Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Simple fuzzer for nlohmann/json library #27

Merged
merged 8 commits into from
Oct 13, 2016
Merged

Conversation

vitalybuka
Copy link
Contributor

No description provided.

std::stringstream s;
s << json::parse(data, data + size);
try {
auto j2 = json::parse(s.str());
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

are you trying to parse twice? why?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To check that it can serialize parsed data, and found the bug nlohmann/json#329

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

alright. are you compiling with assert on?

If yes (i.e. if it crashes) then it is ok. Otherwise please hide the output by default.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This technique finds tons of bugs and should be used for all libraries that has round-trip, i.e. serialization/deserialization, compression/decompression, conversion to/from, etc. Bugs found by such checks (logical bugs) are often valued by developers more than crashes on insane data.

I would also suggest to compare results of first and second parse. That catches losing, doubling, altering (e.g. double/missed escape/unescape) data during round-trip.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should serialize j2 back then. S2 should be equal to s (but not data). Let's assert this as well.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done

@mikea mikea merged commit c4a84d8 into google:master Oct 13, 2016
mhlakhani added a commit to mhlakhani/oss-fuzz that referenced this pull request Sep 25, 2019
The corresponding github issue has more detail on the problem.

This code is not in any meaningful state to be committed -- it contains a bunch of hacks I had to make to get the binary to build statically -- libunwind is linked in (via folly), and I had to make changes to include liblzma as well (which libunwind depends on) as well as fix some other unrelated cmake issues.

Reproduction instructions below.

First, build it with `python infra/helper.py build_fuzzers --sanitizer address proxygen`
Then, check the build (it will pass): `python infra/helper.py check_build proxygen`
Then, run the fuzzer: `python infra/helper.py run_fuzzer proxygen ProxygenHTTP1xFuzzer`

