In this commit, we add a new config options that allows one to start
`btcd` in an operating mode that disables the stall detection. This can
be useful in simnet/regtest integration tests settings where it's
important that `btcd` holds on to its possibly sole connection to the
only other node in the test harness.
A new config flag has been added to gate this behavior, which is off by
default.
To run integration tests with btcd on Windows in non-interactive
environments (such as the Travis build with Windows machines), we
need to make sure we can still spawn a child process instead of only a
windows background service.
This commit modifies the existing block selection logic to limit
preferentially by weight instead of serialized block size, and also to
adhere to the new sig-op cost limits which are weighted according to
the witness discount.
The btclog package has been changed to defining its own logging
interface (rather than seelog's) and provides a default implementation
for callers to use.
There are two primary advantages to the new logger implementation.
First, all log messages are created before the call returns. Compared
to seelog, this prevents data races when mutable variables are logged.
Second, the new logger does not implement any kind of artifical rate
limiting (what seelog refers to as "adaptive logging"). Log messages
are outputted as soon as possible and the application will appear to
perform much better when watching standard output.
Because log rotation is not a feature of the btclog logging
implementation, it is handled by the main package by importing a file
rotation package that provides an io.Reader interface for creating
output to a rotating file output. The rotator has been configured
with the same defaults that btcd previously used in the seelog config
(10MB file limits with maximum of 3 rolls) but now compresses newly
created roll files. Due to the high compressibility of log text, the
compressed files typically reduce to around 15-30% of the original
10MB file.
This contains a bit of cleanup and additional logic to improve the
recently-added ability to specify additional checkpoints via the
--addcheckpoint option.
In particular:
- Improve error messages in the checkpoint parsing
- Correct the mergeCheckpoints function to weed out duplicate height
checkpoints while using the most-recently provided one as described by
its comment
- Add an assertion to blockchain.New that the provided checkpoints are
sorted as required
- Keep comments to 80 columns and use two spaces after periods in them to
be consistent with the rest of the code base
- Make the entry in doc.go match the actual btcd -h output
If a host is down and doesn't send a TCP RST, the net.Dial function
blocks until the OS times out the connection. Convert to using
DialTimeout with a 30 second default timeout.
This commit modifies the `ConnManager` to use the `net.Add` interface
through the package instead of a plain string to represent and
manipulate addresses. This change makes the package much more general as
users of the package can possibly utilize custom implementations of the
`net.Addr` interface to establish connections.
More precisely, the `ConnReq` struct has been modified to use a net.Addr
instance explicitly, and the `DialFunc` type has also been modified to
take a `net.Addr` directly. This latter change gives functions that
adhere to the `DialFunc` type more flexibility as to exactly how the
connection is established.
Additionally, the `connmgr.Config.GetNewAddress` configuration option
now directly returns a `net.Addr. This change allows the `connmgr` to be
decoupled from all DNS queries which allows callers to preferentially
select more secure methods like performing DNS lookups over a Tor proxy.
The current max orphan transaction size causes problems with dependent
transaction relay due to its artificially small size in relation to the
max standard transaction size.
Consequently, this modifies the orphan transaction policy by increasing
the max size of each orphan to the same value allowed for standard
non-orphan transactions and reducing the default max allowed number of
orphans to 100.
From a memory usage standpoint, the worst case max mem usage prior to
this change was 5MB plus structure and tracking overhead (1000 max
orphans * 5KB max each). With this, that is raised to 10MB (100 max
orphans * 100KB max each) in the worst case.
It is important to note that the values were originally implemented as a
naive means to control the size of the orphan pool before several of the
recent enhancements which more aggressively remove orphans from the
orphan pool were added, so they needed to be evaluated again.
For a very long time prior to recent changes, the orphan pool would
quickly reach the max allowed worst-case usage and effectively stay
there forever whereas with more recent changes, the actual run-time
orphan pool usage is usually much smaller.
Finally, as another point in favor of this change, as the network has
evolved, nodes have generally become better about orphan management and
as such missing ancestors will typically either be broadcast or mined
fairly quickly resulting in fewer overall orphans.
Avoid compatibility issues with software that relies on the behavior of
bitcoind's JSON-RPC implementation.
The JSON-RPC 1.0 spec defines that notifications must have their "id"
set to null and states that notifications do not have a response.
A JSON-RPC 2.0 notification is a request with "json-rpc":"2.0", and
without an "id" member. The specification states that notifications
must not be responded to. JSON-RPC 2.0 permits the null value as a
valid request id, therefore such requests are not notifications.
