76c60d7b31 test: validation:block_connected tracepoint test (0xb10c)
260e28ece8 test: utxocache:* tracepoint tests (0xb10c)
34b27bac68 test: net:in/out_message tracepoint tests (0xb10c)
c934087b62 test: checks for tracepoint tests (0xb10c)
Pull request description:
This adds functional tests for the USDT tracepoints added in https://github.com/bitcoin/bitcoin/pull/22006 and https://github.com/bitcoin/bitcoin/pull/22902. This partially fixes#23296. The tests **are probably skipped** on most systems as these tests require:
- a Linux system with a kernel that supports BPF (and available kernel headers)
- that Bitcoin Core is compiled with tracepoints for USDT support (default when compiled with depends)
- [bcc](https://github.com/iovisor/bcc) installed
- the tests are run with a privileged user that is able to e.g. do BPF syscalls and load BPF maps
The tests are not yet run in our CI as the CirrusCI containers lack the required permissions (see https://github.com/bitcoin/bitcoin/issues/23296#issuecomment-1024920845). Running the tests in a VM in the CI could work, but I haven't experimented with this yet. The priority was to get the actual tests done first to ensure the tracepoints work as intended for the v23.0 release. Running the tracepoint tests in the CI is planned as the next step to finish #23296.
The tests can, however, be run against e.g. release candidates by hand. Additionally, they provide a starting point for tests for future tracepoints. PRs adding new tracepoint should include tests. This makes reviewing these PRs easier.
The tests require privileges to execute BPF sycalls (`CAP_SYS_ADMIN` before Linux kernel 5.8 and `CAP_BPF` and `CAP_PERFMON` on 5.8+) and permissions to `/sys/kernel/debug/tracing/`. It's currently recommended to run the tests in a virtual machine (or on a VPS) where it's sensible to use the `root` user to gain these privileges. Never run python scripts you haven't carefully reviewed with `root` permissions! It's unclear if a non-root user can even gain the required privileges. This needs more experimenting.
The goal here is to test the tracepoint interface to make sure the [documented interface](https://github.com/bitcoin/bitcoin/blob/master/doc/tracing.md#tracepoint-documentation) does not break by accident. The tracepoints expose implementation details. This means we also need to rely on implementation details of Bitcoin Core in these functional tests to trigger the tracepoints. An example is the test of the `utxocache:flush` tracepoint: On Bitcoin Core shutdown, the UTXO cache is flushed twice. The corresponding tracepoint test expects two flushes, too - if not, the test fails. Changing implementation details could cause these tests to fail and the tracepoint API to break. However, we purposefully treat the tracepoints only as [**semi-stable**](https://github.com/bitcoin/bitcoin/blob/master/doc/tracing.md#semi-stable-api). The tracepoints should not block refactors or changes to other internals.
ACKs for top commit:
jb55:
tACK 76c60d7b31
laanwj:
Tested ACK 76c60d7b31
Tree-SHA512: 9a63d945c68102e59d751bd8d2805ddd7b37185408fa831d28a9cb6641b701961389b55f216c475df7d4771154e735625067ee957fc74f454ad7a7921255364c
7b00595d33 build: stop overriding user CXXFLAGS (fanquake)
3e2ef23c3e build: stop overriding user LDFLAGS (fanquake)
35c3fd43c3 build: stop overriding user CPPFLAGS (fanquake)
bc7cc57607 doc: explain why we clear CXXFLAGS with enable-debug (fanquake)
Pull request description:
Historically our build system has hijacked `CXXFLAGS` and friends, and this has always been a source of complaints from users and developers. With this PR, we move away from using `CXXFLAGS`, `CPPFLAGS` and `LDFLAGS`, and instead use `CORE_*FLAGS` variables for our flags / options, leaving autoconfs `FLAG` vars to the user.
Note that there are currently two cases where we will at least clear `CXXFLAGS` (if not alreaddy overridden by the user), when doing debugging or when coverage is enabled, to avoid Autoconfs `-g -O2` CXXFLAG default.
