84934bf70e11fe4cda1cfda60113a54895d4fdd5 multiprocess: Add echoipc RPC method and test (Russell Yanofsky)
7d76cf667eff512043a28d4407cc89f58796c42b multiprocess: Add comments and documentation (Russell Yanofsky)
ddf7ecc8dfc64cf121099fb047e1ac871de94f4c multiprocess: Add bitcoin-node process spawning support (Russell Yanofsky)
10afdf0280fa93bfffb0a7665c60dc155cd84514 multiprocess: Add Ipc interface implementation (Russell Yanofsky)
745c9cebd50fea1664efef571dc1ee1bddc96102 multiprocess: Add Ipc and Init interface definitions (Russell Yanofsky)
5d62d7f6cd48bbc4e9f37ecc369f38d5e1e0036c Update libmultiprocess library (Russell Yanofsky)
Pull request description:
This PR is part of the [process separation project](https://github.com/bitcoin/bitcoin/projects/10).
---
This PR adds basic process spawning and IPC method call support to `bitcoin-node` executables built with `--enable-multiprocess`[*].
These changes are used in https://github.com/bitcoin/bitcoin/pull/10102 to let node, gui, and wallet functionality run in different processes, and extended in https://github.com/bitcoin/bitcoin/pull/19460 and https://github.com/bitcoin/bitcoin/pull/19461 after that to allow gui and wallet processes to be started and stopped independently and connect to the node over a socket.
These changes can also be used to implement new functionality outside the `bitcoin-node` process like external indexes or pluggable transports (https://github.com/bitcoin/bitcoin/pull/18988). The `Ipc::spawnProcess` and `Ipc::serveProcess` methods added here are entry points for spawning a child process and serving a parent process, and being able to make bidirectional, multithreaded method calls between the processes. A simple example of this is implemented in commit "Add echoipc RPC method and test."
Changes in this PR aside from the echo test were originally part of #10102, but have been split and moved here for easier review, and so they can be used for other applications like external plugins.
Additional notes about this PR can be found at https://bitcoincore.reviews/19160
[*] Note: the `--enable-multiprocess` feature is still experimental, and not enabled by default, and not yet supported on windows. More information can be found in [doc/multiprocess.md](https://github.com/bitcoin/bitcoin/blob/master/doc/multiprocess.md)
ACKs for top commit:
fjahr:
re-ACK 84934bf70e11fe4cda1cfda60113a54895d4fdd5
ariard:
ACK 84934bf. Changes since last ACK fixes the silent merge conflict about `EnsureAnyNodeContext()`. Rebuilt and checked again debug command `echoipc`.
Tree-SHA512: 52a948b5e18a26d7d7a09b83003eaae9b1ed2981978c36c959fe9a55abf70ae6a627c4ff913a3428be17400a3dace30c58b5057fa75c319662c3be98f19810c6
fa0074e2d82928016a43ca408717154a1c70a4db scripted-diff: Bump copyright headers (MarcoFalke)
Pull request description:
Needs to be done because no one has removed the years yet
ACKs for top commit:
practicalswift:
ACK fa0074e2d82928016a43ca408717154a1c70a4db
Tree-SHA512: 210e92acd7d400b556cf8259c3ec9967797420cfd19f0c2a4fa54cb2b3d32ad9ae27e771269201e7d554c0f4cd73a8b1c1a42c9f65d8685ca4d52e5134b071a3
7e698732836121912f179b7c743a72dd6fdffa72 sync: remove DEBUG_LOCKCONTENTION preprocessor directives (Jon Atack)
9b08006bc502e67956d6ab518388fad6397cac8d log, sync: improve lock contention logging and add time duration (Jon Atack)
3f4c6b87f1098436693c4990f2082515ec0ece26 log, timer: add timing macro in usec LOG_TIME_MICROS_WITH_CATEGORY (Jon Atack)
b7a17444e0746c562ae97b26eba431577947b06a log, sync: add LOCK logging category, apply it to lock contention (Jon Atack)
Pull request description:
To enable lock contention logging, `DEBUG_LOCKCONTENTION` has to be defined at compilation. Once built, the logging is not limited to a category and is high frequency, verbose and in all-caps. With these factors combined, it seems likely to be rarely used.
