stub/amd64-win64.pe.h matches upx-v4.2.4

so that anti-virus checkers of apps for 64-bit Windows that are newly-
compressed by upx should not complain about new matches with malware.
This costs a few bytes and a small fraction of cycles in NRV2
de-compressors, but is considered to be a tolerable concession
to the cesspool environment of software for Microsoft Windows.
	modified:   ../misc/testsuite/upx_testsuite_1-expected_sha256sums.sh
	modified:   p_w64pe_amd64.cpp   see code comments
	modified:   stub/amd64-win64.pe.h
	modified:   stub/src/arch/amd64/lzma_d.S   see code comments
	modified:   stub/src/arch/amd64/nrv2b_d.S
	modified:   stub/src/arch/amd64/nrv2d_d.S
	modified:   stub/src/arch/amd64/nrv2e_d.S
	modified:   stub/tmp/amd64-win64.pe.bin.dump

misc/testsuite/upx_testsuite_1-expected_sha256sums.sh

@@ -48,7 +48,7 @@ a0950546dc17fca9437219431d8ddb0249ce5b08e899e0c799a87ac982adee70 *powerpc64le-li
 expected_sha256sums__t110_compress_ucl_nrv2b_3_no_filter="\
 e4f91dbcb627c0c3fbc4d04afb98f9f1671062b7a997cf0e90826caf33f05cf8 *amd64-linux.elf/upx-3.91
 bda0ed1fac238de8472f8248a3d923e0cec78389e88ec7e72262bad241c88b62 *amd64-linux.elf/upx-3.95
-be68107b41e5ac9fdfe919f6e26b0ed83807dbe1c8f20d9e391a836be445c4fd *amd64-win64.pe/upx-3.95.exe
+1bd70a3cd3b34e635a3c947cacd46c4d66bf07e0fd350cc6d1f3cf60483f75b9 *amd64-win64.pe/upx-3.95.exe
 1beaafe2a32c90a82dc122aa00a0af5e0a00862a7444d5d1274c9303289115a5 *arm-linux.elf/upx-3.95
 94422a91473a467dfb534afa3f20247bc38691aeb386a42fb3e0fd18d3d0ae08 *arm-wince.pe/upx-3.91.exe
 16a4e2ca0db11335534c1a951f3fe61a7013fb6495c794b97e1995864f1d662e *arm64-linux.elf/upx-3.95
@@ -71,7 +71,7 @@ e72f0ff8c367d4940aaf49a61e83369ad83972b8835b071f97eb1cd9e1e8f356 *mipsel-linux.e
 expected_sha256sums__t120_compress_ucl_nrv2d_3_no_filter="\
 0a269fe9119f85366905b80c7d583ad9834e65556429c5ce868aefc82c3be4d9 *amd64-linux.elf/upx-3.91
 c6ab1d511a0a19875b474564b49b964c4d17b4e8730ab209778d7c2985fcee8b *amd64-linux.elf/upx-3.95
-246079a063b098268c2c0c6237d7a4d0b7e18eaace45eff33a78400e1595f066 *amd64-win64.pe/upx-3.95.exe
+0314c6974f29d180e46e41a1004d46a3993fe96f03de7f78d45ad89ad3b2177e *amd64-win64.pe/upx-3.95.exe
 f45844cb8a88cb79efcabcb942e9e24a1a08f5b90b5a5c2094384390760c0677 *arm-linux.elf/upx-3.95
 d7839b7dee866794e4ca3df32e2b1084b35aa22dee5f060fa14b5ca09c612ba3 *arm-wince.pe/upx-3.91.exe
 323762dd4e77c42b2b6288339a52cd8b15527441dab101299eda29ec27d452e2 *arm64-linux.elf/upx-3.95
