a47df13471 [qa] Test disconnect block failure -> shutdown (Suhas Daftuar)
4433ed0f73 [validation] Crash if disconnecting a block fails (Suhas Daftuar)
Pull request description:
If we're unable to disconnect a block during normal operation, then that is a
failure of our local system (such as disk failure) or the chain that we are on
(eg CVE-2018-17144), but cannot be due to failure of the (more work) chain that
we're trying to validate.
We should abort rather than stay on a less work chain.
Fixes#14341.
ACKs for top commit:
practicalswift:
utACK a47df13471
TheBlueMatt:
utACK a47df13471. Didn't bother to review the test in detail, it looked fine. Debated whether invalidateblock should ever crash the node, but *not* crashing in the case of hitting a pruned block (which is the only change here) is clearly better, even if there are other cases I'd argue we should crash in.
ryanofsky:
utACK a47df13471. Only change since last review is new comment.
promag:
ACK a47df1347, it takes awhile to quit (RPC connection timeouts) but that's unrelated - hope to fix that soon.
fanquake:
ACK a47df13471
Tree-SHA512: 4dec8cef6e7dbbe513c138fc5821a7ceab855e603ece3c16185b51a3830ab7ebbc844a28827bf64e75326f45325991dcb672f13bd7baede53304f27289c4af8d
-BEGIN VERIFY SCRIPT-
sed -i 's/inline std::string _(const char\* psz)/inline bilingual_str _(const char\* psz)/' src/util/translation.h
sed -i 's/return G_TRANSLATION_FUN ? (G_TRANSLATION_FUN)(psz) : psz;/return bilingual_str{psz, G_TRANSLATION_FUN ? (G_TRANSLATION_FUN)(psz) : psz};/' src/util/translation.h
sed -i 's/\b_("\([^"]\|\\"\)*")/&.translated/g' $(git grep --files-with-matches '\b_("' src)
echo Hard cases - multiline strings.
sed -i 's/"Visit %s for further information about the software.")/&.translated/g' src/init.cpp
sed -i "s/\"Only rebuild the block database if you are sure that your computer's date and time are correct\")/&.translated/g" src/init.cpp
sed -i 's/" restore from a backup.")/&.translated/g' src/wallet/db.cpp
sed -i 's/" or address book entries might be missing or incorrect.")/&.translated/g' src/wallet/wallet.cpp
echo Special case.
sed -i 's/_(COPYRIGHT_HOLDERS)/&.translated/' src/util/system.cpp test/lint/lint-format-strings.py
-END VERIFY SCRIPT-
50cede3f5a [mempool] Allow one extra single-ancestor transaction per package (Matt Corallo)
Pull request description:
This implements the proposed policy change from [1], which allows
certain classes of contract protocols involving revocation
punishments to use CPFP. Note that some such use-cases may still
want some form of one-deep package relay, though even this alone
may greatly simplify some lightning fee negotiation.
[1] https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2018-November/016518.html
ACKs for top commit:
ajtowns:
ACK 50cede3f5a -- looked over code again, compared with previous commit, compiles, etc.
sdaftuar:
ACK 50cede3f5a
ryanofsky:
utACK 50cede3f5a. Changes since last review: adding EXTRA_DESCENDANT_TX_SIZE_LIMIT constant, changing max ancestor size from 1,000,000 to nLimitAncestorSize constant (101,000), fixing test comment and getting rid of unused test node.
Tree-SHA512: b052c2a0f384855572b4579310131897b612201214b5abbb225167224e4f550049e300b471dbf320928652571e92ca2d650050b7cf39ac92b3bc1d2bcd386c1c
This implements the proposed policy change from [1], which allows
certain classes of contract protocols involving revocation
punishments to use CPFP. Note that some such use-cases may still
want some form of one-deep package relay, though even this alone
may greatly simplify some lightning fee negotiation.
[1] https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2018-November/016518.html
Separate out the management of chain-agnostic block metadata from any given
CChainState instance. This allows us to avoid duplicating data like
`mapBlockIndex` unnecessarily for multiple chainstates.
This also adds a CChainState constructor that accepts and sets m_blockman.
Ultimately this reference will point to a BlockMan instance that
is shared across CChainStates.
