mirror of https://github.com/vacp2p/nim-libp2p.git
Moving idle timeout to Connection to enable across all connection streams (#307)
* move idle timeout logic to connection * more informative logs * more informative logs
This commit is contained in:
parent
5f0637c49a
commit
cf2b42b914
|
@ -19,8 +19,8 @@ macro checkFutures*[T](futs: seq[Future[T]], exclude: untyped = []): untyped =
|
|||
if res.failed:
|
||||
let exc = res.readError()
|
||||
# We still don't abort but warn
|
||||
warn "A future has failed, enable trace logging for details", error=exc.name
|
||||
trace "Exception message", msg=exc.msg
|
||||
warn "A future has failed, enable trace logging for details", error = exc.name
|
||||
trace "Exception message", msg= exc.msg, stack = getStackTrace()
|
||||
else:
|
||||
quote do:
|
||||
for res in `futs`:
|
||||
|
|
|
@ -46,8 +46,6 @@ logScope:
|
|||
type
|
||||
LPChannel* = ref object of BufferStream
|
||||
id*: uint64 # channel id
|
||||
timeout: Duration # channel timeout if no activity
|
||||
activity: bool # reset every time data is sent or received
|
||||
name*: string # name of the channel (for debugging)
|
||||
conn*: Connection # wrapped connection used to for writing
|
||||
initiator*: bool # initiated remotely or locally flag
|
||||
|
@ -57,7 +55,6 @@ type
|
|||
msgCode*: MessageType # cached in/out message code
|
||||
closeCode*: MessageType # cached in/out close code
|
||||
resetCode*: MessageType # cached in/out reset code
|
||||
timerTaskFut: Future[void] # the current timer instanse
|
||||
|
||||
proc open*(s: LPChannel) {.async, gcsafe.}
|
||||
|
||||
|
@ -79,11 +76,6 @@ template withEOFExceptions(body: untyped): untyped =
|
|||
except LPStreamIncompleteError as exc:
|
||||
trace "incomplete message", exc = exc.msg
|
||||
|
||||
proc cleanupTimer(s: LPChannel) {.async.} =
|
||||
## cleanup timers
|
||||
if not isNil(s.timerTaskFut) and not s.timerTaskFut.finished:
|
||||
s.timerTaskFut.cancel()
|
||||
|
||||
proc closeMessage(s: LPChannel) {.async.} =
|
||||
logScope:
|
||||
id = s.id
|
||||
|
@ -150,7 +142,6 @@ proc closeRemote*(s: LPChannel) {.async.} =
|
|||
|
||||
# call to avoid leaks
|
||||
await procCall BufferStream(s).close() # close parent bufferstream
|
||||
await s.cleanupTimer()
|
||||
|
||||
trace "channel closed on EOF"
|
||||
except CancelledError as exc:
|
||||
|
@ -193,8 +184,6 @@ method reset*(s: LPChannel) {.base, async, gcsafe.} =
|
|||
s.isEof = true
|
||||
s.closedLocal = true
|
||||
|
||||
await s.cleanupTimer()
|
||||
|
||||
except CancelledError as exc:
|
||||
raise exc
|
||||
except CatchableError as exc:
|
||||
|
@ -222,7 +211,6 @@ method close*(s: LPChannel) {.async, gcsafe.} =
|
|||
await s.closeMessage().wait(2.minutes)
|
||||
if s.atEof: # already closed by remote close parent buffer immediately
|
||||
await procCall BufferStream(s).close()
|
||||
await s.cleanupTimer()
|
||||
except CancelledError as exc:
|
||||
await s.reset()
|
||||
raise exc
|
||||
|
@ -235,60 +223,16 @@ method close*(s: LPChannel) {.async, gcsafe.} =
|
|||
s.closedLocal = true
|
||||
asyncCheck closeInternal()
|
||||
|
||||
proc timeoutMonitor(s: LPChannel) {.async.} =
|
||||
## monitor the channel for innactivity
|
||||
##
|
||||
## if the timeout was hit, it means that
|
||||
## neither incoming nor outgoing activity
|
||||
## has been detected and the channel will
|
||||
## be reset
|
||||
##
|
||||
|
||||
logScope:
|
||||
id = s.id
|
||||
initiator = s.initiator
|
||||
name = s.name
|
||||
oid = $s.oid
|
||||