@@ -94,7 +94,7 @@ a5e89be848c503ad5610800d13e363243340dbc3751e4fdaa06405c1097d5b9c *mipsel-linux.e
 expected_sha256sums__t130_compress_ucl_nrv2e_3_no_filter="\
 d9adcdf406ed22382c408670c6663d4bc420dc7a3a027d07e8f843441ae26c56 *amd64-linux.elf/upx-3.91
 675fd00c9cb548a2c6587832703407425b300076fb17883c34234f2441de0bb6 *amd64-linux.elf/upx-3.95
-6caf4df19cfe55fc33b31b4e77b6ce0a5665d090de9d4210c1994b62049c66c5 *amd64-win64.pe/upx-3.95.exe
+97bb5953405f3413d9ff9bdc272caf2858697168b3e85dab54213da356b5ea0f *amd64-win64.pe/upx-3.95.exe
 7b754f16dabe92a2265062dbf0850a6a1041264116dd6a2bc38565e93c29a27d *arm-linux.elf/upx-3.95
 8c73f5acfc70e3c518be09978af87b9bec4906240cb457a166da0eba05bea4db *arm-wince.pe/upx-3.91.exe
 876d844e5217897ae2e36a14e302c0808f7ea41eaba66f4b7e151d1fb5298d14 *arm64-linux.elf/upx-3.95
@@ -140,7 +140,7 @@ c45deba5313a02ba8269ab7d148c4ce33d0c9f3e8c6d6ff528d6535a647fe945 *mipsel-linux.e
 expected_sha256sums__t150_compress_ucl_2_all_filters="\
 f2a11bf41482521238f715ad1df1b8b3c9ae917f2b99aa25017a67db0e53f712 *amd64-linux.elf/upx-3.91
 3cea8f06d1bf0e7c8802f254e304dad55d826c07c54cea91121da208e571f151 *amd64-linux.elf/upx-3.95
-fa8ba95a5fcc1a0d591e9b50d26772b3334815ffbf95d0127b4daa6c74ecc28d *amd64-win64.pe/upx-3.95.exe
+434732efb16b89114162a1787a28c128e2c11edb9835411f2b707cac5cd475ae *amd64-win64.pe/upx-3.95.exe
 536109132c83eaf7f0c7b10d268e52928f2e031219b568cbae842e03eda074b8 *arm-linux.elf/upx-3.95
 93e2ead9e54732a1731896681d415e3ce60b7420ad7f3c57822ad7dbd061cdd3 *arm-wince.pe/upx-3.91.exe
 cde130452ae6f94865767e7d27740a0d87bd6b07813b44e2b210c6eb444b597f *arm64-linux.elf/upx-3.95
@@ -186,7 +186,7 @@ bce0a4cf0d72898dccb0a0a9c82c6ee5934be8e99907e4ccbb489aa2bf803fd1 *mips-linux.elf
 expected_sha256sums__t170_compress_all_methods_no_lzma_5_no_filter="\
 885c3dbbc87c83258986005978d4863febce98c290aed5d868b87901b7a4a452 *amd64-linux.elf/upx-3.91
 2ee7f4c38df398b3b25fe339bf4a815c166eddb27bb58954ba9bce4661d7927f *amd64-linux.elf/upx-3.95
-fb362f2b3ea02f9c12de7e8f19f73ea6cfebfeb7b8c316003121a9ad13d3d8ae *amd64-win64.pe/upx-3.95.exe
+5fccb0f5044baaa74d176efb25eefa3b5810dc737ed0f9d38084d275f7ebe7b8 *amd64-win64.pe/upx-3.95.exe
 46d75aa6fa1ecec6c6911b5a6631267fac9819a5a457ad6de7da6c7c478443a0 *arm-linux.elf/upx-3.95
 56be0b4046accdfadbe534ecb6e6452b9e3a18bb9eac39da0ee8893c31d24416 *arm-wince.pe/upx-3.91.exe
 294b251518756d20a82580aca1002b03ca71dcb299943f87f225ec157ab252e1 *arm64-linux.elf/upx-3.95

