Remove request multiplexer (#3624)

* WIP: Get rid of request multiplexer.

* WIP

* Receiver for handling of incoming requests.

* Get rid of useless `Fault` abstraction.

The things the type system let us do are not worth getting abstracted in
its own type. Instead error handling is going to be merely a pattern.

* Make most things compile again.

* Port availability distribution away from request multiplexer.

* Formatting.

* Port dispute distribution over.

* Fixup statement distribution.

* Handle request directly in collator protocol.

+ Only allow fatal errors at top level.

* Use direct request channel for availability recovery.

* Finally get rid of request multiplexer

Fixes #2842 and paves the way for more back pressure possibilities.

* Fix overseer and statement distribution tests.

* Fix collator protocol and network bridge tests.

* Fix tests in availability recovery.

* Fix availability distribution tests.

* Fix dispute distribution tests.

* Add missing dependency

* Typos.

* Review remarks.

* More remarks.
This commit is contained in:
Robert Klotzner
2021-08-12 13:11:36 +02:00
committed by GitHub
parent ecf71233c3
commit 55154a8d37
51 changed files with 1509 additions and 1746 deletions
@@ -18,11 +18,11 @@
//! Error handling related code and Error/Result definitions.
use polkadot_node_network_protocol::PeerId;
use polkadot_node_subsystem_util::runtime;
use polkadot_primitives::v1::{CandidateHash, Hash};
use polkadot_subsystem::SubsystemError;
use thiserror::Error;
use polkadot_node_subsystem_util::{runtime, unwrap_non_fatal, Fault};
use thiserror::Error;
use crate::LOG_TARGET;
@@ -34,29 +34,25 @@ pub type NonFatalResult<T> = std::result::Result<T, NonFatal>;
pub type FatalResult<T> = std::result::Result<T, Fatal>;
/// Errors for statement distribution.
#[derive(Debug, Error)]
#[derive(Debug, Error, derive_more::From)]
#[error(transparent)]
pub struct Error(pub Fault<NonFatal, Fatal>);
impl From<NonFatal> for Error {
fn from(e: NonFatal) -> Self {
Self(Fault::from_non_fatal(e))
}
}
impl From<Fatal> for Error {
fn from(f: Fatal) -> Self {
Self(Fault::from_fatal(f))
}
pub enum Error {
/// Fatal errors of dispute distribution.
Fatal(Fatal),
/// Non fatal errors of dispute distribution.
NonFatal(NonFatal),
}
impl From<runtime::Error> for Error {
fn from(o: runtime::Error) -> Self {
Self(Fault::from_other(o))
match o {
runtime::Error::Fatal(f) => Self::Fatal(Fatal::Runtime(f)),
runtime::Error::NonFatal(f) => Self::NonFatal(NonFatal::Runtime(f)),
}
}
}
/// Fatal runtime errors.
/// Fatal errors.
#[derive(Debug, Error)]
pub enum Fatal {
/// Requester channel is never closed.
@@ -112,9 +108,13 @@ pub enum NonFatal {
///
/// We basically always want to try and continue on error. This utility function is meant to
/// consume top-level errors by simply logging them.
pub fn log_error(result: Result<()>, ctx: &'static str) -> FatalResult<()> {
if let Some(error) = unwrap_non_fatal(result.map_err(|e| e.0))? {
tracing::debug!(target: LOG_TARGET, error = ?error, ctx)
pub fn log_error(result: Result<()>, ctx: &'static str) -> std::result::Result<(), Fatal> {
match result {
Err(Error::Fatal(f)) => Err(f),
Err(Error::NonFatal(error)) => {
tracing::warn!(target: LOG_TARGET, error = ?error, ctx);
Ok(())
},
Ok(()) => Ok(()),
}
Ok(())
}