peer = $s.conn.peerInfo
|
||||
|
||||
try:
|
||||
while true:
|
||||
await sleepAsync(s.timeout)
|
||||
|
||||
if s.closed or s.atEof:
|
||||
return
|
||||
|
||||
if s.activity:
|
||||
s.activity = false
|
||||
continue
|
||||
|
||||
break
|
||||
|
||||
# reset channel on innactivity timeout
|
||||
trace "channel timed out, resetting"
|
||||
await s.reset()
|
||||
except CancelledError as exc:
|
||||
raise exc
|
||||
except CatchableError as exc:
|
||||
trace "exception in timeout", exc = exc.msg
|
||||
|
||||
method initStream*(s: LPChannel) =
|
||||
if s.objName.len == 0:
|
||||
s.objName = "LPChannel"
|
||||
|
||||
s.timeoutHandler = proc() {.async, gcsafe.} =
|
||||
trace "idle timeout expired, resetting LPChannel"
|
||||
await s.reset()
|
||||
|
||||
procCall BufferStream(s).initStream()
|
||||
|
||||
method readOnce*(s: LPChannel,
|
||||
pbytes: pointer,
|
||||
nbytes: int):
|
||||
Future[int] =
|
||||
s.activity = true
|
||||
procCall BufferStream(s).readOnce(pbytes, nbytes)
|
||||
|
||||
method write*(s: LPChannel, msg: seq[byte]): Future[void] =
|
||||
s.activity = true
|
||||
procCall BufferStream(s).write(msg)
|
||||
|
||||
proc init*(
|
||||
L: type LPChannel,
|
||||
id: uint64,
|
||||
|
@ -339,7 +283,6 @@ proc init*(
|
|||
when chronicles.enabledLogLevel == LogLevel.TRACE:
|
||||
chann.name = if chann.name.len > 0: chann.name else: $chann.oid
|
||||
|
||||
chann.timerTaskFut = chann.timeoutMonitor()
|
||||
trace "created new lpchannel"
|
||||
|
||||
return chann
|
||||
|
|
|
@ -16,7 +16,7 @@ logScope:
|
|||
topics = "muxer"
|
||||
|
||||
const
|
||||
DefaultChanTimeout* = 1.minutes
|
||||
DefaultChanTimeout* = 5.minutes
|
||||
|
||||
type
|
||||
StreamHandler* = proc(conn: Connection): Future[void] {.gcsafe.}
|
||||
|
|
|
@ -220,6 +220,8 @@ method readOnce*(s: BufferStream,
|
|||
var index = 0
|
||||
var size = min(nbytes, s.len)
|
||||
let output = cast[ptr UncheckedArray[byte]](pbytes)
|
||||
|
||||
s.activity = true # reset activity flag
|
||||
while s.len() > 0 and index < size:
|
||||
output[index] = s.popFirst()
|
||||
inc(index)
|
||||
|
@ -243,6 +245,7 @@ method write*(s: BufferStream, msg: seq[byte]) {.async.} =
|
|||
if isNil(s.writeHandler):
|
||||
raise newNotWritableError()
|
||||
|
||||
s.activity = true # reset activity flag
|
||||
await s.writeHandler(msg)
|
||||
|
||||
# TODO: move pipe routines out
|
||||
|
|
|
@ -14,18 +14,28 @@ import connection
|
|||
logScope:
|
||||
topics = "chronosstream"
|
||||
|
||||
type ChronosStream* = ref object of Connection
|
||||
const
|
||||
DefaultChronosStreamTimeout = 10.minutes
|
||||
|
||||
type
|
||||
ChronosStream* = ref object of Connection
|
||||
client: StreamTransport
|
||||
|
||||
method initStream*(s: ChronosStream) =
|
||||
if s.objName.len == 0:
|
||||
s.objName = "ChronosStream"
|
||||
|
||||
s.timeoutHandler = proc() {.async, gcsafe.} =
|
||||
trace "idle timeout expired, closing ChronosStream"
|
||||
await s.close()
|
||||
|
||||
procCall Connection(s).initStream()
|
||||
|
||||
proc newChronosStream*(client: StreamTransport): ChronosStream =
|
||||
new result
|
||||
result.client = client
|
||||
proc init*(C: type ChronosStream,
|
||||
client: StreamTransport,
|
||||
timeout = DefaultChronosStreamTimeout): ChronosStream =
|
||||
result = C(client: client,
|
||||
timeout: timeout)
|
||||
result.initStream()
|
||||
|
||||
template withExceptions(body: untyped) =
|
||||
|
@ -48,6 +58,7 @@ method readOnce*(s: ChronosStream, pbytes: pointer, nbytes: int): Future[int] {.
