Support crc32c

.github/workflows/haskell-ci.yml

@@ -1,16 +1,16 @@
 # This GitHub workflow config has been generated by a script via
 #
-#   haskell-ci 'github' 'digest.cabal'
+#   haskell-ci 'github' '--distribution' 'focal' '--submodules' 'digest.cabal'
 #
 # To regenerate the script (for example after adjusting tested-with) run
 #
 #   haskell-ci regenerate
 #
 # For more information, see https://github.com/haskell-CI/haskell-ci
 #
-# version: 0.15.20221107
+# version: 0.16.6
 #
-# REGENDATA ("0.15.20221107",["github","digest.cabal"])
+# REGENDATA ("0.16.6",["github","--distribution","focal","--submodules","digest.cabal"])
 #
 name: Haskell-CI
 on:
@@ -23,7 +23,7 @@ jobs:
     timeout-minutes:
       60
     container:
-      image: buildpack-deps:bionic
+      image: buildpack-deps:focal
     continue-on-error: ${{ matrix.allow-failure }}
     strategy:
       matrix:
@@ -55,10 +55,10 @@ jobs:
           apt-get update
           apt-get install -y --no-install-recommends gnupg ca-certificates dirmngr curl git software-properties-common libtinfo5
           mkdir -p "$HOME/.ghcup/bin"
-          curl -sL https://downloads.haskell.org/ghcup/0.1.18.0/x86_64-linux-ghcup-0.1.18.0 > "$HOME/.ghcup/bin/ghcup"
+          curl -sL https://downloads.haskell.org/ghcup/0.1.19.2/x86_64-linux-ghcup-0.1.19.2 > "$HOME/.ghcup/bin/ghcup"
           chmod a+x "$HOME/.ghcup/bin/ghcup"
           "$HOME/.ghcup/bin/ghcup" install ghc "$HCVER" || (cat "$HOME"/.ghcup/logs/*.* && false)
-          "$HOME/.ghcup/bin/ghcup" install cabal 3.6.2.0 || (cat "$HOME"/.ghcup/logs/*.* && false)
+          "$HOME/.ghcup/bin/ghcup" install cabal 3.10.1.0 || (cat "$HOME"/.ghcup/logs/*.* && false)
         env:
           HCKIND: ${{ matrix.compilerKind }}
           HCNAME: ${{ matrix.compiler }}
@@ -74,7 +74,7 @@ jobs:
           echo "HC=$HC" >> "$GITHUB_ENV"
           echo "HCPKG=$HOME/.ghcup/bin/$HCKIND-pkg-$HCVER" >> "$GITHUB_ENV"
           echo "HADDOCK=$HOME/.ghcup/bin/haddock-$HCVER" >> "$GITHUB_ENV"
-          echo "CABAL=$HOME/.ghcup/bin/cabal-3.6.2.0 -vnormal+nowrap" >> "$GITHUB_ENV"
+          echo "CABAL=$HOME/.ghcup/bin/cabal-3.10.1.0 -vnormal+nowrap" >> "$GITHUB_ENV"
           HCNUMVER=$(${HC} --numeric-version|perl -ne '/^(\d+)\.(\d+)\.(\d+)(\.(\d+))?$/; print(10000 * $1 + 100 * $2 + ($3 == 0 ? $5 != 1 : $3))')
           echo "HCNUMVER=$HCNUMVER" >> "$GITHUB_ENV"
           echo "ARG_TESTS=--enable-tests" >> "$GITHUB_ENV"
@@ -124,16 +124,17 @@ jobs:
       - name: install cabal-plan
         run: |
           mkdir -p $HOME/.cabal/bin
-          curl -sL https://github.com/haskell-hvr/cabal-plan/releases/download/v0.6.2.0/cabal-plan-0.6.2.0-x86_64-linux.xz > cabal-plan.xz
-          echo 'de73600b1836d3f55e32d80385acc055fd97f60eaa0ab68a755302685f5d81bc  cabal-plan.xz' | sha256sum -c -
+          curl -sL https://github.com/haskell-hvr/cabal-plan/releases/download/v0.7.3.0/cabal-plan-0.7.3.0-x86_64-linux.xz > cabal-plan.xz
+          echo 'f62ccb2971567a5f638f2005ad3173dba14693a45154c1508645c52289714cb2  cabal-plan.xz' | sha256sum -c -
           xz -d < cabal-plan.xz > $HOME/.cabal/bin/cabal-plan
           rm -f cabal-plan.xz
           chmod a+x $HOME/.cabal/bin/cabal-plan
           cabal-plan --version
       - name: checkout
-        uses: actions/checkout@v2
+        uses: actions/checkout@v3
         with:
           path: source
+          submodules: "true"
       - name: initial cabal.project for sdist
         run: |
           touch cabal.project
@@ -166,8 +167,8 @@ jobs:
         run: |
           $CABAL v2-build $ARG_COMPILER $ARG_TESTS $ARG_BENCH --dry-run all
           cabal-plan
-      - name: cache
-        uses: actions/cache@v2
+      - name: restore cache
+        uses: actions/cache/restore@v3
         with:
           key: ${{ runner.os }}-${{ matrix.compiler }}-${{ github.sha }}
           path: ~/.cabal/store
@@ -188,8 +189,14 @@ jobs:
           ${CABAL} -vnormal check
       - name: haddock
         run: |
-          $CABAL v2-haddock --haddock-all $ARG_COMPILER --with-haddock $HADDOCK $ARG_TESTS $ARG_BENCH all
+          $CABAL v2-haddock --disable-documentation --haddock-all $ARG_COMPILER --with-haddock $HADDOCK $ARG_TESTS $ARG_BENCH all
       - name: unconstrained build
         run: |
           rm -f cabal.project.local
           $CABAL v2-build $ARG_COMPILER --disable-tests --disable-benchmarks all
+      - name: save cache
+        uses: actions/cache/save@v3
+        if: always()
+        with:
+          key: ${{ runner.os }}-${{ matrix.compiler }}-${{ github.sha }}
+          path: ~/.cabal/store