This commit can be decomposed into smaller commits if necessary.
f874e14cd3 [build]: check std::system for -[alert|block|wallet]notify (Sjors Provoost)
cc3ad56ff2 [build] MSVC: set HAVE_SYSTEM for desktop apps (Sjors Provoost)
c1c91bb78d [build] detect std::system or ::wsystem (Sjors Provoost)
Pull request description:
Platforms such as iOs and Universal Windows Platform do not support launching a process through system().
ACKs for top commit:
laanwj:
code review ACK f874e14cd3
Tree-SHA512: 16bb4a8fa1896046ccb22a46c8985e1aa45f5b11ecf5539eb2299e9a58f1a5b085c0c12cb6939c7493d93abce7e84fadcbfc73374c887db63da6d00c08aa476d
fa2b083c3f [test] Add test to check mempool consistency in case of reorgs (MarcoFalke)
fabeb1f613 validation: Add missing mempool locks (MarcoFalke)
fa0c9dbf91 txpool: Make nTransactionsUpdated atomic (MarcoFalke)
Pull request description:
Take the mempool read lock during reorgs, so that we don't accidentally read an inconsistent mempool.
ACKs for top commit:
laanwj:
code review ACK fa2b083c3f
ryanofsky:
utACK fa2b083c3f [EDIT: was ~e284e422e75189794e24fe482819d8b1407857c3~, from bad copy and paste]. Changes since last review: rebase after #15976, adding vTxHashes lock annotation, adding new commit dropping mempool lock for nTransactionsUpdated and making it atomic to avoid deadlock between mempool lock and g_best_block_mutex
Tree-SHA512: cfe7777993589087753e000e3736d79d320dca412383fb77b56bef8946a04049722bf888c11b6f722adf677165185c7e58b4a269f7c5fa25e84dda375f6c8a7d
0959d37e3e Don't use global (external) symbols for symbols that are used in only one translation unit (practicalswift)
Pull request description:
Don't use global (external) symbols for symbols that are used in only one translation unit.
Before:
```
$ for SYMBOL in $(nm src/bitcoind | grep -E ' [BD] ' | c++filt | cut -f3- -d' ' | grep -v @ | grep -v : | sort | grep '[a-z]' | sort -u | grep -vE '(^_|typeinfo|vtable)'); do
REFERENCES=$(git grep -lE "([^a-zA-Z]|^)${SYMBOL}([^a-zA-Z]|\$)" -- "*.cpp" "*.h")
N_REFERENCES=$(wc -l <<< "${REFERENCES}")
if [[ ${N_REFERENCES} > 1 ]]; then
continue
fi
echo "Global symbol ${SYMBOL} is used in only one translation unit: ${REFERENCES}"
done
Global symbol g_chainstate is used in only one translation unit: src/validation.cpp
Global symbol g_ui_signals is used in only one translation unit: src/ui_interface.cpp
Global symbol instance_of_cmaincleanup is used in only one translation unit: src/validation.cpp
Global symbol instance_of_cnetcleanup is used in only one translation unit: src/net.cpp
Global symbol instance_of_cnetprocessingcleanup is used in only one translation unit: src/net_processing.cpp
Global symbol pindexBestForkBase is used in only one translation unit: src/validation.cpp
Global symbol pindexBestForkTip is used in only one translation unit: src/validation.cpp
$
```
After:
```
$ for SYMBOL in $(nm src/bitcoind | grep -E ' [BD] ' | c++filt | cut -f3- -d' ' | grep -v @ | grep -v : | sort | grep '[a-z]' | sort -u | grep -vE '(^_|typeinfo|vtable)'); do
REFERENCES=$(git grep -lE "([^a-zA-Z]|^)${SYMBOL}([^a-zA-Z]|\$)" -- "*.cpp" "*.h")
N_REFERENCES=$(wc -l <<< "${REFERENCES}")
if [[ ${N_REFERENCES} > 1 ]]; then
continue
fi
echo "Global symbol ${SYMBOL} is used in only one translation unit: ${REFERENCES}"
done
$
```
♻️ Think about future generations: save the global namespace from unnecessary pollution! ♻️
ACKs for commit 0959d3:
Empact:
ACK 0959d37e3e
MarcoFalke:
ACK 0959d37e3e
hebasto:
ACK 0959d37e3e
promag:
ACK 0959d37.
