Merge pull request #1452 from tomt1664/minisketch_patches

Minisketch upstream patches

Author: delta1

configure.ac

@@ -1004,22 +1004,6 @@ AC_CHECK_DECLS([bswap_16, bswap_32, bswap_64],,,
                  #include <byteswap.h>
                  #endif])
 
-AC_MSG_CHECKING([for __builtin_clzl])
-AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[ ]], [[
- (void) __builtin_clzl(0);
-  ]])],
- [ AC_MSG_RESULT([yes]); have_clzl=yes; AC_DEFINE([HAVE_BUILTIN_CLZL], [1], [Define this symbol if you have __builtin_clzl])],
- [ AC_MSG_RESULT([no]); have_clzl=no;]
-)
-
-AC_MSG_CHECKING([for __builtin_clzll])
-AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[ ]], [[
-  (void) __builtin_clzll(0);
-  ]])],
- [ AC_MSG_RESULT([yes]); have_clzll=yes; AC_DEFINE([HAVE_BUILTIN_CLZLL], [1], [Define this symbol if you have __builtin_clzll])],
- [ AC_MSG_RESULT([no]); have_clzll=no;]
-)
-
 dnl Check for malloc_info (for memory statistics information in getmemoryinfo)
 AC_MSG_CHECKING([for getmemoryinfo])
 AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <malloc.h>]],
@@ -1848,7 +1832,6 @@ AM_CONDITIONAL([LIQUID], [test "$liquid_build" = "yes"])
 
 dnl for minisketch
 AM_CONDITIONAL([ENABLE_CLMUL], [test "$enable_clmul" = "yes"])
-AM_CONDITIONAL([HAVE_CLZ], [test "$have_clzl$have_clzll" = "yesyes"])
 
 AC_DEFINE([CLIENT_VERSION_MAJOR], [_CLIENT_VERSION_MAJOR], [Major version])
 AC_DEFINE([CLIENT_VERSION_MINOR], [_CLIENT_VERSION_MINOR], [Minor version])

src/Makefile.minisketch.include

@@ -12,10 +12,6 @@ LIBMINISKETCH_CPPFLAGS += -DHAVE_CLMUL
 MINISKETCH_LIBS += $(LIBMINISKETCH_CLMUL)
 endif
 
-if HAVE_CLZ
-LIBMINISKETCH_CPPFLAGS += -DHAVE_CLZ
-endif
-
 EXTRA_LIBRARIES += $(MINISKETCH_LIBS)
 
 minisketch_libminisketch_clmul_a_SOURCES = $(MINISKETCH_FIELD_CLMUL_SOURCES_INT) $(MINISKETCH_FIELD_CLMUL_HEADERS_INT)

src/minisketch/.cirrus.yml

@@ -36,17 +36,6 @@ env_matrix_snippet: &ENV_MATRIX_VALGRIND
       TESTRUNS: 1
       BUILD:
 
-env_matrix_snippet: &ENV_MATRIX_SAN
-  - env:
-      ENABLE_FIELDS: 28
-  - env:
-      BUILD: distcheck
-  - env:
-      CXXFLAGS:  "-fsanitize=undefined -fno-omit-frame-pointer"
-      LDFLAGS: "-fsanitize=undefined -fno-omit-frame-pointer"
-      UBSAN_OPTIONS: "print_stacktrace=1:halt_on_error=1"
-      BENCH: no
-
 env_matrix_snippet: &ENV_MATRIX_SAN_VALGRIND
   - env:
       ENABLE_FIELDS: "11,64,37"
@@ -72,9 +61,9 @@ task:
     << : *ENV_MATRIX_SAN_VALGRIND
   matrix:
     - env:
-        CC: gcc
+        CXX: g++
     - env:
-        CC: clang
+        CXX: clang++ -gdwarf-4
   << : *MERGE_BASE
   test_script:
     - ./ci/cirrus.sh
@@ -92,30 +81,45 @@ task:
     << : *ENV_MATRIX_VALGRIND
   matrix:
     - env:
-        CC: i686-linux-gnu-gcc
+        CXX: i686-linux-gnu-g++
     - env:
-        CC: clang --target=i686-pc-linux-gnu -isystem /usr/i686-linux-gnu/include
+        CXX: clang++ --target=i686-linux-gnu -gdwarf-4
+        CXXFLAGS: -g -O2 -isystem /usr/i686-linux-gnu/include -isystem /usr/i686-linux-gnu/include/c++/10/i686-linux-gnu
   test_script:
     - ./ci/cirrus.sh
   << : *CAT_LOGS
 
