[fuzz] Add raw dictionary content fuzzer

diff --git a/tests/fuzz/Makefile b/tests/fuzz/Makefile
index 571ab553bd7b616b8f0303af3ac4fd033810958f..3c7dc09e22ff7b94c0534c4d6f04bffba558bd0f 100644 (file)
--- a/tests/fuzz/Makefile
+++ b/tests/fuzz/Makefile
@@ -41,8 +41,10 @@ FUZZ_LDFLAGS := -pthread $(LDFLAGS)
 FUZZ_ARFLAGS := $(ARFLAGS)
 FUZZ_TARGET_FLAGS = $(FUZZ_CPPFLAGS) $(FUZZ_CXXFLAGS) $(FUZZ_LDFLAGS)
 
+FUZZ_ROUND_TRIP_FLAGS := -DFUZZING_ASSERT_VALID_SEQUENCE
+
 FUZZ_HEADERS := fuzz_helpers.h fuzz.h zstd_helpers.h fuzz_data_producer.h
-FUZZ_SRC := $(PRGDIR)/util.c fuzz_helpers.c zstd_helpers.c fuzz_data_producer.c
+FUZZ_SRC := $(PRGDIR)/util.c ./fuzz_helpers.c ./zstd_helpers.c ./fuzz_data_producer.c
 
 ZSTDCOMMON_SRC := $(ZSTDDIR)/common/*.c
 ZSTDCOMP_SRC   := $(ZSTDDIR)/compress/*.c
@@ -56,9 +58,25 @@ FUZZ_SRC       := \
        $(ZSTDCOMP_SRC) \
        $(ZSTDDICT_SRC) \
        $(ZSTDLEGACY_SRC)
-
-FUZZ_OBJ := $(patsubst %.c,%.o, $(wildcard $(FUZZ_SRC)))
-
+FUZZ_SRC := $(wildcard $(FUZZ_SRC))
+
+FUZZ_D_OBJ1 := $(subst $(ZSTDDIR)/common/,d_lib_common_,$(FUZZ_SRC))
+FUZZ_D_OBJ2 := $(subst $(ZSTDDIR)/compress/,d_lib_compress_,$(FUZZ_D_OBJ1))
+FUZZ_D_OBJ3 := $(subst $(ZSTDDIR)/decompress/,d_lib_decompress_,$(FUZZ_D_OBJ2))
+FUZZ_D_OBJ4 := $(subst $(ZSTDDIR)/dictBuilder/,d_lib_dictBuilder_,$(FUZZ_D_OBJ3))
+FUZZ_D_OBJ5 := $(subst $(ZSTDDIR)/legacy/,d_lib_legacy_,$(FUZZ_D_OBJ4))
+FUZZ_D_OBJ6 := $(subst $(PRGDIR)/,d_prg_,$(FUZZ_D_OBJ5))
+FUZZ_D_OBJ7 := $(subst $\./,d_fuzz_,$(FUZZ_D_OBJ6))
+FUZZ_DECOMPRESS_OBJ := $(FUZZ_D_OBJ7:.c=.o)
+
+FUZZ_RT_OBJ1 := $(subst $(ZSTDDIR)/common/,rt_lib_common_,$(FUZZ_SRC))
+FUZZ_RT_OBJ2 := $(subst $(ZSTDDIR)/compress/,rt_lib_compress_,$(FUZZ_RT_OBJ1))
+FUZZ_RT_OBJ3 := $(subst $(ZSTDDIR)/decompress/,rt_lib_decompress_,$(FUZZ_RT_OBJ2))
+FUZZ_RT_OBJ4 := $(subst $(ZSTDDIR)/dictBuilder/,rt_lib_dictBuilder_,$(FUZZ_RT_OBJ3))
+FUZZ_RT_OBJ5 := $(subst $(ZSTDDIR)/legacy/,rt_lib_legacy_,$(FUZZ_RT_OBJ4))
+FUZZ_RT_OBJ6 := $(subst $(PRGDIR)/,rt_prg_,$(FUZZ_RT_OBJ5))
+FUZZ_RT_OBJ7 := $(subst $\./,rt_fuzz_,$(FUZZ_RT_OBJ6))
+FUZZ_ROUND_TRIP_OBJ := $(FUZZ_RT_OBJ7:.c=.o)
 