Tree-SHA512: 722f66bb50450f19b57e8a8fbe949f30cd651eb8564e5787cbb772a539bf3a288c048dc49e655fd73ece6a46f6dafade515ec4004729bf2b3ab83117b7c5d153
At this point there is no reasonable excuse to disable opt-in RBF,
and, unlike when this option was added, there are now significant
issues created when disabling it (in the form of compact block
reconstruction failures). Further, it breaks a lot of modern wallet
behavior.
67f4e9c522 Include core_io.h from core_read.cpp (practicalswift)
eca9767673 Make reasoning about dependencies easier by not including unused dependencies (practicalswift)
Pull request description:
Make reasoning about dependencies easier by not including unused dependencies.
Please note that the removed headers are _not_ "transitively included" by other still included headers. Thus the removals are real.
As an added bonus this change means less work for the preprocessor/compiler. At least 51 393 lines of code no longer needs to be processed:
```
$ git diff -u HEAD~1 | grep -E '^\-#include ' | cut -f2 -d"<" | cut -f1 -d">" | \
sed 's%^%src/%g' | xargs cat | wc -l
51393
```
Note that 51 393 is the lower bound: the real number is likely much higher when taking into account transitively included headers :-)
ACKs for commit 67f4e9:
Tree-SHA512: 0c8868aac59813f099ce53d5307eed7962dd6f2ff3546768ef9e5c4508b87f8210f1a22c7e826c3c06bebbf28bdbfcf1628ed354c2d0fdb9a31a42cefb8fdf13
403e677c9 refactoring: IsInitialBlockDownload -> CChainState (James O'Beirne)
3ccbc376d refactoring: FlushStateToDisk -> CChainState (James O'Beirne)
4d6688603 refactoring: introduce ChainstateActive() (James O'Beirne)
d7c97edee move-only: make the CChainState interface public (James O'Beirne)
Pull request description:
This is part of the [assumeutxo project](https://github.com/bitcoin/bitcoin/projects/11):
Parent PR: #15606
Issue: #15605
Specification: https://github.com/jamesob/assumeutxo-docs/tree/2019-04-proposal/proposal
---
This changeset starts moving functionality intimately related to CChainState into methods. Parameterizing these functions by a particular CChainState is necessary for the use of multiple chainstates simultaneously (e.g. for asynchronous background validation).
In this change, we
- make the CChainState interface public - since other units will start to invoke its methods directly,
- introduce `::ChainstateActive()`, the CChainState equivalent for `::ChainActive()`,
- and move `IsInitialBlockDownload()` and `FlushStateToDisk()` into methods on CChainState.
Independent of assumeutxo, these changes better encapsulate chainstate behavior and allow easier use from a testing context.
There are more methods that we'll move in the future, but they require other substantial changes (i.e. moving ownership of the `CCoinsView*` hierarchy into CChainState) so we'll save them for future PRs.
---
The first move-only commit is most easily reviewed with `git diff ... --color-moved=dimmed_zebra`.
ACKs for commit 403e67:
Empact:
utACK 403e677c9e no need to address my nits herein
Sjors:
utACK 403e677
ryanofsky:
utACK 403e677c9e. Only change since previous review is removing global state comment as suggested.
MarcoFalke:
utACK 403e677c9e, though the diff still seems a bit bloated with some unnecessary changes in the second commit.
promag:
utACK 403e677 and rebased with current [master](c7cfd20a7).
Tree-SHA512: 6fcf260bb2dc201361170c0b4547405366f5f331fcc3a2bac29b24442814b7b244ca1b58aac5af716885f9a130c343b544590dff780da0bf835c7c5b3ccb2257
If we're unable to disconnect a block during normal operation, then that is a
failure of our local system (such as disk failure) or the chain that we are on
(eg CVE-2018-17144), but cannot be due to failure of the (more work) chain that
we're trying to validate.
We should abort rather than stay on a less work chain.
1609809fb2 validation: Hold cs_main when reading chainActive in RewindBlockIndex (practicalswift)
Pull request description:
Fixes#15980.
