Commit graph

137 commits

Author SHA1 Message Date
John Tromp
34646ddf51 [T4] Rename all shiftsize / cuckoo_size to edge_bits and change value for T4 (#1752)
* replace all size_shift / cuckoo_size by edge_bits and change some constants for T4
* replace remaining occurrences of sizeshift
2018-10-15 16:14:23 -07:00
Antioch Peverell
86c1d7683b
The Header MMR (One MMR To Rule Them All) (#1716) (#1747)
* header MMR in use within txhashset itself
works with fast sync
not yet in place for initial header sync

* add the (currently unused) sync_head mmr

* use sync MMR during fast sync
rebuild header MMR after we validate full txhashset after download

* support missing header MMR (rebuild as necessary) for legacy nodes

* rename to HashOnly

* cleanup backend.append()

* simplify vec_backend to match simpler append api
2018-10-15 19:24:01 +01:00
Ignotus Peverell
43f4f92730
[T4] Secondary proof of work difficulty adjustments (#1709)
* First pass at secondary proof of work difficulty adjustments
* Core and chain test fixes
* Next difficulty calc now needs a height. Scaling calculation fixes. Setting scaling on mined block.
* Change factor to u32 instead of u64.
* Cleanup structs used by next_difficulty
* Fix header size calc with u32 scaling
2018-10-13 13:57:01 -07:00
Antioch Peverell
4dff68ab61 Remove kernel sum header (#1723)
We maintain this locally in block_sums in the db (sums can be rebuilt from local data)
2018-10-12 18:15:29 -07:00
Ignotus Peverell
805cc24e73
Removed kernel sum from header. Fixes #1568 2018-10-13 00:57:08 +00:00
Yeastplume
92f826a917
[T4 ONLY] T4 PoW changes (#1663)
* T4 PoW changes

* rustfmt

* adjust sizeshift depending on pow type during ser/deser

* update block size tests for cuckatoo sizeshift
2018-10-10 10:09:44 +01:00
Gary Yu
fa656ca44a
chain difficulty_iter don't need a WriteTransaction batch (#1710) 2018-10-10 16:49:07 +08:00
Antioch Peverell
770aca9a2d
check_known_mmr is the source of our deadlock - blitzing it for now (#1691) 2018-10-08 16:44:05 +01:00
Antioch Peverell
bcf41438dc
Simplify chain, remove head. (#1657)
* Go back to db for head and header_head
* Use batch consistently for db access
* Pass active txhashset and batch in the ctx
* Only update head in db if total work increases
* Only updated header_head if total work (on header chain) increases
2018-10-05 08:29:33 +01:00
Antioch Peverell
0635945740
cleanup verifier_cache (add it to block ctx) (#1638)
rename process_block_no_orphans -> process_block_single
2018-10-02 16:13:02 +01:00
Antioch Peverell
6d67cbdecb perf: update sync_head once per batch of headers (#1629)
perf: update sync_head once per batch of headers
2018-10-02 11:13:26 +08:00
Gary Yu
d8ca684260
fix: in case of all known, update header_head and sync_head (#1619)
* fix: in case of all known for Headers received, update sync_head to the last header

* refactor: push this logic down into pipe
2018-10-01 21:58:50 +08:00
Antioch Peverell
73ddd1d01d
no need to pass sync_head around in ctx (#1608) 2018-09-29 16:19:19 +01:00
Antioch Peverell
fd8c03679a
use sync_head consistently in sync_block_headers (#1604)
* use sync_head consistently in sync_block_headers

* rustfmt
2018-09-29 08:47:50 +01:00
Antioch Peverell
2cad812b29
use an lru_cache for the block_hashes_cache (#1602) 2018-09-28 17:27:31 +01:00
Yeastplume
e64f4fbcd1
Rust Cuckatoo for verifier and test-miner (#1558)
* cuck placeholder

* rustfmt

* cuckatoo, early days

* rustfmt

* data structures are in place, siphash key creation is consistent with @tromp

* solver in place, (not yet working)

* cuckatoo test solver working with test nonce

* rustfmt

* update solver to remove adjacency list removals

* verifier functioning

* rustfmt

* Proper error handing in Cuckatoo module, couple of tests

* modify cuckoo/cuckatoo solvers and verifiers to function identically, in advance of trait refactoring

* rustfmt

* refactor PoW context into trait, default to using cuckoo context

* rustfmt

* create macros for integer casting/unwraps

* don't instantiate structs when just verifying, add test validation vector for cuckatoo 29