 .PHONY: default all clean cleanall
 
@@ -75,45 +93,88 @@ FUZZ_TARGETS :=       \
        dictionary_decompress \
        zstd_frame_info \
        simple_compress \
-       dictionary_loader
+       dictionary_loader \
+       raw_dictionary_round_trip
 
 all: $(FUZZ_TARGETS)
 
-%.o: %.c
-       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $^ -c -o $@
+rt_lib_common_%.o: $(ZSTDDIR)/common/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+rt_lib_compress_%.o: $(ZSTDDIR)/compress/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+rt_lib_decompress_%.o: $(ZSTDDIR)/decompress/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+rt_lib_dictBuilder_%.o: $(ZSTDDIR)/dictBuilder/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+rt_lib_legacy_%.o: $(ZSTDDIR)/legacy/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+rt_prg_%.o: $(PRGDIR)/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+rt_fuzz_%.o: %.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $(FUZZ_ROUND_TRIP_FLAGS) $< -c -o $@
+
+d_lib_common_%.o: $(ZSTDDIR)/common/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+d_lib_compress_%.o: $(ZSTDDIR)/compress/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+d_lib_decompress_%.o: $(ZSTDDIR)/decompress/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+d_lib_dictBuilder_%.o: $(ZSTDDIR)/dictBuilder/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+d_lib_legacy_%.o: $(ZSTDDIR)/legacy/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+d_prg_%.o: $(PRGDIR)/%.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+d_fuzz_%.o: %.c
+       $(CC) $(FUZZ_CPPFLAGS) $(FUZZ_CFLAGS) $< -c -o $@
+
+simple_round_trip: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_simple_round_trip.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_simple_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
 
-simple_round_trip: $(FUZZ_HEADERS) $(FUZZ_OBJ) simple_round_trip.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) simple_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
+stream_round_trip: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_stream_round_trip.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_stream_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
 
-stream_round_trip: $(FUZZ_HEADERS) $(FUZZ_OBJ) stream_round_trip.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) stream_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
+block_round_trip: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_block_round_trip.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_block_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
 
-block_round_trip: $(FUZZ_HEADERS) $(FUZZ_OBJ) block_round_trip.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) block_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
+simple_decompress: $(FUZZ_HEADERS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_simple_decompress.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_simple_decompress.o $(LIB_FUZZING_ENGINE) -o $@
 
-simple_decompress: $(FUZZ_HEADERS) $(FUZZ_OBJ) simple_decompress.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) simple_decompress.o $(LIB_FUZZING_ENGINE) -o $@
+stream_decompress: $(FUZZ_HEADERS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_stream_decompress.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_stream_decompress.o $(LIB_FUZZING_ENGINE) -o $@
 
-stream_decompress: $(FUZZ_HEADERS) $(FUZZ_OBJ) stream_decompress.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) stream_decompress.o $(LIB_FUZZING_ENGINE) -o $@
+block_decompress: $(FUZZ_HEADERS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_block_decompress.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_block_decompress.o $(LIB_FUZZING_ENGINE) -o $@
 
-block_decompress: $(FUZZ_HEADERS) $(FUZZ_OBJ) block_decompress.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) block_decompress.o $(LIB_FUZZING_ENGINE) -o $@
+dictionary_round_trip: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_dictionary_round_trip.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_dictionary_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
 
-dictionary_round_trip: $(FUZZ_HEADERS) $(FUZZ_OBJ) dictionary_round_trip.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) dictionary_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
+raw_dictionary_round_trip: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_raw_dictionary_round_trip.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_raw_dictionary_round_trip.o $(LIB_FUZZING_ENGINE) -o $@
 
