- Modify `SipHash_32b` benchmark to use `FastRandomContext` for generating initial values.
- Cycle through and modify each byte of the `uint256` value to ensure no part of it can be optimized away.
The lack of "recursion" (where the method call overwrites the used inputs partially) and the systematic modification of each input byte makes the benchmark usage more reliable and thorough.
d7fd766feb test: added test to assert TX decode rpc error on submitpackage rpc (kevkevinpal)
Pull request description:
This PR adds coverage for this line https://github.com/bitcoin/bitcoin/blob/master/src/rpc/mempool.cpp#L996
If you run the following you will get no results for `submitpackage`
`grep -nri "TX decode failed" ./test/functional`
ACKs for top commit:
achow101:
ACK d7fd766feb
instagibbs:
reACK d7fd766feb
tdb3:
ACK d7fd766feb
rkrux:
reACK d7fd766feb
Tree-SHA512: e92e0e2621a4efab35625d8da3ac61ccb7fa65c378aa977112bc132fd3b42431f8c3ceb081f7c9903ed2833c229042b65bdb11444e1d6367354ae65dc7504231
a1b3ccae4b init: warn, don't error, when '-upnp' is set (Antoine Poinsot)
Pull request description:
It prevented the GUI from starting when its settings.json had the -upnp option set. This also doesn't prevent the node from running, so this error didn't need to be fatal.
Thanks to Sjors for bringing attention to what i broke and to Maflcko for suggesting a simple short term fix.
Fixes https://github.com/bitcoin-core/gui/issues/843.
ACKs for top commit:
maflcko:
lgtm ACK a1b3ccae4b
kevkevinpal:
Concept ACK [a1b3cca](a1b3ccae4b)
achow101:
ACK a1b3ccae4b
tdb3:
ACK a1b3ccae4b
Tree-SHA512: ceb1513bf532698e5143d64430a065f39626ef0d2708103ffc8ab7f81e8393f488af2350c5a299bc80f966add82a3951b4d81ae8b0e3070c0d15c94e8db4badd
54d07dd37d ci: Do not error on unused-member-function in test each commit (Sergi Delgado Segura)
Pull request description:
After https://github.com/bitcoin/bitcoin/pull/31045, an unused method in a commit will trigger a compilation error, even if that method is later used in a following commit within the same PR.
Do not enforce unused-member-function in test each commit.
Close#31180
ACKs for top commit:
kevkevinpal:
lgtm ACK [54d07dd](54d07dd37d)
maflcko:
lgtm ACK 54d07dd37d
TheCharlatan:
lgtm ACK 54d07dd37d
Tree-SHA512: 79723ca604d44b065b08793b442de9049a689cc5a4df508299a91966456bf77e24a0f3febaf3114fd23b061716554b70ee866bf594283960d952366ea2096dbc
It prevented the GUI from starting when its settings.json had the -upnp option set. This also doesn't
prevent the node from running, so this error didn't need to be fatal.
Thanks to Sjors for bringing attention to what i broke and to Maflcko for suggestion a simple short
term fix.
bbbbaa0d9a Fix unsigned integer overflows in interpreter (MarcoFalke)
Pull request description:
Unsigned integer overflow is well defined by the language and in some cases even useful or necessary. However, I think that it should be avoided in interpreter, as it makes the code harder to read and requires the whole file to be suppressed in the sanitizer. This puts more burden on reviewers to check that any changes to interpreter that involve unsigned integer overflow are sane.
This patch involves a few changes:
* Evaluate the addition in 64-bit "space". Previously, the first argument was `size_t` (unsigned, 32-bit or 64-bit, depending on platform) and the second was `int` (32-bit on all supported platforms). Thus the addition was done in 32-bit or 64-bit "unsigned space". Now the addition is done in 64-bit "signed space" on all platforms. This is safe because signed integer overflow (UB) isn't expected here with 64-bit integers.
* Clarify that the value passed to the "stack macros" always fits in an `int64_t`. This is done with the C++11 syntax `int64_t{i}`, which fails to compile if `i` needs to be narrowed to fit into `int64_t`.
* Explicitly convert the result of the addition to `size_t`. This isn't needed, because the called function already converts the value (see https://en.cppreference.com/w/cpp/container/vector/operator_at), however I have a slight preference for the explicit cast. (Happy to remove if reviewers prefer without)
The patch does not change the bitcoind binary on my 64-bit system with `clang++ -O2`. However, it does change with gcc.