 task:
-  name: "x86_64: macOS Catalina"
+  name: "arm64: macOS Monterey"
   macos_instance:
-    image: catalina-base
+    image: ghcr.io/cirruslabs/macos-monterey-base:latest
   env:
-    # Cirrus gives us a fixed number of 12 virtual CPUs.
-    MAKEFLAGS: -j13
-  matrix:
-    << : *ENV_MATRIX_SAN
+    # Cirrus gives us a fixed number of 4 virtual CPUs.
+    MAKEFLAGS: -j5
   matrix:
     - env:
-        CC: gcc-9
+        CXX: g++-11
+      # Homebrew's gcc for arm64 has no libubsan.
+      matrix:
+        - env:
+            ENABLE_FIELDS: 28
+        - env:
+            BUILD: distcheck
     - env:
-        CC: clang
+        CXX: clang++
+      matrix:
+        - env:
+            ENABLE_FIELDS: 28
+        - env:
+            BUILD: distcheck
+        - env:
+            CXXFLAGS:  "-fsanitize=undefined -fno-omit-frame-pointer"
+            LDFLAGS: "-fsanitize=undefined -fno-omit-frame-pointer"
+            UBSAN_OPTIONS: "print_stacktrace=1:halt_on_error=1"
+            BENCH: no
   brew_script:
     - brew update
-    - brew install automake libtool gcc@9
+    - brew install automake libtool gcc@11
   << : *MERGE_BASE
   test_script:
     - ./ci/cirrus.sh
@@ -128,13 +132,11 @@ task:
     cpu: 4
     memory: 2G
   env:
-    EXEC_CMD: qemu-s390x -L /usr/s390x-linux-gnu
+    EXEC_CMD: qemu-s390x
     HOST: s390x-linux-gnu
     BUILD:
   << : *MERGE_BASE
   test_script:
-    # https://sourceware.org/bugzilla/show_bug.cgi?id=27008
-    - rm /etc/ld.so.cache
     - ./ci/cirrus.sh
   << : *CAT_LOGS
 
@@ -146,6 +148,7 @@ task:
     memory: 2G
   env:
     EXEC_CMD: wine
+    EXEC_EXT: .exe
     HOST: x86_64-w64-mingw32
     BUILD:
   << : *MERGE_BASE

