b7aa717cdd refactor: gui, simplify boost signals disconnection (furszy)
f3a612f901 gui: guard accessing a nullptr 'clientModel' (furszy)
Pull request description:
Fixing #800.
During shutdown, already queue events dispatched from the backend such
'numConnectionsChanged' and 0networkActiveChanged' could try to access
the clientModel object, which might not exist because we manually delete
it inside 'BitcoinApplication::requestShutdown()'.
This happen because boost does not clears the queued events when they arise
concurrently with the signal disconnection (see https://www.boost.org/doc/libs/1_55_0/doc/html/signals2/thread-safety.html).
From the docs:
1) "Note that since we unlock the connection's mutex before executing its associated slot, it is possible a slot will still be executing after it has been disconnected by a [connection::disconnect](https://www.boost.org/doc/libs/1_55_0/doc/html/boost/signals2/connection.html#idp89761576-bb)(), if the disconnect was called concurrently with signal invocation."
2) "The fact that concurrent signal invocations use the same combiner object means you need to insure any custom combiner you write is thread-safe"
So, we need to guard `clientModel` before accessing it at the handler side.
ACKs for top commit:
hebasto:
re-ACK b7aa717cdd
Tree-SHA512: f1a21d69248628f6a13556a9438c9e4ea9f0a3678aab09ddfe836e78e4eee405a6730d37d39f1445068ada3a110b655b619cf0e090fc2d0cdf99bed061364aeb
86b7f28d6c serialization: use internal endian conversion functions (Cory Fields)
432b18ca8d serialization: detect byteswap builtins without autoconf tests (Cory Fields)
297367b3bb crypto: replace CountBits with std::bit_width (Cory Fields)
52f9bba889 crypto: replace non-standard CLZ builtins with c++20's bit_width (Cory Fields)
Pull request description:
This replaces #28674, #29036, and #29057. Now ready for testing and review.
Replaces platform-specific endian and byteswap functions. This is especially useful for kernel, as it means that our deep serialization code no longer requires bitcoin-config.h.
I apologize for the size of the last commit, but it's hard to avoid making those changes at once.
All platforms now use our internal functions rather than libc or platform-specific ones, with the exception of MSVC.
Sadly, benchmarking showed that not all compilers are capable of detecting and optimizing byteswap functions, so compiler builtins are instead used where possible. However, they're now detected via macros rather than autoconf checks.
This[ matches how libc++ implements std::byteswap for c++23](https://github.com/llvm/llvm-project/blob/main/libcxx/include/__bit/byteswap.h#L26).
I suggest we move/rename `compat/endian.h`, but I left that out of this PR to avoid bikeshedding.
#29057 pointed out some irregularities in benchmarks. After messing with various compilers and configs for a few weeks with these changes, I'm of the opinion that we can't win on every platform every time, so we should take the code that makes sense going forward. That said, if any real-world slowdowns are caused here, we should obviously investigate.
ACKs for top commit:
maflcko:
ACK 86b7f28d6c📘
fanquake:
ACK 86b7f28d6c - we can finish pruning out the __builtin_clz* checks/usage once the minisketch code has been updated. This is more good cleanup pre-CMake & for the kernal.
Tree-SHA512: 715a32ec190c70505ffbce70bfe81fc7b6aa33e376b60292e801f60cf17025aabfcab4e8c53ebb2e28ffc5cf4c20b74fe3dd8548371ad772085c13aec8b7970e
f8a06f7a02 doc: remove references to disable-asm option now that it's gone (Cory Fields)
376f0f6d07 build: remove confusing and inconsistent disable-asm option (Cory Fields)
Pull request description:
1. It didn't actually disable asm usage in our code. Regardless of the setting, asm is used in random.cpp and support/cleanse.cpp.
2. The value wasn't forwarded to libsecp as a user might have reasonably expected.
3. We now have the DISABLE_OPTIMIZED_SHA256 define which is what disable-asm actually did in practice.
If there is any desire, we can hook DISABLE_OPTIMIZED_SHA256 up to a new configure option that actually does what it says.