Bitcoin Core serves requests with "id":null or even an absent "id", and
responds to such requests with "id":null in the response.
Btcd does not respond to any request without and "id" or with "id":null,
regardless the indicated JSON-RPC protocol version.
In order to avoid compatibility issues with software relying on
Core's behavior, this commit implements "quirks mode" as follows:
- quirks mode can be enabled via configuration (disabled by default)
- If no JSON-RPC version is indicated in the request, accept and
respond to request with "id":null
- If no JSON-RPC version is indicated in the request, accept and
respond to requests without an "id" member
- In both cases above, use "id":null in the response
- Do not respond to request without an "id" or with "id":null when
JSON-RPC version is indicated in the request (process as notification)
This commit introduces package connmgr which contains connection
management related functionality.
The following is an overview of the features the package provides:
- Maintain fixed number of outbound connections
- Optional connect-only mode
- Retry persistent connections with increasing back-off
- Source peers from DNS seeds
- Use Tor to resolve DNS
- Dynamic ban scores
- Test coverage
In addition, btcd has been refactored to make use of the new package by
extending the connection manager to work with the server to source and
maintain peer connections. The following is a broad overview of the
changes to integrate the package:
- Simplify peer state by removing pending, retry peers
- Refactor to remove retries which are now handled by connmgr
- Use callback to add addresses sourced from the DNS seed
Finally the following connection-related things have been improved as a
part of this refactor:
- Fixes 100% cpu usage when network is down (#129)
- Fixes issues with max peers (#577)
- Simplify outbound peer connections management
This corrects the case where any errors that might have happened when
creating the config file were not being displayed due to the logging
system not being initialized yet.
While here, consistently use fmt.Fprint{f,ln} throughout the loadConfig
function even after the logging system is initialized since it will
help prevent copy/paste errors such as the one that induced this change
to begin with.
This commit adds two new cli flags: one for accepting non-std
transactions, and the other for rejecting non-std transactions.
The two flag are rejected when using concurrently. Config parsing is
set up such that, the desired policy expressed via the config always
overrides the policy set by default for a particular chain.
The doc.go files and the sample-btcd.conf file have been updated to document
the new flags exposing further policy control.
This does the minimum work necessary to refactor the mempool code into
its own package. The idea is that separating this code into its own
package will greatly improve its testability, allow independent
benchmarking and profiling, and open up some interesting opportunities
for future development related to the memory pool.
There are likely some areas related to policy that could be further
refactored, however it is better to do that in future commits in order
to keep the changeset as small as possible during this refactor.
Overview of the major changes:
- Create the new package
- Move several files into the new package:
- mempool.go -> mempool/mempool.go
- mempoolerror.go -> mempool/error.go
- policy.go -> mempool/policy.go
- policy_test.go -> mempool/policy_test.go
- Update mempool logging to use the new mempool package logger
- Rename mempoolPolicy to Policy (so it's now mempool.Policy)
- Rename mempoolConfig to Config (so it's now mempool.Config)
- Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc)
- Rename txMemPool to TxPool (so it's now mempool.TxPool)
- Move defaultBlockPrioritySize to the new package and export it
- Export DefaultMinRelayTxFee from the mempool package
- Export the CalcPriority function from the mempool package
- Introduce a new RawMempoolVerbose function on the TxPool and update
the RPC server to use it
- Update all references to the mempool to use the package.
- Add a skeleton README.md
Profiles discovered that lookups into the signature cache included an
expensive comparison to the stored `sigInfo` struct. This lookup had the
potential to be more expensive than directly verifying the signature
itself!
In addition, evictions were rather expensive because they involved
reading from /dev/urandom, or equivalent, for each eviction once the
signature cache was full as well as potentially iterating over every
item in the cache in the worst-case.
To remedy this poor performance several changes have been made:
* Change the lookup key to the fixed sized 32-byte signature hash
* Perform a full equality check only if there is a cache hit which
results in a significant speed up for both insertions and existence
checks
* Override entries in the case of a colliding hash on insert Add an
* .IsEqual() method to the Signature and PublicKey types in the
btcec package to facilitate easy equivalence testing
* Allocate the signature cache map with the max number of entries in
order to avoid unnecessary map re-sizes/allocations
* Optimize evictions from the signature cache Delete the first entry
* seen which is safe from manipulation due to
the pre image resistance of the hash function
* Double the default maximum number of entries within the signature
cache due to the reduction in the size of a cache entry
* With this eviction scheme, removals are effectively O(1)
Fixes#575.