ACKs for top commit:
hebasto:
ACK 7b00595d33
Tree-SHA512: bda936a7aa8f98a1bf1552306845cb4bbab54e19a7a0b9ce3210e10fef70db146e9fe42a0cc8c50b2908506771b5b96f39c334e41323b70ec878e4010373096c
0d01272cd8 build: don't use Boost multi_index serialization (fanquake)
Pull request description:
We don't use the serialization or archiving facilities of multi_index.
So globally disable support, which gives a minor improvement in build
time, i.e less preprocessing work, given we don't link any Boost libs.
See: https://www.boost.org/doc/libs/1_78_0/libs/multi_index/doc/tutorial/creation.html
> Serialization capabilities are automatically provided by just linking with the appropriate Boost.Serialization library module: it is not necessary to explicitly include any header from Boost.Serialization, apart from those declaring the type of archive used in the process. If not used, however, serialization support can be disabled by globally defining the macro BOOST_MULTI_INDEX_DISABLE_SERIALIZATION. Disabling serialization for Boost.MultiIndex can yield a small improvement in build times, and may be necessary in those defective compilers that fail to correctly process Boost.Serialization headers.
ACKs for top commit:
MarcoFalke:
cr ACK 0d01272cd8
Tree-SHA512: 87c664a2f142dc6b8f8598341f9829be3fda8cf614d73cc9a894c8033ee40c6daa9b50f4049ecb1f1e3aaf342568d9a5f5c65af1e04c36ee3a9cb46eca95767b
d4ba2b2cbc compat: remove strnlen back-compat code (fanquake)
Pull request description:
This was needed for mingw (not mingw-w64), and some older versions of
macOS, which we no-longer support.
ACKs for top commit:
hebasto:
ACK d4ba2b2cbc
Tree-SHA512: d1beb9df58464feea3076091361d7d46e4a8901e347644a5fa6f24e052ca24ee0c7c0dd3f2a3d682b0204bf50430fa89eac62121691ea08af6dcf6b907bdec87
532c64a726 build: Fix Boost.Process test for Boost 1.78 (Hennadii Stepanov)
Pull request description:
Rebased #24415 with Luke's suggestion.
Fixes#24413.
ACKs for top commit:
hebasto:
ACK 532c64a726, tested on Mac mini (M1, 2020) + macOS Monterey 12.3 (21E230).
Tree-SHA512: 74f779695f6bbc45a2b7341a1402f747cc0d433d74825c7196cb9f156db0c0299895365f01665bd0bff12a8ebb5ea33a29b9a52f5eac0007ec35d1dca6544705
5a157eb370 Bugfix: configure: Only avoid -isystem for exact /usr/include path (Luke Dashjr)
556ee6f2fa Bugfix: configure: Quote SUPPRESS_WARNINGS sufficiently to preserve brackets (Luke Dashjr)
Pull request description:
The regex includes `[/ ]` which is supposed to match either a forward slash or a space, but m4 treats the brackets as special characters and effectively strips them out, leading to -isystem /usr/include paths except for in the typical scenario where it is the final parameter in the flag string.
ACKs for top commit:
hebasto:
ACK 5a157eb370, tested on Ubuntu 22.04 with clang 14.0.
vasild:
ACK 5a157eb370
Tree-SHA512: 5c8c282b647b7853b8fad1b5b473703c4a0635073d2685a8ac984151046e2c6a859e6972465419d27356dd29a47f21a2a3a6ad402ec434fe1f9882e5a35f0749
999982b06c build: Add --enable-c++20 option (MarcoFalke)
fae679065e Add CSerializedNetMsg::Copy() helper (MarcoFalke)
fabb7c4ba6 Make fs.h C++20 compliant (MarcoFalke)
fae2220f4e scheduler: Capture ‘this’ explicitly in lambda (MarcoFalke)
Pull request description:
This is for CI and devs only and doesn't change that C++17 is the standard we are currently using. The option `--enable-c++20` allows CI to check that the C++17 code in the repo is also valid C++20. (There are some cases where valid C++17 doesn't compile under C++20).
Also, it allows developers to easily play with C++20 in the codebase.