Additionally, this is one of the last (THE last?) remaining uses of autoconf defines in our crypto code. As such it seems like low-hanging fruit.
ACKs for top commit:
fanquake:
ACK f8a06f7a02
Tree-SHA512: 4a99c2130225acbe9dc7399ed572a04ca155cbfa3eef8178a632ba533017d264691e6482cceb1d8f9c5d768619d99a2466dea4b82b27b18b872bceae91b92fbb
6ee3997d03 test: removes unnecessary check from validation_tests (Sergi Delgado Segura)
Pull request description:
An unnecessary check was added to the block mutation tests in #29412 where IsBlockMutated is returning true for the invalid reasons: we try to check mutation via transaction duplication, but the merkle root is not updated before the check, therefore the check fails because the provided root and the computed root differ, but not because the block contains the same transaction twice.
Notice that a proper check to test the duplication case is added a few lines later, so this check is just meaningless and can be removed. Check https://github.com/bitcoin/bitcoin/pull/29412#discussion_r1506490281 for context.
ACKs for top commit:
maflcko:
ACK 6ee3997d03
dergoegge:
utACK 6ee3997d03
BrandonOdiwuor:
utACK 6ee3997d03
Tree-SHA512: e4627668091dda5f589e4c15edac39dc84aabc9b34b8f7fadbf512beb7111d5477e1b69567a34b4a657e48ba66dfb864db5ff37c9bbe3ff24cd32931b2dd89e6
1. It didn't actually disable asm usage in our code. Regardless of the setting,
asm is used in random.cpp and support/cleanse.cpp.
2. The value wasn't forwarded to libsecp as a user might have reasonably
expected.
3. We now have the DISABLE_OPTIMIZED_SHA256 define which is what disable-asm
actually did in practice.
If there is any desire, we can hook DISABLE_OPTIMIZED_SHA256 up to a new
configure option that actually does what it says.
e073f1dfda test: make sure keypool sizes do not change on `getrawchangeaddress`/`getnewaddress` failures (UdjinM6)
367bb7a80c wallet: Avoid updating `ReserveDestination::nIndex` when `GetReservedDestination` fails (UdjinM6)
Pull request description:
I think the expected behaviour of `getrawchangeaddress` and `getnewaddress` RPCs is that their failure should not affect keypool in any way. At least that's how legacy wallets work, you can confirm this behaviour by running `wallet_keypool.py --legacy-wallet` on master with e073f1dfda applied on top. However running `wallet_keypool.py --descriptors` on the same commit results in the following failure:
```
File "/path/to/bitcoin/test/functional/test_framework/test_framework.py", line 131, in main
self.run_test()
File "/path/to/bitcoin/test/functional/wallet_keypool.py", line 114, in run_test
assert_equal(kp_size_before, kp_size_after)
File "/path/to/bitcoin/test/functional/test_framework/util.py", line 57, in assert_equal
raise AssertionError("not(%s)" % " == ".join(str(arg) for arg in (thing1, thing2) + args))
AssertionError: not([18, 24] == [19, 24])
```
This happens because we pass `nIndex` (which is a class member) into `GetReservedDestination` and since it's passed by reference we get an updated value back, so `nIndex` won't be equal `-1` anymore, no matter if the function failed or succeeded. This means that `ReturnDestination` (called by dtor of `ReserveDestination`) will try to return something we did not actually reserve.
The fix is to simply use a temporary variable instead of a class member and only update `nIndex` when `op_address` actually has value, basically do it the same way we do for other class members (`address` and `fInternal`) already.
ACKs for top commit:
achow101:
ACK e073f1dfda
josibake:
ACK e073f1dfda
Tree-SHA512: 1128288a60dd4d8f306ef6f7ac66cdfeae3c9cc35c66ecada2d78fa61ac759f2a757b70fc3976ba8b5081200942b58dfabc184c01ccf911af40ba8c145344651
An unnecessary check was added to the block mutation tests
in #29412 where IsBlockMutated is returning true for the invalid
reasons: we try to check mutation via transaction duplication,
but the merkle root is not updated before the check, therefore
the check fails because the provided root and the computed root
differ, but not because the block contains the same transaction twice.