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "external/crc32c"]
+	path = external/crc32c
+	url = https://github.com/google/crc32c.git

Data/Digest/CRC32C.hs

@@ -0,0 +1,72 @@
+{-# LANGUAGE CPP               #-}
+{-# LANGUAGE FlexibleInstances #-}
+{-# LANGUAGE MagicHash         #-}
+{-# LANGUAGE UnliftedFFITypes  #-}
+
+module Data.Digest.CRC32C
+  ( CRC32C
+  , crc32c
+  , crc32cUpdate
+  ) where
+
+import qualified Data.ByteString                as BS
+import qualified Data.ByteString.Lazy           as BL
+import qualified Data.ByteString.Short          as BSS
+import           Data.ByteString.Unsafe         (unsafeUseAsCStringLen)
+import           Data.Word
+import           Foreign.C.Types
+import           Foreign.Ptr
+import           GHC.Exts                       (ByteArray#)
+import           System.IO.Unsafe               (unsafeDupablePerformIO)
+
+#if !MIN_VERSION_bytestring(0, 11, 1)
+import qualified Data.ByteString.Short.Internal as BSS
+#endif
+
+class CRC32C a where
+  -- | Compute CRC32C checksum
+  crc32c :: a -> Word32
+  crc32c = crc32cUpdate 0
+
+  -- | Given the CRC32C checksum of a string, compute CRC32C of its
+  -- concatenation with another string (t.i., incrementally update
+  -- the CRC32C hash value)
+  crc32cUpdate :: Word32 -> a -> Word32
+
+instance CRC32C BS.ByteString where
+  crc32c bs = unsafeDupablePerformIO $
+    unsafeUseAsCStringLen bs $ \(ptr, len) ->
+      crc32c_value (castPtr ptr) (fromIntegral len)
+
+  crc32cUpdate cks bs = unsafeDupablePerformIO $
+    unsafeUseAsCStringLen bs $ \(ptr, len) ->
+      crc32c_extend cks (castPtr ptr) (fromIntegral len)
+
+instance CRC32C BL.ByteString where
+  crc32cUpdate = BL.foldlChunks crc32cUpdate
+
+instance CRC32C [Word8] where
+  crc32cUpdate n = (crc32cUpdate n) . BL.pack
+
+instance CRC32C BSS.ShortByteString where
+  crc32c sbs@(BSS.SBS ba#) = unsafeDupablePerformIO $
+    -- Must be unsafe ffi
+    crc32c_value' ba# (fromIntegral $ BSS.length sbs)
+
+  crc32cUpdate cks sbs@(BSS.SBS ba#) = unsafeDupablePerformIO $
+    -- Must be unsafe ffi
+    crc32c_extend' cks ba# (fromIntegral $ BSS.length sbs)
+
+-------------------------------------------------------------------------------
+
+foreign import ccall unsafe "crc32c/crc32c.h crc32c_value"
+  crc32c_value :: Ptr Word8 -> CSize -> IO Word32
+
+foreign import ccall unsafe "crc32c/crc32c.h crc32c_extend"
+  crc32c_extend :: Word32 -> Ptr Word8 -> CSize -> IO Word32
+
+foreign import ccall unsafe "crc32c/crc32c.h crc32c_value"
+  crc32c_value' :: ByteArray# -> CSize -> IO Word32
+
+foreign import ccall unsafe "crc32c/crc32c.h crc32c_extend"
+  crc32c_extend' :: Word32 -> ByteArray# -> CSize -> IO Word32

