... unless they're UNCONDITIONAL_RELAY_DELAY old, or there has been
a response to a MEMPOOL request in the mean time.
This is accomplished using a rolling Bloom filter for the last
3500 announced transactions. The probability of seeing more than 100
broadcast events (which can be up to 35 txids each) in 2 minutes for
an outbound peer (where the average frequency is one per minute), is
less than 1 in a million.
This constant is set to 2 minutes, rather than 15. This is still many times
larger than the transaction broadcast interval (2s for outbound, 5s for
inbound), so it should be acceptable for peers to know what our contents of
the mempool was that long ago.
fa8a341b88 wallet: Replace CDataStream& with CDataStream&& where appropriate (MarcoFalke)
fa021e9a5b wallet: Remove confusing double return value ret+success (MarcoFalke)
Pull request description:
The keys and values are only to be used once because their memory is set
to zero. Make that explicit by moving the bytes into the lower level
methods.
ACKs for top commit:
sipa:
utACK fa8a341b88
ryanofsky:
Code review ACK fa8a341b88. Nice changes.
Tree-SHA512: 5c0218bae0f3cd2a07346f1bbf4ad232e5dde7ef2f807d82cc6cfd208d11fe60c8b0f37e7986087b52fbfc79cdfd33c3c8a5822b3d4d9a44d1c6b09e354fc424
e8a2822119 [net] Don't try to take cs_inventory before deleting CNode (John Newbery)
3556227ddd [net] Make cs_inventory a non-recursive mutex (John Newbery)
344e831de5 [net processing] Remove PushBlockInventory and PushBlockHash (John Newbery)
Pull request description:
- Remove PushBlockInventory() and PushBlockHash(). These are one-line functions that can easy be inlined into the calling code. Doing so also allows us to eliminate the one place that cs_inventory is recursively locked.
- Make cs_inventory a nonrecursive mutex
- Remove a redundant TRY_LOCK of cs_inventory when deleting CNode.
ACKs for top commit:
sipa:
utACK e8a2822119
MarcoFalke:
ACK e8a2822119🍬
hebasto:
re-ACK e8a2822119
Tree-SHA512: dbc721d102cdef7b5827a8f2549daf8b54f543050266999a7ea56c9f36618565b71e31ce0beb1209ba2db43d15388be173355a03fb6db8ad24e2475b145050bd
2ad58381ff Clean up separated ban/discourage interface (Pieter Wuille)
b691f2df5f Replace automatic bans with discouragement filter (Pieter Wuille)
Pull request description:
This patch improves performance and resource usage around IP addresses that are banned for misbehavior. They're already not actually banned since #14929, as connections from them are still allowed, but they are preferred for eviction if the inbound connection slots are full.
Stop treating these like manually banned IP ranges, and instead just keep them in a rolling Bloom filter of misbehaving nodes, which isn't persisted to disk or exposed through the ban framework. The effect remains the same: preferred for eviction, avoided for outgoing connections, and not relayed to other peers.
Also change the name of this mechanism to "discouraged" to better reflect reality.
ACKs for top commit:
naumenkogs:
utACK 2ad58381ff
amitiuttarwar:
code review ACK 2ad58381ff
jonatack:
ACK 2ad5838 per changes since last review `git range-diff 3276c14 1f7e0ca 2ad5838`
jnewbery:
Code review ACK 2ad58381ff
Tree-SHA512: 5dedef401d9cbfa026812651303e6286223563dbeed7a10766ed536ac9e3f29ed4bd0df29cc6deadceeb35cbe9f066346add14ef0833958ca9f93d123fe7aab5
40506bf93f test: Test gettxouttsetinfo hash_type option (Fabian Jahr)
f17a4d1c4d rpc: Add hash_type NONE to gettxoutsetinfo (Fabian Jahr)
a712cf6f68 rpc: gettxoutsetinfo can specify hash_type (only legacy option for now) (Fabian Jahr)
605884ef21 refactor: Extract GetBogoSize function (Fabian Jahr)
Pull request description:
This is another intermediate part of the Coinstats Index (tracked in #18000).
