From 382303a99d35b8628d02f6b0f4ad4f72a5d8293e Mon Sep 17 00:00:00 2001 From: Nigel Tao Date: Sun, 14 May 2023 23:13:43 +1000 Subject: [PATCH] fuzz: add jpeg_fuzzer.c Updates #42 --- fuzz/c/std/jpeg_fuzzer.c | 88 ++++++++++++ fuzz/c/std/pixel_swizzler_fuzzer.c | 208 ++++++++++++++++++++++++++--- fuzz/c/std/seed_corpora.txt | 1 + 3 files changed, 281 insertions(+), 16 deletions(-) create mode 100644 fuzz/c/std/jpeg_fuzzer.c diff --git a/fuzz/c/std/jpeg_fuzzer.c b/fuzz/c/std/jpeg_fuzzer.c new file mode 100644 index 00000000..c8182eb9 --- /dev/null +++ b/fuzz/c/std/jpeg_fuzzer.c @@ -0,0 +1,88 @@ +// Copyright 2023 The Wuffs Authors. +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// https://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +// ---------------- + +// Silence the nested slash-star warning for the next comment's command line. +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wcomment" + +/* +This fuzzer (the fuzz function) is typically run indirectly, by a framework +such as https://github.com/google/oss-fuzz calling LLVMFuzzerTestOneInput. + +When working on the fuzz implementation, or as a coherence check, defining +WUFFS_CONFIG__FUZZLIB_MAIN will let you manually run fuzz over a set of files: + +gcc -DWUFFS_CONFIG__FUZZLIB_MAIN jpeg_fuzzer.c +./a.out ../../../test/data/*.jpeg +rm -f ./a.out + +It should print "PASS", amongst other information, and exit(0). +*/ + +#pragma clang diagnostic pop + +// Wuffs ships as a "single file C library" or "header file library" as per +// https://github.com/nothings/stb/blob/master/docs/stb_howto.txt +// +// To use that single file as a "foo.c"-like implementation, instead of a +// "foo.h"-like header, #define WUFFS_IMPLEMENTATION before #include'ing or +// compiling it. +#define WUFFS_IMPLEMENTATION + +#if defined(WUFFS_CONFIG__FUZZLIB_MAIN) +// Defining the WUFFS_CONFIG__STATIC_FUNCTIONS macro is optional, but when +// combined with WUFFS_IMPLEMENTATION, it demonstrates making all of Wuffs' +// functions have static storage. +// +// This can help the compiler ignore or discard unused code, which can produce +// faster compiles and smaller binaries. Other motivations are discussed in the +// "ALLOW STATIC IMPLEMENTATION" section of +// https://raw.githubusercontent.com/nothings/stb/master/docs/stb_howto.txt +#define WUFFS_CONFIG__STATIC_FUNCTIONS +#endif // defined(WUFFS_CONFIG__FUZZLIB_MAIN) + +// Defining the WUFFS_CONFIG__MODULE* macros are optional, but it lets users of +// release/c/etc.c choose which parts of Wuffs to build. That file contains the +// entire Wuffs standard library, implementing a variety of codecs and file +// formats. Without this macro definition, an optimizing compiler or linker may +// very well discard Wuffs code for unused codecs, but listing the Wuffs +// modules we use makes that process explicit. Preprocessing means that such +// code simply isn't compiled. +#define WUFFS_CONFIG__MODULES +#define WUFFS_CONFIG__MODULE__BASE +#define WUFFS_CONFIG__MODULE__JPEG + +// If building this program in an environment that doesn't easily accommodate +// relative includes, you can use the script/inline-c-relative-includes.go +// program to generate a stand-alone C file. +#include "../../../release/c/wuffs-unsupported-snapshot.c" +#include "../fuzzlib/fuzzlib.c" +#include "../fuzzlib/fuzzlib_image_decoder.c" + +const char* // +fuzz(wuffs_base__io_buffer* src, uint64_t