_for_LE(x) x
+#else
+  #define MY_rev64_for_LE(x) vrev64q_u8(x)
+#endif
+
+#define LOAD_128_64(_p)       vld1q_u64(_p)
+#define LOAD_128_8(_p)        vld1q_u8 (_p)
+#define STORE_128_64(_p, _v)  vst1q_u64(_p, _v)
+
+#define LOAD_SHUFFLE(m, k) \
+    m = vreinterpretq_u64_u8( \
+        MY_rev64_for_LE( \
+        LOAD_128_8(data + (k) * 16))); \
+
+// K array must be aligned for 16-bytes at least.
+extern
+MY_ALIGN(64)
+const UInt64 SHA512_K_ARRAY[80];
+#define K SHA512_K_ARRAY
+
+#define NN(m0, m1, m4, m5, m7)
+#define SM(m0, m1, m4, m5, m7) \
+    m0 = vsha512su1q_u64(vsha512su0q_u64(m0, m1), m7, vextq_u64(m4, m5, 1));
+
+#define R2(k, m0,m1,m2,m3,m4,m5,m6,m7, a0,a1,a2,a3, OP) \
+    OP(m0, m1, m4, m5, m7) \
+    t = vaddq_u64(m0, vld1q_u64(k)); \
+    t = vaddq_u64(vextq_u64(t, t, 1), a3); \
+    t = vsha512hq_u64(t, vextq_u64(a2, a3, 1), vextq_u64(a1, a2, 1)); \
+    a3 = vsha512h2q_u64(t, a1, a0); \
+    a1 = vaddq_u64(a1, t); \
+
+#define R8(k,     m0,m1,m2,m3,m4,m5,m6,m7, OP) \
+    R2 ( (k)+0*2, m0,m1,m2,m3,m4,m5,m6,m7, a0,a1,a2,a3, OP ) \
+    R2 ( (k)+1*2, m1,m2,m3,m4,m5,m6,m7,m0, a3,a0,a1,a2, OP ) \
+    R2 ( (k)+2*2, m2,m3,m4,m5,m6,m7,m0,m1, a2,a3,a0,a1, OP ) \
+    R2 ( (k)+3*2, m3,m4,m5,m6,m7,m0,m1,m2, a1,a2,a3,a0, OP ) \
+
+#define R16(k, OP) \
+    R8 ( (k)+0*2, m0,m1,m2,m3,m4,m5,m6,m7, OP ) \
+    R8 ( (k)+4*2, m4,m5,m6,m7,m0,m1,m2,m3, OP ) \
+
+
+void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks);
+#ifdef ATTRIB_SHA512
+ATTRIB_SHA512
+#endif
+void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks)
+{
+  v128_64 a0, a1, a2, a3;
+
+  if (numBlocks == 0)
+    return;
+  a0 = LOAD_128_64(&state[0]);
+  a1 = LOAD_128_64(&state[2]);
+  a2 = LOAD_128_64(&state[4]);
+  a3 = LOAD_128_64(&state[6]);
+  do
+  {
+    v128_64 a0_save, a1_save, a2_save, a3_save;
+    v128_64 m0, m1, m2, m3, m4, m5, m6, m7;
+    v128_64 t;
+    unsigned i;
+    const UInt64 *k_ptr;
+    
+    LOAD_SHUFFLE (m0, 0)
+    LOAD_SHUFFLE (m1, 1)
+    LOAD_SHUFFLE (m2, 2)
+    LOAD_SHUFFLE (m3, 3)
+    LOAD_SHUFFLE (m4, 4)
+    LOAD_SHUFFLE (m5, 5)
+    LOAD_SHUFFLE (m6, 6)
+    LOAD_SHUFFLE (m7, 7)
+
+    a0_save = a0;
+    a1_save = a1;
+    a2_save = a2;
+    a3_save = a3;
+    
+    R16 ( K, NN )
+    k_ptr = K + 16;
+    for (i = 0; i < 4; i++)
+    {
+      R16 ( k_ptr, SM )
+      k_ptr += 16;
+    }
+    
+    a0 = vaddq_u64(a0, a0_save);
+    a1 = vaddq_u64(a1, a1_save);
+    a2 = vaddq_u64(a2, a2_save);
+    a3 = vaddq_u64(a3, a3_save);
+
+    data += 128;
+  }
+  while (--numBlocks);
+
+  STORE_128_64(&state[0], a0);
+  STORE_128_64(&state[2], a1);
+  STORE_128_64(&state[4], a2);
+  STORE_128_64(&state[6], a3);
+}
+
+#endif // USE_HW_SHA
+
+#endif // MY_CPU_ARM_OR_ARM64
+
+
+#if !defined(USE_HW_SHA) && defined(Z7_USE_HW_SHA_STUB)
+// #error Stop_Compiling_UNSUPPORTED_SHA
+// #include <stdlib.h>
+// We can compile this file with another C compiler,
+// or we can compile asm version.