Hold `cs_main` when reading `chainActive` (via `::ChainActive()`) in `RewindBlockIndex`.
ACKs for commit 160980:
MarcoFalke:
utACK 1609809fb2
Tree-SHA512: 54f180ab391f92f04950735c2bb337f0b7495826d2096f7a0f9a2da50bc29d08747f404a0495e33ca4edd4c842efbab4c4730d5e1a8b9da3e1249cf884268f4b
fa86c8aec6 init: Remove dead code in LoadChainTip (MarcoFalke)
Pull request description:
`LoadChainTip` sets `::ChainActive()` based on `pcoinsTip`'s best block. `LoadChainTip` is never called when that block is null, so we can remove all code from within that method that is only executed when that block is null.
Fixes#15967 Inconsistent locking behavior in LoadChainTip
ACKs for commit fa86c8:
promag:
utACK fa86c8aec6.
practicalswift:
utACK fa86c8aec6
Empact:
utACK fa86c8aec6
laanwj:
utACK fa86c8aec6
ryanofsky:
utACK fa86c8aec6. LoadChainTip isn't called currently when pcoinsTip best block is null due to this line:
jamesob:
utACK fa86c8aec6
Tree-SHA512: 8961c0e579800a52038ac5655478468852faac055299b64d6cfdf0c213d3bf09669c4889467d09d93457f6c8b073967bb0475a137f77ddd3a3a3c03ad90001c4
Also renames global methods for clarity:
- ::FlushStateToDisk() -> CChainState::ForceFlushStateToDisk()
- This performs an unconditional flush.
- ::PruneAndFlush() -> CChainState::PruneAndFlush()
along with DisconnectResult, and CBlockIndexWorkComparator.
The CChainState interface needs to be known to the rest of the system because
many global functions will move to CChainState methods. This is to allow
other parts of the system to be parameterized per chainstate instance
instead of assuming a single global.
62d50ef308 Add LOCKS_EXCLUDED(cs_main) to LimitValidationInterfaceQueue(...) which does AssertLockNotHeld(cs_main) (practicalswift)
Pull request description:
This PR adds compile-time checking for negative locking requirements that follow from the run-time locking requirement `AssertLockNotHeld(cs_main)` in `LimitValidationInterfaceQueue(...)`.
Changes:
* Add `LOCKS_EXCLUDED(cs_main)` to `LimitValidationInterfaceQueue(...)` which does `AssertLockNotHeld(cs_main)`
* Add `LOCKS_EXCLUDED(cs_main)` to `CChainState::ActivateBestChain(…)`, `CChainState:: InvalidateBlock(…)` and `CChainState::RewindBlockIndex(…)` which all call `LimitValidationInterfaceQueue(...)` which does `AssertLockNotHeld(cs_main)`
* Add `LOCKS_EXCLUDED(cs_main)` to `InvalidateBlock(…)` which calls `CChainState::InvalidateBlock(...)` which in turn calls `LimitValidationInterfaceQueue(...)` which does `AssertLockNotHeld(cs_main)`
* Add `LOCKS_EXCLUDED(cs_main)` to `RewindBlockIndex(…)` which calls `CChainState::RewindBlockIndex(...)` which in turn calls `LimitValidationInterfaceQueue(...)` which does `AssertLockNotHeld(cs_main)`
ACKs for commit 62d50e:
MarcoFalke:
utACK 62d50ef308
Tree-SHA512: 73d092ccd08c851ae3c5d60370c369fc030c5793f5507e2faccb6f91c851ddc0ce059fbea3899f2856330d7a8c78f2ac6a2988e8268b03154f946be9e60e3be1
Add LOCKS_EXCLUDED(cs_main) to functions calling LimitValidationInterfaceQueue(...) which does AssertLockNotHeld(cs_main)
Add LOCKS_EXCLUDED(cs_main) to functions calling CChainState::InvalidateBlock(...) which calls LimitValidationInterfaceQueue(...) which in turn does AssertLockNotHeld(cs_main)
Add LOCKS_EXCLUDED(cs_main) to functions calling CChainState::RewindBlockIndex(...) which calls LimitValidationInterfaceQueue(...) which in turn does AssertLockNotHeld(cs_main)
486c1eea86 refactoring: remove unused chainActive (James O'Beirne)
631940aab2 scripted-diff: replace chainActive -> ::ChainActive() (James O'Beirne)
a3a609079c refactoring: introduce unused ChainActive() (James O'Beirne)
1b6e6fcfd2 rename: CChainState.chainActive -> m_chain (James O'Beirne)
Pull request description:
This is part of the assumeutxo project:
Parent PR: #15606
Issue: #15605
Specification: https://github.com/jamesob/assumeutxo-docs/tree/2019-04-proposal/proposal
---
This change refactors the `chainActive` reference into a `::ChainActive()` call. It also distinguishes `CChainState`'s `CChain` data member as `m_chain` instead of the current `chainActive`, which makes it easily confused with the global data.