ACKs for top commit:
achow101:
ACK bbbbaa0d9a
ismaelsadeeq:
Code review ACK bbbbaa0d9a
hebasto:
ACK bbbbaa0d9a, I have reviewed the code and it looks OK.
Tree-SHA512: 0e9cbc6a0afd3db0d1d9489fd5e32ff856217604abde370add1f01c2cae8c526f2afedeb372997217c3a70ab0f8f56442e8230f87456f8e21c9abcb7c6578f7c
2957ca9611 build: have "make test" depend on "make all" (Cory Fields)
Pull request description:
See [Upstream docs](https://cmake.org/cmake/help/latest/variable/CMAKE_SKIP_TEST_ALL_DEPENDENCY.html) for specifics.
Unfortunately, this **seems to have no effect when directly executing `ctest`** :(
This brings the test -> hack -> test cycle more inline with how it worked with autotools.
With `CMAKE_SKIP_TEST_ALL_DEPENDENCY` set to FALSE, `make test` will trigger a rebuild, ensuring that test binaries are current before running them.
To test:
```
cmake -S . -B build
make -C build -j24
touch src/primitives/transaction.cpp
make -C build test ARGS=-j24
```
Without this commit, the above will not rebuild before running tests.
ACKs for top commit:
tdb3:
ACK 2957ca9611
itornaza:
ACK 2957ca9611
laanwj:
ACK 2957ca9611
Tree-SHA512: 46e6d53fd9c28b2f47b34c42ae0b061b01ef7614cdf5ee7d49fb9d3413ca92983d8a996379a9067ac8a8cfc16b230946401e68cf114bd4e3c5c02db2c55ec6bb
e60cecc811 doc: add release note for 31156 (Martin Zumsande)
fc7dfb3df5 test: Don't enforce BIP94 on regtest unless specified by arg (Martin Zumsande)
Pull request description:
The added arg `-test=bip94` is only used in a functional test for BIP94. This is done because the default regtest consensus rules should follow mainnet, not testnet.
Fixes#31137.
ACKs for top commit:
achow101:
ACK e60cecc811
tdb3:
cr and light test ACK e60cecc811
rkrux:
tACK e60cecc811
BrandonOdiwuor:
utACK e60cecc811
laanwj:
Code review ACK e60cecc811
Tree-SHA512: ca2f322f89d8808dfc3565fe020d2615cfcc110e188a02128ad7108fef51c735b33d55b5e6a70c505d78f7291f3c635dc7dfbcd78be1348d4d6e483883be4216
559a8dd9c0 key: clear out secret data in `DecodeExtKey` (Sebastian Falbesoner)
Pull request description:
Same as in `DecodeSecret`, we should also clear out the secret data from the vector resulting from the Base58Check parsing for xprv keys. Note that the if condition is needed in order to avoid UB, see #14242 (commit d855e4cac8).
ACKs for top commit:
davidgumberg:
utACK 559a8dd9c0
achow101:
ACK 559a8dd9c0
tdb3:
cr ACK 559a8dd9c0
laanwj:
Code review ACK 559a8dd9c0
Tree-SHA512: c22499fe2899a9a5a58159ec55e94cf961570d8af06358d4a6d1943d567be9b88657af90d060d3083985ea957886a4f91bb762a2fcf3311007e7a535b42b0fde
57529ac4db test: set P2PConnection.p2p_connected_to_node in peer_connect_helper() (Vasil Dimov)
22cd0e888c test: support WTX INVs from P2PDataStore and fix a comment (Vasil Dimov)
ebe42c00aa test: extend the SOCKS5 Python proxy to actually connect to a destination (Vasil Dimov)
ba621ffb9c test: improve debug log message from P2PConnection::connection_made() (Vasil Dimov)
Pull request description:
If requested, make the SOCKS5 Python proxy redirect connections to a set of given destinations. Actually act as a real proxy, connecting the client to a destination, except that the destination is not what the client asked for.
This would enable us to "connect" to Tor addresses from the functional tests.
Plus a few other minor improvements in the test framework as individual commits.
---
These changes are part of https://github.com/bitcoin/bitcoin/pull/29415 but they make sense on their own and would be good to have them, regardless of the fate of #29415. Also, if this is merged, that would reduce the size of #29415, thus the current standalone PR.
ACKs for top commit:
jonatack:
Approach ACK 57529ac4db
achow101:
ACK 57529ac4db
tdb3:
CR and test ACK 57529ac4db
mzumsande:
Code review / tested ACK 57529ac4db
Tree-SHA512: a2892c97bff2d337b37455c409c6136cb62423ce6cc32b197b36f220c1eec9ca046b599135b9a2603c0eb6c1ac4d9795e73831ef0f04378aeea8b245ea733399
3a4a788ee0 init: Correct coins db cache size setting (TheCharlatan)
Pull request description:
The chainstate caches are currently re-balanced on startup even in the non-assumeutxo case, leading to the database being needlessly re-opened and its cache re-allocated.
Similar to `InitCoinsCache` and `m_coinstip_cache_size_bytes`, the `m_coinsdb_cache_size_bytes` should be set in `InitCoinsDB`.
Together with only conservatively setting the cache values when a assumeutxo chainstate is present, this allows for skipping the cache re-balance during initialization in the normal non-assumeutxo case.
Before:
```
2024-10-09T21:22:17Z Checking all blk files are present...
2024-10-09T21:22:17Z Initializing chainstate Chainstate [ibd] @ height -1 (null)
2024-10-09T21:22:17Z Opening LevelDB in /home/drgrid/.bitcoin/signet/chainstate
2024-10-09T21:22:17Z Opened LevelDB successfully
2024-10-09T21:22:17Z Using obfuscation key for /home/drgrid/.bitcoin/signet/chainstate: b0a6f4e95fd05c92
2024-10-09T21:22:17Z Loaded best chain: hashBestChain=0000000e119967d4937dad58456885ae43fb1761db686947e2f8e168c9a39a4f height=216852 date=2024-10-09T21:06:16Z progress=0.999989
2024-10-09T21:22:17Z Opening LevelDB in /home/drgrid/.bitcoin/signet/chainstate
2024-10-09T21:22:17Z Opened LevelDB successfully
2024-10-09T21:22:17Z Using obfuscation key for /home/drgrid/.bitcoin/signet/chainstate: b0a6f4e95fd05c92
2024-10-09T21:22:17Z [Chainstate [ibd] @ height 216852 (0000000e119967d4937dad58456885ae43fb1761db686947e2f8e168c9a39a4f)] resized coinsdb cache to 8.0 MiB
2024-10-09T21:22:17Z [Chainstate [ibd] @ height 216852 (0000000e119967d4937dad58456885ae43fb1761db686947e2f8e168c9a39a4f)] resized coinstip cache to 440.0 MiB
2024-10-09T21:22:17Z init message: Verifying blocks…
```
After:
```
2024-10-09T21:21:37Z Checking all blk files are present...
2024-10-09T21:21:37Z Initializing chainstate Chainstate [ibd] @ height -1 (null)
2024-10-09T21:21:37Z Opening LevelDB in /home/drgrid/.bitcoin/signet/chainstate
2024-10-09T21:21:37Z Opened LevelDB successfully
2024-10-09T21:21:37Z Using obfuscation key for /home/drgrid/.bitcoin/signet/chainstate: b0a6f4e95fd05c92
2024-10-09T21:21:37Z Loaded best chain: hashBestChain=0000012c12b48011a7d9150ce96ed6a44bbf32b09eeecaff4a667789dda2a566 height=216850 date=2024-10-09T20:37:05Z progress=0.999971
2024-10-09T21:21:37Z init message: Verifying blocks…
```
The change may also be verified by looking at the `feature_assumeutxo.py` functional test debug logs.
ACKs for top commit:
fjahr:
utACK 3a4a788ee0
achow101:
ACK 3a4a788ee0
laanwj:
Code review ACK 3a4a788ee0
BrandonOdiwuor:
Code Review ACK 3a4a788ee0
Tree-SHA512: 87878d0d196bb426370d4b4bd180ca52a34017a0799ecea651c2532461fd2927b0f7cc8182276a7d9bb1fe0ede7d0ad677e3714ca22f321917d711c643acc578
0ea84bc362 test: explicitly check boolean verbosity is disallowed (tdb3)
7a2e6b68cd doc: add rpc guidance for boolean verbosity avoidance (tdb3)
698f302df8 rpc: disallow boolean verbosity in getorphantxs (tdb3)
63f5e6ec79 test: add entry and expiration time checks (tdb3)
808a708107 rpc: add entry time to getorphantxs (tdb3)
56bf302714 refactor: rename rpc_getorphantxs to rpc_orphans (tdb3)
7824f6b077 test: check that getorphantxs is hidden (tdb3)
ac68fcca70 rpc: disallow undefined verbosity in getorphantxs (tdb3)
Pull request description:
Implements follow-up suggestions from #30793.
- Now disallows undefined verbosity levels (below and above valid values) (https://github.com/bitcoin/bitcoin/pull/30793#discussion_r1786093549)
- Disallows boolean verbosity (https://github.com/bitcoin/bitcoin/pull/30793#discussion_r1788273274) and adds guidance to developer-notes
- Checks that `getorphantxs` is a hidden rpc (https://github.com/bitcoin/bitcoin/pull/30793#discussion_r1786107786)
- Adds a test for `expiration` time
- Adds `entry` time to the returned orphan objects (verbosity >=1) to relieve the user from having to calculate it from `expiration`. Also adds associated test. (https://github.com/bitcoin/bitcoin/pull/30793#discussion_r1743687732)
- Minor cleanup (blank line removal and log message move) (https://github.com/bitcoin/bitcoin/pull/30793#discussion_r1786092641)
Included a commit to rename the test to a more generic `get_orphans` to better accommodate future orphanage-related RPCs (e.g. `getorphanangeinfo`). Can drop the refactor commit from this PR if people feel strongly about it.
ACKs for top commit:
achow101:
ACK 0ea84bc362
glozow:
utACK 0ea84bc362
rkrux:
tACK 0ea84bc362
itornaza:
tACK 0ea84bc362
Tree-SHA512: e48a088f333ebde132923072da58e970461e74362d0acebbc799c3043d5727cdf5f28e82b43cb38bbed27c603df6710695dba91ff0695e623ad168e985dce08e
0f4bc63585 [fuzz] txdownloadman and txdownload_impl (glozow)
699643f23a [unit test] MempoolRejectedTx (glozow)
fa584cbe72 [p2p] add TxDownloadOptions bool to make TxRequestTracker deterministic (glozow)
f803c8ce8d [p2p] filter 1p1c for child txid in recent rejects (glozow)
5269d57e6d [p2p] don't process orphan if in recent rejects (glozow)
2266eba43a [p2p] don't find 1p1cs for reconsiderable txns that are AlreadyHaveTx (glozow)
fa7027d0fc [refactor] add CheckIsEmpty and GetOrphanTransactions, remove access to TxDownloadMan internals (glozow)
969b07237b [refactor] wrap {Have,Get}TxToReconsider in txdownload (glozow)
f150fb94e7 [refactor] make AlreadyHaveTx and Find1P1CPackage private to TxDownloadImpl (glozow)
1e08195135 [refactor] move new tx logic to txdownload (glozow)
257568eab5 [refactor] move invalid package processing to TxDownload (glozow)
c4ce0c1218 [refactor] move invalid tx processing to TxDownload (glozow)
c6b21749ca [refactor] move valid tx processing to TxDownload (glozow)
a8cf3b6e84 [refactor] move Find1P1CPackage to txdownload (glozow)
f497414ce7 [refactor] put peerman tasks at the end of ProcessInvalidTx (glozow)
6797bc42a7 [p2p] restrict RecursiveDynamicUsage of orphans added to vExtraTxnForCompact (glozow)
798cc8f5aa [refactor] move Find1P1CPackage into ProcessInvalidTx (glozow)
416fbc952b [refactor] move new orphan handling to ProcessInvalidTx (glozow)
c8e67b9169 [refactor] move ProcessInvalidTx and ProcessValidTx definitions down (glozow)
3a41926d1b [refactor] move notfound processing to txdownload (glozow)
042a97ce7f [refactor] move tx inv/getdata handling to txdownload (glozow)
58e09f244b [p2p] don't log tx invs when in IBD (glozow)
288865338f [refactor] rename maybe_add_extra_compact_tx to first_time_failure (glozow)
f48d36cd97 [refactor] move peer (dis)connection logic to TxDownload (glozow)
f61d9e4b4b [refactor] move AlreadyHaveTx to TxDownload (glozow)
84e4ef843d [txdownload] add read-only reference to mempool (glozow)
af918349de [refactor] move ValidationInterface functions to TxDownloadManager (glozow)
f6c860efb1 [doc] fix typo in m_lazy_recent_confirmed_transactions doc (glozow)
5f9004e155 [refactor] add TxDownloadManager wrapping TxOrphanage, TxRequestTracker, and bloom filters (glozow)
Pull request description:
Part of #27463.