This patch:
- adds a `lock` logging category
- adds a timing macro in microseconds, `LOG_TIME_MICROS_WITH_CATEGORY`
- updates `BCLog::LogMsg()` to omit irrelevant decimals for microseconds and skip unneeded code and math
- improves the lock contention logging, drops the all-caps, and displays the duration in microseconds
- removes the conditional compilation directives
- allows lock contentions to be logged on startup with `-debug=lock` or at run time with `bitcoin-cli logging '["lock"]'`
```
$ bitcoind -signet -debug=lock
2021-09-01T12:40:01Z LockContention: cs_vNodes, net.cpp:1920 started
2021-09-01T12:40:01Z LockContention: cs_vNodes, net.cpp:1920 completed (4μs)
2021-09-01T12:40:01Z LockContention: cs_vNodes, net.cpp:1302 started
2021-09-01T12:40:01Z LockContention: cs_vNodes, net.cpp:1302 completed (4μs)
2021-09-01T12:40:02Z LockContention: cs_vNodes, net.cpp:2242 started
2021-09-01T12:40:02Z LockContention: cs_vNodes, net.cpp:2242 completed (20μs)
2021-09-01T12:43:04Z LockContention: ::cs_main, validation.cpp:4980 started
2021-09-01T12:43:04Z LockContention: ::cs_main, validation.cpp:4980 completed (3μs)
$ bitcoin-cli -signet logging
"lock": true,
$ bitcoin-cli -signet logging [] '["lock"]'
"lock": false,
$ bitcoin-cli -signet logging '["lock"]'
"lock": true,
```
I've tested this with Clang 13 and GCC 10.2.1, on Debian, with and without `--enable-debug`.
ACKs for top commit:
hebasto:
re-ACK 7e698732836121912f179b7c743a72dd6fdffa72, added a contention duration to the log message since my [previous](https://github.com/bitcoin/bitcoin/pull/22736#pullrequestreview-743764606) review.
theStack:
re-ACK 7e698732836121912f179b7c743a72dd6fdffa72 🔏⏲️
Tree-SHA512: c4b5eb88d3a2c051acaa842b3055ce30efde1f114f61da6e55fcaa27476c1c33a60bc419f7f5ccda532e1bdbe70815222ec2b2b6d9226f29c8e94e598aacfee7
fa77de2baa40ee828c850ef4068c76cc3619e87b rpc: Assert that RPCArg names are equal to CRPCCommand ones (misc) (MarcoFalke)
fa50bdc755489b2e291ea5ba0e39e44a20c6c6de rpc: Limit echo to 10 args (MarcoFalke)
fa89ca9b5bd334813fd7e7edb202c56b35076e8d refactor: Use C++11 range based for loops to simplify rpc code (MarcoFalke)
fa459bdc87bbb050ca1c8d469023a96ed798540e rpc: Treat all args after a hidden arg as hidden as well (MarcoFalke)
Pull request description:
This is split out from #18531 to just touch the RPC methods in misc. Description from the main pr:
### Motivation
RPCArg names in the rpc help are currently only used for documentation. However, in the future they could be used to teach the server the named arguments. Named arguments are currently registered by the `CRPCCommand`s and duplicate the RPCArg names from the documentation. This redundancy is fragile, and has lead to errors in the past (despite having linters to catch those kind of errors). See section "bugs found" for a list of bugs that have been found as a result of the changes here.
### Changes
The changes here add an assert in the `CRPCCommand` constructor that the RPCArg names are identical to the ones in the `CRPCCommand`.
### Future work
> Here or follow up, makes sense to also assert type of returned UniValue?
