update the lzma lib to 2409

stefankueng · Dec 7, 2024 · 404172d · 404172d
1 parent d4d6f4f
commit 404172d
Show file tree

Hide file tree

Showing 169 changed files with 7,285 additions and 2,782 deletions.
diff --git a/lzma/7Zip.vcxproj b/lzma/7Zip.vcxproj
@@ -274,8 +274,12 @@
     <ClCompile Include="CPP\7zip\Compress\ZstdDecoder.cpp" />
     <ClCompile Include="CPP\Common\CksumReg.cpp" />
     <ClCompile Include="CPP\Common\LzFindPrepare.cpp" />
+    <ClCompile Include="CPP\Common\Md5Reg.cpp" />
     <ClCompile Include="CPP\Common\Sha1Prepare.cpp" />
     <ClCompile Include="CPP\Common\Sha256Prepare.cpp" />
+    <ClCompile Include="CPP\Common\Sha3Reg.cpp" />
+    <ClCompile Include="CPP\Common\Sha512Prepare.cpp" />
+    <ClCompile Include="CPP\Common\Sha512Reg.cpp" />
     <ClCompile Include="CPP\Common\Xxh64Reg.cpp" />
     <ClCompile Include="CPP\Windows\PropVariantConv.cpp" />
     <ClCompile Include="C\7zBuf2.c" />
@@ -356,6 +360,7 @@
       <Optimization Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">MaxSpeed</Optimization>
       <Optimization Condition="'$(Configuration)|$(Platform)'=='Release|x64'">MaxSpeed</Optimization>
     </ClCompile>
+    <ClCompile Include="C\Md5.c" />
     <ClCompile Include="C\MtCoder.c">
     </ClCompile>
     <ClCompile Include="C\MtDec.c">
@@ -398,6 +403,9 @@
     </ClCompile>
     <ClCompile Include="C\Sha256Opt.c">
     </ClCompile>
+    <ClCompile Include="C\Sha3.c" />
+    <ClCompile Include="C\Sha512.c" />
+    <ClCompile Include="C\Sha512Opt.c" />
     <ClCompile Include="C\Sort.c">
       <Optimization Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">MaxSpeed</Optimization>
       <Optimization Condition="'$(Configuration)|$(Platform)'=='Release|x64'">MaxSpeed</Optimization>
@@ -812,6 +820,7 @@
     <ClInclude Include="C\Lzma2Enc.h" />
     <ClInclude Include="C\LzmaDec.h" />
     <ClInclude Include="C\LzmaEnc.h" />
+    <ClInclude Include="C\Md5.h" />
     <ClInclude Include="C\MtCoder.h" />
     <ClInclude Include="C\MtDec.h" />
     <ClInclude Include="C\Ppmd.h" />
@@ -821,6 +830,8 @@
     <ClInclude Include="C\RotateDefs.h" />
     <ClInclude Include="C\Sha1.h" />
     <ClInclude Include="C\Sha256.h" />
+    <ClInclude Include="C\Sha3.h" />
+    <ClInclude Include="C\Sha512.h" />
     <ClInclude Include="C\Sort.h" />
     <ClInclude Include="CPP\7zip\Bundles\Format7zF\StdAfx.h" />
     <ClInclude Include="C\SwapBytes.h" />

diff --git a/lzma/7Zip.vcxproj.filters b/lzma/7Zip.vcxproj.filters
@@ -1041,6 +1041,30 @@
     <ClCompile Include="CPP\7zip\Common\FileStreams.cpp">
       <Filter>7zip Common</Filter>
     </ClCompile>
+    <ClCompile Include="C\Sha3.c">
+      <Filter>C</Filter>
+    </ClCompile>
+    <ClCompile Include="C\Sha512.c">
+      <Filter>C</Filter>
+    </ClCompile>
+    <ClCompile Include="C\Sha512Opt.c">
+      <Filter>C</Filter>
+    </ClCompile>
+    <ClCompile Include="C\Md5.c">
+      <Filter>C</Filter>
+    </ClCompile>
+    <ClCompile Include="CPP\Common\Md5Reg.cpp">
+      <Filter>Common</Filter>
+    </ClCompile>
+    <ClCompile Include="CPP\Common\Sha3Reg.cpp">
+      <Filter>Common</Filter>
+    </ClCompile>
+    <ClCompile Include="CPP\Common\Sha512Prepare.cpp">
+      <Filter>Common</Filter>
+    </ClCompile>
+    <ClCompile Include="CPP\Common\Sha512Reg.cpp">
+      <Filter>Common</Filter>
+    </ClCompile>
   </ItemGroup>
   <ItemGroup>
     <ResourceCompile Include="CPP\7zip\Bundles\Format7zF\resource.rc" />
@@ -1755,6 +1779,15 @@
     <ClInclude Include="CPP\7zip\Common\FileStreams.h">
       <Filter>7zip Common</Filter>
     </ClInclude>
+    <ClInclude Include="C\Sha3.h">
+      <Filter>C</Filter>
+    </ClInclude>
+    <ClInclude Include="C\Sha512.h">
+      <Filter>C</Filter>
+    </ClInclude>
+    <ClInclude Include="C\Md5.h">
+      <Filter>C</Filter>
+    </ClInclude>
   </ItemGroup>
   <ItemGroup>
     <CustomBuild Include="CPP\7zip\Archive\Archive2.def">

diff --git a/lzma/Asm/x86/LzFindOpt.asm b/lzma/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
diff --git a/lzma/Asm/x86/LzmaDecOpt.asm b/lzma/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
diff --git a/lzma/Asm/x86/Sha1Opt.asm b/lzma/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

diff --git a/lzma/Asm/x86/Sha256Opt.asm b/lzma/Asm/x86/Sha256Opt.asm
@@ -1,5 +1,5 @@
 ; Sha256Opt.asm -- SHA-256 optimized code for SHA-256 x86 hardware instructions
-; 2022-04-17 : 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

diff --git a/lzma/C/7zDec.c b/lzma/C/7zDec.c
@@ -1,5 +1,5 @@
 /* 7zDec.c -- Decoding from 7z folder
-2024-03-01 : Igor Pavlov : Public domain */
+: Igor Pavlov : Public domain */
 
 #include "Precomp.h"
 
@@ -312,8 +312,9 @@ static BoolInt IS_MAIN_METHOD(UInt32 m)
     case k_PPMD:
   #endif
       return True;
+    default:
+      return False;
   }
-  return False;
 }
 
 static BoolInt IS_SUPPORTED_CODER(const CSzCoderInfo *c)

diff --git a/lzma/C/7zVersion.h b/lzma/C/7zVersion.h
@@ -1,7 +1,7 @@
 #define MY_VER_MAJOR 24
-#define MY_VER_MINOR 06
+#define MY_VER_MINOR 9
 #define MY_VER_BUILD 0
-#define MY_VERSION_NUMBERS "24.06"
+#define MY_VERSION_NUMBERS "24.09"
 #define MY_VERSION MY_VERSION_NUMBERS
 
 #ifdef MY_CPU_NAME
@@ -10,7 +10,7 @@
   #define MY_VERSION_CPU MY_VERSION
 #endif
 
-#define MY_DATE "2024-05-26"
+#define MY_DATE "2024-11-29"
 #undef MY_COPYRIGHT
 #undef MY_VERSION_COPYRIGHT_DATE
 #define MY_AUTHOR_NAME "Igor Pavlov"