Sjors suggested [here](https://github.com/bitcoin/bitcoin/pull/18000#issuecomment-641423019) that the part of the changes in #19145 that don't rely on the new `hash_type` muhash, i.e. that are for `hash_type=none`, could be merged separately from everything involving muhash. So these changes are extracted from #19145 here and can be merged without any other requirements.
Building the index with no UTXO set hash is still valuable because `gettxoutsetinfo` can still be used to audit the `total_amount` for example. By itself this PR is not a huge improvement, `hash_type=none` is speeding up `gettxoutsetinfo` by about 10%, but it enables the implementation of an index on top of it in a follow-up and that means large parts of the index code of Coinstats Index can be merged while reviews for the hashing algorithm might take longer.
ACKs for top commit:
MarcoFalke:
ACK 40506bf93f 🖨
Sjors:
tACK 40506bf93f
Tree-SHA512: 3964c2b8eed427511b1aa9b2ef285dff27dc4d1537d72c3911e435b6e6b40912232da4acb3a09bd19a0372ddffa44103388d8a650169d95a4a727b970d210add
This commit includes a short comment in doc/fuzzing.md that gives
guidance on compiling Bitcoin Core with AFL instrumentation using
afl-gcc and afl-g++.
a76dafa51d ci: Add tsan suppression for race in BerkeleyBatch (Hennadii Stepanov)
Pull request description:
A temporary workaround for #19448.
Top commit has no ACKs.
Tree-SHA512: 47b83ff373e710bc9ba8c3661f9850a14417436028c42eb7765d21337ef25faaac4cf8cf93be844ae592d40264934d7d2f6b7ba0ab6c7209fc0da8fc13067769
d8e9ca66d1 walletdb: Move Rewrite into BerkeleyDatabase (Andrew Chow)
91d109156d walletdb: Move PeriodicFlush into WalletDatabase (Andrew Chow)
8f1bcf8b7b walletdb: Combine VerifyDatabaseFile and VerifyEnvironment (Andrew Chow)
Pull request description:
The `BerkeleyBatch` class has 4 static functions that operate on `BerkeleyDatabase` or `BerkeleyEnvironment`. It doesn't make sense for these to be standalone nor for them to be static functions. So instead, move them from `BerkeleyBatch` into `BerkeleyDatabase` and make them member functions instead of static.
`BerkeleyBatch::VerifyEnvironment` and `BerkeleyBatch::VerifyDatabaseFile` are combined into a single `BerkeleyDatabase::Verify` function that operates on that `BerkeleyDatabase` object.
`BerkeleyBatch::Rewrite` and `BerkeleyBatch::PeriodicFlush` both took a `BerkeleyDatabase` as an argument and did stuff on it. So we just make it a member function so it doesn't need to take a database as an argument.
Part of #18971
ACKs for top commit:
MarcoFalke:
re-ACK d8e9ca66d1 only change is test fixup 🤞
promag:
Code review ACK d8e9ca66d1, good stuff.
Tree-SHA512: 9847e55b13d98bf4e5636cc14bc3f5351d56737f7e320fafffaed128606240765599e5400382c5aecac06690f7e36265ca3e1031f3f6d8a9688f6d5cb1bacd2a
961e667600 Remove cached directories and associated script blocks from appveyor CI configuration. (Aaron Clauson)
Pull request description:
Appveyor CI jobs have been failing in the last 24 hours due to a seemingly corrupted cache, see #19440.
It's possible that the appveyor cache issue is related to the[ recent update](https://www.appveyor.com/updates/2020/07/03/) of the Visual Studio 2019 image
PR #19431 changes the "save cache or error" to false in an attempt to avoid a failing CI job from potentially corrupting the cache. In theory the only way a PR could affect the cache is if the `vcpkg` install list changed. That happens very rarely and did not happen in the last 24 hours and so was not the cause of the current cache problems.
I have done some testing with appveyor build jobs on my own fork and found that installing the `vcpkg` dependencies from scratch and doing a full build can now be done in just under 60 minutes. This is the first time in over 5 months I have been able to build Bitcoin Core on appveyor. Either the new Visual Studio 2019 image has dramatically reduced the build time or appveyor images have had their CPU increased.