The active chain must be obtained via function because its reference will be swapped at some point during runtime after loading a UTXO snapshot.
This change, though lengthy, should be pretty easy to review since most of it is contained within a scripted-diff. Once merged, the parent PR should be easier to review.
ACKs for commit 486c1e:
Sjors:
utACK 486c1ee
promag:
utACK 486c1ee.
practicalswift:
utACK 486c1eea86
Tree-SHA512: 06ed8f9e77f2d25fc9bea0ba86436d80dbbce90a1e8be23e37ec4eeb26060483e60b4a5c4fba679cb1867f61e3921c24abeb9cabdfb4d0a9b1c4ddd77b17456a
0ff1c2a838 Separate reason for premature spends (coinbase/locktime) (Suhas Daftuar)
54470e767b Assert validation reasons are contextually correct (Suhas Daftuar)
2120c31521 [refactor] Update some comments in validation.cpp as we arent doing DoS there (Matt Corallo)
12dbdd7a41 [refactor] Drop unused state.DoS(), state.GetDoS(), state.CorruptionPossible() (Matt Corallo)
aa502b88d1 scripted-diff: Remove DoS calls to CValidationState (Matt Corallo)
7721ad64f4 [refactor] Prep for scripted-diff by removing some \ns which annoy sed. (Matt Corallo)
5e78c5734b Allow use of state.Invalid() for all reasons (Matt Corallo)
6b34bc6b6f Fix handling of invalid headers (Suhas Daftuar)
ef54b486d5 [refactor] Use Reasons directly instead of DoS codes (Matt Corallo)
9ab2a0412e CorruptionPossible -> BLOCK_MUTATED (Matt Corallo)
6e55b292b0 CorruptionPossible -> TX_WITNESS_MUTATED (Matt Corallo)
7df16e70e6 LookupBlockIndex -> CACHED_INVALID (Matt Corallo)
c8b0d22698 [refactor] Drop redundant nDoS, corruptionPossible, SetCorruptionPossible (Matt Corallo)
34477ccd39 [refactor] Add useful-for-dos "reason" field to CValidationState (Matt Corallo)
6a7f8777a0 Ban all peers for all block script failures (Suhas Daftuar)
7b999103e2 Clean up banning levels (Matt Corallo)
b8b4c80146 [refactor] drop IsInvalid(nDoSOut) (Matt Corallo)
8818729013 [refactor] Refactor misbehavior ban decisions to MaybePunishNode() (Matt Corallo)
00e11e61c0 [refactor] rename stateDummy -> orphan_state (Matt Corallo)
f34fa719cf Drop obsolete sigops comment (Matt Corallo)
Pull request description:
This is a rebase of #11639 with some fixes for the last few comments which were not yet addressed.
The original PR text, with some strikethroughs of text that is no longer correct:
> This cleans up an old main-carryover - it made sense that main could decide what DoS scores to assign things because the DoS scores were handled in a different part of main, but now validation is telling net_processing what DoS scores to assign to different things, which is utter nonsense. Instead, we replace CValidationState's nDoS and CorruptionPossible with a general ValidationInvalidReason, which net_processing can handle as it sees fit. I keep the behavior changes here to a minimum, but in the future we can utilize these changes for other smarter behavior, such as disconnecting/preferring to rotate outbound peers based on them providing things which are invalid due to SOFT_FORK because we shouldn't ban for such cases.