+// So we can generate real code instead of this stub function.
+// #include "Sha512.h"
+// #if defined(_MSC_VER)
+#pragma message("Sha512 HW-SW stub was used")
+// #endif
+void Z7_FASTCALL Sha512_UpdateBlocks   (UInt64 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks);
+void Z7_FASTCALL Sha512_UpdateBlocks_HW(UInt64 state[8], const Byte *data, size_t numBlocks)
+{
+  Sha512_UpdateBlocks(state, data, numBlocks);
+  /*
+  UNUSED_VAR(state);
+  UNUSED_VAR(data);
+  UNUSED_VAR(numBlocks);
+  exit(1);
+  return;
+  */
+}
+#endif
+
+
+#undef K
+#undef RND2
+#undef MY_rev64_for_LE
+#undef NN
+#undef NNN
+#undef LOAD_128
+#undef STORE_128
+#undef LOAD_SHUFFLE
+#undef SM1
+#undef SM2
+#undef SM
+#undef R2
+#undef R4
+#undef R16
+#undef PREPARE_STATE
+#undef USE_HW_SHA
+#undef ATTRIB_SHA512
+#undef USE_VER_MIN
+#undef Z7_USE_HW_SHA_STUB

C/Sort.c

@@ -1,141 +1,268 @@
 /* Sort.c -- Sort functions
-2014-04-05 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
 #include "Sort.h"
+#include "CpuArch.h"
 
-#define HeapSortDown(p, k, size, temp) \
+#if (  (defined(__GNUC__) && (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))) \
-  { for (;;) { \
+    || (defined(__clang__) && Z7_has_builtin(__builtin_prefetch)) \
-    size_t s = (k << 1); \
+    )
-    if (s > size) break; \
+// the code with prefetch is slow for small arrays on x86.
-    if (s < size && p[s + 1] > p[s]) s++; \
+// So we disable prefetch for x86.
-    if (temp >= p[s]) break; \
+#ifndef MY_CPU_X86
-    p[k] = p[s]; k = s; \
+  // #pragma message("Z7_PREFETCH : __builtin_prefetch")
-  } p[k] = temp; }
+  #define Z7_PREFETCH(a)  __builtin_prefetch((a))
+#endif
 
-void HeapSort(UInt32 *p, size_t size)
+#elif defined(_WIN32) // || defined(_MSC_VER) && (_MSC_VER >= 1200)
-{
-  if (size <= 1)
-    return;
-  p--;
-  {
-    size_t i = size / 2;
-    do
-    {
-      UInt32 temp = p[i];
-      size_t k = i;
-      HeapSortDown(p, k, size, temp)
-    }
-    while (--i != 0);
-  }
-  /*
-  do
-  {
-    size_t k = 1;
-    UInt32 temp = p[size];
-    p[size--] = p[1];
-    HeapSortDown(p, k, size, temp)
-  }
-  while (size > 1);
-  */
-  while (size > 3)
-  {
-    UInt32 temp = p[size];
-    size_t k = (p[3] > p[2]) ? 3 : 2;
-    p[size--] = p[1];
-    p[1] = p[k];
-    HeapSortDown(p, k, size, temp)
-  }
-  {
-    UInt32 temp = p[size];
-    p[size] = p[1];
-    if (size > 2 && p[2] < temp)
-    {
-      p[1] = p[2];
-      p[2] = temp;
-    }
-    else
-      p[1] = temp;
-  }
-}
 
-void HeapSort64(UInt64 *p, size_t size)
+#include "7zWindows.h"
-{
+
-  if (size <= 1)
+// NOTE: CLANG/GCC/MSVC can define different values for _MM_HINT_T0 / PF_TEMPORAL_LEVEL_1.