src/p_w64pe_amd64.cpp

@@ -89,11 +89,13 @@ void PackW64PeAmd64::buildLoader(const Filter *ft) {
         addLoader("PEISDLL0");
     if (isefi)
         addLoader("PEISEFI0");
-    addLoader(isdll ? "PEISDLL1" : "", "PEMAIN01",
+    addLoader(isdll ? "PEISDLL1" : "", "PEMAIN01",  // outer "enter" (push rbp; mov rsp,rbp)
               icondir_count > 1 ? (icondir_count == 2 ? "PEICONS1" : "PEICONS2") : "",
               tmp_tlsindex ? "PETLSHAK" : "", "PEMAIN02",
               // ph.first_offset_found == 1 ? "PEMAIN03" : "",
               M_IS_LZMA(ph.method)    ? "LZMA_HEAD,LZMA_ELF00,LZMA_DEC20,LZMA_TAIL"
+                  // LZMA_ELF00 has inner "enter" (push rbp; mov rsp,rbp)
+                  // LZMA_TAIL  has inner 'leave' (mov rbp,rsp; pop rbp)
               : M_IS_NRV2B(ph.method) ? "NRV_HEAD,NRV2B"
               : M_IS_NRV2D(ph.method) ? "NRV_HEAD,NRV2D"
               : M_IS_NRV2E(ph.method) ? "NRV_HEAD,NRV2E"
@@ -126,7 +128,7 @@ void PackW64PeAmd64::buildLoader(const Filter *ft) {
     if (use_tls_callbacks)
         addLoader("PETLSC");
 
-    addLoader("PEMAIN20");
+    addLoader("PEMAIN20");  // outer "leave" (mov rbp,rsp; pop rbp)
     if (use_clear_dirty_stack)
         addLoader("CLEARSTACK");
     addLoader("PEMAIN21");

src/stub/src/arch/amd64/lzma_d.S

@@ -53,7 +53,7 @@ section LZMA_ELF00
 #define M_LZMA          14
         cmpb $ M_LZMA,methb; jne not_lzma
 #endif
-        pushq %rbp; movq %rsp,%rbp  // we use alloca
+        pushq %rbp; movq %rsp,%rbp  // we use alloca; 'leave' is in LZMA_DEC30
 
 //LzmaDecode(  // from lzmaSDK/C/7zip/Compress/LZMA_C/LzmaDecode.h
 //      %arg1= &CLzmaDecoderState,
@@ -124,19 +124,22 @@ section LZMA_DEC30
 #endif
         movl szSizeT(%rbx),%ecx; addq %rcx,%rsi  //  inSizeProcessed
         movl        (%rbx),%edx; addq %rdx,%rdi  // outSizeProcessed
-        leave  // movl %ebp,%rsp; popq %rbp
+        leave  //  %rsp = %rbp;  %rbp = * %rsp ++;  "enter" is in LZMA_ELF00
-#ifdef WINDOWS_BACK
-        jmp eof_lzma_win
-#else
-        jmp eof_lzma
-#endif // WINDOWS_BACK
 
-#ifndef NO_METHOD_CHECK
+#ifdef WINDOWS_BACK  //{  Must match stub/amd64-win64.pe.h of upx-v4.2.4.
-not_lzma:
+    // amd64-win64.pe.S has unconditional "#define NO_METHOD_CHECK"
+    // because p_w64pe_amd64.cpp assumes only one method, and generates it;
+    // thus there is no need to check the method, nor to recover from a mismatch.
+    // In such a case LZMA_DEC30 ends with the [inner] 'leave' above, and
+    // falls into section PEMAIN20, which immediately has an [outer] 'leave'.
+#else  //}{  ndef WINDOWS_BACK
+        jmp eof_lzma  # lzma de-compression is done
+    #ifndef NO_METHOD_CHECK  //{
+    not_lzma:  # recover from not matching M_LZMA
         push %rdi; pop %rsi  # src = arg1
         # fall into next daisy-chain method
-#endif // NO_METHOD_CHECK
+    #endif  //}
-eof_lzma_win:
+#endif  //}
 
