* 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>
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>
Fixed InterleavingTest that was using the wrong MetaData.Response constructor.
Fixed handling of HEAD methods in HttpTransportOverHTTP2.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Updates after review.
Now the Content-Length header is generated by HpackEncoder based on
MetaData.contentLength, so that the MetaData.HttpFields are not modified.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Introduced RateControl.Factory to create instances of RateControl
for each connection.
Modified relevant XML files and added distribution test for h2.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Reduce the number of RateControl fields, instead using common field in
HeaderParser.
Avoid null checking rateControl by having a NO_RATE_CONTROL static
HPack does not emit field with empty header name.
Apply rate control to any header parsing issue resulting in
session/stream failure
Signed-off-by: Greg Wilkins <gregw@webtide.com>
Implemented rate control for HTTP/2 frames using a single RateControl
object to avoid that each individual vulnerability is within limits,
but combined they still overload the server.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
The client reset wakes up threads blocked in
writes, but these may again attempt to write,
therefore blocking again.
Now we detect that the stream is not writable
and mark the transport as failed, so that
writes fail immediately without blocking.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Removing Legacy Method Separators
* Restyling branch `jetty-9.4.x`
* Applying changes highlighted by checkstyle
* Applying XML restyling
* Fixing XML codestyle for IntelliJ
* Fixing XML style mistakes
* Revert "Applying XML restyling"
* Updating checkstyle for XML codestyle
* Reformatting pom.xml files
* Fixed empty string from line wraps
* Update intellij style to not do expression relative formatting. Reformatted code based on that.
* Increasing line split on Eclipse IDE Formatter to 512
* Restoring setting on internal default value.
+ IntelliJ will not export settings on things that set to their
internal default values.
We want to keep those values as a hedge against future default
value changes in future releases of IntelliJ.
* Fixing intellij codestyle
* do not allow single line simple methods
* misc checkstyle fixes
* re-exported with correct name and all values
Signed-off-by: Joakim Erdfelt <joakim.erdfelt@gmail.com>
Signed-off-by: Greg Wilkins <gregw@webtide.com>
+ Deprecated B64Code
+ All code that isn't B64CodeTest is now using java.util.Base64
+ B64CodeTest is updated to confirm change to java.util.Base64
is possible without change in behavior. Just have to make
sure you use the appropriate Encoder / Decoder for the task
at hand (default vs mime vs url)
Signed-off-by: Joakim Erdfelt <joakim.erdfelt@gmail.com>
Client and server need to to treat an incoming RST_STREAM frame
differently, testing whether the stream is closed for respectively
local and remote streams.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Cleanup the dump implementation
* improved the clarity of utility methods for dump and updated most dump methods
* fixed upgrade filter dump
* Improved dump after review
* Moved dumpObjects to Dumpable
* implemented dumpBeans with dumpObjects
* less verbose dump
* Dump streams
* fixed dump test
Signed-off-by: Greg Wilkins <gregw@webtide.com>
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
+ Changes needed for new Junit 5
+ Migrating from Vintage junit API to Jupiter junit API
+ Relies on SNAPSHOT jetty-test-helper
- this will be a formal release once this issue has been
resolved satisfactory
+ Have jenkins always pull latest SNAPSHOT for each build
+ Adding jetty.snapshots repository
+ Using surefire 2.22.0 per advice from junit
+ Ensuring <reuseForks>true</reuseForks> to work around issue junit-team/junit5#801
+ Disabling <forkMode>always</forkMode> in maven-surefire-plugin
due to bug https://github.com/junit-team/junit5/issues/801
+ OSGi tests must remain at vintage due to PaxExam
+ Moving from vintage TestingDir to jupiter WorkDir
+ Fixing imports to use jupiter, not vintage
+ Migrating vintage ExpectedException to jupiter assertThrows
+ Migrating vintage TestName to jupiter TestInfo
+ Migrating @RunWith(Parameterized.class)
to @ParameterizedTest with Argument Sources
+ Migrating assertTrue(val.contains(needle))
to assertThat(val, containsString(needle))
+ Aligning junit versions per recommendations from @sormuras
+ Adjusting parameter order change for assertEquals()
+ Test LifeCycle Annotation Migration
junit 4 | junit 5 / jupiter
------------ | -----------
@Before | @BeforeEach
@After | @AfterEach
@BeforeClass | @BeforeAll
@AfterClass | @AfterAll
Signed-off-by: Joakim Erdfelt <joakim.erdfelt@gmail.com>
Signed-off-by: olivier lamy <oliver.lamy@gmail.com>
Now HTTP/2 sessions are not added to the Jetty component tree,
but rather just held by HTTP2SessionContainer that is added to
the Jetty container tree at startup.
HTTP2SessionContainer uses a concurrent Set to hold HTTP/2 sessions
to have good add/remove performance.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Simone Bordet
Lachlan Roberts
Ludovic Orban
Joakim Erdfelt
olivier lamy
康智冬
Greg Wilkins
Olivier Lamy