Commit Graph

10090 Commits

Author SHA1 Message Date
Norman Maurer
a63faa4fa1
Use netty-jni-util and so remove a lot of duplication (#10735)
Motivation:

We had a lot of duplication in our jni code which was mostly due macros but also related to how we support shading. By using netty-jni-util we can share all the code between netty and netty-tcnative ( and possible other jni based netty projects in the future).

Modifications:

- Use netty-jni-util and re-use its macros / functions
- Remove duplicated code
- Adjust build files

Result:

Less code duplication for JNI
2020-10-29 16:36:07 +01:00
Aayush Atharva
e9b28e76a3
Change variable name of Http2Headers (#10743)
Motivation:
`addHttp2ToHttpHeaders(int streamId, Http2Headers sourceHeaders, FullHttpMessage destinationMessage, boolean addToTrailer)` 
should match 
`addHttp2ToHttpHeaders(int streamId, Http2Headers inputHeaders, HttpHeaders outputHeaders, HttpVersion httpVersion, boolean isTrailer, boolean isRequest)`.

However, the `Http2Headers` variable name is different.

Modification:
Changed `sourceHeaders` to `inputHeaders`.

Result:
Variable and JavaDoc naming now correct.
2020-10-29 10:40:49 +01:00
Aayush Atharva
8a10adaa55
Make Http2ToHttpHeaderTranslator method package-private. (#10736)
Motivation:
Http2ToHttpHeaderTranslator is a private class but translateHeaders(Iterable<Entry<CharSequence, CharSequence>>) is public but it should be package-private.

Modification:
Removed public.

Result:
Correct access modifer.
2020-10-29 10:39:02 +01:00
Chris Vest
96da45de2d
Fix explicitly little-endian accessors in SwappedByteBuf (#10747)
Motivation:
Some buffers implement ByteBuf#order(order) by wrapping themselves in a SwappedByteBuf.
The SwappedByteBuf is then responsible for swapping the byte order on accesses.
The explicitly little-endian accessor methods, however, should not be swapped to big-endian, but instead remain explicitly little-endian.

Modification:
The SwappedByteBuf was passing through calls to e.g. writeIntLE, to the big-endian equivalent, e.g. writeInt.
This has been changed so that these calls delegate to their explicitly little-endian counterpart.

Result:
This makes all buffers that make use of SwappedByteBuf for their endian-ness configuration, consistent with all the buffers that use other implementation strategies.
In the end, all buffers now behave exactly the same, when using their explicitly little-endian accessor methods.
2020-10-29 10:35:47 +01:00
Norman Maurer
b4edf21bd3
Errors.throwConnectException(...) should be public (#10745)
Motivation:

ddebc1027d missed to make Errors.throwConnectException(...) public

Modifications:

Make method public

Result:

Be able to use Errors.throwConnectException(...) from other module
2020-10-29 07:44:49 +01:00
Norman Maurer
d371b1bbaa
Rethrow NoSuchMethodError with more hints about incompatible native library versions (#10740)
Motivation:

03aafb9cff did ensure we fail while loading a natibe library which is not compatible. While this is great it is still sometimes hard for people to understand what NoSuchMethodError means in this context.

Modifications:

If possible rethrow the NoSuchMethodError and provide some more hints about multiple versions of the shared library

Result:

Easier to understand for people why loading fails
2020-10-28 19:22:08 +01:00
Norman Maurer
d4defc30b8
Add PlatformDependent* methods that are needed by io_uring (#10744)
Motivation:

ddebc10 added a few adjustments that are needed by io_uring that we will add as an incubator repository. Unfortunally we missed the changed in PlatformDependent*.

Modifications:

Add missing methods

Result:

Be able to compile io_uring code against core netty
2020-10-28 19:20:52 +01:00
Norman Maurer
ddebc1027d
Make changes to prepare for io_uring incubator repository (#10741)
Motivation:

During the last few month we did develop an io_uring based transport which shows very promising performance numbers. To give it more time to bake we will develop it outside of netty in an "incubator" module which will make it clear to users what to expect and also allow us to seperate its release cycle. While the implementation of it is very self contained there are few small adjustments that need to be made in netty itself to allow us to reuse code.

