Motivation:
There are some use cases when a client may only be willing to read from a channel once
its previous write is finished (eg: serial dispatchers in Finagle). In this case, a
connection with SslHandler installed and ctx.channel().config().isAutoRead() == false
will stall in 100% of cases no matter what order of "channel active", "write", "flush"
events was.
The use case is following (how Finagle serial dispatchers work):
1. Client writeAndFlushes and waits on a write-promise to perform read() once it's satisfied.
2. A write-promise will only be satisfied once SslHandler finishes with handshaking and
sends the unencrypted queued message.
3. The handshaking process itself requires a number of read()s done by a client but the
SslHandler doesn't request them explicitly assuming that either auto-read is enabled
or client requested at least one read() already.
4. At this point a client will stall with NEED_UNWRAP status returned from underlying engine.
Modifiations:
Always request a read() on NEED_UNWRAP returned from engine if
a) it's handshaking and
b) auto read is disabled and
c) it wasn't requested already.
Result:
SslHandler is now completely tolerant of whether or not auto-read is enabled and client
is explicitly reading a channel.
Motivation:
We should throw a more helpful exception when a non PKCS#8 key is used by the user.
Modifications:
Change exception message to give a hint what is wrong.
Result:
Easier for user to understand whats wrong with their used key.
Motivation:
See https://github.com/netty/netty/issues/3411.
Backport perf improvements on 4.0 and make AsyncHttpClient DNS modules
backports easier to maintain.
Modifications:
Cherry-picked b7415a3307
Result:
Reuse a thread local ArrayList to avoid allocations.
Motivation:
SSLContext.buildTrustManagerFactory(...) builds a KeyStore to
initialize the TrustManagerFactory from an array of X509Certificates,
assuming that array is a chain and that each certificate will have a
unique Subject Distinguised Name.
However, the collection of certificates used as trust anchors is generally
not a chain (it is an unordered collection), and it is legitimate for it
to contain multiple certificates with the same Subject DN.
The existing code uses the Subject DN as the alias name when filling in
the `KeyStore`, thereby overwriting other certificates with the same
Subject DN in this collection, so some certificates may be discarded.
In addition, the code related to building trust managers can take an array of
X509Certificate instances to use as trust anchors. The variable name is
usually trustCertChain, and the documentation refers to them as a "chain".
However, while it makes sense to talk about a "chain" from a keymanager
point of view, these certificates are just an unordered collection in a
trust manager. (There is no chaining requirement, having the Subject DN
matching its predecessor's Issuer DN.)
This can create confusion to for users not used with PKI concepts.
Modifications:
SSLContext.buildTrustManagerFactory(...) now uses a distinct alias for each
array (simply using a counter, since this name is never used for reference
later). This patch also includes a unit test with CA certificates using the
same Subject DN.
Also renamed trustCertChain into trustCertCollection, and changed the
references to "chain" in the Javadoc.
Result:
Each loaded certificate now has a unique identifier when loaded, so it is
now possible to use multiple certificates with the same Subject DN as
trust anchors.
Hopefully, renaming the parameter should also reduce confusion around PKI
concepts.
Motivation:
We need to ensure we call ctx.flush() before closing the actual channel when an handshake failure took place. If we miss to do so we may not send all pending data to the remote peer which also include SSL alerts.
Modifications:
Ensure we call ctx.flush() before ctx.close() on a handshake error.
Result:
All pending data (including SSL alerts) are written to the remote peer on a handshake error.
Motivation:
We need to ensure we do all checks inside of the try / catch block so we free native memory that was allocated in the constructor of the super class in a timely manner.
Modifications:
Move all checks inside of the try block.
Result:
Correctly release native memory (and not depend on the finalizer) when a check in the constructors fails
Motivation:
A user may use a private key which is encrypted with an empty password. Because of this we should only handle a null password in a special way.
Modifications:
- Correctly handle private key that is encrypted with empty password.
- Make OpenSsl*Context implementions consistent in terms of initialization in the constructor.
Result:
Correctly support private key that is encrypted with empty password.
Motivation:
We want to allow the use of an uber jar that contains shared dynamic libraries for all platforms (including fedora).
Modifications:
Modified OpenSsl to try and load the fedora library if the OS is Linux and the platform specified library fails before using the default lib.
Result:
True uber support.
Motivation:
We want to allow the use of an uber jar that contains the shared libraries for all platforms.
Modifications:
Modified OpenSsl to first check for a platform-specific lib before using the default lib.
Result:
uber support.
Motivation:
Depending on the actual CertificateException we should set the correct alert type so it will be sent back to the remote peer and so make it easier for them to fix it.
Modification:
Correctly set the alert and not always just use a general alert.
Result:
It's easier for the remote peer to fix the problems.
Motivation:
Commit 108dc23cab introduced a workaround due to a JDK crash when GCM cipher was used during an unwrap operation. Attempting to reproduce this issue with the latest JDK (1.8.0_72-b15) demonstrate that this issue no longer exists while it can be reliably reproduced on earlier JDKs (1.8.0_25-b17 and earlier)
Modifications:
- Remove the copy-to-heap-buffer workaround for JDK engine
Result:
Fixes https://github.com/netty/netty/issues/3256
Motivation:
netty-tcnative-1.1.33.Fork was released, we should upgrade. Also we should skip renegotiate tests if boringssl is used because boringssl does not support renegotiation.
