* Bumped the rate control rate from 50 events/s to 128.
* Added rate control for all CONTINUATION frames.
* Added rate control for invalid PUSH_PROMISE frames.
* Added rate control for RST_STREAM frames.
* Added rate control for all SETTINGS frames.
* Fixed growth of header block accumulation buffer.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Now allowing to specify a negative value for AbstractHTTP2ServerConnectionFactory.streamIdleTimeout, while 0 implies to use the default value (from the EndPoint).
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#8405 - onAllDataRead() is called twice under h2 if the stream times out
Per Servlet semantic, HTTP/2 stream timeout should be ignored.
The code was trying to fail the read via `_contentDemander.onTimeout()`, but
then it was still calling `onContentProducible()`, which was returning `true`
because the state of the read was IDLE (all the request content was read) and
the request was suspended.
Now the code checks if the read was really failed; if it is not, then
`onContentProducible()` is not called and so the idle timeout is ignored.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Implemented a few required error handlings.
* Changed `Parser.init()` to directly take the listener, rather than wrapping it.
The reason for this change was to be able to reconfigure the Parser upon receiving a SETTINGS frame.
* Initially setting the encoder and decoder max table capacity at the default of 4096, as per spec.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#8678 - Jetty client receives GO_AWAY and continue to send traffic on same connection
* Now upon receiving the GOAWAY, the connection is removed from the pool, so it cannot be used by new requests.
* HTTP2Session.removeStream() now happens _after_ notifying HEADERS and DATA events, although the Stream state change still happens before.
This is necessary to avoid that a "close" event is notified before a "headers" or "data" event.
With these changes, the race window of a client acquiring a connection while the server is closing it is reduced, but it is impossible to close it completely.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
All `ByteBufferPool` can now be accessed as `RetainableByteBufferPools`.
Users now need to configure only a single buffer pool and there is just the additional retained parameter that needs consideration.
Default buffer pool has been changed to logarithmic, but we may wish to review that before next release.
Default factor size has been increased to 4096.
* Fixes#7348 - Slow CONNECT request causes NPE (#7349)
Added NPE guard in `HttpReceiverOverHTTP.onUpgradeFrom()`.
Expanded logic in `HttpReceiverOverHTTP.parse()` to return true in case of CONNECT + 200.
Fixed `ProxyConnection.toConnectionString()` to avoid NPEs.
Fixed `HttpClientTest.testCONNECTWithHTTP10()` logic
after changes to fix this issue.
Now a tunneled connection is not put back into the connection pool,
and if applications explicitly want to use it, they must re-enable
fill interest, similarly to what should be done after upgrade+101.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
(cherry picked from commit 5eb7b70df7)
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Changed order of entries in module-info.java to be canonical
(cherry picked from commit 02691171d5)
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Incremented rateControl.maxEventsPerSecond to 50.
Fixed mistakes in the http2c module.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
(cherry picked from commit 784293aa6d)
This is important in tests that check that streams have been removed from sessions after counting down a latch in the notification.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Fixed initial session recv window update: it was wrong if the initial value was less than the default value (65535).
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Reimplemented close/idle_timeout/stop/onGoAway/input_shutdown following more closely the specification.
In particular, the semantic of sending a GOAWAY is now to:
* stop creation of new both local and remote streams
* record the last processed stream
* continue processing streams that are pending
This means that a GOAWAY is "graceful" in the sense that it allows for streams to be completed by applications.
The semantic of stop() and idle timeout is harsher: for pending streams a RST_STREAM is sent to the other peer and they are failed locally.
Added support for GOAWAY with 2^31-1 lastStreamId.
Added support for a peer to send and receive multiple GOAWAY frames.
Reviewed the stream creation/destruction mechanism so that when the last stream completes after a GOAWAY, proper actions can be run to tear down the connection.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Make EOF and errors be special content.
Transition to a much simplified FSM by using the needContent() / produceContent() model.
Implement blocking on top of async, this way there is only one FSM.
(Milestone 6)
Signed-off-by: Ludovic Orban <lorban@bitronix.be>
* Fixes#3766 - Introduce HTTP/2 API to batch frames.
Introduced Stream.FrameList to hold HEADERS+DATA+HEADERS frames.
These are often used by the client and by the server when the
request/response content is known and FrameList will allow to
send them in a single TCP write, rather than multiple ones.
Rewritten HttpSenderOverHTTP2.sendHeaders() and
HttpTransportOverHTTP2.sendHeaders() to take advantage of
FrameList.
Now using ConcurrentHashMap as a client context, because
with DEBUG logging enabled it may be access concurrently.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Replaced relevant usages of synchronized with AutoLock.
* Made AutoLock serializable since classes that use it may be stored in the HttpSession.
* Added convenience methods to AutoLock to execute lambdas with the lock held.
* Introduced AutoLock.WithCondition to use a Lock and a Condition together.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Fixed test failure after merge.
Made sure the lastStreamId is updated when resetting a new
stream since it has been seen by the server and handled.
This avoids that a new stream arriving during shutdown is
interpreted as a connection failure, therefore failing all
pending streams - we want them to complete.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Fixes#4967 - Possible buffer corruption in HTTP/2 session failures
Partially reverted the changes introduced in #4855, because they
were working only when sends were synchronous.
Introduced ByteBufferPool.remove(ByteBuffer) to fix the issue.
Now when a concurrent failure happens while frames are being
generated or sent, the buffer is discarded instead of being
recycled, therefore resolving the buffer corruption.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Issue #4965 - WINDOW_UPDATE for locally failed stream should not close the HTTP/2 session.
Improved HTTP2Session.onWindowUpdate() code to correctly check whether
the stream is already closed, and if so, just drop the WINDOW_UPDATE.
Refactored onResetForUnknownStream() to base class.
Other small refactorings to improve logging.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#4855 - Occasional h2spec failures on CI
In case of bad usage of the HTTP/2 API, we don't want to close()
the stream but just fail the callback, because the stream
may be performing actions triggered by a legit API usage.
In case of a call to `AsyncListener.onError()`, applications may decide to call
AsyncContext.complete() and that would be a correct usage of the Servlet API.
This case was not well handled and was wrongly producing a WARN log with an
`IllegalStateException`.
Completely rewritten `HttpTransportOverHTTP2.TransportCallback`.
The rewrite handles correctly asynchronous failures that now are executed
sequentially (and not concurrently) with writes.
If a write is in progress, the failure will just change the state and at the
end of the write a check on the state will determine what actions to take.
A session failure is now handled in HTTP2Session by first failing all the
streams - which notifies the Stream.Listeners - and then failing the session
- which notifies the Session.Listener.
The stream failures are executed concurrently by dispatching each one to a
different thread; this means that the stream failure callbacks are executed
concurrently (likely sending RST_STREAM frames).
The session failure callback is completed only when all the stream failure
callbacks have completed, to ensure that a GOAWAY frame is processed after
all the RST_STREAM frames.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Simone Bordet
gregw
Lachlan
Joakim Erdfelt
Ludovic Orban
Jan Bartel