-    return;
+// For example, clang-cl can generate "prefetcht2" instruction for
-  p--;
+// PreFetchCacheLine(PF_TEMPORAL_LEVEL_1) call.
-  {
+// But we want to generate "prefetcht0" instruction.
-    size_t i = size / 2;
+// So for CLANG/GCC we must use __builtin_prefetch() in code branch above
-    do
+// instead of PreFetchCacheLine() / _mm_prefetch().
-    {
+
-      UInt64 temp = p[i];
+// New msvc-x86 compiler generates "prefetcht0" instruction for PreFetchCacheLine() call.
-      size_t k = i;
+// But old x86 cpus don't support "prefetcht0".
-      HeapSortDown(p, k, size, temp)
+// So we will use PreFetchCacheLine(), only if we are sure that
-    }
+// generated instruction is supported by all cpus of that isa.
-    while (--i != 0);
+#if defined(MY_CPU_AMD64) \
-  }
+    || defined(MY_CPU_ARM64) \
-  /*
+    || defined(MY_CPU_IA64)
-  do
+// we need to use additional braces for (a) in PreFetchCacheLine call, because
-  {
+// PreFetchCacheLine macro doesn't use braces:
-    size_t k = 1;
+//   #define PreFetchCacheLine(l, a)  _mm_prefetch((CHAR CONST *) a, l)
-    UInt64 temp = p[size];
+  // #pragma message("Z7_PREFETCH : PreFetchCacheLine")
-    p[size--] = p[1];
+  #define Z7_PREFETCH(a)  PreFetchCacheLine(PF_TEMPORAL_LEVEL_1, (a))
-    HeapSortDown(p, k, size, temp)
+#endif
-  }
+
-  while (size > 1);
+#endif // _WIN32
-  */
+
-  while (size > 3)
+
-  {
+#define PREFETCH_NO(p,k,s,size)
-    UInt64 temp = p[size];
+
-    size_t k = (p[3] > p[2]) ? 3 : 2;
+#ifndef Z7_PREFETCH
-    p[size--] = p[1];
+  #define SORT_PREFETCH(p,k,s,size)
-    p[1] = p[k];
+#else
-    HeapSortDown(p, k, size, temp)
+
-  }
+// #define PREFETCH_LEVEL 2  // use it if cache line is 32-bytes
-  {
+#define PREFETCH_LEVEL 3  // it is fast for most cases (64-bytes cache line prefetch)
-    UInt64 temp = p[size];
+// #define PREFETCH_LEVEL 4  // it can be faster for big array (128-bytes prefetch)
-    p[size] = p[1];
+
-    if (size > 2 && p[2] < temp)
+#if PREFETCH_LEVEL == 0
-    {
+
-      p[1] = p[2];
+  #define SORT_PREFETCH(p,k,s,size)
-      p[2] = temp;
+
-    }
+#else // PREFETCH_LEVEL != 0
-    else
-      p[1] = temp;
-  }
-}
 
 /*
-#define HeapSortRefDown(p, vals, n, size, temp) \
+if  defined(USE_PREFETCH_FOR_ALIGNED_ARRAY)
-  { size_t k = n; UInt32 val = vals[temp]; for (;;) { \
+    we prefetch one value per cache line.
-    size_t s = (k << 1); \
+    Use it if array is aligned for cache line size (64 bytes)
-    if (s > size) break; \
+    or if array is small (less than L1 cache size).