The check is meaningless so it can be removed.
d8087adc7e [test] IsBlockMutated unit tests (dergoegge)
1ed2c98297 Add transaction_identifier::size to allow Span conversion (dergoegge)
1ec6bbeb8d [validation] Cache merkle root and witness commitment checks (dergoegge)
5bf4f5ba32 [test] Add regression test for #27608 (dergoegge)
49257c0304 [net processing] Don't process mutated blocks (dergoegge)
2d8495e080 [validation] Merkle root malleation should be caught by IsBlockMutated (dergoegge)
66abce1d98 [validation] Introduce IsBlockMutated (dergoegge)
e7669e1343 [refactor] Cleanup merkle root checks (dergoegge)
95bddb930a [validation] Isolate merkle root checks (dergoegge)
Pull request description:
This PR proposes to check for mutated blocks early as a defense-in-depth mitigation against attacks leveraging mutated blocks.
We introduce `IsBlockMutated` which catches all known forms of block malleation and use it to do an early mutation check whenever we receive a `block` message.
We have observed attacks that abused mutated blocks in the past, which could have been prevented by simply not processing mutated blocks (e.g. https://github.com/bitcoin/bitcoin/pull/27608 for which a regression test is included in this PR).
ACKs for top commit:
achow101:
ACK d8087adc7e
maflcko:
ACK d8087adc7e🏄
fjahr:
Code review ACK d8087adc7e
sr-gi:
Code review ACK d8087adc7e
Tree-SHA512: 618ff4ea7f168e10f07504d3651290efbb1bb2ab3b838ffff3527c028caf6c52dedad18d04d3dbc627977479710930e200f2dfae18a08f627efe7e64a57e535f
51bc1c7126 test: Remove Windows-specific code from `system_tests/run_command` (Hennadii Stepanov)
Pull request description:
The removed code has been dead since https://github.com/bitcoin/bitcoin/pull/28967.
Required as a precondition for replacing Boost.Process with [cpp-subprocess](https://github.com/bitcoin/bitcoin/pull/28981) to make diff for this code meaningful and reviewable.
The plan is to reintroduce Windows-specific code in this test simultaneously with enabling Windows support in cpp-subprocess.
ACKs for top commit:
Sjors:
utACK 51bc1c7126
theStack:
Code-review ACK 51bc1c7126
Tree-SHA512: 0e3875c4dc20564332555633daf2227223b10dc3d052557635eced2734575d1e0252fb19e46ea6e6c47a15c51c345f70b6d437e33435abcd0e4fcf29edb50887
During shutdown, already queue events dispatched from the backend such
'numConnectionsChanged' and 'networkActiveChanged' could try to access
the clientModel object, which might not exist because we manually delete
it inside 'BitcoinApplication::requestShutdown()'.
These replace our platform-specific mess in favor of c++20 endian detection
via std::endian and internal byteswap functions when necessary.
They no longer rely on autoconf detection.
Rather than a complicated set of tests to decide which bswap functions to
use, always prefer the compiler built-ins when available.
These builtins and fallbacks can all be removed once we're using c++23, which
adds std::byteswap.
This code has been dead since https://github.com/bitcoin/bitcoin/pull/28967.
Required as a precondition for replacing Boost.Process with
cpp-subprocess to make diff for this code meaningful and reviewable.
The plan is to reintroduce Windows-specific code in this test
simultaneously with enabling Windows support in cpp-subprocess.
We preemptively perform a block mutation check before further processing
a block message (similar to early sanity checks on other messsage
types). The main reasons for this change are as follows:
- `CBlock::GetHash()` is a foot-gun without a prior mutation check, as
the hash returned only commits to the header but not to the actual
transactions (`CBlock::vtx`) contained in the block.
- We have observed attacks that abused mutated blocks in the past, which
could have been prevented by simply not processing mutated blocks
(e.g. https://github.com/bitcoin/bitcoin/pull/27608).