Modifications:

- AbstractChannel: Add method which can be used when a write fails and remove final from one method
- IovArray: Allow to create an IovArray from a ByteBuf instance
- FileDescriptor: Allow to reuse mark close logic via sub-class

Result:

Be able to reuse netty core classes in io_uring incubator repository
2020-10-28 15:31:02 +01:00
Norman Maurer
701eed8762
Unregister all native methods on unload for kqueue (#10738)
Motivation:

03aafb9cff did ensure we unregister all native methods when we unload / or fail during load of the native library. Unfortunally it missed to unregister in one case for kqueue.

Modifications:

Add unregister calls to the unload function

Result:

Correctly unregister in all cases
2020-10-28 13:08:22 +01:00
Aayush Atharva
fe5a9f3e32
Upgrade Conscrypt to 2.5.1 (#10732)
Motivation:
Conscrypt 2.5.1 is available and it's a good idea to upgrade to the latest version.

Modification:
Upgraded Conscrypt 2.4.0 to 2.5.1

Result:
Newer Conscrypt version.
2020-10-26 19:40:39 +01:00
Aayush Atharva
ee3b9a5f7b
Fix possible NPEs and IndexOutOfBoundsExceptions in HTTP/2 Codec (#10640)
Motivation:

There are possible NPEs and IndexOutOfBoundsExceptions in HTTP/2 code. 

Modification:
Fixed possible NPEs and IOOBEs

Result:
Better code
2020-10-26 14:41:49 +01:00
Artem Smotrakov
065c39611e
Better hash algorithm in FingerprintTrustManagerFactory (#10683)
Motivation:

FingerprintTrustManagerFactory can only use SHA-1 that is considered
insecure.

Modifications:

- Updated FingerprintTrustManagerFactory to accept a stronger hash algorithm.
- Deprecated the constructors that still use SHA-1.
- Added a test for FingerprintTrustManagerFactory.

Result:

A user can now configure FingerprintTrustManagerFactory to use a
stronger hash algorithm.
The compiler shows a warning if the code still uses one of the
unsafe constructors.

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2020-10-26 14:29:26 +01:00
Norman Maurer
03aafb9cff
Unregister all previous registered native methods if loading of native code fails… (#10719)
Motivation:

It's important to unload all previous registered native methods when there is a failure during loading the native lib. Failing to do so may lead to an "invalid state" and so may segfault the JVM when trying to call a native method that was previous loaded.

This was observed when two versions of netty-tcnative were on the classpath which had different requirements in terms of linking.

Something like this was reported in he hs log:

```
Instructions: (pc=0x0000000116413bf0)
0x0000000116413bd0:
[error occurred during error reporting (printing registers, top of stack, instructions near pc), id 0xb]

Register to memory mapping:

RAX=0x0000000116413bf0 is an unknown value
RBX={method} {0x000000011422e708} 'aprMajorVersion' '()I' in 'io/netty/internal/tcnative/Library'
RCX=0x000000000000000a is an unknown value
RDX=0x000000000000000a is an unknown value
```

Modifications:

- Unregister previous registered native methods on failure
- Unregister previous registered native methods on on unload of the native lib

Result:

No more segfault caused by invalid state when loading of the native lib fails in between. In this case the user will receive an error now like:
2020-10-26 14:15:04 +01:00
Aayush Atharva
321f0f0e66
Upgrade Slf4j to 1.7.30 (#10712)
Motivation:
SLF4J 1.7.30 is the latest version in 1.7.x and we should upgrade to it from 1.7.21.