digest.cabal

@@ -1,51 +1,132 @@
-name:            digest
-version:         0.0.1.7
-x-revision: 1
-copyright:       (c) 2009 Eugene Kirpichov
-license:         BSD2
-license-file:    LICENSE
-author:          Eugene Kirpichov <ekirpichov@gmail.com>
-maintainer:      Eugene Kirpichov <ekirpichov@gmail.com>
-category:        Cryptography
-synopsis:        CRC32 and Adler32 hashes for bytestrings
-description:     This package provides efficient hash implementations for  
-                 strict and lazy bytestrings. For now, CRC32 and Adler32 are supported; 
-                 they are implemented as FFI bindings to efficient code from zlib.
-stability:       provisional
-build-type:      Simple
-cabal-version:   >= 1.10
-tested-with:
-    GHC==8.10.7
-  , GHC==9.0.2
-  , GHC==9.2.5
-  , GHC==9.4.3
+cabal-version:      2.4
+name:               digest
+version:            0.0.1.7
+x-revision:         1
+copyright:          (c) 2009 Eugene Kirpichov
+license:            BSD-2-Clause
+license-file:       LICENSE
+author:             Eugene Kirpichov <ekirpichov@gmail.com>
+maintainer:         Eugene Kirpichov <ekirpichov@gmail.com>
+category:           Cryptography
+synopsis:           CRC32 and Adler32 hashes for bytestrings
+description:
+  This package provides efficient hash implementations for
+  strict and lazy bytestrings. For now, CRC32 and Adler32 are supported;
+  they are implemented as FFI bindings to efficient code from zlib.
 
+stability:          provisional
+build-type:         Simple
+tested-with:        GHC ==8.10.7 || ==9.0.2 || ==9.2.5 || ==9.4.3
 extra-source-files:
+  CHANGELOG.md
+  external/crc32c/include/crc32c/crc32c.h
+  external/crc32c/LICENSE
+  external/crc32c/src/*.h
+  include/crc32c/crc32c_config.h
   testing/trivial-reference.c
   testing/trivial.expected
   testing/trivial.hs
-  CHANGELOG.md
 
 flag pkg-config
   default:     True
   manual:      True
   description: Use @pkg-config(1)@ to locate @zlib@ library.
 
+-- TODO: auto detect
+flag have_builtin_prefetch
+  default:     False
+  manual:      True
+  description: The cxx compiler has the __builtin_prefetch intrinsic.
+
+-- TODO: auto detect
+flag have_mm_prefetch
+  default:     False
+  manual:      True
+  description:
+    Targeting X86 and the compiler has the _mm_prefetch intrinsic.
+
+-- TODO: auto detect
+flag have_sse42
+  default:     False
+  manual:      True
+  description:
+    Can be enabled to improve performance of CRC32C if targeting X86 and
+    the compiler has the _mm_crc32_u{8,32,64} intrinsics.
+
+-- TODO: auto detect
+flag have_arm64_crc32c
+  default:     False
+  manual:      True
+  description:
+    Targeting ARM and the compiler has the __crc32c{b,h,w,d} and the
+    vmull_p64 intrinsics.
+
+-- TODO: auto detect
+flag have_strong_getauxval
+  default:     False
+  manual:      True
+  description:
+    The system libraries have the getauxval function in the <sys/auxv.h> header.
+    Should be true on Linux and Android API level 20+.
+
+-- TODO: auto detect
+flag have_weak_getauxval
+  default:     False
+  manual:      True
+  description:
+    The compiler supports defining getauxval as a weak symbol.
+    Should be true for any compiler that supports __attribute__((weak)).
+
 source-repository head
-  type: git
+  type:     git
   location: https://github.com/TeofilC/digest
 