-    if (s < size && vals[p[s + 1]] > vals[p[s]]) s++; \
-    if (val >= vals[p[s]]) break; \
-    p[k] = p[s]; k = s; \
-  } p[k] = temp; }
 
-void HeapSortRef(UInt32 *p, UInt32 *vals, size_t size)
+if !defined(USE_PREFETCH_FOR_ALIGNED_ARRAY)
+    we perfetch all cache lines that can be required.
+    it can be faster for big unaligned arrays.
+*/
+  #define USE_PREFETCH_FOR_ALIGNED_ARRAY
+
+// s == k * 2
+#if 0 && PREFETCH_LEVEL <= 3 && defined(MY_CPU_X86_OR_AMD64)
+  // x86 supports (lea r1*8+offset)
+  #define PREFETCH_OFFSET(k,s)  ((s) << PREFETCH_LEVEL)
+#else
+  #define PREFETCH_OFFSET(k,s)  ((k) << (PREFETCH_LEVEL + 1))
+#endif
+
+#if 1 && PREFETCH_LEVEL <= 3 && defined(USE_PREFETCH_FOR_ALIGNED_ARRAY)
+  #define PREFETCH_ADD_OFFSET   0
+#else
+  // last offset that can be reqiured in PREFETCH_LEVEL step:
+  #define PREFETCH_RANGE        ((2 << PREFETCH_LEVEL) - 1)
+  #define PREFETCH_ADD_OFFSET   PREFETCH_RANGE / 2
+#endif
+
+#if PREFETCH_LEVEL <= 3
+
+#ifdef USE_PREFETCH_FOR_ALIGNED_ARRAY
+  #define SORT_PREFETCH(p,k,s,size) \
+  { const size_t s2 = PREFETCH_OFFSET(k,s) + PREFETCH_ADD_OFFSET; \
+    if (s2 <= size) { \
+      Z7_PREFETCH((p + s2)); \
+  }}
+#else /* for unaligned array */
+  #define SORT_PREFETCH(p,k,s,size) \
+  { const size_t s2 = PREFETCH_OFFSET(k,s) + PREFETCH_RANGE; \
+    if (s2 <= size) { \
+      Z7_PREFETCH((p + s2 - PREFETCH_RANGE)); \
+      Z7_PREFETCH((p + s2)); \
+  }}
+#endif
+
+#else // PREFETCH_LEVEL > 3
+
+#ifdef USE_PREFETCH_FOR_ALIGNED_ARRAY
+  #define SORT_PREFETCH(p,k,s,size) \
+  { const size_t s2 = PREFETCH_OFFSET(k,s) + PREFETCH_RANGE - 16 / 2; \
+    if (s2 <= size) { \
+      Z7_PREFETCH((p + s2 - 16)); \
+      Z7_PREFETCH((p + s2)); \
+  }}
+#else /* for unaligned array */
+  #define SORT_PREFETCH(p,k,s,size) \
+  { const size_t s2 = PREFETCH_OFFSET(k,s) + PREFETCH_RANGE; \
+    if (s2 <= size) { \
+      Z7_PREFETCH((p + s2 - PREFETCH_RANGE)); \
+      Z7_PREFETCH((p + s2 - PREFETCH_RANGE / 2)); \
+      Z7_PREFETCH((p + s2)); \
+  }}
+#endif
+
+#endif // PREFETCH_LEVEL > 3
+#endif // PREFETCH_LEVEL != 0
+#endif // Z7_PREFETCH
+
+
+#if defined(MY_CPU_ARM64) \
+    /* || defined(MY_CPU_AMD64) */ \
+    /* || defined(MY_CPU_ARM) && !defined(_MSC_VER) */
+  // we want to use cmov, if cmov is very fast:
+  // - this cmov version is slower for clang-x64.
+  // - this cmov version is faster for gcc-arm64 for some fast arm64 cpus.
+  #define Z7_FAST_CMOV_SUPPORTED
+#endif
+ 
+#ifdef Z7_FAST_CMOV_SUPPORTED
+  // we want to use cmov here, if cmov is fast: new arm64 cpus.
+  // we want the compiler to use conditional move for this branch
+  #define GET_MAX_VAL(n0, n1, max_val_slow)  if (n0 < n1) n0 = n1;
+#else
+  // use this branch, if cpu doesn't support fast conditional move.