Sure, but let's not get ahead of ourselves. I am going to submit any further works as follow-ups, including:
* Removing the CRPCCommand arguments, now that they are asserted to be equal and thus redundant
* Removing all python regex linters on the args, now that RPCMan can be used to generate any output, including the cli.cpp table
* Auto-formatting and sanity checking the RPCExamples with RPCMan
* Checking passed-in json in self-check. Removing redundant checks
* Checking returned json against documentation to avoid regressions or false documentation
* Compile the RPC documentation at compile-time to ensure it doesn't change at runtime and is completely static
### Bugs found
* The assert identified issue #18607
* The changes itself fixed bug #19250
ACKs for top commit:
laanwj:
Code review ACK fa77de2baa40ee828c850ef4068c76cc3619e87b
fjahr:
tested ACK fa77de2baa40ee828c850ef4068c76cc3619e87b
theStack:
ACK fa77de2baa
ryanofsky:
Code review ACK fa77de2baa40ee828c850ef4068c76cc3619e87b. Pretty straightfoward changes
Tree-SHA512: badae1606518c0b55ce2c0bb9025d14f05556532375eb20fd6f3bfadae1e5e6568860bff8599d037e655bf1d23f1f464ca17f4db10a6ab3d502b6e9e61c7b3d3
d841301010914203fb5ef02627c76fad99cb11f1 test: Add docstring to wait_until() in util.py to warn about its usage (Seleme Topuz)
1343c86c7cc1fc896696b3ed87c12039e4ef3a0c test: Update wait_until usage in tests not to use the one from utils (Seleme Topuz)
Pull request description:
Replace global (from [test_framework/util.py](https://github.com/bitcoin/bitcoin/blob/master/test/functional/test_framework/util.py#L228)) `wait_until()` usages with the ones provided by `BitcoinTestFramework` and `P2PInterface` classes.
The motivation behind this change is that the `util.wait_until()` expects a timeout, timeout_factor and lock and it is not aware of the context of the test framework. `BitcoinTestFramework` offers a `wait_until()` which has an understandable amount of default `timeout` and a shared `timeout_factor`. Moreover, on top of these, `mininode.wait_until()` also has a shared lock.
closes#19080
ACKs for top commit:
MarcoFalke:
ACK d841301010914203fb5ef02627c76fad99cb11f1 🦆
kallewoof:
utACK d841301010914203fb5ef02627c76fad99cb11f1
Tree-SHA512: 81604f4cfa87fed98071a80e4afe940b3897fe65cf680a69619a93e97d45f25b313c12227de7040e19517fa9c003291b232f1b40b2567aba0148f22c23c47a88
fad0ce59e9 tests: Fail if RPC has been added without tests (MarcoFalke)
Pull request description:
Need to be run with --coverage
ACKs for commit fad0ce:
ryanofsky:
utACK fad0ce59e9154f9b7e61907a71c740a942c60282. New comment in travis.yml is the only change since last review.
Tree-SHA512: b53632dfe9865ec06991bfcba2fd67238bebbb866b355f09624eaf233257b2bca902caac6c24abb358b2f4c1c43f28ca75e30982765911e1a117102df65276d9
f13ad1cae0 modify test for memory locked in case locking pages failed at some point (Adam Jonas)
2fa85ebd1c add rpc_misc.py, mv test getmemoryinfo, add test mallocinfo (Adam Jonas)
Pull request description:
Creating the `rpc_misc.py` functional test file to add space for adding tests to a file that doesn't have a lot of coverage.
- Removing the `getmemoryinfo()` smoke test from wallet basic rather than moving it to keep the wallet decoupled. Feel like testing for reasonable memory allocation values should suffice.
- Adding coverage for `mallocinfo()`. Introduced standard lib XML parser since the function exports an XML string that describes the current state of the memory-allocation implementation in the caller.
Tree-SHA512: ced30115622916c88d1e729969ee331272ec9f2881eb36dee4bb7331bf633a6810a57fed63a0cfaf86de698edb5162e6a035efd07c89ece1df56b69d61288072