This PR removes all use of dependency caching from the appveyor CI config. The trade-off is the 15 minutes saved on each build from having the dependencies cached versus the hours maintainers need to spend investigating when the CI jobs start failing.
ACKs for top commit:
MarcoFalke:
ACK 961e667600
Tree-SHA512: 788c7efbfe6e044739ec41b08df30e24e26bfe0f31d1f5695e7243222a2eb649a2b5fd0254a9238fd416661dc05f737b0545d39feea7aa0da2236fffd7683a1b
fab80fef61 refactor: Remove unused EnsureChainman (MarcoFalke)
fa34587f1c scripted-diff: Replace EnsureChainman with Assert in unit tests (MarcoFalke)
fa6ef701ad util: Add Assert identity function (MarcoFalke)
fa457fbd33 move-only: Move NDEBUG compile time check to util/check (MarcoFalke)
Pull request description:
The utility function is primarily useful to dereference pointer types, which are known to be not null at that time.
For example, the ArgsManager is known to exist when the wallets are started: https://github.com/bitcoin/bitcoin/pull/18923/files#diff-fdb2a1a1d8bc790fcddeb6cf5a42ac55R503 . Instead of silently relying on that assumption, `Assert` can be used to abort the program and avoid UB should the assumption ever be violated.
ACKs for top commit:
promag:
Tested ACK fab80fef61.
ryanofsky:
Code review ACK fab80fef61
Tree-SHA512: 830fba10152ba17d47c4dd42809c7e26f9fe6d38e17a2d5b3f054fd644a5c4c9841286ac421ec9bb28cea9f5faeb659740fcf00de6cc589d423fee7694c42d16
This patch improves performance and resource usage around IP
addresses that are banned for misbehavior. They're already not
actually banned, as connections from them are still allowed,
but they are preferred for eviction if the inbound connection
slots are full.
Stop treating these like manually banned IP ranges, and instead
just keep them in a rolling Bloom filter of misbehaving nodes,
which isn't persisted to disk or exposed through the ban
framework. The effect remains the same: preferred for eviction,
avoided for outgoing connections, and not relayed to other peers.
Also change the name of this mechanism to better reflect reality;
they're not banned, just discouraged.
Contains release notes and several interface improvements by
John Newbery.
When loading descriptor caches, we would accidentally reinitialize the
descriptor cache when seeing that one already exists. This should have
only been initializing the cache when one does not exist. However this
code itself is unnecessary as the act of looking up the cache to add to
it will initialize it if it didn't already exist.
This issue could be hit by trying to load a wallet that had imported a
multisig descriptor. The wallet would fail to load.
A test has been added to wallet_importdescriptors.py to catch this case.
Another test case has also been added to check that loading a wallet
with only single key descriptors works.
fa8e6df282 ci: Run tsan ci config on cirrus (MarcoFalke)
Pull request description:
Fixesbitcoin-core/gui#12
Copied description from #19321:
Currently it is not possible to use travis in forked repositories due to the 50 minute limit on builds. A fresh build (uncached) of the thread sanitizer config takes more than 50 minutes.
One approach to fix this could be to throw away tests until the run time is less than 50 minutes. However, the risk of being blind of failures in the thrown away tests is not worth the gain. Also, to detect them, one has to run the tsan configuration nightly and failures could only be detected post-merge.
Another approach would be to ask travis support to raise the limit for a forked repository. This is a tedious and manual one-by-one process, so I'd rather not.
Finally, a different ci provider can be used, since the config files are designed to be platform-agnostic. This is what I picked.
I kept all settings identical to the travis machine for now. Both providers run in the google cloud, so this should be a "move-only".
ACKs for top commit:
fanquake:
ACK fa8e6df282 - my understanding is that test coverage remains the same. Just swapping providers to work-around the Travis time-limit in other repos.
Tree-SHA512: 26fed248a4f743107160d3b9e5df57fa0be280fd065ae6fece83d254f59d58ccf3e11a245519d158da109c47b053f62ee8756215008541973c65dc28c4efb748
fa0dfdf447 refactor: Remove confusing BlockIndex global (MarcoFalke)
Pull request description:
The global `::BlockIndex()` is problematic for several reasons:
* It returns a mutable reference to the block tree, without the appropriate lock annotation (`m_block_index` is guarded by `cs_main`). The current code is fine, but in the future this might lead to accidental races and data corruption.
* The rpc server shouldn't rely on node globals, but rather a context that is passed in to the RPC method.
* Tests might want to spin up their own block tree, and thus should also not rely on a single global.
Fix all issues by removing the global
ACKs for top commit:
promag:
Code review ACK fa0dfdf447.
jonatack:
re-ACK fa0dfdf
Tree-SHA512: 8f158fc5e1c67e73588a21c25677b3fa0fe442313b13ec24b87054806c59607d6ba0c062a865ce3e0ee568706bd0d1faa84febda21aff5bcd65dab172f74c52f
a8d39b8840 doc: explain why passing -mlinker-version is required (fanquake)
Pull request description:
I have been down a 🐇 hole. Closes#19359.
When Clang is compiled, [a check is run](https://github.com/llvm/llvm-project/blob/release/8.x/clang/CMakeLists.txt#L353) to define `HOST_LINK_VERSION` as the output of `$CMAKE_LINKER -v`. Note the this is the version of the linker being used to compile Clang itself.. and this check is only run when compiling Clang for macOS.
In the Clang driver, if `HOST_LINK_VERSION` has been defined, there is some additional runtime functionality. An `-mlinker-version` argument, with the value of `HOST_LINK_VERSION` [will be added to the linker arguments](89de0d8dfb/clang/lib/Driver/Driver.cpp (L382)), if `-mlinker-version` has not been passed in by the user.
This is a bit weird, as by default, you are setting `-mlinker-version` to the version of the linker that was used to build the Clang binary, not the linker which will be used when compiling. The commit which introduced the functionality, 628fcf4e3b, described it as a "hack", that should be replaced. However, that was 10 years ago, and the behaviour is still here.
In the Darwin driver, [a check is done](89de0d8dfb/clang/lib/Driver/ToolChains/Darwin.cpp (L208)) for the `-mlinker-version` argument. If there is no argument, the version will default to `0`. Given the above, this should never happen when using Clang for macOS. A series of comparisons are then performed, to check whether the linker version is modern enough to enable certain features, like [`-demangle`](89de0d8dfb/clang/lib/Driver/ToolChains/Darwin.cpp (L215)).
### What this means
#### macOS
A Clang compiled for macOS, i.e `clang+llvm-8.0.0-x86_64-apple-darwin`, will have `HOST_LINKER_VERSION` set to the version of the linker used to compile Clang itself.
At runtime, `-mlinker-version=HOST_LINKER_VERSION` will be added to the linker args, if `-mlinker-version` wasn't passed in. In the Darwin driver, additional arguments, like `-demangle`, will be added to the linker arguments, because `HOST_LINKER_VERSION` was likely some very modern version of `lld` or `ld64`.
#### Linux (cross compilation in depends)
A Clang compiled for Linux, i.e `clang+llvm-8.0.0-x86_64-linux-gnu-ubuntu-14.04`, which we now use for macOS builds in depends, will behave differently. As it's built for Linux, `HOST_LINKER_VERSION` was not defined at compile time, and there will be no default behaviour of appending `-mlinker-version=HOST_LINKER_VERSION` to the linker args. Thus, unless you pass in `-mlinker-version` yourself, when the version checks are done in the Darwin driver, no modern linker features will be enabled, as the version will have defaulted to `0`.
Therefore, it's important that we continue to pass `-mlinker-version="our LD64 version"` as part of our compilation flags, if we want to have "modern" linker features enabled for our macOS builds.
#### Summary
[Clang 8](https://releases.llvm.org/download.html#8.0.0). Building a macOS binary. Link line with path arguments trimmed.
| | default behaviour | `-mlinker-version=100` (`-demangle threshold`) | `-mlinker-version=530` |
| - | --------------- | --------------------- | ---------------------- |
| macOS Clang | `-demangle -lto_library ../libLTO.dylib -no_deduplicate -dynamic -arch x86_64 -macosx_version_min 10.15.0 -o a.out ../test-b8b9b3.o -lc++ -lSystem ../libclang_rt.osx.a` | `-demangle -dynamic -arch x86_64 -macosx_version_min 10.15.0 -o a.out ../test-a66966.o -lc++ -lSystem ../libclang_rt.osx.a` | same as default |
| Linux Clang | `-dynamic -arch x86_64 -macosx_version_min 10.12.0 -o a.out ../test-bfce57.o -lc++ -lSystem` | `-demangle -dynamic -arch x86_64 -macosx_version_min 10.12.0 -o a.out ../test-a846a3.o -lc++ -lSystem` | `-demangle -lto_library ../libLTO.dylib -no_deduplicate -dynamic -arch x86_64 -macosx_version_min 10.12.0 -o a.out ../test-de0280.o -lc++ -lSystem` |
Note: Most links here are pointing to the 8.x branch of LLVM/Clang, as we are using that version in depends.
Note: To add a little more confusion, you wont see `-mlinker-version X` in your compile flags, you'll see [`-target-linker-version X`](431daedee4/clang/lib/Driver/ToolChains/Clang.cpp (L4777)).
ACKs for top commit:
laanwj:
ACK a8d39b8840
Tree-SHA512: 92f93079a5e59a0d561e74336b5cb03e3bf5a34437f5850283b9128c7624494b8285ec16290b1fa8103fe87f8789a53ce44b17902b8c1db5fde24d74b76fb168
92bc268e4a build: Detect missed pkg-config early (Hennadii Stepanov)
1739eb23d8 build: Drop unused use_pkgconfig variable (Hennadii Stepanov)
a661449a2e build: Drop use_pkgconfig check for libmultiprocess check (Hennadii Stepanov)
90b95e7929 build: Drop dead non-pkg-config code for libevent check (Hennadii Stepanov)
44a14afbb8 build: Drop dead non-pkg-config code for qrencode check (Hennadii Stepanov)
10cbae0c39 build: Drop dead non-pkg-config code for ZMQ check (Hennadii Stepanov)
06cfc9cadf build: Fix indentation in UNIVALUE check (Hennadii Stepanov)
6fd2118e77 build: Drop dead non-pkg-config code for UNIVALUE check (Hennadii Stepanov)
e9edbe4dbd build: Always use pkg-config (Hennadii Stepanov)
9e2e753b06 build: Always define ZMQ_STATIC for MinGW (Hennadii Stepanov)
Pull request description:
This PR:
- is based on #18297 (already merged)
- drops all of the non-pkg-config paths from the `configure` script
Ref: #17768
ACKs for top commit:
fanquake:
ACK 92bc268e4a. I re-gitian-built. There are a couple follow-ups that I'll PR shortly. Thanks for addressing my feedback above. I took too long to get back to this.
laanwj:
ACK 92bc268e4a
Tree-SHA512: 83c2d9cf03518867a1ebf7e26a8fc5b6dd8962ef983fe0d84e0c7eb74717f4c36a834da02faf0e503ffd87167005351671cf040c0d4ddae57ee152a6ff84012b
54b5eb2b14 tests: Add std::locale::global to list of locale dependent functions in lint-locale-dependence.sh (practicalswift)
Pull request description:
Add `std::locale::global` to list of locale dependent functions in `lint-locale-dependence.sh`.
We currently flag `setlocale(...)` as locale dependent, but prior to this commit we didn't flag
`std::locale::global(...)` as such.
In addition to setting the global C++ locale `std::locale::global(...)` also does the equivalent of `std::setlocale(LC_ALL, ...);`.