ACKs for top commit:
ryanofsky:
Code review ACK 999982b06c. Since last review was rebased, and enum-conversion change was dropped, and CSerializedNetMsg copy workaround was added
fanquake:
utACK 999982b06c
Tree-SHA512: afc95ba03ea2b937017fc8e2b1449379cd2b6f7093c430d2e344c665a00c51e402d6651cbcbd0be8118ea1e54c3a86e67d2021d19ba1d4da67168e9fcb6b6f83
The regex includes [/ ] which is supposed to match either a forward slash or a
space, but m4 treats the brackets as special characters and effectively strips
them out, leading to -isystem /usr/include paths except for in the typical
scenario where it is the final parameter in the flag string.
2c03cec2ff ci: Build bitcoin-chainstate (Carl Dong)
095aa6ca37 build: Add example bitcoin-chainstate executable (Carl Dong)
Pull request description:
Part of: #24303
This PR introduces an example/demo `bitcoin-chainstate` executable using said library which can print out information about a datadir and take in new blocks on stdin.
Please read the commit messages for more details.
-----
#### You may ask: WTF?! Why is `index/*.cpp`, etc. being linked in?
This PR is meant only to capture the state of dependencies in our consensus engine as of right now. There are many things to decouple from consensus, which will be done in subsequent PRs. Listing the files out right now in `bitcoin_chainstate_SOURCES` is purely to give us a clear picture of the task at hand, it is **not** to say that these dependencies _belongs_ there in any way.
### TODO
1. Clean up `bitcoin-chainstate.cpp`
It is quite ugly, with a lot of comments I've left for myself, I should clean it up to the best of my abilities (the ugliness of our init/shutdown might be the upper bound on cleanliness here...)
ACKs for top commit:
ajtowns:
ACK 2c03cec2ff
ryanofsky:
Code review ACK 2c03cec2ff. Just rebase, comments, formatting change since last review
MarcoFalke:
re-ACK 2c03cec2ff 🏔
Tree-SHA512: 86e7fb5718caa577df8abc8288c754f4a590650d974df9d2f6476c87ed25c70f923c4db651c6963f33498fc7a3a31f6692b9a75cbc996bf4888c5dac2f34a13b
On the master branch, bump to 23.99 (pre-24.0).
Tree-SHA512: 1e3b0cee8a2b5080170b59a4c445a3c1b69b99152e8eec7eba7080ab447cc6f9c6bd8f69df2b18ee9416de44a6ed88009a200ad26e89275f6230339330d12314
774323e378 ci: Force `--enable-external-signer` to prevent future regressions (Hennadii Stepanov)
69978858a4 build: Fix Boost.Process check for Boost 1.73 and older (Hennadii Stepanov)
2199ef79cb build: Fix a non-portable use of `test` (Hennadii Stepanov)
d436c488d4 build, refactor: Replace tabs with spaces (Hennadii Stepanov)
Pull request description:
On master (5f44c5c428) Boost.Process check false fails without the `-lpthread` flag.
```
$ grep -C 2 pthread_detach config.log
/usr/bin/ld: /tmp/cczCQfQv.o: in function `boost::asio::detail::posix_global_impl<boost::asio::system_context>::~posix_global_impl()':
conftest.cpp:(.text._ZN5boost4asio6detail17posix_global_implINS0_14system_contextEED2Ev[_ZN5boost4asio6detail17posix_global_implINS0_14system_contextEED5Ev]+0xa3): undefined reference to `pthread_join'
/usr/bin/ld: conftest.cpp:(.text._ZN5boost4asio6detail17posix_global_implINS0_14system_contextEED2Ev[_ZN5boost4asio6detail17posix_global_implINS0_14system_contextEED5Ev]+0xc4): undefined reference to `pthread_detach'
collect2: error: ld returned 1 exit status
configure:26674: $? = 1
```
Not required for Boost 1.74+.
ACKs for top commit:
laanwj:
Code review ACK 774323e378, is a bugfix/workaround, seems fine to merge last minute for 23.0.
Tree-SHA512: 2a9d4b67fd8910e107af972d8c223286b7c933bc310616f86c8b6d8c903438916980fc76bd7e37f2698f6ce5361dc706cbf2933d1ac2c081bcabe1b83ca7d6b6
For testing the USDT tracepoint API in the functional tests we
require:
- that we are on a Linux system*
- that Bitcoin Core is compiled with tracepoints
- that bcc and the the Python bcc module [0] is installed
- that we run the tests with the required permissions**
otherwise we skip the tests.