src/minisketch/README.md

@@ -203,8 +203,8 @@ Some improvements that are still TODO:
 * <a name="myfootnote4">[4]</a> Bhaskar Biswas, Vincent Herbert. *Efficient Root Finding of Polynomials over Fields of Characteristic 2.* 2009. hal-00626997. [[URL]](https://hal.archives-ouvertes.fr/hal-00626997) [[PDF]](https://hal.archives-ouvertes.fr/hal-00626997/document)
 * <a name="myfootnote6">[6]</a> Eppstein, David, Michael T. Goodrich, Frank Uyeda, and George Varghese. *What's the difference?: efficient set reconciliation without prior context.* ACM SIGCOMM Computer Communication Review, vol. 41, no. 4, pp. 218-229. ACM, 2011. [[PDF]](https://www.ics.uci.edu/~eppstein/pubs/EppGooUye-SIGCOMM-11.pdf)
 * <a name="myfootnote7">[7]</a> Goodrich, Michael T. and Michael Mitzenmacher. *Invertible bloom lookup tables.* 2011 49th Annual Allerton Conference on Communication, Control, and Computing (Allerton) (2011): 792-799. [[PDF]](https://arxiv.org/pdf/1101.2245.pdf)
-* <a name="myfootnote8">[8]</a> Maxwell, Gregory F. *[Blocksonly mode BW savings, the limits of efficient block xfer, and better relay](https://bitcointalk.org/index.php?topic=1377345.0)* Bitcointalk 2016, *[Technical notes on mempool synchronizing relay](https://people.xiph.org/~greg/mempool_sync_relay.txt)* #bitcoin-wizards 2016.
-* <a name="myfootnote9">[9]</a> Maxwell, Gregory F. *[Block network coding](https://en.bitcoin.it/wiki/User:Gmaxwell/block_network_coding)* Bitcoin Wiki 2014, *[Technical notes on efficient block xfer](https://people.xiph.org/~greg/efficient.block.xfer.txt)* #bitcoin-wizards 2015.
+* <a name="myfootnote8">[8]</a> Maxwell, Gregory F. *[Blocksonly mode BW savings, the limits of efficient block xfer, and better relay](https://bitcointalk.org/index.php?topic=1377345.0)* Bitcointalk 2016, *[Technical notes on mempool synchronizing relay](https://nt4tn.net/tech-notes/2016.mempool_sync_relay.txt)* #bitcoin-wizards 2016.
+* <a name="myfootnote9">[9]</a> Maxwell, Gregory F. *[Block network coding](https://en.bitcoin.it/wiki/User:Gmaxwell/block_network_coding)* Bitcoin Wiki 2014, *[Technical notes on efficient block xfer](https://nt4tn.net/tech-notes/201512.efficient.block.xfer.txt)* #bitcoin-wizards 2015.
 * <a name="myfootnote10">[10]</a> Ruffing, Tim, Moreno-Sanchez, Pedro, Aniket, Kate, *P2P Mixing and Unlinkable Bitcoin Transactions* NDSS Symposium 2017 [[URL]](https://eprint.iacr.org/2016/824) [[PDF]](https://eprint.iacr.org/2016/824.pdf)
 * <a name="myfootnote11">[11]</a> Y. Misky, A. Trachtenberg, R. Zippel. *Set Reconciliation with Nearly Optimal Communication Complexity.* Cornell University, 2000. [[URL]](https://ecommons.cornell.edu/handle/1813/5803) [[PDF]](https://ecommons.cornell.edu/bitstream/handle/1813/5803/2000-1813.pdf)
 * <a name="myfootnote12">[12]</a> Itoh, Toshiya, and Shigeo Tsujii. "A fast algorithm for computing multiplicative inverses in GF (2m) using normal bases." Information and computation 78, no. 3 (1988): 171-177. [[URL]](https://www.sciencedirect.com/science/article/pii/0890540188900247)

src/minisketch/ci/cirrus.sh

@@ -7,7 +7,7 @@ export LC_ALL=C
 
 env >> test_env.log
 
-$CC -v || true
+$CXX -v || true
 valgrind --version || true
 
 ./autogen.sh
@@ -32,10 +32,10 @@ then
 fi
 
 if [ -n "$EXEC_CMD" ]; then
-    $EXEC_CMD ./test $TESTRUNS
-    $EXEC_CMD ./test-verify $TESTRUNS
+    $EXEC_CMD "./test$EXEC_EXT" $TESTRUNS
+    $EXEC_CMD "./test-verify$EXEC_EXT" $TESTRUNS
 fi
 
 if [ "$BENCH" = "yes" ]; then
-    $EXEC_CMD ./bench
+    $EXEC_CMD "./bench$EXEC_EXT"
 fi

src/minisketch/ci/linux-debian.Dockerfile

@@ -8,10 +8,10 @@ RUN apt-get update
 RUN apt-get install --no-install-recommends --no-upgrade -y \
         git ca-certificates \
         make automake libtool pkg-config dpkg-dev valgrind qemu-user \
-        gcc g++ clang libc6-dbg \
+        gcc g++ clang libclang-rt-dev libc6-dbg \
         gcc-i686-linux-gnu g++-i686-linux-gnu libc6-dev-i386-cross libc6-dbg:i386 \
-        g++-s390x-linux-gnu gcc-s390x-linux-gnu libc6-dev-s390x-cross libc6-dbg:s390x \
-        wine g++-mingw-w64-x86-64
+        g++-s390x-linux-gnu libstdc++6:s390x gcc-s390x-linux-gnu libc6-dev-s390x-cross libc6-dbg:s390x \
+        wine wine64 g++-mingw-w64-x86-64
 