 // vi:ts=8:et
 

src/stub/src/arch/amd64/nrv2b_d.S

@@ -45,7 +45,11 @@ lit_n2b:
         incq %rsi; movb %dl,(%rdi)
         incq %rdi
 top_n2b:
+#ifdef WINDOWS_BACK
+        movb (%rsi),%dl  # prefetch: literal, or bottom 8 bits of offset
+#else
         movzbl (%rsi),%edx  # prefetch: literal, or bottom 8 bits of offset
+#endif
         jnextb1yp lit_n2b
         lea 1(lenq),off  # [len= 0] off= 1
 offmore_n2b:
@@ -53,9 +57,15 @@ offmore_n2b:
         jnextb0np offmore_n2b
 
         subl $ 3,off; jc len_n2b  # use previous offset
+#ifdef WINDOWS_BACK
+        shll $ 8,off; movzbl %dl,%edx
+        orl %edx,off; incq %rsi
+        xorl $~0,off; jz EOF
+#else
         shll $ 8,off; incq %rsi
         orl %edx,off
         xorl $~0,off; jz EOF
+#endif
         movslq off,dispq  # XXX: 2GB
 len_n2b:
         lea 1(lenq),off  # [len= 0] off= 1
@@ -73,11 +83,17 @@ gotlen_n2b:
 bot_n2b:  # In: 0==len
         jmp top_n2b
 
-#ifndef NO_METHOD_CHECK
+#ifndef NO_METHOD_CHECK  //{
 not_n2b:
         push %rdi; pop %rsi  # src = arg1
+#ifdef WINDOWS_BACK
+        # fall into 'EOF'
 #endif
+#endif  //}
+
+#ifndef WINDOWS_BACK
         # fall into daisy chain
+#endif
 /*
 vi:ts=8:et:nowrap
 */

src/stub/src/arch/amd64/nrv2d_d.S

@@ -45,7 +45,11 @@ lit_n2d:
         incq %rsi; movb %dl,(%rdi)
         incq %rdi
 top_n2d:
+#ifdef WINDOWS_BACK
+        movb (%rsi),%dl  // speculate: literal, or bottom 8 bits of offset
+#else
         movzbl (%rsi),%edx  // speculate: literal, or bottom 8 bits of offset
+#endif
         jnextb1yp lit_n2d
         lea 1(lenq),off  // [len= 0] off= 1
         jmp getoff_n2d
@@ -58,9 +62,15 @@ getoff_n2d:
         jnextb0np off_n2d
 
         subl $ 3,off; jc offprev_n2d
+#ifdef WINDOWS_BACK
+        shll $ 8,off; movzbl %dl,%edx
+        orl %edx,off; incq %rsi
+        xorl $~0,off; jz eof
+#else
         shll $ 8,off; incq %rsi
         orl %edx,off
         xorl $~0,off; jz EOF
+#endif
         sarl     off  // Carry= original low bit
         movslq   off,dispq  // XXX: 2GB
         jmp len_n2d-2  // CHEAT: assume tail of getnextb is "adcl len,len"

src/stub/src/arch/amd64/nrv2e_d.S

@@ -45,7 +45,11 @@ lit_n2e:
         incq %rsi; movb %dl,(%rdi)
         incq %rdi
 top_n2e:
+#ifdef WINDOWS_BACK
+        movb (%rsi),%dl  # speculate: literal, or bottom 8 bits of offset
+#else
         movzbl (%rsi),%edx  # speculate: literal, or bottom 8 bits of offset
+#endif
         jnextb1yp lit_n2e
         lea 1(lenq),off  # [len= 0] off= 1
         jmp getoff_n2e
@@ -58,9 +62,15 @@ getoff_n2e:
         jnextb0np off_n2e
 
         subl $ 3,off; jc offprev_n2e
+#ifdef WINDOWS_BACK
+        shll $ 8,off; movzbl %dl,%edx
+        orl %edx,off; incq %rsi
+        xorl $~0,off; jz eof
+#else
         shll $ 8,off; incq %rsi
         orl %edx,off
         xorl $~0,off; jz EOF
+#endif
         sarl off  # Carry= original low bit
         movslq off,dispq  # XXX: 2GB
         jc lenlast_n2e