Modification:
Changed 1.7.21 to 1.7.30

Result:
Newer version of SLF4J
2020-10-26 13:03:12 +01:00
Norman Maurer
f5185ed73b
Release message before notify promise (#10726)
Motivation:

We should preferable always release the message before we notify the promise. Thhis has a few advantages:

 - Release memory as soon as possible
 - Listeners observe the "more correct" reference count

Modifications:

Release message before fail the promises

Result:

Faster releasing of resources. This came up in https://github.com/netty/netty/issues/10723
2020-10-26 12:59:20 +01:00
Roman Puchkovskiy
4ecd78e104
Fix native image build on modern GraalVM versions for the cases when the program uses netty-dns (#10630)
Motivation:

Since GraalVM version 19.3.0, instances of java.net.InetAddress (and its subclasses Inet4Address and Inet6Address) are not allowed in native image heap (that is, they cannot be stored in static fields of classes initialized at build time or be reachable through static fields of such classes). When building a native image, it makes sense to initialize at build time as many classes as possible.
But some fields of some classes in Netty (for example, NetUtil.LOCALHOST4) contain InetAddress instances. If a program is using code path that makes it possible to reach such fields at build time initialization, it becomes impossible to build a native image initializing core Netty classes initialized at runtime. An example of such a program is a client that uses netty-dns.

Modifications:

- Add netty-testsuite-native-image-client Maven module to test that such an example program can be built after the corresponding fixes
- Add native-image.properties to resolver-dns module to move initialization of some classes to runtime (some of them are parsing configuration during initialization, so it makes no sense to initialize them at build time; for others, it's needed to avoid InetAddress reachability at build time)
- Add substitutions for NetUtil.LOCALHOST4, NetUtil.LOCALHOST6 and NetUtil.LOCALHOST to overcome the InetAddress-related prohibition
- Extract some initialization code from NetUtil to NetUtilInitializations to allow it to be used by the substitutions

Result:

A client program using netty-dns with --initialize-at-build-time=io.netty builds successfully
2020-10-26 08:34:31 +01:00
greenjustin
090e9a7271
Allow EventLoops to rethrow Error (#10694)
Motivation:

Thread.stop() works by producing a ThreadDeath error in the target thread. EventLoops swallow all Throwables, which makes them effectively unkillable. This is effectively a memory leak, for our application. Beside this we should also just regrow all `Error` as there is almost no way to recover.

Modification:

Edit the EventLoops that swallow Throwables to instead rethrow Error.

Result:

`EventLoop` can crash if `Error` is thrown
2020-10-24 14:56:33 +02:00
Norman Maurer
d58d1add34 Fix checkstyle errors introduced by 33de96f448 2020-10-24 14:50:00 +02:00
Andrey Mizurov
33de96f448
Provide new client and server websocket handshake exceptions (#10646)
Motivation:

At the moment we have only one base `WebSocketHandshakeException` for handling WebSocket upgrade issues.
Unfortunately, this message contains only a string message about the cause of the failure, which is inconvenient in handling.

Modification:

Provide new `WebSocketClientHandshakeException` with `HttpResponse` field  and `WebSocketServerHandshakeException` with `HttpRequest` field both of them without content for avoid reference counting 
problems. 

Result:

More information for more flexible handling.

Fixes #10277 #4528 #10639.
2020-10-24 14:41:11 +02:00
Aayush Atharva
6ae7f9e1a8
Fix JavaDoc of Http2Headers (#10711)
Motivation:
Http2Headers has JavaDoc error which says Sets the {@link PseudoHeaderName#AUTHORITY} header or {@code null} if there is no such header however it should be Sets the {@link PseudoHeaderName#AUTHORITY} header in Http2Headers#authority(CharSequence) methods because it only sets CharSequence.

This is true for all setters in Http2Headers.

Modification:
Fixed all JavaDoc errors.

Result:
Better JavaDoc.
2020-10-23 15:35:43 +02:00
Artem Smotrakov
e5951d46fc
Enable nohttp check during the build (#10708)
Motivation:

HTTP is a plaintext protocol which means that someone may be able
to eavesdrop the data. To prevent this, HTTPS should be used whenever
possible. However, maintaining using https:// in all URLs may be
difficult. The nohttp tool can help here. The tool scans all the files
in a repository and reports where http:// is used.