It will fail in about ~30 seconds with the below trace. If the fuzzer goes down another path, I've had it reproduce easily by pulling up the shell, running `base64 -d`, pushing that input to a file, and then running the fuzzer directly on that file.

```
terminating with uncaught exception of type folly::ConversionError: Non-digit character found: "OPY"
AddressSanitizer:DEADLYSIGNAL
=================================================================
==11==ERROR: AddressSanitizer: ABRT on unknown address 0x00000000000b (pc 0x7fc166eb7428 bp 0x000000000001 sp 0x7ffcff0a6c78 T0)
SCARINESS: 10 (signal)
    #0 0x7fc166eb7427 in gsignal (/lib/x86_64-linux-gnu/libc.so.6+0x35427)
    google#1 0x7fc166eb9029 in abort (/lib/x86_64-linux-gnu/libc.so.6+0x37029)
    google#2 0x9ab7fa in abort_message (/out/ProxygenHTTP1xFuzzer+0x9ab7fa)
    google#3 0x9b009d in demangling_terminate_handler() (/out/ProxygenHTTP1xFuzzer+0x9b009d)
    google#4 0x9ab282 in std::__terminate(void (*)()) (/out/ProxygenHTTP1xFuzzer+0x9ab282)
    google#5 0x9adebd in __cxxabiv1::call_terminate(bool, _Unwind_Exception*) (/out/ProxygenHTTP1xFuzzer+0x9adebd)
    google#6 0x9ade60 in __cxxabiv1::scan_eh_tab(__cxxabiv1::(anonymous namespace)::scan_results&, _Unwind_Action, bool, _Unwind_Exception*, _Unwind_Context*) (/out/ProxygenHTTP1xFuzzer+0x9ade60)
    google#7 0x9ad5c6 in __gxx_personality_v0 (/out/ProxygenHTTP1xFuzzer+0x9ad5c6)
    google#8 0x7fc16725c262 in _Unwind_RaiseException (/lib/x86_64-linux-gnu/libgcc_s.so.1+0x10262)
    google#9 0x9acfd6 in __cxa_throw (/out/ProxygenHTTP1xFuzzer+0x9acfd6)
    google#10 0x5cedab in void folly::throw_exception<folly::ConversionError>(folly::ConversionError&&) /src/proxygen/proxygen/_build/deps/include/folly/lang/Exception.h:36:3
    google#11 0x5eea03 in _ZZN5folly2toItEENSt3__19enable_ifIXntsr3std7is_sameINS_5RangeIPKcEET_EE5valueES7_E4typeES6_ENKUlNS_14ConversionCodeEE_clESA_ /src/proxygen/proxygen/_build/deps/include/folly/Conv.h:1581:26
    google#12 0x5eea03 in _ZN5folly15expected_detail30expected_detail_ExpectedHelper14ExpectedHelper12thenOrThrow_IRNS0_15ExpectedStorageINS_5RangeIPKcEENS_14ConversionCodeELNS0_11StorageTypeE1EEENS_6detail18CheckTrailingSpaceEZNS_2toItEENSt3__19enable_ifIXntsr3std7is_sameIS8_T_EE5valueESI_E4typeES8_EUlS9_E_NS_8ExpectedINS_4UnitES9_EEvLb0ELi0EEET2_OSI_OT0_OT1_ /src/proxygen/proxygen/_build/deps/include/folly/Expected.h:610:5
    google#13 0x5ed846 in _ZNR5folly8ExpectedINS_5RangeIPKcEENS_14ConversionCodeEE11thenOrThrowINS_6detail18CheckTrailingSpaceEZNS_2toItEENSt3__19enable_ifIXntsr3std7is_sameIS4_T_EE5valueESD_E4typeES4_EUlS5_E_EEDTclclsr3stdE7declvalISD_EEclL_ZNSB_7declvalIRS4_EEDTclsr3std3__1E9__declvalISD_ELi0EEEvEEEEOSD_OT0_ /src/proxygen/proxygen/_build/deps/include/folly/Expected.h:1226:16
    google#14 0x5ed846 in _ZN5folly2toItEENSt3__19enable_ifIXntsr3std7is_sameINS_5RangeIPKcEET_EE5valueES7_E4typeES6_ /src/proxygen/proxygen/_build/deps/include/folly/Conv.h:1579:8
    google#15 0x5ed445 in proxygen::ParseURL::parseAuthority() /src/proxygen/proxygen/lib/utils/ParseURL.cpp:155:15
    google#16 0x5ec5e9 in proxygen::ParseURL::parseNonFully() /src/proxygen/proxygen/lib/utils/ParseURL.cpp:140:8
    google#17 0x5ea87a in proxygen::ParseURL::parse() /src/proxygen/proxygen/lib/utils/ParseURL.cpp:96:5
    google#18 0x59a99c in proxygen::ParseURL::init(folly::Range<char const*>) /src/proxygen/proxygen/lib/utils/ParseURL.h:34:5
    google#19 0x58add1 in proxygen::ParseURL::ParseURL(folly::Range<char const*>) /src/proxygen/proxygen/lib/utils/ParseURL.h:28:5
    google#20 0x58add1 in proxygen::ParseURL proxygen::HTTPMessage::setURL<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >&&) /src/proxygen/proxygen/lib/http/HTTPMessage.h:202:14
    google#21 0x5795f0 in proxygen::HTTP1xCodec::onHeadersComplete(unsigned long) /src/proxygen/proxygen/lib/http/codec/HTTP1xCodec.cpp:976:31
    google#22 0x58e2a3 in proxygen::HTTP1xCodec::onHeadersCompleteCB(proxygen::http_parser*, char const*, unsigned long) /src/proxygen/proxygen/lib/http/codec/HTTP1xCodec.cpp:1315:19
    google#23 0x5f2222 in proxygen::http_parser_execute(proxygen::http_parser*, proxygen::http_parser_settings const*, char const*, unsigned long) /src/proxygen/proxygen/external/http_parser/http_parser_cpp.cpp:1868:17
    google#24 0x577783 in proxygen::HTTP1xCodec::onIngress(folly::IOBuf const&) /src/proxygen/proxygen/lib/http/codec/HTTP1xCodec.cpp:200:26
    google#25 0x55b728 in unsigned long proxygen::parse<proxygen::HTTP1xCodec>(proxygen::HTTP1xCodec*, unsigned char const*, unsigned int, int, std::__1::function<bool ()>) /src/proxygen/proxygen/lib/http/codec/test/TestUtils.h:57:23
    google#26 0x55a6ef in LLVMFuzzerTestOneInput /src/proxygen/proxygen/fuzzers/ProxygenHTTP1xFuzzer.cpp:23:3
    google#27 0x460771 in fuzzer::Fuzzer::ExecuteCallback(unsigned char const*, unsigned long) /src/llvm/projects/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:554:15
    google#28 0x45fe95 in fuzzer::Fuzzer::RunOne(unsigned char const*, unsigned long, bool, fuzzer::InputInfo*, bool*) /src/llvm/projects/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:470:3
    google#29 0x462247 in fuzzer::Fuzzer::MutateAndTestOne() /src/llvm/projects/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:696:19
    google#30 0x462fe5 in fuzzer::Fuzzer::Loop(std::Fuzzer::vector<fuzzer::SizedFile, fuzzer::fuzzer_allocator<fuzzer::SizedFile> >&) /src/llvm/projects/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:832:5
    google#31 0x450dd8 in fuzzer::FuzzerDriver(int*, char***, int (*)(unsigned char const*, unsigned long)) /src/llvm/projects/compiler-rt/lib/fuzzer/FuzzerDriver.cpp:825:6
    google#32 0x47b442 in main /src/llvm/projects/compiler-rt/lib/fuzzer/FuzzerMain.cpp:19:10
    google#33 0x7fc166ea282f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2082f)
    google#34 0x424468 in _start (/out/ProxygenHTTP1xFuzzer+0x424468)

AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: ABRT (/lib/x86_64-linux-gnu/libc.so.6+0x35427) in gsignal
==11==ABORTING
MS: 3 CrossOver-ChangeBit-EraseBytes-; base unit: 467e3fbf38770044dea566169b84c60cb269c89a
0x43,0x4f,0x50,0x59,0x20,0x43,0x3a,0x4f,0x50,0x59,0x2f,0x4f,0xff,0xa,0x43,0x4f,0x50,0x59,0x20,0x2f,0x4f,0xff,0xa,0x43,0x4f,0x50,0x59,0x20,0x2f,0x6f,0x20,0xff,0x50,0x59,0x0,0x43,0xcf,0x50,0x59,0xff,0x4f,0x50,0x33,
COPY C:OPY/O\xff\x0aCOPY /O\xff\x0aCOPY /o \xffPY\x00C\xcfPY\xffOP3
artifact_prefix='./'; Test unit written to ./crash-2cd12847d55ebf08ec2eaee4e814c52e545e7f92
Base64: Q09QWSBDOk9QWS9P/wpDT1BZIC9P/wpDT1BZIC9vIP9QWQBDz1BZ/09QMw==
```

This seems wrong to me since the code is inside a try/catch block: https://github.com/facebook/proxygen/blob/master/proxygen/lib/utils/ParseURL.cpp#L159-L164

I can confirm that the same input does not crash the fuzzer in the version being built on oss-fuzz trunk (and indeed it would have also reported this as a test-case by now given it's been running over a day)
DavidKorczynski pushed a commit that referenced this pull request Jul 9, 2024
DavidKorczynski pushed a commit that referenced this pull request Jul 9, 2024
Fixes #27

---------

Co-authored-by: Jim Choi <jimchoi@google.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants