2811f40f30 ci: only run USDT interface tests on CirrusCI (0xb10c)
Pull request description:
As mentioned in #26571, the task running the USDT interface tests fail when run in docker. cc7335edc8 in #25528 added that the tests are run in a **VM** in Cirrus CI. Running them locally in docker containers might not work:
- We use [bcc] as tracing toolkit which requires the kernel headers to compile the BPF bytecode. As docker containers use the hosts kernel and don't run their own, there is a potential for mismatches between kernel headers available in the container and the host kernel. This results in a failure loading the BPF byte code.
- Privilges are required to load the BPF byte code into the kernel. Normally, the docker containers aren't run with these.
- We currently use an untrusted third-party PPA to install the bpfcc-tools package on Ubuntu 22.04. Using this on a local dev system could be a security risk.
To not hinder the ASan + LSan + UBSan part of the CI task, the USDT tests are disabled on non-CirrusCI runs.
[bcc]: https://github.com/iovisor/bcc
Top commit has no ACKs.
Tree-SHA512: 7c159b59630b36d5ed927b501fa0ad6f54ff3d98d0902f975cc8122b4147a7f828175807d40a470a85f0f3b6eeb79307d3465a287dbd2f3d75b803769ad0b6e7
As mentioned in #26571, the task running the USDT interface tests
fail when run in docker. cc7335edc8
in #25528 added that the tests are run in a **VM** in Cirrus CI.
Running them locally in docker containers might not work:
- We use [bcc] as tracing toolkit which requires the kernel headers
to compile the BPF bytecode. As docker containers use the hosts
kernel and don't run their own, there is a potential for mismatches
between kernel headers available in the container and the host
kernel. This results in a failure loading the BPF byte code.
- Privilges are required to load the BPF byte code into the kernel.
Normally, the docker containers aren't run with these.
- We currently use an untrusted third-party PPA to install the
bpfcc-tools package on Ubuntu 22.04. Using this on a local dev
system could be a security risk.
To not hinder the ASan + LSan + UBSan part of the CI task, the USDT
tests are disabled on non-CirrusCI runs.
[bcc]: https://github.com/iovisor/bcc
fad1c55301 lint: Skip COMMIT_RANGE if no CIRRUS_PR (MarcoFalke)
Pull request description:
It doesn't make sense to run this for non-PRs, because:
* There are known whitespace "violations" in previous commits, so the lint may fail
* Once the changes are merged, it is too late to fix them up (force pushes are illegal)
* It isn't possible to determine which commits to run on if there is no reference branch (target branch of the pull request)
Moreover, the test fails on non-master:
* https://github.com/bitcoin/bitcoin/runs/8664441400
Fix all issues by skipping it.
ACKs for top commit:
hebasto:
ACK fad1c55301, also tested in my personal Cirrus account.
Tree-SHA512: be15f00e2b2a9069583833545883e0e5968a33d2455dad59e6fb47c1102b4dd16ef932e9ba945e29e9d941e6c17bd531a02c66b0491097801be6bda476875537
54dd8f51ce ci: use ci_exec_root for clang install (josibake)
Pull request description:
fixes a bug introduced in #25900 ; see https://github.com/bitcoin/bitcoin/pull/25900#issuecomment-1327311069
the general idea of #25900 was to use a non-root user as much as possible to avoid modifying the user's local filesystem. however, it appears the root user is needed to correctly install clang.
ACKs for top commit:
hebasto:
ACK 54dd8f51ce, tested on Ubuntu 22.04.
Tree-SHA512: beb01d4b6127fbba3c8d18e85cf7ec7d1b2ec93ea05c475ab51bcaa04ef1b0591d886f1a7e0732c5ae86806013f022c0b44027380d2b0cfb1bfdc843e40f99b4
849f20a6d3 ci: create and use non-root user for docker image (josibake)
Pull request description:
Previously, everything in the ci docker image ran as the root user. This would lead to certain directories (`ci/scratch`, `depends`) being owned by `root` after running the ci locally which would lead to annoying behavior such as subsequent guix builds failing due to `depends/` being owned by root.
This PR adds a non-root user in the container and chowns the mounted working directory. All the `docker exec` commands now run as the non-root user, except for the few that still need to run as root (mainly, installing packages).
To test this I checked out a fresh copy of the repo, applied my changes, ran the CI, and verified all the local file permissions were unchanged after the CI was finished running.