src/stub/tmp/amd64-win64.pe.bin.dump

@@ -13,46 +13,46 @@ Idx Name          Size      VMA               LMA               File off  Algn
   8 PEMAIN02      01  0  0  08c  2**0  CONTENTS
   9 PEMAIN03      0  0  0  08d  2**0  CONTENTS
  10 NRV_HEAD      060  0  0  08d  2**0  CONTENTS
- 11 NRV2B         081  0  0  0ed  2**0  CONTENTS
+ 11 NRV2B         083  0  0  0ed  2**0  CONTENTS
- 12 NRV2D         08f  0  0  016e  2**0  CONTENTS
+ 12 NRV2D         091  0  0  0170  2**0  CONTENTS
- 13 NRV2E         0a8  0  0  01fd  2**0  CONTENTS
+ 13 NRV2E         0aa  0  0  0201  2**0  CONTENTS
- 14 LZMA_HEAD     014  0  0  02a5  2**0  CONTENTS
+ 14 LZMA_HEAD     014  0  0  02ab  2**0  CONTENTS
- 15 LZMA_ELF00    05a  0  0  02b9  2**0  CONTENTS
+ 15 LZMA_ELF00    05a  0  0  02bf  2**0  CONTENTS
- 16 LZMA_DEC20    0a01  0  0  0313  2**0  CONTENTS
+ 16 LZMA_DEC20    0a01  0  0  0319  2**0  CONTENTS
- 17 LZMA_DEC30    016  0  0  0d14  2**0  CONTENTS
+ 17 LZMA_DEC30    014  0  0  0d1a  2**0  CONTENTS
- 18 LZMA_TAIL     02  0  0  0d2a  2**0  CONTENTS
+ 18 LZMA_TAIL     02  0  0  0d2e  2**0  CONTENTS
- 19 PEMAIN10      01  0  0  0d2c  2**0  CONTENTS
+ 19 PEMAIN10      01  0  0  0d30  2**0  CONTENTS
- 20 PETLSHAK2     04  0  0  0d2d  2**0  CONTENTS
+ 20 PETLSHAK2     04  0  0  0d31  2**0  CONTENTS
- 21 PECTTPOS      07  0  0  0d31  2**0  CONTENTS
+ 21 PECTTPOS      07  0  0  0d35  2**0  CONTENTS
- 22 PECTTNUL      03  0  0  0d38  2**0  CONTENTS
+ 22 PECTTNUL      03  0  0  0d3c  2**0  CONTENTS
- 23 PEFILTER49    052  0  0  0d3b  2**0  CONTENTS
+ 23 PEFILTER49    052  0  0  0d3f  2**0  CONTENTS
- 24 PEIMPORT      034  0  0  0d8d  2**0  CONTENTS
+ 24 PEIMPORT      034  0  0  0d91  2**0  CONTENTS
- 25 PEIBYORD      02  0  0  0dc1  2**0  CONTENTS
+ 25 PEIBYORD      02  0  0  0dc5  2**0  CONTENTS
- 26 PEK32ORD      012  0  0  0dc3  2**0  CONTENTS
+ 26 PEK32ORD      012  0  0  0dc7  2**0  CONTENTS
- 27 PEIMORD1      0a  0  0  0dd5  2**0  CONTENTS
+ 27 PEIMORD1      0a  0  0  0dd9  2**0  CONTENTS
- 28 PEIMPOR2      021  0  0  0ddf  2**0  CONTENTS
+ 28 PEIMPOR2      021  0  0  0de3  2**0  CONTENTS
- 29 PEIERDLL      0b  0  0  0e00  2**0  CONTENTS
+ 29 PEIERDLL      0b  0  0  0e04  2**0  CONTENTS
- 30 PEIEREXE      06  0  0  0e0b  2**0  CONTENTS
+ 30 PEIEREXE      06  0  0  0e0f  2**0  CONTENTS
- 31 PEIMDONE      04  0  0  0e11  2**0  CONTENTS
+ 31 PEIMDONE      04  0  0  0e15  2**0  CONTENTS
- 32 PERELOC1      07  0  0  0e15  2**0  CONTENTS
+ 32 PERELOC1      07  0  0  0e19  2**0  CONTENTS
- 33 PERELOC2      04  0  0  0e1c  2**0  CONTENTS
+ 33 PERELOC2      04  0  0  0e20  2**0  CONTENTS