 library
-  exposed-modules: Data.Digest.CRC32,
-                   Data.Digest.Adler32
-  default-extensions: CPP, ForeignFunctionInterface
-  default-language: Haskell2010
-  build-depends: 
-    base < 5
-    , bytestring >= 0.9 && < 0.13
-  includes:        zlib.h
-  ghc-options:     -Wall
-  if flag(pkg-config) && !os(windows) && !os(freebsd)
+  exposed-modules:
+    Data.Digest.Adler32
+    Data.Digest.CRC32
+    Data.Digest.CRC32C
+
+  default-extensions:
+    CPP
+    ForeignFunctionInterface
+
+  default-language:   Haskell2010
+  build-depends:
+    , base        >=4.12 && <5
+    , bytestring  >=0.10 && <0.13
+
+  includes:           zlib.h
+  include-dirs:       include external/crc32c/include
+  cxx-options:        -std=c++11
+  cxx-sources:
+    external/crc32c/src/crc32c.cc
+    external/crc32c/src/crc32c_portable.cc
+
+  if flag(have_builtin_prefetch)
+    cxx-options: -DHAVE_BUILTIN_PREFETCH
+
+  if flag(have_mm_prefetch)
+    cxx-options: -DHAVE_MM_PREFETCH
+
+  if (arch(x86_64) && flag(have_sse42))
+    cxx-options: -DHAVE_SSE42 -msse4.2
+    cxx-sources: external/crc32c/src/crc32c_sse42.cc
+
+  if (arch(aarch64) && flag(have_arm64_crc32c))
+    cxx-options: -DHAVE_ARM64_CRC32C
+    cxx-sources: external/crc32c/src/crc32c_arm64.cc
+
+  if flag(have_strong_getauxval)
+    cxx-options: -DHAVE_STRONG_GETAUXVAL
+
+  if flag(have_weak_getauxval)
+    cxx-options: -DHAVE_WEAK_GETAUXVAL
+
+  ghc-options:        -Wall
+
+  if ((flag(pkg-config) && !os(windows)) && !os(freebsd))
     pkgconfig-depends: zlib
+
   else
     build-depends: zlib

include/crc32c/crc32c_config.h

@@ -0,0 +1,49 @@
+// Also see: external/crc32c/src/crc32c_config.h.in
+#ifndef CRC32C_CRC32C_CONFIG_H_
+#define CRC32C_CRC32C_CONFIG_H_
+
+// From GHC
+#include <MachDeps.h>
+
+#ifdef WORDS_BIGENDIAN
+// Define to 1 if building for a big-endian platform.
+#define BYTE_ORDER_BIG_ENDIAN 1
+#else
+#define BYTE_ORDER_BIG_ENDIAN 0
+#endif
+
+// Set by cabal flag 'have_builtin_prefetch'
+//
+// Define to 1 if the compiler has the __builtin_prefetch intrinsic.
+//#define HAVE_BUILTIN_PREFETCH 0
+
+// Set by cabal flag 'have_mm_prefetch'
+//
+// Define to 1 if targeting X86 and the compiler has the _mm_prefetch intrinsic.
+//#define HAVE_MM_PREFETCH 0
+
+// Set by cabal flag 'have_sse42'
+//
+// Define to 1 if targeting X86 and the compiler has the _mm_crc32_u{8,32,64}
+// intrinsics.
+//#define HAVE_SSE42 0
+
+// Set by cabal flag 'have_arm64_crc32c'
+//
+// Define to 1 if targeting ARM and the compiler has the __crc32c{b,h,w,d} and
+// the vmull_p64 intrinsics.
+//#define HAVE_ARM64_CRC32C 0
+
+// Set by cabal flag 'have_strong_getauxval'
+//
+// Define to 1 if the system libraries have the getauxval function in the
+// <sys/auxv.h> header. Should be true on Linux and Android API level 20+.
+//#define HAVE_STRONG_GETAUXVAL 0
+
+// Set by cabal flag 'have_weak_getauxval'
+//
+// Define to 1 if the compiler supports defining getauxval as a weak symbol.
+// Should be true for any compiler that supports __attribute__((weak)).
+//#define HAVE_WEAK_GETAUXVAL 0
+
+#endif  // CRC32C_CRC32C_CONFIG_H_