|
|||
|
||||
withExceptions:
|
||||
result = await s.client.readOnce(pbytes, nbytes)
|
||||
s.activity = true # reset activity flag
|
||||
|
||||
method write*(s: ChronosStream, msg: seq[byte]) {.async.} =
|
||||
if s.closed:
|
||||
|
@ -60,6 +71,7 @@ method write*(s: ChronosStream, msg: seq[byte]) {.async.} =
|
|||
var written = 0
|
||||
while not s.client.closed and written < msg.len:
|
||||
written += await s.client.write(msg[written..<msg.len])
|
||||
s.activity = true # reset activity flag
|
||||
|
||||
if written < msg.len:
|
||||
raise (ref LPStreamClosedError)(msg: "Write couldn't finish writing")
|
||||
|
|
|
@ -20,12 +20,19 @@ logScope:
|
|||
|
||||
const
|
||||
ConnectionTrackerName* = "libp2p.connection"
|
||||
DefaultConnectionTimeout* = 1.minutes
|
||||
|
||||
type
|
||||
TimeoutHandler* = proc(): Future[void] {.gcsafe.}
|
||||
|
||||
Direction* {.pure.} = enum
|
||||
None, In, Out
|
||||
|
||||
Connection* = ref object of LPStream
|
||||
activity*: bool # reset every time data is sent or received
|
||||
timeout*: Duration # channel timeout if no activity
|
||||
timerTaskFut: Future[void] # the current timer instanse
|
||||
timeoutHandler*: TimeoutHandler # timeout handler
|
||||
peerInfo*: PeerInfo
|
||||
observedAddr*: Multiaddress
|
||||
dir*: Direction
|
||||
|
@ -35,6 +42,7 @@ type
|
|||
closed*: uint64
|
||||
|
||||
proc setupConnectionTracker(): ConnectionTracker {.gcsafe.}
|
||||
proc timeoutMonitor(s: Connection) {.async, gcsafe.}
|
||||
|
||||
proc getConnectionTracker*(): ConnectionTracker {.gcsafe.} =
|
||||
result = cast[ConnectionTracker](getTracker(ConnectionTrackerName))
|
||||
|
@ -58,21 +66,23 @@ proc setupConnectionTracker(): ConnectionTracker =
|
|||
result.isLeaked = leakTransport
|
||||
addTracker(ConnectionTrackerName, result)
|
||||
|
||||
proc init*(C: type Connection,
|
||||
peerInfo: PeerInfo,
|
||||
dir: Direction): Connection =
|
||||
result = C(peerInfo: peerInfo, dir: dir)
|
||||
result.initStream()
|
||||
|
||||
method initStream*(s: Connection) =
|
||||
if s.objName.len == 0:
|
||||
s.objName = "Connection"
|
||||
|
||||
procCall LPStream(s).initStream()
|
||||
s.closeEvent = newAsyncEvent()
|
||||
|
||||
doAssert(isNil(s.timerTaskFut))
|
||||
s.timerTaskFut = s.timeoutMonitor()
|
||||
|
||||
inc getConnectionTracker().opened
|
||||
|
||||
method close*(s: Connection) {.async.} =
|
||||
## cleanup timers
|
||||
if not isNil(s.timerTaskFut) and not s.timerTaskFut.finished:
|
||||
s.timerTaskFut.cancel()
|
||||
|
||||
if not s.isClosed:
|
||||
await procCall LPStream(s).close()
|
||||
inc getConnectionTracker().closed
|
||||
|
@ -83,3 +93,50 @@ proc `$`*(conn: Connection): string =
|
|||
|
||||
func hash*(p: Connection): Hash =
|
||||
cast[pointer](p).hash
|
||||
|
||||
proc timeoutMonitor(s: Connection) {.async, gcsafe.} =
|
||||
## monitor the channel for innactivity
|
||||
##
|
||||
## if the timeout was hit, it means that
|
||||
## neither incoming nor outgoing activity
|
||||
## has been detected and the channel will
|
||||
## be reset
|
||||
##
|
||||
|
||||
logScope:
|
||||
oid = $s.oid
|
||||
|
||||
try:
|
||||
while true:
|
||||
await sleepAsync(s.timeout)
|
||||
|
||||
if s.closed or s.atEof:
|
||||
return
|
||||
|
||||
if s.activity:
|
||||
s.activity = false
|
||||
continue
|
||||
|
||||
break
|
||||
|
||||
# reset channel on innactivity timeout
|
||||
trace "Connection timed out"
|
||||
if not(isNil(s.timeoutHandler)):
|
||||
await s.timeoutHandler()
|
||||
|
||||
except CancelledError as exc:
|
||||
raise exc
|
||||
except CatchableError as exc:
|
||||
trace "exception in timeout", exc = exc.msg
|
||||
|
||||
proc init*(C: type Connection,
|
||||
peerInfo: PeerInfo,
|
||||
dir: Direction,
|
||||
timeout: Duration = DefaultConnectionTimeout,
|
||||
timeoutHandler: TimeoutHandler = nil): Connection =
|
||||
result = C(peerInfo: peerInfo,
|
||||
dir: dir,
|
||||
timeout: timeout,
|
||||
timeoutHandler: timeoutHandler)
|
||||
|
||||
result.initStream()
|
||||
|
|
|
@ -63,7 +63,7 @@ proc connHandler*(t: TcpTransport,
|
|||
client: StreamTransport,
|
||||
initiator: bool): Connection =
|
||||
trace "handling connection", address = $client.remoteAddress
|
||||
let conn: Connection = Connection(newChronosStream(client))
|
||||
let conn: Connection = Connection(ChronosStream.init(client))
|
||||
conn.observedAddr = MultiAddress.init(client.remoteAddress).tryGet()
|
||||
if not initiator:
|
||||
if not isNil(t.handler):
|
||||
|
|
Loading…
Reference in New Issue