*: We currently only support tracepoints on Linux. Tracepoints are
not compiled on other platforms.
**: Currently, we check for root permissions via getuid == 0. It's
unclear if it's even possible to run the tests a non-root user
with e.g. CAP_BPF, CAP_PERFMON, and access to /sys/kernel/debug/
tracing/. Anyone running these tests as root should carefully
review them first and then run them in a disposable VM.
[0]: https://github.com/iovisor/bcc/blob/master/INSTALL.md
faef344f84 Print enable_fuzz_binary in configure (MarcoFalke)
Pull request description:
A *disabled* `enable_fuzz` on current master does *not* mean the the fuzz binary is not compiled. This is confusion, so fix it.
* `enable_fuzz` toggles compilation flags for fuzzing and disables all other target. There is no need to print this in the configure result, because the compilation flags are already printed. Also, all other targets are already printed as `no`.
* `enable_fuzz_binary` does what it says it does and is currently not printed. So print it.
ACKs for top commit:
hebasto:
ACK faef344f84, tested on Linux Mint 20.2 (x86_64):
Tree-SHA512: 9b02b05c4b9c5fc92cf3487497392690303c36eace5e217f18b4349f059b5a23a7c0e0d030fb6fa7bbad83e927576a5e81c00099164f9ed8e185c0969dc17689
abc057c603 build: Add Boost.Process usage check (Hennadii Stepanov)
Pull request description:
This PR adds a check that Boost.Process can be used without linking any libraries (header-only).
Disable the functionality if that is not the case.
Fixesbitcoin/bitcoin#24314.
ACKs for top commit:
fanquake:
ACK abc057c603
Tree-SHA512: ed2a32b1f751ec6f88cc7220766edd4cdab93c1d0515c892aa3094ee8d5b13ef569830d6e7a7a00c0197b117585dc526d00d943cc99a1f8c8a66ac4e20fe2061
The bitcoin-chainstate executable serves to surface the dependencies
required by a program wishing to use Bitcoin Core's consensus engine as
it is right now.
More broadly, the _SOURCES list serves as a guiding "North Star" for the
libbitcoinkernel project: as we decouple more and more modules of the
codebase from our consensus engine, this _SOURCES list will grow shorter
and shorter. One day, only what is critical to our consensus engine will
remain. Right now, it's "the minimal list of files to link in to even
use our consensus engine".
[META] In a future commit the libbitcoinkernel library will be extracted
from bitcoin-chainstate, and the libbitcoinkernel library's
_SOURCES list will be the list that we aim to shrink.
0c49e52b22 build: remove unneeded getentropy detection (HAVE_GETENTROPY) (Sebastian Falbesoner)
5cd15ffdce random: use arc4random on OpenBSD (Sebastian Falbesoner)
Pull request description:
Inspired by a discussion on obtaining randomness on various OSes in a secp256k1 PR (https://github.com/bitcoin-core/secp256k1/pull/748#discussion_r524605472, see also https://bitcoincore.reviews/libsecp256k1-748), I think it makes sense to follow best practices and use `arc4random_buf` rather than `getentropy` on OpenBSD in our random module.
The [getentropy(2) man page](https://man.openbsd.org/getentropy.2) states:
```
getentropy() is not intended for regular code; please use the
arc4random(3) family of functions instead.
```
The [arc4random(3) man page](https://man.openbsd.org/arc4random.3) states:
```
Use of these functions is encouraged for almost all random number
consumption because the other interfaces are deficient in either quality,
portability, standardization, or availability.
```
On the linked PR discussion worries about using RC4 internally has been expressed (see https://security.stackexchange.com/questions/85601/is-arc4random-secure-enough/172905#172905), but this would only affect users of OpenBSD <5.5, using a version that was released more than 8 years ago.
ACKs for top commit:
laanwj:
Tested ACK 0c49e52b22
Tree-SHA512: b5ed3d0718962c5a3839db9a28f93d08a0ac93094cc664f83bc4cf1cfad25049e6240b7b81fe06b71e6a3a0ca24a2c337eab088abec5470ad014e10c04fdb216
