cpython3: enable building of more modules

Signed-off-by: Adam Korczynski <adam@adalogics.com>

Author: AdamKorcz

projects/cpython3/Dockerfile

@@ -18,8 +18,10 @@ FROM gcr.io/oss-fuzz-base/base-builder
 RUN apt-get update && \
     apt-get install -y build-essential libncursesw5-dev \
 	libreadline-dev libssl-dev libgdbm-dev \
-	libc6-dev libsqlite3-dev tk-dev libbz2-dev \
-	zlib1g-dev libffi-dev
+	libc6-dev libsqlite3-dev libbz2-dev \
+	zlib1g-dev libffi-dev libmpdec-dev \
+	liblzma-dev libzstd-dev uuid-dev \
+	libgdbm-compat-dev
 
 RUN git clone https://github.com/python/cpython.git cpython3
 WORKDIR cpython3

projects/cpython3/build.sh

@@ -30,7 +30,12 @@ CFLAGS="${CFLAGS} -UNDEBUG"
 # We use some internal CPython API.
 CFLAGS="${CFLAGS} -IInclude/internal/"
 
+# Build all stdlib C extension modules statically into libpython
+# rather than as shared .so files, so they are linked into the fuzzers.
+export MODULE_BUILDTYPE=static
+
 FLAGS=()
+FLAGS+=("--disable-test-modules")
 case $SANITIZER in
   address)
     FLAGS+=("--with-address-sanitizer")
@@ -49,19 +54,41 @@ case $SANITIZER in
 esac
 ./configure "${FLAGS[@]:-}" --prefix $OUT
 
+# When building modules statically, HACL* crypto object files must be linked
+# via static archives (.a) instead of raw .o files to avoid duplicate symbols
+# across _blake2, _sha*, _hmac modules that share HACL* objects.
+# Only patch the MODULE_ dependency lines, not the LIBHACL variable definitions.
+sed -i '/^MODULE__/s/LIB_SHARED/LIB_STATIC/g' Makefile
+
 # We use altinstall to avoid having the Makefile create symlinks
 make -j$(nproc) altinstall
 
+# When modules are statically linked into libpython, the fuzzer binaries
+# need the system libraries that those modules depend on (e.g. -lsqlite3,
+# -lz, -lssl). These are tracked in the Makefile's MODLIBS variable but
+# are not included by python-config --ldflags.
+PYTHON=$(ls $OUT/bin/python3.* | grep -v config | head -1)
+MODLIBS=$($PYTHON -c "import sysconfig; print(sysconfig.get_config_var('MODLIBS'))")
+
 FUZZ_DIR=Modules/_xxtestfuzz
 for fuzz_test in $(cat $FUZZ_DIR/fuzz_tests.txt)
 do
+  # Template fuzzers have their own .c file; original targets use fuzzer.c
+  if [ -f "$FUZZ_DIR/${fuzz_test}.c" ]; then
+    FUZZ_SOURCE="$FUZZ_DIR/${fuzz_test}.c"
+    FUZZ_DEFINES=""
+  else
+    FUZZ_SOURCE="$FUZZ_DIR/fuzzer.c"
+    FUZZ_DEFINES="-D _Py_FUZZ_ONE -D _Py_FUZZ_$fuzz_test"
+  fi
+
   # Build (but don't link) the fuzzing stub with a C compiler
-  $CC $CFLAGS $($OUT/bin/python*-config --cflags) $FUZZ_DIR/fuzzer.c \
-    -D _Py_FUZZ_ONE -D _Py_FUZZ_$fuzz_test -c -Wno-unused-function \
+  $CC $CFLAGS $($OUT/bin/python*-config --cflags) $FUZZ_SOURCE \
+    $FUZZ_DEFINES -c -Wno-unused-function \
     -o $WORK/$fuzz_test.o
   # Link with C++ compiler to appease libfuzzer
   $CXX $CXXFLAGS -rdynamic $WORK/$fuzz_test.o -o $OUT/$fuzz_test \
-    $LIB_FUZZING_ENGINE $($OUT/bin/python*-config --ldflags --embed)
+    $LIB_FUZZING_ENGINE $($OUT/bin/python*-config --ldflags --embed) $MODLIBS
 
   # Zip up and copy any seed corpus
   if [ -d "${FUZZ_DIR}/${fuzz_test}_corpus" ]; then