* rustfmt

* don't init cuckoo structs if just validating

* test fix

* ensure BH hashing for POW is only done within miner/validators
2018-09-28 11:53:14 +01:00
Antioch Peverell
5a83989cf6
explicitly handle the header when processing blocks (#1600) 2018-09-27 15:12:08 +01:00
Antioch Peverell
4b0fdc2499
rework to only pass a single block_context to pipeline (#1599) 2018-09-27 11:44:50 +01:00
Antioch Peverell
56f84cc2f4
pass batch to process_block and process_block_header (#1594)
* use a batch consistently when processing blocks and headers

* rustfmt
2018-09-27 09:35:25 +01:00
Antioch Peverell
acec59e249
introduce rewindable_kernel_view (readonly but rewindable kernel MMR view) (#1587)
* introduce rewindable_kernel_view
cleanup header in extension

* cleanup and docs/comments

* txhashset does not need to be mutable here

* pull validate_kernel_history out into fn
2018-09-26 09:59:00 +01:00
Antioch Peverell
11f2d7b6d4
UTXOView (readonly, minimal, output only txhashset extension) (#1584)
* make the utxo view explicit

* introduce utxo_view (readonly, minimal txhashset extension)

* rustfmt

* cleanup

* cleanup

* rustfmt

* cleanup build warnings, comments etc.

* rustfmt

* utxo_view even more readonly now

* cleanup

* refactor pmmr, split out readonly_pmmr and backend

* rustfmt
2018-09-25 11:01:19 +01:00
Antioch Peverell
e72d8b58e4
[WIP] txpool (tx validation) using block sums for full validation (#1567)
tx validation (txpool) rework/simplify
2018-09-24 09:24:10 +01:00
Antioch Peverell
f042f67fcd
Reintroduce block sums, verify full kernel sums per block (#1559)
* block_sums and full kernel sum verification

* rustfmt

* add docs/comments

* docs

* rustfmt

* comment on fact total_kernel_sum is redundant now

* make sure we setup block_sums correctly on a fork

* rustfmt

* replace those asserts with errors

* rustfmt
2018-09-20 09:19:32 +01:00
Gary Yu
e1c8dc5a3a
fix: reset sync_head to header_head on initial transition to HeaderSync (#1531)
* fix for bug #1524, reset sync_head to header_head on initial transition to HeaderSync.
* using @antiochp fix solution in #1539
2018-09-20 07:29:24 +08:00
Ignotus Peverell
9cc7aed713
Dual proof of work, first step (#1534)
* Move Proof and Difficulty types to pow module. Difficulty scaling calculation.
* Diffculty scaling and size shift for 2nd PoW
* Backport e7eb26ee
* Test compilation fixes
* Scaling only need to go one way, as @tromp pointed out
2018-09-18 15:12:57 -07:00
Ignotus Peverell
ecf20602d5
Hard fork tryout: better PoW encapsulation in block header (#1478)
* Improve encapsulation with ProofOfWork struct
* Add dual pow scaling factor, fix test
* Fix pre_pow serialization, chain tests
* Adjust header serialized size calc
* Hard fork handling, version-based serialization
2018-09-10 15:36:57 -07:00
Antioch Peverell
77765796ab
improve "check known" steps during process_block() (#1475)
* document what we do during pipe::apply_block()

* rustfmt

* wip

* rustfmt

* wip

* additional check_known_store and check_known_mmr checks in process_block

* rustfmt

* cleanup coinbase maturity check in process_block

* consolidate the "check in store" logic
add TODOs around the 50 block OldBlock logic

* rustfmt

* cleanup
2018-09-05 10:51:29 +01:00
Antioch Peverell
b06b112afb
document/comment pipe::process_block() (#1472)
* document what we do during pipe::apply_block()

* rustfmt
2018-09-03 16:55:09 +01:00
Antioch Peverell
eae0ab6b2a
take write lock on txhashet earlier when processing blocks (#1456)
* take write lock on txhashet earlier when processing blocks

* no txhashset lock in process_block_header
2018-09-03 11:09:53 +01:00
Antioch Peverell
939d391e0e
cache rangeproof and kernel signature verification results (#1419)
* wip - cache rangeproof verification

* rustfmt

* rustfmt

* rename to ok_verifier

* rustfmt

* wire in caching verifier in more places
skip rangeproof and kernel signature verification during deserialization