Thus the functionality of `std::locale::global(...)` is a superset of `setlocale(...)` :)
ACKs for top commit:
MarcoFalke:
ACK 54b5eb2b14, fine with me
Tree-SHA512: bcf2f1c765add6ed09c3debca968b75eeea81602503f109c0f76ec98635911d453f4834a39e741703c3d470f123178e8952191a9b1a3429394b99c07765dcf1f
84d295e513 tests: Check that segwit inputs in psbt have both UTXO types (Andrew Chow)
4600479058 psbt: always put a non_witness_utxo and don't remove it (Andrew Chow)
5279d8bc07 psbt: Allow both non_witness_utxo and witness_utxo (Andrew Chow)
72f6bec1da rpc: show both UTXOs in decodepsbt (Andrew Chow)
Pull request description:
Due to recent changes to hardware wallets, the full previous transaction will need to be provided for segwit inputs. Since some software may be checking for the existence of a `witness_utxo` to determine whether to produce a segwit signature, we keep that field to ease the transition.
Because all of the sanity checks implemented by the `IsSane` functions were related to having mixed segwit and non-segwit data in a PSBT, those functions are removed as those checks are no longer proper.
Some tests are updated/removed to accommodate this and a simple test added to check that both UTXOs are being added to segwit inputs.
As discussed in the wallet IRC meeting, our own signer will not require `non_witness_utxo` for segwit inputs.
ACKs for top commit:
Sjors:
utACK 84d295e513 (didn't retest compared to 836d6fc, but fortunately HWI's CI tracks our master branch, with a bunch of hardware wallet simulators)
ryanofsky:
Code review re-ACK 84d295e513. No changes since last review, but now I understand the context better. I think it would good to improve the comments as suggested https://github.com/bitcoin/bitcoin/pull/19215#discussion_r447889473 and maybe refer to
meshcollider:
utACK 84d295e513
Tree-SHA512: ccc1fd3c16ac3859f5aca4fa489bd40f68be0b81bbdc4dd51188bbf28827a8642dc8b605a37318e5f16cf40f1c4910052dace2f27eca21bb58435f02a443e940
870f0cd2a0 build: Add MemorySanitizer (MSan) in Travis to detect use of uninitialized memory (practicalswift)
Pull request description:
Add MemorySanitizer (MSan) in Travis to detect use of uninitialized memory.
First UBSan, then ASan followed by TSan... and now: yes, the wait is over -- **MSan is finally here!** :)
Some historical context:
* 2017: Continuous compilation with Clang Thread Safety analysis enabled (#10866, #10923)
* 2018: Continuous testing with trapping on signed integer overflows (`-ftrapv`) (#12686)
* 2018: Continuous testing of use of locale dependent functions (#13041)
* 2018: Continuous testing of format strings (#13705)
* 2018: Continuous compilation with MSVC `TreatWarningAsError` (#14151)
* 2018: Continuous testing under UndefinedBehaviorSanitizer – UBSan (#14252, #14673, #17006)
* 2018: Continuous testing under AddressSanitizer – ASan (#14794, #17205, #17674)
* 2018: Continuous testing under ThreadSanitizer – TSan (#14829)
* 2019: Continuous testing in an unsigned char environment (`-funsigned-char`) (#15134)
* 2019: Continuous compile-time testing of assumptions we're making (#15391)
* 2019: Continuous testing of fuzz test cases under Valgrind (#17633, #18159, #18166)
* 2020: Finally... MemorySanitizer – MSAN! :)
What is the next step? What tools should we add to CI to keep bugs from entering `master`? :)
ACKs for top commit:
MarcoFalke:
ACK 870f0cd2a0
Tree-SHA512: 38327c8b75679d97d469fe42e704cacd1217447a5a603701dd8a58ee50b3be2c10248f8d68a479ed081c0c4b254589d3081c9183f991640b06ef689061f75578
fa12d8d3ed ci: Add tsan suppression for race in wallet (MarcoFalke)
Pull request description:
Workaround to fix#19417 (Intermittent CI failure)
Top commit has no ACKs.
Tree-SHA512: 2d68783d6db1bf425ce830cb23eab2f7fa3b9ee18cfb08665e4187196af571547206646dc6dfac0b4444e3dc6c4c13ae45efb09607d2d50df20a3d0a4eec98bd