hash) { + wuffs_jpeg__decoder dec; + wuffs_base__status status = wuffs_jpeg__decoder__initialize( + &dec, sizeof dec, WUFFS_VERSION, + (hash & 1) ? WUFFS_INITIALIZE__LEAVE_INTERNAL_BUFFERS_UNINITIALIZED : 0); + hash = wuffs_base__u64__rotate_right(hash, 1); + if (!wuffs_base__status__is_ok(&status)) { + return wuffs_base__status__message(&status); + } + return fuzz_image_decoder( + src, hash, + wuffs_jpeg__decoder__upcast_as__wuffs_base__image_decoder(&dec)); +} diff --git a/fuzz/c/std/pixel_swizzler_fuzzer.c b/fuzz/c/std/pixel_swizzler_fuzzer.c index e826e23f..aafe60a0 100644 --- a/fuzz/c/std/pixel_swizzler_fuzzer.c +++ b/fuzz/c/std/pixel_swizzler_fuzzer.c @@ -94,28 +94,37 @@ const wuffs_base__pixel_blend blends[] = { WUFFS_BASE__PIXEL_BLEND__SRC_OVER, }; -// allocate_guarded_page allocates (2 * page_size) bytes of memory. The first -// page has read|write permissions. The second page has no permissions, so that +size_t // +round_up_to_pagesize(size_t n) { + size_t ps = getpagesize(); + if (ps <= 0) { + return n; + } + return ((n + (ps - 1)) / ps) * ps; +} + +// allocate_guarded_pages allocates (2 * len) bytes of memory. The first half +// has read|write permissions. The second half has no permissions, so that // attempting to read or write to it will cause a segmentation fault. const char* // -allocate_guarded_page(uint8_t** ptr_out, int page_size) { +allocate_guarded_pages(uint8_t** ptr_out, size_t len) { void* ptr = - mmap(NULL, 2 * page_size, PROT_NONE, MAP_ANONYMOUS | MAP_PRIVATE, -1, 0); + mmap(NULL, 2 * len, PROT_NONE, MAP_ANONYMOUS | MAP_PRIVATE, -1, 0); if (ptr == MAP_FAILED) { return "fuzz: internal error: mmap failed"; } - if (mprotect(ptr, page_size, PROT_READ | PROT_WRITE)) { + if (mprotect(ptr, len, PROT_READ | PROT_WRITE)) { return "fuzz: internal error: mprotect failed"; } *ptr_out = ptr; return NULL; } -// fuzz tests that, regardless of the randomized inputs, calling -// wuffs_base__pixel_swizzler__swizzle_interleaved_from_slice will not crash -// the fuzzer (e.g. due to reads or write past buffer bounds). +// fuzz_swizzle_interleaved_from_slice tests that, regardless of the randomized +// inputs, calling wuffs_base__pixel_swizzler__swizzle_interleaved_from_slice +// will not crash the fuzzer (e.g. due to reads or write past buffer bounds). const char* // -fuzz(wuffs_base__io_buffer* src, uint64_t hash) { +fuzz_swizzle_interleaved_from_slice(wuffs_base__io_buffer* src, uint64_t hash) { uint8_t dst_palette_array[1024]; uint8_t src_palette_array[1024]; if ((src->meta.wi - src->meta.ri) < 2048) { @@ -149,14 +158,14 @@ fuzz(wuffs_base__io_buffer* src, uint64_t hash) { return wuffs_base__status__message(&status); } - int page_size = getpagesize(); - if (page_size < 0x100) { - return "fuzz: internal error: page_size is too small"; + static size_t alloc_size = 0; + if (alloc_size == 0) { + alloc_size = round_up_to_pagesize(0x100); } static uint8_t* dst_alloc = NULL; if (!dst_alloc) { - const char* z = allocate_guarded_page(&dst_alloc, page_size); + const char* z = allocate_guarded_pages(&dst_alloc, alloc_size); if (z != NULL) { return z; } @@ -164,7 +173,7 @@ fuzz(wuffs_base__io_buffer* src, uint64_t hash) { static uint8_t* src_alloc = NULL; if (!src_alloc) { - const char* z = allocate_guarded_page(&src_alloc, page_size); + const char* z = allocate_guarded_pages(&src_alloc, alloc_size); if (z != NULL) { return z; } @@ -176,14 +185,20 @@ fuzz(wuffs_base__io_buffer* src, uint64_t hash) { return "fuzz: not enough data"; } wuffs_base__slice_u8 dst_slice = - wuffs_base__make_slice_u8(dst_alloc + page_size - dst_len, dst_len); + wuffs_base__make_slice_u8(dst_alloc + alloc_size - dst_len, dst_len); memcpy(dst_slice.ptr, src->data.ptr + src->meta.ri, dst_len); src->meta.ri += dst_len; wuffs_base__slice_u8 src_slice = - wuffs_base__make_slice_u8(src_alloc + page_size - src_len, src_len); + wuffs_base__make_slice_u8(src_alloc + alloc_size - src_len, src_len); memcpy(src_slice.ptr, src->data.ptr + src->meta.ri, src_len); src->meta.ri += src_len; + // When manually testing this program, enabling this code should lead to a + // segmentation fault. +#if 0 + src_slice.ptr[src_slice.len]++; +#endif + // Calling etc__swizzle_interleaved_from_slice should not crash, whether for // reading/writing out of bounds or for other reasons. wuffs_base__pixel_swizzler__swizzle_interleaved_from_slice( @@ -191,3 +206,164 @@ fuzz(wuffs_base__io_buffer* src, uint64_t hash) { return NULL; } + +const char* // +fuzz_swizzle_ycck(wuffs_base__io_buffer* src, uint64_t hash) { + uint8_t dst_palette_array[1024] = {0}; + wuffs_base__slice_u8 dst_palette = + wuffs_base__make_slice_u8(&dst_palette_array[0], 1024); + + size_t num_pixfmts = sizeof(pixfmts) / sizeof(pixfmts[0]); + wuffs_base__pixel_format dst_pixfmt = wuffs_base__make_pixel_format( + pixfmts[(0xFF & (hash >> 0)) % num_pixfmts]); + + uint32_t width = (63 & (hash >> 8)) + 1; + uint32_t height = (63 & (hash >> 14)) + 1; + + uint32_t width_in_mcus = (3 & (hash >> 20)) + 1; + uint32_t height_in_mcus = (3 & (hash >> 22)) + 1; + + uint32_t allow_hv3 = 1 & (hash >> 23); + uint32_t possible_hv_values[2][4] = { + {1, 1, 2, 4}, + {1, 1, 3, 3}, + }; + uint32_t h0 = possible_hv_values[allow_hv3][3 & (hash >> 25)]; + uint32_t h1 = possible_hv_values[allow_hv3][3 & (hash >> 27)]; + uint32_t h2 = possible_hv_values[allow_hv3][3 & (hash >> 29)]; + uint32_t v0 = possible_hv_values[allow_hv3][3 & (hash >> 31)]; + uint32_t v1 = possible_hv_values[allow_hv3][3 & (hash >> 33)]; + uint32_t v2 = possible_hv_values[allow_hv3][3 & (hash >> 35)]; + // TODO: spend a hash bit for triangle_filter_for_2to1. + + uint32_t width0 = 8 * width_in_mcus * h0; + uint32_t width1 = 8 * width_in_mcus * h1; + uint32_t width2 = 8 * width_in_mcus * h2; + uint32_t height0 = 8 * height_in_mcus * v0; + uint32_t height1 = 8 * height_in_mcus * v1; + uint32_t height2 = 8 * height_in_mcus * v2; + + uint32_t hmax = wuffs_base__u32__max(h0, wuffs_base__u32__max(h1, h2)); + uint32_t vmax = wuffs_base__u32__max(v0, wuffs_base__u32__max(v1, v2)); + width = wuffs_base__u32__min(width, 8 * width_in_mcus * hmax); + height = wuffs_base__u32__min(height, 8 * height_in_mcus * vmax); + + static size_t dst_alloc_size = 0; + if (dst_alloc_size == 0) { + dst_alloc_size = round_up_to_pagesize(8 * 64 * 64); + } + + static uint8_t* dst_alloc = NULL; + if (!dst_alloc) { + const char* z = allocate_guarded_pages(&dst_alloc, dst_alloc_size); + if (z != NULL) { + return z; + } + } + + wuffs_base__pixel_buffer dst_pixbuf; + wuffs_base__pixel_config dst_pixcfg; + wuffs_base__pixel_config__set(&dst_pixcfg, dst_pixfmt.repr, 0, width, height); + uint64_t dst_pixbuf_len = wuffs_base__pixel_config__pixbuf_len(&dst_pixcfg); + if (dst_pixbuf_len > dst_alloc_size) { + return "fuzz: internal error: dst_alloc_size is too