ACKs for top commit:
hebasto:
ACK 849f20a6d3, tested on Ubuntu 22.04 by running commands as follows:
Tree-SHA512: 734dca0f36157fce5fab243b4ff657fc17ba980e8e4e4644305f41002ff21bd5cef02c306ea1e0b5c841d4c07c095e8e4be16722e6a38c890717c60a3f5ec62a
b89530483d util: move threadinterrupt into util (fanquake)
Pull request description:
Alongside thread and threadnames. It's part of libbitcoin_util.
ACKs for top commit:
ryanofsky:
Code review ACK b89530483d. No changes since last review other than rebase
theuni:
ACK b89530483d.
Tree-SHA512: 0421f4d1881ec295272446804b27d16bf63e6b62b272f8bb52bfecde9ae6605e8109ed16294690d3e3ce4b15cc5e7c4046f99442df73adb10bdf069d3fb165aa
Running all commands as the root user in the docker image
will change local file permissions in the ci and depends directory.
Add a non-root user to the container and use this user whenever
possible when running docker exec commands.
da16893474 ci: Use `macos-ventura-xcode:14.1` image for "macOS native" task (Hennadii Stepanov)
702836530f ci: Make `getopt` path architecture agnostic (Hennadii Stepanov)
Pull request description:
The "macOS native" CI task always uses the recent OS image.
This PR updates it up to the recent macOS release.
Cirrus Labs [stopped](https://github.com/bitcoin/bitcoin/pull/25160#issuecomment-1162829773) updating macOS images for `x86_64`, therefore, an `arm64` image been used.
Also `make test-security-check` has been dropped as it ["isn't even expected to pass"](https://github.com/bitcoin/bitcoin/issues/26386#issuecomment-1290318628) on `arm64` in CI.
ACKs for top commit:
Sjors:
utACK da16893
Tree-SHA512: 36785d33b7f11b3cdbc53bcfbf97d88bf821fad248c825982dd9f8e3413809a4ef11190eaf950e60fdf479b62ff66920c35d9ea42d534723f015742eec7e19b6
fa6054e952 ci: Allow PIP_PACKAGES on centos (MacroFake)
fac085a05c ci: Remove unused package (MacroFake)
Pull request description:
This was added in 7fc5e865b9 but I can't see a reason why this should be forbidden.
This is also needed for other changes (bumping the minimum python version).
ACKs for top commit:
hebasto:
re-ACK fa6054e952
Tree-SHA512: e8ead9ee00079024eb1e8c6e7b31c78cf2a3392159b444765c2ea9a58bed2a7550bf71083210692a45bb8ed7896cb882b72bf70baa13a2384864b2b510b73005
21a9e94dbb ci: remove hardcoded tag list from ci scripts (josibake)
d530ba390e doc: update test/README.md (josibake)
614d4682ba script: default to necessary tags in get_previous_releases.py (josibake)
Pull request description:
Almost every time I need to use this script, I forget the tag list is needed and that a specific set of tags is needed for the backwards compatibility tests to work. I end up wasting time reading through the script and googling to find the tag list before remembering it is in `test/README.md`
I assume (hope) I'm not the only one this happens to, so I figured it would make more sense to have the script default to downloading/building the necessary tags. This has the added benefit of making the script the source of truth: the script already needs to be updated with the SHA256_SUM of the binary for every new tag that is added, so it makes sense to use `SHA256_SUMS` list as the necessary tag list. This means there is less risk of the README and the script drifting (i.e updating the readme with a new tag and forgetting to update the script, or updating the script and forgetting to update the README). Now all that needs to happen is to update the `SHA256_SUMS` list in the script and everything Just Works (TM)
ACKs for top commit:
Sjors:
re-tACK 21a9e94dbb
Tree-SHA512: 97b488227a89a6827584edd251820a7074fad75dfd7f26f1aa5f858e2521d2e02effd0f11e6dc4676e1155d3d5aba6ff94a4b58ffef80dc201376afd5927deb9
Also:
- Make DEFAULT_MAX_SIG_CACHE_SIZE into constexpr
DEFAULT_MAX_SIG_CACHE_BYTES to utilize the compile-time integer
arithmetic overflow checking available to constexpr.
- Fix comment (MiB instead of MB) for DEFAULT_MAX_SIG_CACHE_BYTES.