-dictionary_decompress: $(FUZZ_HEADERS) $(FUZZ_OBJ) dictionary_decompress.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) dictionary_decompress.o $(LIB_FUZZING_ENGINE) -o $@
+dictionary_decompress: $(FUZZ_HEADERS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_dictionary_decompress.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_dictionary_decompress.o $(LIB_FUZZING_ENGINE) -o $@
 
-simple_compress: $(FUZZ_HEADERS) $(FUZZ_OBJ) simple_compress.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) simple_compress.o $(LIB_FUZZING_ENGINE) -o $@
+simple_compress: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_simple_compress.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_simple_compress.o $(LIB_FUZZING_ENGINE) -o $@
 
-zstd_frame_info: $(FUZZ_HEADERS) $(FUZZ_OBJ) zstd_frame_info.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) zstd_frame_info.o $(LIB_FUZZING_ENGINE) -o $@
+zstd_frame_info: $(FUZZ_HEADERS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_zstd_frame_info.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_DECOMPRESS_OBJ) d_fuzz_zstd_frame_info.o $(LIB_FUZZING_ENGINE) -o $@
 
-dictionary_loader: $(FUZZ_HEADERS) $(FUZZ_OBJ) dictionary_loader.o
-       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_OBJ) dictionary_loader.o $(LIB_FUZZING_ENGINE) -o $@
+dictionary_loader: $(FUZZ_HEADERS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_dictionary_loader.o
+       $(CXX) $(FUZZ_TARGET_FLAGS) $(FUZZ_ROUND_TRIP_OBJ) rt_fuzz_dictionary_loader.o $(LIB_FUZZING_ENGINE) -o $@
 
 libregression.a: $(FUZZ_HEADERS) $(PRGDIR)/util.h $(PRGDIR)/util.c regression_driver.o
        $(AR) $(FUZZ_ARFLAGS) $@ regression_driver.o
@@ -136,7 +197,6 @@ regressiontest: corpora
        $(PYTHON) ./fuzz.py regression all
 
 clean:
-       @$(MAKE) -C $(ZSTDDIR) clean
        @$(RM) *.a *.o
        @$(RM) simple_round_trip stream_round_trip simple_decompress \
            stream_decompress block_decompress block_round_trip \

diff --git a/tests/fuzz/fuzz.py b/tests/fuzz/fuzz.py
index 52d41d2be37f2acb78eb7b7144ad08929140f7fd..29d2695573329841d42a0450bccf087db21e82f4 100755 (executable)
--- a/tests/fuzz/fuzz.py
+++ b/tests/fuzz/fuzz.py
@@ -57,6 +57,7 @@ TARGET_INFO = {
     'zstd_frame_info': TargetInfo(InputType.COMPRESSED_DATA),
     'simple_compress': TargetInfo(InputType.RAW_DATA),
     'dictionary_loader': TargetInfo(InputType.DICTIONARY_DATA),
+    'raw_dictionary_round_trip': TargetInfo(InputType.RAW_DATA),
 }
 TARGETS = list(TARGET_INFO.keys())
 ALL_TARGETS = TARGETS + ['all']

