9220a0fdd0 tests: Add one specialized ProcessMessage(...) fuzzing binary per message type for optimal results when using coverage-guided fuzzing (practicalswift)
fd1dae10b4 tests: Add fuzzing harness for ProcessMessage(...) (practicalswift)
Pull request description:
Add fuzzing harness for `ProcessMessage(...)`. Enables high-level fuzzing of the P2P layer.
All code paths reachable from this fuzzer can be assumed to be reachable for an untrusted peer.
Seeded from thin air (an empty corpus) this fuzzer reaches roughly 20 000 lines of code.
To test this PR:
```
$ make distclean
$ ./autogen.sh
$ CC=clang CXX=clang++ ./configure --enable-fuzz \
--with-sanitizers=address,fuzzer,undefined
$ make
$ src/test/fuzz/process_message
…
```
Worth noting about this fuzzing harness:
* To achieve a reasonable number of executions per seconds the state of the fuzzer is unfortunately not entirely reset between `test_one_input` calls. The set-up (`FuzzingSetup` ctor) and tear-down (`~FuzzingSetup`) work is simply too costly to be run on every iteration. There is a trade-off to handle here between a.) achieving high executions/second and b.) giving the fuzzer a totally blank slate for each call. Please let me know if you have any suggestion on how to improve this situation while maintaining >1000 executions/second.
* To achieve optimal results when using coverage-guided fuzzing I've chosen to create one specialised fuzzing binary per message type (`process_message_addr`, `process_message_block`, `process_message_blocktxn `, etc.) and one general fuzzing binary (`process_message`) which handles all messages types. The latter general fuzzer can be seeded with inputs generated by the former specialised fuzzers.
Happy fuzzing friends!
ACKs for top commit:
MarcoFalke:
ACK 9220a0fdd0🏊
Tree-SHA512: c314ef12b0db17b53cbf3abfb9ecc10ce420fb45b17c1db0b34cabe7c30e453947b3ae462020b0c9f30e2c67a7ef1df68826238687dc2479cd816f0addb530e5
4444edc2e6 ci: Enable all functional tests in valgrind (MarcoFalke)
Pull request description:
The travis timeout for our repo has been bumped to 2h, so we can run all tests in valgrind now
ACKs for top commit:
practicalswift:
ACK 4444edc2e6 -- regarding the three disabled cases (`feature_abortnode`, `feature_block` and `rpc_bind`): not a big deal since MSan will take care of those once #18288 is merged. More is more :)
Tree-SHA512: ea2f798112911b6d1f3d88cfcdf0a7cdb698687248343703d6fe55da144542c961c15d888bffb41672c10aa76765615cb7c7ff93d468bfad3c51f962f24e7abb
f1b4503114 bumpfee test: exit loop at proper time with new fee value being compared (Gregory Sanders)
2e4edc68f9 Add some test logging to wallet_bumpfee.py (Gregory Sanders)
Pull request description:
In the loop we accidentally used `origfee` which is not the value to check, and also allowed the loop to exit too early since the new fee must be strictly greater than `0.0005`.
Also converted/added a bunch of logging from comments.
Resolves https://github.com/bitcoin/bitcoin/issues/17716
ACKs for top commit:
MarcoFalke:
ACK f1b4503114🏈
Tree-SHA512: eb73297fc82b09b9ec08d85ba3f0bec662119d0ff63ccf5d978a7bad6a674b5915f5ed021ec42f72a732c9ee7af43212d1de87361f50a970df7755caec96f6d8
ffff9dcdc3 test: Explain why test logging should be used (MarcoFalke)
Pull request description:
Background is that some tests don't have any `self.log` call at all. Thus there are no "anchor points" and those tests are hard to debug because the logs can't easily be parsed by a human.
ACKs for top commit:
jonatack:
ACK ffff9dcdc3
instagibbs:
ACK ffff9dcdc3
fanquake:
re-ACK ffff9dcdc3
Tree-SHA512: 08d962e85c4892c2a0c58feb5dc697c680a9d68e41a79417da6fcd415e0c5c735c4533a985cf225bb89deb5ca717d9bedf990657958079185804caa512b10f5a
cbf2d75d8f qa: Add getdescriptorinfo functional test (João Barbosa)
Pull request description:
The `getdescriptorinfo` RPC was added in #15368, this PR adds some tests.
Top commit has no ACKs.
Tree-SHA512: 5bf3fb5842b975089821c7ac52202ecb23df255f655862646eb532e38e335ff963f8973bcf5b8bba386183281dc9bfe7279ba1cf25fd518c9a45fb45a9243e4d
d484279a46 test: add logging to wallet_listsinceblock.py (Jon Atack)
Pull request description:
This is the first commit from #17535.
Top commit has no ACKs.