 # Run a dummy command in wine to make it set up configuration
 RUN wine true || true

src/minisketch/configure.ac

@@ -104,11 +104,6 @@ esac
 AX_CHECK_COMPILE_FLAG([-Wall],[WARN_CXXFLAGS="$WARN_CXXFLAGS -Wall"],,[[$CXXFLAG_WERROR]])
 AX_CHECK_COMPILE_FLAG([-fvisibility=hidden],[CXXFLAGS="$CXXFLAGS -fvisibility=hidden"],[],[$CXXFLAG_WERROR])
 
-## Some compilers (gcc) ignore unknown -Wno-* options, but warn about all
-## unknown options if any other warning is produced. Test the -Wfoo case, and
-## set the -Wno-foo case if it works.
-AX_CHECK_COMPILE_FLAG([-Wshift-count-overflow],[NOWARN_CXXFLAGS="$NOWARN_CXXFLAGS -Wno-shift-count-overflow"],,[[$CXXFLAG_WERROR]])
-
 if test "x$use_ccache" != "xno"; then
   AC_MSG_CHECKING(if ccache should be used)
   if test x$CCACHE = x; then
@@ -119,14 +114,10 @@ if test "x$use_ccache" != "xno"; then
     fi
   else
     use_ccache=yes
-    CC="$ac_cv_path_CCACHE $CC"
     CXX="$ac_cv_path_CCACHE $CXX"
   fi
   AC_MSG_RESULT($use_ccache)
 fi
-if test "x$use_ccache" = "xyes"; then
-    AX_CHECK_COMPILE_FLAG([-Qunused-arguments],[NOWARN_CXXFLAGS="$NOWARN_CXXFLAGS -Qunused-arguments"],,[[$CXXFLAG_WERROR]])
-fi
 
 VERIFY_DEFINES=-DMINISKETCH_VERIFY
 RELEASE_DEFINES=

src/minisketch/include/minisketch.h

@@ -5,7 +5,8 @@
 #include <stdlib.h>
 
 #ifdef _MSC_VER
-#  include <compat.h>
+#  include <BaseTsd.h>
+   typedef SSIZE_T ssize_t;
 #else
 #  include <unistd.h>
 #endif

src/minisketch/src/bench.cpp

@@ -62,13 +62,11 @@ int main(int argc, char** argv) {
             if (!states[0]) {
                 printf("         -\t");
             } else {
-                double total = 0.0;
                 for (auto& state : states) {
                     auto start = std::chrono::steady_clock::now();
                     minisketch_decode(state, 2 * syndromes, roots.data());
                     auto stop = std::chrono::steady_clock::now();
                     std::chrono::duration<double> dur(stop - start);
-                    total += dur.count();
                     benches.push_back(dur.count());
                 }
                 std::sort(benches.begin(), benches.end());
@@ -98,15 +96,13 @@ int main(int argc, char** argv) {
             if (!states[0]) {
                 printf("         -\t");
             } else {
-                double total = 0.0;
                 for (auto& state : states) {
                     auto start = std::chrono::steady_clock::now();
                     for (auto val : data) {
                         minisketch_add_uint64(state, val);
                     }
                     auto stop = std::chrono::steady_clock::now();
                     std::chrono::duration<double> dur(stop - start);
-                    total += dur.count();
                     benches.push_back(dur.count());
                 }
                 std::sort(benches.begin(), benches.end());

src/minisketch/src/int_utils.h

@@ -7,13 +7,16 @@
 #ifndef _MINISKETCH_INT_UTILS_H_
 #define _MINISKETCH_INT_UTILS_H_
 
+#include <stdint.h>
 #include <stdlib.h>
 
 #include <limits>
 #include <algorithm>
 #include <type_traits>
 
-#ifdef _MSC_VER
+#if defined(__cpp_lib_int_pow2) && __cpp_lib_int_pow2 >= 202002L
+#  include <bit>
+#elif defined(_MSC_VER)
 #  include <intrin.h>
 #endif
 
@@ -54,11 +57,10 @@ class BitWriter {
     int offset = 0;
     unsigned char* out;
 