diff --git a/tests/fuzz/raw_dictionary_round_trip.c b/tests/fuzz/raw_dictionary_round_trip.c
new file mode 100644 (file)
index 0000000..08e5fd9
--- /dev/null
+++ b/tests/fuzz/raw_dictionary_round_trip.c
@@ -0,0 +1,115 @@
+/*
+ * Copyright (c) 2016-2020, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/**
+ * This fuzz target performs a zstd round-trip test (compress & decompress) with
+ * a raw content dictionary, compares the result with the original, and calls
+ * abort() on corruption.
+ */
+
+#include <stddef.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include "fuzz_helpers.h"
+#include "zstd_helpers.h"
+#include "fuzz_data_producer.h"
+
+static ZSTD_CCtx *cctx = NULL;
+static ZSTD_DCtx *dctx = NULL;
+
+static size_t roundTripTest(void *result, size_t resultCapacity,
+                            void *compressed, size_t compressedCapacity,
+                            const void *src, size_t srcSize,
+                            const void *dict, size_t dictSize,
+                            FUZZ_dataProducer_t *producer)
+{
+    ZSTD_dictContentType_e const dictContentType = ZSTD_dct_rawContent;
+    int const refPrefix = FUZZ_dataProducer_uint32Range(producer, 0, 1) != 0;
+    size_t cSize;
+
+    FUZZ_setRandomParameters(cctx, srcSize, producer);
+    /* Disable checksum so we can use sizes smaller than compress bound. */
+    FUZZ_ZASSERT(ZSTD_CCtx_setParameter(cctx, ZSTD_c_checksumFlag, 0));
+    if (refPrefix)
+        FUZZ_ZASSERT(ZSTD_CCtx_refPrefix_advanced(
+            cctx, dict, dictSize,
+            ZSTD_dct_rawContent));
+    else 
+        FUZZ_ZASSERT(ZSTD_CCtx_loadDictionary_advanced(
+            cctx, dict, dictSize,
+            (ZSTD_dictLoadMethod_e)FUZZ_dataProducer_uint32Range(producer, 0, 1),
+            ZSTD_dct_rawContent));
+    cSize = ZSTD_compress2(cctx, compressed, compressedCapacity, src, srcSize);
+    FUZZ_ZASSERT(cSize);
+
+    if (refPrefix)
+        FUZZ_ZASSERT(ZSTD_DCtx_refPrefix_advanced(
+            dctx, dict, dictSize,
+            dictContentType));
+    else
+        FUZZ_ZASSERT(ZSTD_DCtx_loadDictionary_advanced(
+            dctx, dict, dictSize,
+            (ZSTD_dictLoadMethod_e)FUZZ_dataProducer_uint32Range(producer, 0, 1),
+            dictContentType));
+    {
+        size_t const ret = ZSTD_decompressDCtx(
+                dctx, result, resultCapacity, compressed, cSize);
+        return ret;
+    }
+}
+
+int LLVMFuzzerTestOneInput(const uint8_t *src, size_t size)
+{
+    /* Give a random portion of src data to the producer, to use for
+    parameter generation. The rest will be used for (de)compression */
+    FUZZ_dataProducer_t *producer = FUZZ_dataProducer_create(src, size);
+    size = FUZZ_dataProducer_reserveDataPrefix(producer);
+
+    uint8_t const* const srcBuf = src;
+    size_t const srcSize = FUZZ_dataProducer_uint32Range(producer, 0, size);
+    uint8_t const* const dictBuf = srcBuf + srcSize;
+    size_t const dictSize = size - srcSize;
+    size_t const decompSize = srcSize;
+    void* const decompBuf = FUZZ_malloc(decompSize);
+    size_t compSize = ZSTD_compressBound(srcSize);
+    void* compBuf;
+    /* Half of the time fuzz with a 1 byte smaller output size.
+     * This will still succeed because we force the checksum to be disabled,
+     * giving us 4 bytes of overhead.
+     */
+    compSize -= FUZZ_dataProducer_uint32Range(producer, 0, 1);
+    compBuf = FUZZ_malloc(compSize);
+
+    if (!cctx) {
+        cctx = ZSTD_createCCtx();
+        FUZZ_ASSERT(cctx);
+    }
+    if (!dctx) {
+        dctx = ZSTD_createDCtx();
+        FUZZ_ASSERT(dctx);
+    }
+
+    {
+        size_t const result =
+            roundTripTest(decompBuf, decompSize, compBuf, compSize, srcBuf, srcSize, dictBuf, dictSize, producer);
+        FUZZ_ZASSERT(result);
+        FUZZ_ASSERT_MSG(result == srcSize, "Incorrect regenerated size");
+        FUZZ_ASSERT_MSG(!FUZZ_memcmp(src, decompBuf, srcSize), "Corruption!");
+    }
+    free(decompBuf);
+    free(compBuf);
+    FUZZ_dataProducer_free(producer);
+#ifndef STATEFUL_FUZZING
+    ZSTD_freeCCtx(cctx); cctx = NULL;
+    ZSTD_freeDCtx(dctx); dctx = NULL;
+#endif
+    return 0;
+}