small"; + } else { + wuffs_base__status status = wuffs_base__pixel_buffer__set_from_slice( + &dst_pixbuf, &dst_pixcfg, + wuffs_base__make_slice_u8(dst_alloc + dst_alloc_size - dst_pixbuf_len, + dst_pixbuf_len)); + if (status.repr) { + return "fuzz: internal error: wuffs_base__pixel_buffer__set_from_slice " + "failed"; + } + } + + static size_t src_alloc_size = 0; + if (src_alloc_size == 0) { + src_alloc_size = round_up_to_pagesize(8 * 4 * 4 * 8 * 4 * 4); + } + + static uint8_t* src_alloc0 = NULL; + static uint8_t* src_alloc1 = NULL; + static uint8_t* src_alloc2 = NULL; + if (!src_alloc0) { + const char* z = allocate_guarded_pages(&src_alloc0, src_alloc_size); + if (z != NULL) { + return z; + } + } + if (!src_alloc1) { + const char* z = allocate_guarded_pages(&src_alloc1, src_alloc_size); + if (z != NULL) { + return z; + } + } + if (!src_alloc2) { + const char* z = allocate_guarded_pages(&src_alloc2, src_alloc_size); + if (z != NULL) { + return z; + } + } + + uint32_t src_len0 = width0 * height0; + uint32_t src_len1 = width1 * height1; + uint32_t src_len2 = width2 * height2; + if ((src_len0 > src_alloc_size) || // + (src_len1 > src_alloc_size) || // + (src_len2 > src_alloc_size)) { + return "fuzz: internal error: src_alloc_size is too small"; + } + + uint8_t s0 = 0x90; + if (src->meta.ri < src->meta.wi) { + s0 = src->data.ptr[src->meta.ri++]; + } + uint8_t s1 = 0x91; + if (src->meta.ri < src->meta.wi) { + s1 = src->data.ptr[src->meta.ri++]; + } + uint8_t s2 = 0x92; + if (src->meta.ri < src->meta.wi) { + s2 = src->data.ptr[src->meta.ri++]; + } + + wuffs_base__slice_u8 src0 = wuffs_base__make_slice_u8( + src_alloc0 + src_alloc_size - src_len0, src_len0); + memset(src0.ptr, s0, src0.len); + wuffs_base__slice_u8 src1 = wuffs_base__make_slice_u8( + src_alloc1 + src_alloc_size - src_len1, src_len1); + memset(src1.ptr, s1, src1.len); + wuffs_base__slice_u8 src2 = wuffs_base__make_slice_u8( + src_alloc2 + src_alloc_size - src_len2, src_len2); + memset(src2.ptr, s2, src2.len); + wuffs_base__slice_u8 src3 = wuffs_base__empty_slice_u8(); + + wuffs_base__pixel_swizzler swizzler = {0}; + wuffs_base__status status = wuffs_base__pixel_swizzler__swizzle_ycck( + &swizzler, &dst_pixbuf, dst_palette, // + width, height, // + src0, src1, src2, src3, // + width0, width1, width2, 0, // + height0, height1, height2, 0, // + width0, width1, width2, 0, // + h0, h1, h2, 0, // + v0, v1, v2, 0, // + false); + if (status.repr) { + return wuffs_base__status__message(&status); + } + return NULL; +} + +const char* // +fuzz(wuffs_base__io_buffer* src, uint64_t hash) { + const char* s0 = fuzz_swizzle_interleaved_from_slice(src, hash); + const char* s1 = fuzz_swizzle_ycck(src, hash); + if (s0 && strstr(s0, "internal error:")) { + return s0; + } + if (s1 && strstr(s1, "internal error:")) { + return s1; + } + return s0 ? s0 : s1; +} diff --git a/fuzz/c/std/seed_corpora.txt b/fuzz/c/std/seed_corpora.txt index 76f4b2de..f2b3882e 100644 --- a/fuzz/c/std/seed_corpora.txt +++ b/fuzz/c/std/seed_corpora.txt @@ -19,6 +19,7 @@ bzip2: test/data/*.bz2 cbor: test/data/*.cbor gif: test/data/*.gif test/data/artificial-gif/*.gif gzip: test/data/*.gz +jpeg: test/data/*.jpeg ../libjpeg_turbo_corpus/*.jpg json: test/data/*.json ../rapidjson_corpus/* ../simdjson_corpus/* ../JSONTestSuite/test_*/*.json png: test/data/*.png test/data/artificial-png/*.png ../pngsuite_corpus/*.png tga: test/data/*.tga