- 34 PERELOC3      030  0  0  0e20  2**0  CONTENTS
+ 34 PERELOC3      030  0  0  0e24  2**0  CONTENTS
- 35 REL64BIG      0a  0  0  0e50  2**0  CONTENTS
+ 35 REL64BIG      0a  0  0  0e54  2**0  CONTENTS
- 36 RELOC64J      02  0  0  0e5a  2**0  CONTENTS
+ 36 RELOC64J      02  0  0  0e5e  2**0  CONTENTS
- 37 PERLOHI0      0a  0  0  0e5c  2**0  CONTENTS
+ 37 PERLOHI0      0a  0  0  0e60  2**0  CONTENTS
- 38 PERELLO0      0b  0  0  0e66  2**0  CONTENTS
+ 38 PERELLO0      0b  0  0  0e6a  2**0  CONTENTS
- 39 PERELHI0      0e  0  0  0e71  2**0  CONTENTS
+ 39 PERELHI0      0e  0  0  0e75  2**0  CONTENTS
- 40 PEDEPHAK      04b  0  0  0e7f  2**0  CONTENTS
+ 40 PEDEPHAK      04b  0  0  0e83  2**0  CONTENTS
- 41 PETLSC        01b  0  0  0eca  2**0  CONTENTS
+ 41 PETLSC        01b  0  0  0ece  2**0  CONTENTS
- 42 PEMAIN20      04  0  0  0ee5  2**0  CONTENTS
+ 42 PEMAIN20      04  0  0  0ee9  2**0  CONTENTS
- 43 CLEARSTACK    010  0  0  0ee9  2**0  CONTENTS
+ 43 CLEARSTACK    010  0  0  0eed  2**0  CONTENTS
- 44 PEMAIN21      0  0  0  0ef9  2**0  CONTENTS
+ 44 PEMAIN21      0  0  0  0efd  2**0  CONTENTS
- 45 PEISDLL9      0f  0  0  0ef9  2**0  CONTENTS
+ 45 PEISDLL9      0f  0  0  0efd  2**0  CONTENTS
- 46 PEISEFI9      02  0  0  0f08  2**0  CONTENTS
+ 46 PEISEFI9      02  0  0  0f0c  2**0  CONTENTS
- 47 PERETURN      04  0  0  0f0a  2**0  CONTENTS
+ 47 PERETURN      04  0  0  0f0e  2**0  CONTENTS
- 48 PEDOJUMP      05  0  0  0f0e  2**0  CONTENTS
+ 48 PEDOJUMP      05  0  0  0f12  2**0  CONTENTS
- 49 PETLSC2       026  0  0  0f13  2**0  CONTENTS
+ 49 PETLSC2       026  0  0  0f17  2**0  CONTENTS
- 50 UPX1HEAD      020  0  0  0f39  2**0  CONTENTS
+ 50 UPX1HEAD      020  0  0  0f3d  2**0  CONTENTS
 SYMBOL TABLE:
 0000000000000000 l    d  NRV_HEAD 0 NRV_HEAD
 0000000000000000 l       PEIMDONE 0 imports_done
@@ -164,15 +164,15 @@ OFFSET           TYPE              VALUE
 
 RELOCATION RECORDS FOR [NRV2B]:
 OFFSET           TYPE              VALUE
-000000000000007b R_X86_64_PC32     NRV_HEAD+0x000000000000001b
+000000000000007d R_X86_64_PC32     NRV_HEAD+0x000000000000001b
 
 RELOCATION RECORDS FOR [NRV2D]:
 OFFSET           TYPE              VALUE
-0000000000000086 R_X86_64_PC32     NRV_HEAD+0x000000000000001b
+0000000000000088 R_X86_64_PC32     NRV_HEAD+0x000000000000001b
 
 RELOCATION RECORDS FOR [NRV2E]:
 OFFSET           TYPE              VALUE
-000000000000009f R_X86_64_PC32     NRV_HEAD+0x000000000000001b
+00000000000000a1 R_X86_64_PC32     NRV_HEAD+0x000000000000001b
 
 RELOCATION RECORDS FOR [LZMA_HEAD]:
 OFFSET           TYPE              VALUE