>
> This is somewhat complementary with, though obviously conflicts heavily with #11523, which added enums in place of DoS scores, as well as a few other cleanups (which are still relevant).
>
> Compared with previous bans, the following changes are made:
>
> Txn with empty vin/vout or null prevouts move from 10 DoS
> points to 100.
> Loose transactions with a dependency loop now result in a ban
> instead of 10 DoS points.
> ~~BIP68-violation no longer results in a ban as it is SOFT_FORK.~~
> ~~Non-SegWit SigOp violation no longer results in a ban as it
> considers P2SH sigops and is thus SOFT_FORK.~~
> ~~Any script violation in a block no longer results in a ban as
> it may be the result of a SOFT_FORK. This should likely be
> fixed in the future by differentiating between them.~~
> Proof of work failure moves from 50 DoS points to a ban.
> Blocks with timestamps under MTP now result in a ban, blocks
> too far in the future continue to not result in a ban.
> Inclusion of non-final transactions in a block now results in a
> ban instead of 10 DoS points.
Note: The change to ban all peers for consensus violations is actually NOT the change I'd like to make -- I'd prefer to only ban outbound peers in those situations. The current behavior is a bit of a mess, however, and so in the interests of advancing this PR I tried to keep the changes to a minimum. I plan to revisit the behavior in a followup PR.
EDIT: One reviewer suggested I add some additional context for this PR:
> The goal of this work was to make net_processing aware of the actual reasons for validation failures, rather than just deal with opaque numbers instructing it to do something.
>
> In the future, I'd like to make it so that we use more context to decide how to punish a peer. One example is to differentiate inbound and outbound peer misbehaviors. Another potential example is if we'd treat RECENT_CONSENSUS_CHANGE failures differently (ie after the next consensus change is implemented), and perhaps again we'd want to treat some peers differently than others.
ACKs for commit 0ff1c2:
jnewbery:
utACK 0ff1c2a838
ryanofsky:
utACK 0ff1c2a838. Only change is dropping the first commit (f3883a321bf4ab289edcd9754b12cae3a648b175), and dropping the temporary `assert(level == GetDoS())` that was in 35ee77f2832eaffce30042e00785c310c5540cdc (now c8b0d22698)
Tree-SHA512: e915a411100876398af5463d0a885920e44d473467bb6af991ef2e8f2681db6c1209bb60f848bd154be72d460f039b5653df20a6840352c5f7ea5486d9f777a3
Though at the moment ChainActive() simply references `g_chainstate.m_chain`,
doing this change now clears the way for multiple chainstate usage and allows
us to script the diff.
-BEGIN VERIFY SCRIPT-
git grep -l "chainActive" | grep -E '(h|cpp)$' | xargs sed -i '/chainActive =/b; /extern CChain& chainActive/b; s/\(::\)\{0,1\}chainActive/::ChainActive()/g'
-END VERIFY SCRIPT-
This can't be a scripted-diff due to the confusion of the global
chainActive and the CChainState member of the same name.
This specific rename makes the following chainActive -> ::ChainActive() diff
scriptable.
This is a first step towards cleaning up our DoS interface - make
validation return *why* something is invalid, and let net_processing
figure out what that implies in terms of banning/disconnection/etc.
Behavior change: peers will now be banned for providing blocks
with premature coinbase spends.
Co-authored-by: Anthony Towns <aj@erisian.com.au>
Suhas Daftuar <sdaftuar@gmail.com>
Compared with previous bans, the following changes are made:
* Txn with empty vin/vout or null prevouts move from 10 DoS
points to 100.
* Loose transactions with a dependency loop now result in a ban
instead of 10 DoS points.
* Many pre-segwit soft-fork errors now result in a ban.
Note: Transactions that violate soft-fork script flags since P2SH do not generally
result in a ban. Also, banning behavior for invalid blocks is dependent on
whether the node is validating with multiple script check threads, due to a long-
standing bug. That inconsistency is still present after this commit.
* Proof of work failure moves from 50 DoS points to a ban.
* Blocks with timestamps under MTP now result in a ban, blocks
too far in the future continue to *not* result in a ban.
* Inclusion of non-final transactions in a block now results in a
ban instead of 10 DoS points.
Co-authored-by: Anthony Towns <aj@erisian.com.au>
