* OnDemand no longer loop until there is a query.
All peer known at the time will be queried, and the query fail if all
return no reply.
Returning the failure is done through an empty Vec of reply (the type
of the oneshot channel remains unchanged).
Before this commit the query were send randomly to any peer until there
is a reply (for a query that got no result it was an issue, for other
queries it was quering multiple times the same peers).
After this commit the first query is random but next queries
follows hashmap iterator order.
Test no_capability was broken by this commit (the pending query was
removed).
* OnDemand no longer loop until there is a query.
All peer known at the time will be queried, and the query fail if all
return no reply.
Returning the failure is done through an empty Vec of reply (the type
of the oneshot channel remains unchanged).
Before this commit the query were send randomly to any peer until there
is a reply (for a query that got no result it was an issue, for other
queries it was quering multiple times the same peers).
After this commit the first query is random but next queries
follows hashmap iterator order.
Test no_capability was broken by this commit (the pending query was
removed). If adding some kind of timeout mechanism it could be restored.
* Comment plus better field names.
* No panick on dropped oneshot channel.
* Use Set to avoid counter heuristic
* Cli option `on_demand_nb_retry` for maximum number of retry when doing
on demand query in light client.
* Missing test update for previous commit
* Add a timeout (only when there is no peer to query), that way we do not
set number of query to minimum current number peer or configured number
of query : that way capability test was restored.
* Adding an error type for on_demand, it helps having variant of error
reported at rpc level : choice of rpc error code error might not be
right.
* Duration as constant is nice
* Switch to duration in main too
* Fix indentation (sorry for that).
* Fix error management (bad merge in previous commit)
* Lots of english corrections, major change on the new command parameters :
- use standard '-' instead of '_'
- renaming nb_retry params to 'on-demand-retry-count'
This PR is fixing deadlock for #8918
It avoids some recursive calls on light_sync by making state check optional for Informant.
The current behavior is to display the information when informant checks if block is major version.
This change a bit the informant behavior, but not on most cases.
To remember where and how this kind of deadlock are likely to happen (not seen with Parkinglot deadlock detection because it uses std condvar), I am adding a description of the deadlock.
Also, for the reviewers there may be better solution than modifying the informant.
### Thread1
- ethcore/sync/light_sync/mod.rs
A call to the light handler through any Io (having a loop of rpc query running on like client makes the dead lock way more likely).
At the end of those calls we systematically call `maintain_sync` method.
Here maintain_sync locks `state` (it is the deadlock cause), with a write purpose
`maintain_sync` -> `begin_search` with the state locked open
`begin_search` -> lightcliennt `flush_queue` method
- ethcore/light/src/client/mod.rs
`flush_queue` -> `flush` on queue (HeaderQueue aka VerificationQueue of headers)
- ethcore/src/verification/queue/mod.rs
Condition there is some unverified or verifying content
`flush` wait on a condvar until the queue is empty. The only way to unlock condvar is that worker is empty and unlock it (so thread 2 is Verification worker).
### Thread2
A verification worker at the end of a verify loop (new block).
- ethcore/src/verification/queue/mod.rs
thread loops on `verify` method.
End of loop condition is_ready -> Import the block immediately
calls `set_sync` on QueueSignal which send a BlockVerified ClientIoMessage in inner channel (IoChannel of ClientIoMessage) using `send_sync`
- util/io/src/service_mio.rs
IoChannel `send_sync` method calls all handlers with `message` method; one of the handlers is ImportBlocks IoHandler (with a single inner Client service field)
- ethcore/light/src/client/service.rs
`message` trigger inner method `import_verified`
- core/light/src/client/mod.rs
`import_verified` at the very end notify the listeners of a new_headers, one of the listeners is Informant `listener` method
- parity/informant.rs
`newHeaders` run up to call to `is_major_importing` on its target (again clinet)
- ethcore/sync/src/light_sync/mod.rs
Here `is_major_importing` tries to get state lock (read purpose only) but cannot because of previous state lock, thus deadlock
* Import the `home` crate in `util/dir`.