-public:
-    BitWriter(unsigned char* output) : out(output) {}
-
     template<int BITS, typename I>
-    inline void Write(I val) {
+    inline void WriteInner(I val) {
+        // We right shift by up to 8 bits below. Verify that's well defined for the type I.
+        static_assert(std::numeric_limits<I>::digits > 8, "BitWriter::WriteInner needs I > 8 bits");
         int bits = BITS;
         if (bits + offset >= 8) {
             state |= ((val & ((I(1) << (8 - offset)) - 1)) << offset);
@@ -77,6 +79,19 @@ class BitWriter {
         offset += bits;
     }
 
+
+public:
+    BitWriter(unsigned char* output) : out(output) {}
+
+    template<int BITS, typename I>
+    inline void Write(I val) {
+        // If I is smaller than an unsigned int, invoke WriteInner with argument converted to unsigned.
+        using compute_type = typename std::conditional<
+            (std::numeric_limits<I>::digits < std::numeric_limits<unsigned>::digits),
+            unsigned, I>::type;
+        return WriteInner<BITS, compute_type>(val);
+    }
+
     inline void Flush() {
         if (offset) {
             *(out++) = state;
@@ -129,17 +144,11 @@ constexpr inline I Mask() { return ((I((I(-1)) << (std::numeric_limits<I>::digit
 /** Compute the smallest power of two that is larger than val. */
 template<typename I>
 static inline int CountBits(I val, int max) {
-#ifdef HAVE_CLZ
+#if defined(__cpp_lib_int_pow2) && __cpp_lib_int_pow2 >= 202002L
+    // c++20 impl
     (void)max;
-    if (val == 0) return 0;
-    if (std::numeric_limits<unsigned>::digits >= std::numeric_limits<I>::digits) {
-        return std::numeric_limits<unsigned>::digits - __builtin_clz(val);
-    } else if (std::numeric_limits<unsigned long>::digits >= std::numeric_limits<I>::digits) {
-        return std::numeric_limits<unsigned long>::digits - __builtin_clzl(val);
-    } else {
-        return std::numeric_limits<unsigned long long>::digits - __builtin_clzll(val);
-    }
-#elif _MSC_VER
+    return std::bit_width(val);
+#elif defined(_MSC_VER)
     (void)max;
     unsigned long index;
     unsigned char ret;
@@ -149,7 +158,17 @@ static inline int CountBits(I val, int max) {
         ret = _BitScanReverse64(&index, val);
     }
     if (!ret) return 0;
-    return index;
+    return index + 1;
+#elif HAVE_CLZ
+    (void)max;
+    if (val == 0) return 0;
+    if (std::numeric_limits<unsigned>::digits >= std::numeric_limits<I>::digits) {
+        return std::numeric_limits<unsigned>::digits - __builtin_clz(val);
+    } else if (std::numeric_limits<unsigned long>::digits >= std::numeric_limits<I>::digits) {
+        return std::numeric_limits<unsigned long>::digits - __builtin_clzl(val);
+    } else {
+        return std::numeric_limits<unsigned long long>::digits - __builtin_clzll(val);
+    }
 #else
     while (max && (val >> (max - 1) == 0)) --max;
     return max;
@@ -175,6 +194,7 @@ class BitsInt {
     }
 
     static constexpr inline bool IsZero(I a) { return a == 0; }
+    static constexpr inline bool IsOne(I a) { return a == 1; }
     static constexpr inline I Mask(I val) { return val & MASK; }
     static constexpr inline I Shift(I val, int bits) { return ((val << bits) & MASK); }
     static constexpr inline I UnsafeShift(I val, int bits) { return (val << bits); }
@@ -233,7 +253,7 @@ template<typename I, int N, typename L, typename F> inline constexpr I GFMul(con
 template<typename I, typename F, int BITS, uint32_t MOD>
 inline I InvExtGCD(I x)
 {
-    if (F::IsZero(x)) return x;
+    if (F::IsZero(x) || F::IsOne(x)) return x;
     I t(0), newt(1);
     I r(MOD), newr = x;
     int rlen = BITS + 1, newrlen = F::Bits(newr, BITS);