* rustfmt

* actually cache verification results via ok verifier

* rustfmt

* cleanup

* pass a cache around, not a verifier

* rustfmt

* cleanup trait shenanigans

* rustfmt

* core tests passing

* rustfmt

* rustfmt

* tests passing

* rustfmt

* verifier cache now takes vecs of data in and out

* rustfmt

* logging + rustfmt

* add docs and comments
2018-08-30 15:44:34 +01:00
Antioch Peverell
bb7bc2c4f2
Refactor rewind single header (#1390)
* refactor txhashset to only take a single header
no need to pass in current head_header each time

* rustfmt

* fix bug introduced in rewind with LMDB

* rustfmt

* add some comments
2018-08-20 20:34:12 +01:00
hashmap
99a66c1960
Consolidate validation in Block and Transaction (#1354)
* Consolidate validation in Block and Transaction

Introduce TransactionBody which is included into block and tx.
Fixes #1333
2018-08-15 23:14:48 +02:00
antiochp
bcebe7b5b3
run rustfmt against chain crate
suspect it has not been running successfully for a while...
2018-08-10 14:56:35 +01:00
Gary Yu
b040aaa434 Fix: old time crate use 0..11 as the month, make mistake when switching to chrono Utc.ymd (#1304) (#1305) 2018-07-31 09:35:57 -07:00
Gary Yu
76f4915bf9 improve: time crate flagged as deprecated, switch to actively maintained chrono crate (#1297)
* improve: time crate flagged as deprecated, switch to actively maintained chrono crate

* improve: complete the switching from deprecated time crate to chrono crate, for all the tests part

* improve: complete switching to chrono crate, for the left tests in 'chain' and 'wallet'
2018-07-30 09:33:28 +01:00
Gary Yu
724b19e648 process_block check_known() check orphans also, to avoid double-processing (#1287) 2018-07-23 21:29:31 -07:00
Ignotus Peverell
f5b03a6d5a
Fix sync hash cache (#1285)
* Refactor hash cache handling to add properly. Only add to cache if the block is definitevely accepted or rejected.
* Dedup of header processing now looks like premature optimization. Very fast anyway. Removing the header hash cache.
2018-07-22 12:43:55 -07:00
Gary Yu
75b72e48ff Fix bug on block sync (#1271) (#1281)
Do not add orphans to the "already seen" cache.
2018-07-22 12:25:56 -07:00
Ignotus Peverell
3d89e86906
Various improvements and cleanup in chain crate (#1265)
* Remove now unnecessary txhashset write lock
* Ring buffer of hashes the chain has already processed
* Specifically report too old blocks as peer should be banned
* Move sync check for block relay, clean TODO
* No use processing transactions when syncing
* Ignore blocks older than horizon in pruning nodes
2018-07-16 21:58:56 +01:00
Antioch Peverell
5c142864ff
txhashset extension error was being silently dropped (#1228) 2018-07-08 12:42:21 +01:00
Ignotus Peverell
26d41fefd0
Fix kernel history verification (#1218)
Only rewind kernels to avoid requiring proper bitmap extraction. Also avoids maintaining bitmap data for kernels by introducing a "prunable" flag on PMMR backend.
2018-07-05 03:31:08 +01:00
Yeastplume
d2a84b7600
Error handling improvements (particularly in chain) (#1208)
* update error handling in chain and other modules to use error/errorkind

* sizeshift errorkind
2018-06-30 23:36:38 +01:00
Ignotus Peverell
1398e0bf85
Accept multiple Cuckoo graph sizes (#1193)
* Added sizeshift to proof of work and encoding as VLQ
* Fix some loose ends to get things to compile
* Rename sizehift to min_sizeshift
  * Pipeline checks for PoW Cuckoo Cycle size shift and adjustment
factor before difficulty comparison.
  * Working delta+VLQ encoding of u64 PoW, unfortunately still a
little larger than u32, at least for Cuckoo30.
* Changed binary encoding of proof of work to a sequence of exact
compact bit representation of each nonce. Somewhat simpler,
shorter and matches the data to be hashed.
* Few fixes based on @tromp feedback:
* Max nonce is 2^(N-1)
* Need a separate constant for reference sizeshift
* Cuckoo implementation now conserves sizeshift. Test providing a
block at higher sizeshift.
* Last small overflow protection
2018-06-29 18:41:28 +01:00
yeastplume
94322c796c rustfmt 2018-06-22 11:45:25 +01:00
yeastplume
8dfc7f3b2f rework block input bitmap logic to avoid batch locks 2018-06-22 11:44:50 +01:00
Ignotus Peverell
d189770080 Wallet LMDB backend (#1151)
* Migrate main node store to LMDB

In preparation to using LMDB as a wallet database, migrate the
node db. There's no point in having 2 key-value stores.
In addition LMDB provides a few advantages as a node db, namely a
much faster build (compared to RocksDb), lesser dependencies and
transactions.

* Migrated p2p store to lmdb, stuff compiles

* More fixes, chain tests starting to pass

* Fixed txhashset rollback messing with block save and general batch delimitation. Chain tests passing.

* rustfmt

* LMDB max map size of 10MB isn't really workable. Half TB seems reasonable.

* Fix wallet tests

* Rather crucial commit was missing

* rustfmt

* Fixing new merged tests following lmdb changes

* rustfmt

* * Make txhashset validation read-only on fast sync to avoid having
a really long open transaction.
* Fix deadlock in new block processing, batch should always be
created within a txhashset lock (when they interact).

* Comment about batch and txhashset interlacing

* Fix store tests to use batch

* Externalize wallet config and seed

* Converted direct read access to file outputs map to an iterator

* Cleaned up and simplified wallet Backend trait:

* No more direct mutable access to internal structures (HashMap)
* Batch interface for all writes
* Remove unneeded read wrapper (read_wallet)

* rustfmt

* First (incomplete) pass at wallet LMDB backend

* Progressing on lmdb backent iml

* Added batch impl for LMDB wallet backend. Pretty much done with it, but not sure how to deal with commit (owned).

* rustfmt

* Wrapping LMDB batch around a refcell to work around borrow rules

* Compilation up to grin chain
2018-06-22 09:08:06 +01:00
Ignotus Peverell
0967a5302b
Kernel sum and MMR sizes in block header (#1163)
* Add kernel commitments sum and kernel and output MMR sizes to block header
* Sum a block without including previous sums, cleanup. Blocks are now summed and validated based on their own totals and not the totals since genesis. This allows to get rid of BlockSum and simplified the setting of a new block's roots, kernel sum and MMR sizes. Fixes #116
* Additional kernel MMR validation to check all prior header roots successively
* Wallet tests fix
2018-06-21 02:30:22 +01:00
Antioch Peverell
028b14d9d9
Replace rm_log with leaf_set (compact bitmap) (#1147)
* only store leaves in rm_log

* cleanup

* commit

* add failing test to cover case where we compact
an already compacted data file

* fix the logic for pruning the data file

* rm_log only contains leaves
prunelist only contains non-leaf subtree roots

* cleanup

* commit

* bitmap impl running in parallel

* rustfmt

* this is better - rewind unremoves outpu pos spent by rewound inputs

* commit

* commit

* pass bitmap around when rewinding

* store and retrieve input bitmap per block from the db

* Replace the RemoveLog with the UTXO set.

* utxo set starting to pass tests

* stuff works...

* split store types out
added some tests to compare the performance of the rm_log and the proposed utxo_set

* pull prune_list out into standalone file

* cleanup, get rid of unused height param

* cleanup and add doc comments

* add support for migrating rm_log to utxo_set

* take snapshot of utxo file during fast sync
implement migration of rm_log -> utxo_set

* rename rewound_copy to snapshot

* fixup pmmr tests to reflect cutoff_pos

* cleanup unused import

* check_compact needs to rewind the utxo_set as appropriate

* fix pool tests

* fixup core tests

* cache block_input_bitmaps via LruCache in store

* cache block header on initial write to db

* rename utxo_set -> leaf_set
and remove references to "spent" in grin_store

* better document the rewind behavior
2018-06-18 11:18:38 -04:00
Quentin Le Sceller
a30ee88236
Cleanup imports (#1161)
Reorganize imports
2018-06-14 08:16:14 -04:00
hashmap
2fa32d15ce Rustify core/src/core (#1122)
Small refactoring of one folder, if it makes sense I could extend the scope.
* Remove some cloning (real and just verbosity in the code)
* Naming conventions like to/into*
* Some Clippy's suggestions

I found that we don't use field init shorthand syntax, so I didn't touch this part, was it discussed before?
2018-06-01 20:41:26 +01:00