This PR does 3 things:
(1) It modularizes transaction download logic into a `TxDownloadManager`. Transaction download logic refers to the process of deciding what transactions to request, download, and validate.[1] There should be no behavior changes. Using `--color_moved=dimmed_zebra -w` may help.
(2) It adds unit and fuzz (🪄) testing for transaction download.
(3) It makes a few small behavioral changes:
- Stop (debug-only) logging tx invs during IBD
- Just like all other transactions, require orphans have RecursiveDynamicUsage < 100k before adding to vExtraTxnForCompact
- Don't return a 1p1c that contains a parent or child in recent rejects. Don't process any orphan already in recent rejects. These cases should not happen in actual node operation; it's just to allow tighter sanity checks during fuzzing.
There are several benefits to this interface, such as:
- Unit test coverage and fuzzing for logic that currently isn't feasible to test as thoroughly (without lots of overhead) and/or currently only lightly tested through `assert_debug_log` (not good) in functional tests.
- When we add more functionality (e.g. package relay messages, more robust orphan handling), the vast majority of it will be within `TxDownloadManager` instead of `PeerManager`, making it easier to review and test. See #28031 for what this looks like.
- `PeerManager` will no longer know anything about / have access to `TxOrphanage`, `TxRequestTracker` or the rejection caches. Its primary interface with `TxDownloadManager` would be much simpler:
- Passing on `ValidationInterface` callbacks
- Telling `txdownloadman` when a peer {connects, disconnects}
- Telling `txdownloadman`when a {transaction, package} is {accepted, rejected} from mempool
- Telling `txdownloadman` when invs, notfounds, and txs are received.
- Getting instructions on what to download.
- Getting instructions on what {transactions, packages, orphans} to validate.
- Get whether a peer `HaveMoreWork` for the `ProessMessages` loop
- (todo) Thread-safety can be handled internally.
[1]: This module is concerned with tx *download*, not upload. It excludes transaction announcements/gossip which happens after we download/accept a transaction. Txreconciliation (erlay) is excluded from this module, as it only relates to deciding which `inv`s to send or helping the other peer decide which `inv`s to send. It is independent from this logic.
ACKs for top commit:
achow101:
light ACK 0f4bc63585
theStack:
ACK 0f4bc63585
instagibbs:
reACK 0f4bc63585
naumenkogs:
ACK 0f4bc63585
Tree-SHA512: 84ab8ef8a0fc705eb829d7f7d6885f28944aaa42b03172f256a42605677b3e783919bb900d4e3b8589f85a0c387dfbd972bcd61d252d44a88c6aaa90e4bf920f
915640e191 depends: zeromq: don't install .pc files and remove patches for them (Cory Fields)
6b8a74463b cmake: Add `FindZeroMQ` module (Hennadii Stepanov)
Pull request description:
This PR introduces the `FindZeroMQ` module, which first attempts to find the `libzmq` library using CMake's `find_package()` and falls back to `pkg_check_modules()` if unsuccessful.
Addresses https://github.com/bitcoin/bitcoin/issues/30876 for the ZeroMQ package.
ACKs for top commit:
fanquake:
ACK 915640e191
Tree-SHA512: 2f17bae21be5d3f280a13425d22f5d1b2e23837a8aaf5ec89c433767509de030a42d598b261e102bdb5b860d8ede98013c124c3d25e081e956d4ee3a81b2584f
This is needed in the next commit to add compile-time checking to strprintf
calls, because bitcoin-cli.cpp uses dynamic width in many format strings.
This change is easiest to review ignoring whitespace.
Co-authored-by: MarcoFalke <*~=`'#}+{/-|&$^_@721217.xyz>
Co-authored-by: Hodlinator <172445034+hodlinator@users.noreply.github.com>
Co-authored-by: l0rinc <pap.lorinc@gmail.com>
c4dc81f9c6 test: Remove dead code from interface_zmq (Fabian Jahr)
Pull request description:
The loop removed here appears to be effectively dead code: In case `get_raw_seq` is behind `zmq_mem_seq` the loop runs and tries to get a more recent (higher) number for `get_raw_seq`. However, the exact number of `get_raw_seq` is asserted in the line above: `assert_equal(get_raw_seq, 6)`. If the loop would actually achieve its purpose this assert would need to be racy. This does not seem to be the case and 6 appears to be the final number. `zmq_mem_seq` however does take some time to catch up (if it were continue to be updated). But this is not handled by the loop and does not seem to be relevant at this point in the test. The backlog is consumed a bit later in another loop that handles this correctly already.