Modifications:
- Upgrade to netty-tcnative-1.1.33.Fork13
- Skip renegotiate tests if boringssl is used.
Result:
Use newest version of netty-tcnative and be able to build if boringssl is used.
Motivation:
Sometimes it's easier to get keys/certificates as `InputStream`s than it is to
get an actual `File`. This is especially true when operating in a container
environment and `getResourceAsInputStream` is the best way to load resources
packaged with an application.
Modifications:
- Add read-from-`InputStream` methods to `PemReader`
- Allow `SslContext` to get keys/certificates from `InputStreams`
- Add `InputStream`-based setters for key/trust managers to `SslContextBuilder`
Result:
Callers may pass an `InputStream` instead of a `File` to `SslContextBuilder`.
Motivation:
OpenSslContext constructor fails with a UnsupportedOperationException if Unsafe is not present on the system.
Modifications:
Make OpenSslContext work also when Unsafe is not present by fallback to using JNI to get the memory address.
Result:
Using OpenSslContext also works on systems without Unsafe.
Motivation:
We need to enable SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER when using OpenSslContext as the memory address of the buffer that is passed to OpenSslEngine.wrap(...) may change during calls and retries. This is the case as
if the buffer is a heap-buffer we will need to copy it to a direct buffer to hand it over to the JNI layer. When not enable this mode we may see errors like: 'error:1409F07F:SSL routines:SSL3_WRITE_PENDING: bad write retry'.
Related to https://github.com/netty/netty-tcnative/issues/100.
Modifications:
Explitict set mode to SSL.SSL_MODE_RELEASE_BUFFERS | SSL.SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER . (SSL.SSL_MODE_RELEASE_BUFFERS was used before implicitly).
Result:
No more 'error:1409F07F:SSL routines:SSL3_WRITE_PENDING: bad write retry' possible when writing heap buffers.
Motivation:
When using SslProvider.OPENSSL we currently not handle SNI on the client side.
Modifications:
Correctly enable SNI when using clientMode and peerHost != null.
Result:
SNI works even with SslProvider.OPENSSL.
Motivation:
As we now can easily build static linked versions of tcnative it makes sense to run our netty build against all of them.
This helps to ensure our code works with libressl, openssl and boringssl.
Modifications:
Allow to specify -Dtcnative.artifactId= and -Dtcnative.version=
Result:
Easy to run netty build against different tcnative flavors.
Motivation:
When an SSL record contains an invalid extension data, SniHandler
currently throws an IndexOutOfBoundsException, which is not optimal.
Modifications:
- Do strict index range checks
Result:
No more unnecessary instantiation of exceptions and their stack traces
Motivation:
Not all SSLEngine implementations permit beginHandshake being called while a handshake is in progress during the initial handshake. We should ensure we only go through the initial handshake code once to prevent unexpected exceptions from being thrown.
Modifications:
- Only call beginHandshake if there is not currently a handshake in progress
Result:
SslHandler's handshake method is compatible with OpenSSLEngineImpl in Android 5.0+ and 6.0+.
Fixes https://github.com/netty/netty/issues/4718
Motivation:
We incorrectly added the trustCertChain as certificate chain when OpenSslClientContext was created. We need to correctly add the keyCertChain.
Modifications:
Correctly add whole keyCertChain.
Result:
SSL client auth is working when usin OpenSslClientContext and more then one cert is contained in the certificate chain.
Motivation:
Attempts to enable SSL protocols which are currently disabled fail when using the OpenSslEngine. Related to https://github.com/netty/netty/issues/4736
Modifications:
Clear out all options that have disabled SSL protocols before attempting to enable any SSL protocol.
Result:
setEnabledProtocols works as expected.
Motivation:
We need to ensure we flush out all pending data when an SslException accours so the remote peer receives all alerts.
Modifications:
Ensure we call ctx.flush() when needed.
Result:
Correctly receive alerts in all cases on the remote peer.
Motivation:
We need to ensure we add the correct handshake error to the SSLHandshakeException before throwing it when failing the
handshake.
Modifications:
Use the correct error string when creating the SSLHandshakeException.
Result:
Correct SSLHandshakeException message included.
Motivation:
As we can only handle handshake commands to parse SNI we should try to skip alert and change cipher spec commands a few times before we fallback to use a default SslContext.
Modifications:
- Use default SslContext if no application data command was received
- Use default SslContext if after 4 commands we not received a handshake command
- Simplify code
- Eliminate multiple volatile fields
- Rename SslConstants to SslUtils
- Share code between SslHandler and SniHandler by moving stuff to SslUtils
Result:
Correct handling of non handshake commands and cleaner code.
Motivation:
Android 5.0 (API version 21) has a bug which not correctly set the bytesConsumed of SSLEngineResult when HandshakeStatus is FINISHED. Because of this we need to special handle the status and so workaround the Android bug.