- Pass in max_size_bytes parameter to InitS*Cache(), modify log line to
no longer allude to maxsigcachesize being split evenly between the two
validation caches.
- Fix possible integer truncation and add a comment.
[META] I've kept the integer types as int64_t in order to not introduce
unintended behaviour changes, in the next commit we will make
them size_t.
It is part of the node library. Also, it won't be moved to the kernel
lib, as it will be pruned of ArgsManager.
-BEGIN VERIFY SCRIPT-
# Move module
git mv src/mempool_args.cpp src/node/
git mv src/mempool_args.h src/node/
# Replacements
sed -i 's:mempool_args\.h:node/mempool_args.h:g' $(git grep -l mempool_args)
sed -i 's:mempool_args\.cpp:node/mempool_args.cpp:g' $(git grep -l mempool_args)
sed -i 's:MEMPOOL_ARGS_H:NODE_MEMPOOL_ARGS_H:g' $(git grep -l MEMPOOL_ARGS_H)
-END VERIFY SCRIPT-
cc7335edc8 ci: run USDT interface test in a VM (0xb10c)
dba6f82342 test: adopt USDT utxocache interface tests (0xb10c)
220a5a2841 test: hook into PID in tracing tests (0xb10c)
Pull request description:
Changes a CI task that runs test the previously not run `test/functional/interface_usdt_*.py` functional tests (added in https://github.com/bitcoin/bitcoin/pull/24358).
This task is run as CirussCI `compute_engine_instance` VM as hooking into the tracepoints is not possible in CirrusCI docker containers (https://github.com/bitcoin/bitcoin/issues/23296#issuecomment-1024920845). We use an unoffical PPA and untrusted `bpfcc-tools` package in the CI as the Ubuntu jammy and Debian bullseye packages are outdated. We hope use an official package when new Ubuntu/Debian releases are available for the use with Google Compute Engine.
We make sure to hook into `bitcoind` binaries in USDT interface tests via their PID, instead of their path. This makes sure multiple functional tests running in parallel don't interfere with each other.
The utxocache USDT interface tests is adopted to a change of the functional test framework that wasn't detected as the tests weren't run in the CI. As the tracepoints expose internals, it can happen that we need to adopt the interface test when internals change. This is a bit awkward, and if it happens to frequently, we should consider generalizing the tests a bit more. For now it's fine, I think.
See the individual commit messages for more details on the changes.
Fixes https://github.com/bitcoin/bitcoin/issues/24782
Fixes https://github.com/bitcoin/bitcoin/issues/23296
I'd like to hear from reviewers:
- Are we OK with using the [`hadret/bpfcc`](https://launchpad.net/~hadret/+archive/ubuntu/bpfcc) PPA for now? There is a clear plan when to drop it and as is currently, it could only impact the newly added VM task.
- ~~Adding a new task increases CI runtime and costs. Should an existing `container` CI task be ported to a VM and reused instead?~~ Yes, see https://github.com/bitcoin/bitcoin/pull/25528#issuecomment-1179509525
ACKs for top commit:
MarcoFalke:
cr ACK cc7335edc8
Tree-SHA512: b7fddccc0a77d82371229d048abe0bf2c4ccaa45906497ef3040cf99e7f05561890aef4c253c40e4afc96bb838c9787fae81c8454c6fd9db583276e005a4ccb3
fa74e726c4 refactor: Make FEELER_SLEEP_WINDOW type safe (std::chrono) (MacroFake)
fa3b3cb9b5 Expose underlying clock in CThreadInterrupt (MacroFake)
Pull request description:
This gets rid of the `value*1000` manual conversion.