+  // it uses slow array access reading:
+  #define GET_MAX_VAL(n0, n1, max_val_slow)  n0 = max_val_slow;
+#endif
+
+#define HeapSortDown(p, k, size, temp, macro_prefetch) \
+{ \
+  for (;;) { \
+    UInt32 n0, n1; \
+    size_t s = k * 2; \
+    if (s >= size) { \
+      if (s == size) { \
+        n0 = p[s]; \
+        p[k] = n0; \
+        if (temp < n0) k = s; \
+      } \
+      break; \
+    } \
+    n0 = p[k * 2]; \
+    n1 = p[k * 2 + 1]; \
+    s += n0 < n1; \
+    GET_MAX_VAL(n0, n1, p[s]) \
+    if (temp >= n0) break; \
+    macro_prefetch(p, k, s, size) \
+    p[k] = n0; \
+    k = s; \
+  } \
+  p[k] = temp; \
+}
+
+
+/*
+stage-1 : O(n) :
+  we generate intermediate partially sorted binary tree:
+  p[0]  : it's additional item for better alignment of tree structure in memory.
+  p[1]
+  p[2]       p[3]
+  p[4] p[5]  p[6] p[7]
+  ...
+  p[x] >= p[x * 2]
+  p[x] >= p[x * 2 + 1]
+  
+stage-2 : O(n)*log2(N):
+  we move largest item p[0] from head of tree to the end of array
+  and insert last item to sorted binary tree.
+*/
+
+// (p) must be aligned for cache line size (64-bytes) for best performance
+
+void Z7_FASTCALL HeapSort(UInt32 *p, size_t size)
 {
-  if (size <= 1)
+  if (size < 2)
     return;
-  p--;
+  if (size == 2)
   {
-    size_t i = size / 2;
+    const UInt32 a0 = p[0];
+    const UInt32 a1 = p[1];
+    const unsigned k = a1 < a0;
+    p[k] = a0;
+    p[k ^ 1] = a1;
+    return;
+  }
+  {
+    // stage-1 : O(n)
+    // we transform array to partially sorted binary tree.
+    size_t i = --size / 2;
+    // (size) now is the index of the last item in tree,
+    // if (i)
+    {
+      do
+      {
+        const UInt32 temp = p[i];
+        size_t k = i;
+        HeapSortDown(p, k, size, temp, PREFETCH_NO)
+      }
+      while (--i);
+    }
+    {
+      const UInt32 temp = p[0];
+      const UInt32 a1 = p[1];
+      if (temp < a1)
+      {
+        size_t k = 1;
+        p[0] = a1;
+        HeapSortDown(p, k, size, temp, PREFETCH_NO)
+      }
+    }
+  }
+
+  if (size < 3)
+  {
+    // size == 2
+    const UInt32 a0 = p[0];
+    p[0] = p[2];
+    p[2] = a0;
+    return;
+  }
+  if (size != 3)
+  {
+    // stage-2 : O(size) * log2(size):
+    // we move largest item p[0] from head to the end of array,
+    // and insert last item to sorted binary tree.
     do
     {
-      UInt32 temp = p[i];
+      const UInt32 temp = p[size];
-      HeapSortRefDown(p, vals, i, size, temp);
+      size_t k = p[2] < p[3] ? 3 : 2;
+      p[size--] = p[0];
+      p[0] = p[1];
+      p[1] = p[k];
+      HeapSortDown(p, k, size, temp, SORT_PREFETCH) // PREFETCH_NO
     }
-    while (--i != 0);
+    while (size != 3);
   }
-  do
   {
-    UInt32 temp = p[size];
+    const UInt32 a2 = p[2];
-    p[size--] = p[1];
+    const UInt32 a3 = p[3];
-    HeapSortRefDown(p, vals, 1, size, temp);
+    const size_t k = a2 < a3;
+    p[2] = p[1];
+    p[3] = p[0];
+    p[k] = a3;
+    p[k ^ 1] = a2;
   }
-  while (size > 1);
 }
-*/