Tree-SHA512: bb4f527a41bca3ffbf69e910311ce7f85dcc7a2be41350b3c653a27f4044f392b7e528f330e9691f497212469f6b16ce263230bb7a919548dd4e3e21cc72142f
d2774c09cf Clear any input_errors for an input after it is signed (Andrew Chow)
dc174881ad Replace GetSigningProvider with GetSolvingProvider (Andrew Chow)
6a9c429084 Move direct calls to MessageSign into new SignMessage functions in CWallet and ScriptPubKeyMan (Andrew Chow)
82a30fade7 Move key and script filling and signing from CWallet::FillPSBT to ScriptPubKeyMan::FillPSBT (Andrew Chow)
3d70dd99f9 Move FillPSBT to be a member of CWallet (Andrew Chow)
a4af324d15 Use CWallet::SignTransaction in CreateTransaction and signrawtransactionwithwallet (Andrew Chow)
f37de92744 Implement CWallet::SignTransaction using ScriptPubKeyMan::SignTransaction (Andrew Chow)
d999dd588c Add SignTransaction function to ScriptPubKeyMan and LegacyScriptPubKeyMan (Andrew Chow)
2c52b59d0a Refactor rawtransaction's SignTransaction into generic SignTransaction function (Andrew Chow)
Pull request description:
Following #17261, the way to sign transactions, PSBTs, and messages was to use `GetSigningProvider()` and get a `SigningProvider` containing the private keys. However this may not be feasible for future `ScriptPubKeyMan`s, such as for hardware wallets. Instead of exporting a `SigningProvider` containing private keys, we need to pass these things into the `ScriptPubKeyMan` (via `CWallet`) so that they can do whatever is needed internally to sign them. This is largely a refactor as the logic of processing transactions, PSBTs, and messages for is moved into `LegacyScriptPubKeyMan` and `CWallet` instead of being handled by the caller (e.g. `signrawtransaction`).
To help with this, I've refactored the 3(!) implementations of a `SignTransaction()` function into one generic one. This function will be called by `signrawtransactionwithkey` and `LegacyScriptPubKeyMan::SignTransaction()`. `CWallet::CreateTransaction()` is changed to call `CWallet::SignTransaction()` which in turn, calls `LegacyScriptPubKeyMan::SignTransaction()`. Other `ScriptPubKeyMan`s may implement `SignTransaction()` differently.
`FillPSBT()` is moved to be a member function of `CWallet` and the `psbtwallet.cpp/h` files removed. It is further split so that `CWallet` handles filling the UTXOs while the `ScriptPubKeyMan` handles adding keys, derivation paths, scripts, and signatures. In the end `LegacyScriptPubKeyMan::FillPSBT` still calls `SignPSBTInput`, but the `SigningProvider` is internal to `LegacyScriptPubKeyMan`. Other `ScriptPubKeyMan`s may do something different.
A new `SignMessage()` function is added to both `CWallet` and `ScriptPubKeyMan`. Instead of having the caller (i.e. `signmessage` or the sign message dialog) get the private key, hash the message, and sign, `ScriptPubKeyMan` will now handle that (`CWallet` passes through to the `ScriptPubKeyMan`s as it does for many functions). This signing code is thus consolidated into `LegacyScriptPubKeyMan::SignMessage()`, though other `ScriptPubKeyMan`s may implement it differently. Additionally, a `SigningError` enum is introduced for the different errors that we expect to see from `SignMessage()`.
Lastly, `GetSigningProvider()` is renamed to `GetPublicSigningProvider()`. It will now only provide pubkeys, key origins, and scripts. `LegacySigningProvider` has it's `GetKey` and `HaveKey` functions changed to only return false. Future implementations should return `HidingSigningProvider`s where private keys are hidden.
Other things like `dumpprivkey` and `dumpwallet` are not changed because they directly need and access the `LegacyScriptPubKeyMan` so are not relevant to future changes.
ACKs for top commit:
instagibbs:
reACK d2774c09cf
Sjors:
re-utACK d2774c09cf
meshcollider:
re-utACK d2774c09cf
Tree-SHA512: 89c83e7e7e9315e283fae145a2264648a9d7f7ace8f3281cb3f44f0b013c988d67ba4fa9726e50c643c0ed921bdd269adaec984840d11acf4a681f3e8a582cc1
46a1ee7732 doc: Add historical release notes for 0.19.1 (Wladimir J. van der Laan)
Pull request description:
Top commit has no ACKs.
Tree-SHA512: 9ff578fbaec5694d63256d65cbe7e9d0ba99ac50803c8192f9b5e90a614dc89778bdb17b03f5f7725397f820e085e7d704deb1d6a58b18cd4e4fc91ea413b6ae
6590395f60 tests: Remove FUZZERS_MISSING_CORPORA (practicalswift)
815c7a6793 tests: Add basic fuzzing harness for CNetAddr/CService/CSubNet related functions (netaddress.h) (practicalswift)
Pull request description:
Add basic fuzzing harness for `CNetAddr`/`CService`/`CSubNet` related functions (`netaddress.h`).
To test this PR:
```
$ make distclean
$ ./autogen.sh
$ CC=clang CXX=clang++ ./configure --enable-fuzz \
--with-sanitizers=address,fuzzer,undefined
$ make
$ src/test/fuzz/netaddress
…
```
Top commit has no ACKs.
Tree-SHA512: 69dc0e391d56d5e9cdb818ac0ac4b69445d0195f714442a06cf662998e38b6e0bbaa635dce78df37ba797feed633e94abba4764b946c1716d392756e7809112d
Make sure that there are no errors set for an input after it is signed.
This is useful for when there are multiple ScriptPubKeyMans. Some may
fail to sign, but one may be able to sign, and after it does, we don't
want there to be any more errors there.
Not all ScriptPubKeyMans will be able to provide private keys,
but pubkeys and scripts should be. So only provide public-only
SigningProviders, i.e. ones that can help with Solving.
Instead of fetching a SigningProvider from ScriptPubKeyMan in order
to fill and sign the keys and scripts for a PSBT, just pass that
PSBT to a new FillPSBT function that does all that for us.