* Replace uses of `env::home_dir()` with `home::home_dir()`.
* `home` uses a 'correct' impl. on windows and the stdlib impl.
of `::home_dir` otherwise.
* Reexport `home::home_dir` from `util/dir`.
* Bump `util/dir` to 0.1.2.
* Added --tx-queue-no-early-reject flag to disable early tx queue rejects because of low gas price
* Fixed failing tests, clarified comments and simplified no_early_reject field name.
* Added test case for the --tx-queue-no-early-reject flag
* parity: fix UserDefaults json parser
* parity: use serde_derive for UserDefaults
* parity: support deserialization of old UserDefault json format
* parity: make UserDefaults serde backwards compatible
* parity: tabify indentation in UserDefaults
* Revert "Replace `std::env::home_dir` with `dirs::home_dir` (#9077)"
This reverts commit 7e779327eb.
* Restore some of the changes
* Update parity-common
* Remove the dapps system from Parity
* Move node-health outside of dapps
* Fix set dapps list test
* Update Cargo.lock
* Deprecate options
* Add _legacy_ prefixes in Dapps
* Fix tests
* Fix deprecatedness of dapps-path
* getting started
* refactor main
* unwrap_or -> unwrap_or_else
* force parity to lower version number to trigger update
* Fix typos
* formating
* some minor refactoring
* enable lints and fix some warnings
* make it compile
* minor tweaks to make it work
* address review comments
* Rename exe to exe_path and minor import changes
* updater: unreleased -> unknown
* Add `debug` configuration to force parity-updater
* Introduce a new feature `test-updater` in order conditionally hardcode
the version number in parity in order to force an update
* This should only be used for debug/dev purposes
* nits
* Pulled latest submodule of `wasm-tests`
* CHECK macro to replace hard-coded checks
* Clean up deprecated options
* typo: FlAG -> FLAG
* Fix tests
* Deprecated non-used ui params default is None
* test: arg_ui_path is not deprecated
* arg_ui_port should be None
* Address grumbles
* Fix tests
* Make work-notify an optional feature
* More optional ethcore features: price-info, stratum
* Make ethcore features part of dependency instead of local features
* Put cfg gate in right location
* Feature gate register function on work-notify
* Store recently rejected transactions.
* Don't cache AlreadyImported rejections.
* Make the size of transaction verification queue dependent on pool size.
* Add a test for recently rejected.
* Fix logging for recently rejected.
* Make rejection cache smaller.
* obsolete test removed
* obsolete test removed
* Construct cache with_capacity.
* Handle graceful shutdown with unwinding
* Fix a race condition
* Avoid double exit deadlock
* typo: fix docs
* Fix ethkey cli compilation
* Fix all other cases panic_hook::set -> panic_hook::set_abort
* struct fields do not need to be public
* Add comments on why exiting AtomicBool is needed
* Add option for user to set max size limit for RPC requests as requested in #8961.
* Add max_payload to tests.
* Change name for max payload option and change value from NUM to MB.
* Fix broken test.
* Fix incorrect indentation.
* new blooms database
* fixed conflict in Cargo.lock
* removed bloomchain
* cleanup in progress
* all tests passing in trace db with new blooms-db
* added trace_blooms to BlockChainDB interface, fixed db flushing
* BlockChainDB no longer exposes RwLock in the interface
* automatically flush blooms-db after every insert
* blooms-db uses io::BufReader to read files, wrap blooms-db into Mutex, cause fs::File is just a shared file handle
* fix json_tests
* blooms-db can filter multiple possibilities at the same time
* removed enum trace/db.rs CacheId
* lint fixes
* fixed tests
* kvdb-rocksdb uses fs-swap crate
* update Cargo.lock
* use fs::rename
* fixed failing test on linux
* fix tests
* use fs_swap
* fixed failing test on linux
* cleanup after swap
* fix tests
* fixed osx permissions
* simplify parity database opening functions
* added migration to blooms-db
* address @niklasad1 grumbles
* fix license and authors field of blooms-db Cargo.toml
* restore blooms-db after snapshot