ACKs for top commit:
l0rinc:
ACK c4dc81f9c6
tdb3:
CR re ACK c4dc81f9c6
Tree-SHA512: 663a1711ba1ce04a3d2e2916e0df7a7bb51069e28bc2644b816a483628c95b5e6c29fc6eacc31a5f72b7d9af11096f3c437ea1dc57eaa1ee9ddce43cc20bacd3
6c9fe7b73e test: Prevent connection attempts to random IPs in p2p_seednodes.py (Martin Zumsande)
bb97b1ffa9 test: fix intermittent timeout in p2p_seednodes.py (Martin Zumsande)
Pull request description:
Fixes#31103
On some CI runs, the seed node timer in `ThreadOpenConnection` was only started *after* the mocktime was set.
Fix this by waiting for the first connection attempt, which happens after the timer was started.
Also I noticed that the "unreachable" connections are not in fact unreachable, so that the functional test could attempt connections
to random IPs on the internet. This was already noted in https://github.com/bitcoin/bitcoin/pull/29605#discussion_r1701616675 but the suggested fix never made it in, so I added it to this PR.
ACKs for top commit:
sr-gi:
tACK [6c9fe7b](6c9fe7b73e)
laanwj:
Code review ACK 6c9fe7b73e
tdb3:
cr and light test ACK 6c9fe7b73e
Tree-SHA512: 021b6d5325eab85d79708b4b137f61723a36f2b8a1faf681463bad2ea5283ea528b5ff1701467a86b035d3a6972750a61ace5020e58b7aa61ecaad97664488c8
70713303b6 scripted-diff: Rename `PACKAGE_*` variables to `CLIENT_*` (Hennadii Stepanov)
332655cb52 build: Rename `PACKAGE_*` variables to `CLIENT_*` (Hennadii Stepanov)
e6e29e3c94 scripted-diff: Clarify "user agent" variable name (Hennadii Stepanov)
Pull request description:
The use of `PACKAGE_NAME` for the project's variable name is problematic, as this name is commonly used in CMake's [interface variables](https://cmake.org/cmake/help/latest/command/find_package.html#config-mode-version-selection). If third-party CMake code handles with scopes improperly, our `PACKAGE_NAME` variable could end up with an unexpected value.
This PR avoids such conflicts by renaming all `PACKAGE_*` variables to `CLIENT_*`.
The code in the master branch works correctly only incidentally. It is definitely broken in https://github.com/bitcoin/bitcoin/pull/30997.
ACKs for top commit:
TheCharlatan:
Re-ACK 70713303b6
Tree-SHA512: f0992fb9a1ac4a41125b223f97bcaae50f521c813f334e606693dd0113a4732f12451bbcfb158df8bed44f34b37dadac478b2f5661e8b2588b401f43ae4bc1a4
The added regtest option -test=bip94 is only used in the functional
test for BIP94.
This is done because the default regtest consensus rules
should aim to follow to mainnet, not testnet.
BCC needs the PID of a bitcoind process to attach to the tracepoints
(instead of the binary path used before) when the tracepoints have a
semaphore.
For reference, we already use the PID in our tracepoint interface
tests. See 220a5a2841.
Before this commit, we would always prepare tracepoint arguments
regardless of the tracepoint being used or not. While we already made
sure not to include expensive arguments in our tracepoints, this
commit introduces gating to make sure the arguments are only prepared
if the tracepoints are actually used. This is a win-win improvement
to our tracing framework. For users not interested in tracing, the
overhead is reduced to a cheap 'greater than 0' compare. As the
semaphore-gating technique used here is available in bpftrace, bcc,
and libbpf, users interested in tracing don't have to change their
tracing scripts while profiting from potential future tracepoints
passing slightly more expensive arguments. An example are mempool
tracepoints that pass serialized transactions. We've avoided the
serialization in the past as it was too expensive.
Under the hood, the semaphore-gating works by placing a 2-byte
semaphore in the '.probes' ELF section. The address of the semaphore
is contained in the ELF note providing the tracepoint information
(`readelf -n ./src/bitcoind | grep NT_STAPSDT`). Tracing toolkits
like bpftrace, bcc, and libbpf increase the semaphore at the address
upon attaching to the tracepoint. We only prepare the arguments and
reach the tracepoint if the semaphore is greater than zero. The
semaphore is decreased when detaching from the tracepoint.
This also extends the "Adding a new tracepoint" documentation to
include information about the semaphores and updated step-by-step
instructions on how to add a new tracepoint.