Modifications:
- Break the unwrap for (;;) loop when HandshakeStatus is FINISHED and bytesConsumed == 0 && bytesProduced == 0.
Result:
SslHandler works with all known version of Android.
Motivation:
There are a few buffer leaks related to how Unpooled.wrapped and Base64.encode is used.
Modifications:
- Fix usages of Bas64.encode to correct leaks
- Clarify interface of Unpooled.wrapped* to ensure reference count ownership is clearly defined.
Result:
Reference count code is more clearly defined and less leaks are possible.
Motivation:
We need to ensure we only add the OpenSslEngine to the OpenSslEngineMap when the handshake is started as otherwise we may produce a memory leak when the OpenSslEngine is created but not actually used. This can for example happen if we encounter a connection refused from the remote peer. In this case we will never remove the OpenSslEngine from the OpenSslEngineMap and so it will never be collected (as we hold a reference). This has as affect that the finalizer will never be run as well.
Modifications:
- Lazy add the OpenSslEngine to the OpenSslEngineMap to elimate possible leak.
- Call OpenSslEngine.shutdown() when SslHandler is removed from the ChannelPipeline to free memory asap in all cases.
Result:
No more memory leak with OpenSslEngine if connection is refused.
Motivation:
We had to add a new profile for each OpenJDK/OracleJDK release to make
Maven choose the correct alpn-boot.jar and npn-boot.jar. As a result,
our pom.xml has a large number of `<profile/>` sections.
Modifications:
- Use jetty-alpn-agent, which chooses the correct alpn-boot.jar and
npn-boot.jar automatically to remove all the nasty profile sections
from pom.xml
- Visit https://github.com/trustin/jetty-alpn-agent for more info
Result:
Cleaner pom.xml
Motivation:
- Javadoc is not correct (#4353)
- WriteTimeoutHandler does not always cancel the timeout task (#2973)
Modifications:
Fix the javadoc and cleanup timeout task in handlerRemoved
Result:
WriteTimeoutHandler's javadoc describes the correct behavior and it will cancel timeout tasks when it's removed.
Motivation:
As we not used Unpooled anymore for allocate buffers in Base64.* methods we need to ensure we realease all the buffers.
Modifications:
Correctly release buffers
Result:
No more buffer leaks
Motivation:
OpenSslEngine now tests ALPN behavior. However it is possible that OpenSSL is present, but the version does not support ALPN. This will result in test failures instead of just skipping the test.
Modifications:
- Skip ALPN tests in OpenSslEngineTest if the version of OpenSSL does not support ALPN
Result:
Tests don't fail due to unsupported feature in OpenSSL.
Motivation:
Currently there are no tests for OpenSSL Engine,
only for JdkSSL engine.
Modifications:
Common methods from `JdkSslEngine` test moved
to `SSLEngineTest`, JdkSslEngine now implements
NPN and ALPN tests.
Result:
OpenSSL Engine is now covered with unit tests.
Motivation:
FileInputStream opened by SelfSignedCertificate wasn't closed.
Modifications:
Use a try-finally to close the opened FileInputStream.
Result:
FileInputStream will be closed properly.
Motivation:
We currently not supported using KeyManagerFactory with OpenSslServerContext and so should throw an exception if the user tries to do so. This will at least not give suprising and hard to debug problems later.
Modifications:
Throw exception if a user tries to construct a OpenSslServerContext with a KeyManagerFactory
Result:
Fail fast if the user tries to use something that is not supported.
Motivation:
We need to ensure we consume all pending data in the BIO on error to correctly send the close notify for the remote peer.
Modifications:
Correctly force the user to call wrap(...) if there is something left in the BIO.
Result:
close_notify is not lost.
Motivation:
When ClientAuth is set via SslContextBuilder we pass it into the OpenSslEngine constructor. Due a bug we missed to call the correct native methods and so never enabled ClientAuth in this case.
Modifications:
Correctly call setClientAuth(...) in the constructor if needed.
Result:
client auth also works when configured via the SslContextBuilder and OPENSSL is used.
Motivation:
We missed to remove a method in SslContext while refactored the implementation. We should remove the method to keep things clean.
Modifications:
Remove unused method.
Result:
Code cleanup.
Motivation:
Use new / non-deprecated APIs for creating SSL Context
in tests, in order to be able to implement OpenSsl
tests with maximum code reuse.
Modifications:
Use `SslContextBuilder.(forServer|forClient)` instead
of deprecated `JdkSslServerContext` constructor.
Use `ApplicationProtocolConfig` instead of Protocol
Negotiator.
Use custom exception type for skipping tests to avoid
swallowing exceptions arising from tests.
Result:
Exceptions from tests aren't swallowed.
Using new APIs allows reusing same test code for
OpenSsl tests.
Motivation:
We should use OneTimeTask where possible to reduce object creation.
Modifications:
Replace Runnable with OneTimeTask
Result:
Less object creation
Motivation:
Child classes of ApplicationProtocolNegotiationHandler may want to override the behavior when a handshake failure is detected.
Modifications:
- Provide a method which can be overriden when a handshake failure is detected.
Result:
Child classes can override ApplicationProtocolNegotiationHandler handshake failure behavior.