ACKs for top commit:
naumenkogs:
utACK fa74e726c4
dergoegge:
Code review ACK fa74e726c4
Tree-SHA512: 90409c05c25f0dd2f1c4dead78f707ebfd78b7d84ea4db9fcefd9c4958a1a3338ac657cd9e99eb8b47d52d4485fa3c947dce4ee1559fb56ae65878685e1ed9a3
cb3e9a1e3f Move {Load,Dump}Mempool to kernel namespace (Carl Dong)
aa30676541 Move DEFAULT_PERSIST_MEMPOOL out of libbitcoinkernel (Carl Dong)
06b88ffb8a LoadMempool: Pass in load_path, stop using gArgs (Carl Dong)
b857ac60d9 test/fuzz: Invoke LoadMempool via CChainState (Carl Dong)
b3267258b0 Move FopenFn to fsbridge namespace (Carl Dong)
ae1e8e3756 mempool: Use NodeClock+friends for LoadMempool (Carl Dong)
f9e8e5719f mempool: Improve comments for [GS]etLoadTried (Carl Dong)
813962da0b scripted-diff: Rename m_is_loaded -> m_load_tried (Carl Dong)
413f4bb52b DumpMempool: Pass in dump_path, stop using gArgs (Carl Dong)
bd4407817e DumpMempool: Use std::chrono instead of weird int64_t arthmetics (Carl Dong)
c84390b741 test/mempool_persist: Test manual savemempool when -persistmempool=0 (Carl Dong)
Pull request description:
This is part of the `libbitcoinkernel` project: #24303, https://github.com/bitcoin/bitcoin/projects/18
-----
This PR moves `{Dump,Load}Mempool` into its own `kernel/mempool_persist` module and introduces `ArgsManager` `node::` helpers in `node/mempool_persist_args`to remove the scattered calls to `GetBoolArg("-persistmempool", DEFAULT_PERSIST_MEMPOOL)`.
More context can be gleaned from the commit messages.
-----
One thing I was reflecting on as I wrote this was that in the long run, I think we should probably invert the validation <-> mempool relationship. Instead of mempool not depending on validation, it might make more sense to have validation not depend on mempool. Not super urgent since `libbitcoinkernel` will include both validation and mempool, but perhaps something for the future.
ACKs for top commit:
glozow:
re ACK cb3e9a1e3f via `git range-diff 7ae032e...cb3e9a1`
MarcoFalke:
ACK cb3e9a1e3f🔒
ryanofsky:
Code review ACK cb3e9a1e3f
Tree-SHA512: 979d7237c3abb5a1dd9b5ad3dbf3b954f906a6d8320ed7b923557f41a4472deccae3e8a6bca0018c8e7a3c4a93afecc502acd1e26756f2054f157f1c0edd939d
Overloading sleep_for is not needed, as
* seconds and minutes can be converted to milliseconds by the compiler,
not needing a duration_cast
* std::condition_variable::wait_for will convert milliseconds to the
duration type of the underlying clock
So simply expose the clock.
Our CI tasks are run by CirrusCI in Docker containers in a Google
Compute Engine based Kubernetes environment. These containers have
limited capabilities - especially CAP_SYS_ADMIN is missing. See
https://github.com/bitcoin/bitcoin/issues/23296#issuecomment-1024920845
We need elevated privileges to hook into the USDT tracepoints. We use a
CirrusCI "compute_engine_instance" (a VM, not a container) where we have
the required privileges. The ubunut-mininmal-2204-lts was choosen with
debian-11 being an alternative. Both pack an outdated 'bpfcc-tools'
package (v0.18.0) from 2020. This version prints warnings to stderr
during BPF bytecode compilation, which causes our functional test runner
to fail. This is fixed in newer verison.
Until debian-12 or a newer Ubuntu release is avaliable as image in GCE
(https://cloud.google.com/compute/docs/images/os-details), we use a
third-party and untrusted PPA that releases up-to-date versions of the
package.
The official iovisor (authors of BCC) PPA is outdated too. An
alternative would be to compile BCC from source in the CI.
Co-authored-by: MacroFake <falke.marco@gmail.com>
These files change infrequently, and not much header shuffling is required.
We don't add everything in src/util/ yet, because IWYU makes some
dubious suggestions, which I'm going to follow up with upstream.
6eb0909cb7 fuzz: add low-level target for txorphanage (chinggg)
Pull request description:
This adds a low-level fuzz target for orphan transaction handling by creating random transactions and calling all functions in `TxOrphanage`.
It cannot simulate real-world `orphan/unorphan` scenarios effectively since it does not maintain any state about the node and the chain. A high-level fuzz target which construct well-designed transaction graphs will be added later.
ACKs for top commit:
MarcoFalke:
review ACK 6eb0909cb7🐈
Tree-SHA512: b4d64f5941df77d13981f75ec170cef6ffabe782797c982ede7f34134be01dc0026dd7c0bee614bc1d64715e90a933d2a8c95974d402e32eaba8e24cc928299e