Modifications:

- Added nohttp (via checkstyle) into the build process.
- Suppressed findings for the websites
  that don't support HTTPS or that are not reachable

Result:

- Prevent using HTTP in the future.
- Encourage users to use HTTPS when they follow the links they found in
  the code.
2020-10-23 14:44:18 +02:00
Norman Maurer
24e6f37745
Add NULL checks to fix possible undefined behavior (#10718)
Motivation:

In some situations we could have end up calling some functions with NULL parameters which in this case could lead to undefined behavior. All of this would have happened during loading of the native lib.

Modifications:

Add NULL check as guards and return early

Result:

Fix some possible undefined behavior
2020-10-23 14:39:27 +02:00
James Yuzawa
726944146b
Use named exceptions in ChannelPool implementations (#10721)
Motivation:

I was collecting stats for failed promises with a FixedChannelPool and I was bucketing by stats using cause.getSimpleName(). After #9152 was released, the introduction of the anonymous classes make getSimpleName() return "" causing confusion.

Modification:

Use named classes in the ChannelPool implementations. I made them private, but I can change that if you think otherwise.

Result:

The SimpleChannelPool fails the promises with a ChannelPoolFullException. The FixedChannelPool fails the promises with an AcquireTimeoutException. Also AcquireTimeoutException is more specific than just a plain TimeoutException, which is also useful for troubleshooting. If you want different class names, please advise.
2020-10-23 14:29:37 +02:00
Norman Maurer
ee4164ab8f
DatagramDnsResponseDecoder should rethrow as CorruptedFrameException (#10714)
Motivation:

DatagramDnsResponseDecoder should rethrow as CorruptedFrameException if an IndexOutOfBoundsException happens.

Modifications:

- Catch IndexOutOfBoundsException and rethrow as CorruptedFrameException
- Add a testcase

Result:

Less noise in the logs
2020-10-22 09:05:46 +02:00
Norman Maurer
ad8fe88abd
We should have a special config that allows to configure half closure for DuplexChannel (#10701)
Motivation:

DuplexChannel allow for half-closure, we should have a special config interface for it as well.

Modifications:

Add DuplexChannelConfig which allows to configure half-closure.

Result:

More consistent types
2020-10-21 15:26:27 +02:00
Stuart Douglas
303baf5c0b
Minor performance improvement in websocket upgrade (#10710)
Motivation:

I noticed WebSocketServerExtensionHandler taking up a non-trivial
amount of CPU time for a non-websocket based menchmark. This attempts
to speed it up.

Modifications:

- It is faster to check for a 101 response than to look at headers,
so an initial response code check is done
- Move all the actual upgrade code into its own method to increase
chance of this method being inlined
- Add an extra contains() check for the upgrade header, to avoid
allocating an iterator if there is no upgrade header

Result:

A small but noticable performance increase.

Signed-off-by: Stuart Douglas <stuart.w.douglas@gmail.com>
2020-10-21 12:09:32 +02:00
Norman Maurer
ffbddcd842
Replace deprecated Assert.assertThat(...) with MatcherAssert.assertThat(...) (#10699)
Motivation:

junit deprecated Assert.assertThat(...)

Modifications:

Use MatcherAssert.assertThat(...) as replacement for deprecated method

Result:

Less deprecation warnings
2020-10-18 14:30:52 +02:00
Norman Maurer
ea2a0899de
Create codeql-analysis.yml (#10696)
Motivation:

Github now allows to run CodeQL during pull request verification. This allows to detect errors / security problems early.

Modification:

Add config

Result:

Fixes https://github.com/netty/netty/issues/10669


Co-authored-by: Artem Smotrakov <artem.smotrakov@sap.com>
2020-10-18 14:25:48 +02:00
Norman Maurer
5924262175
Fix ByteBuf leaks in HaProxyMessageEncoderTest (#10704)
Motivation:

We need to ensure we not leak in tests. We did see some leaks reported related to HaProxyMessageEncoderTest on our CI.

Modifications:

- Use readSlice(...) and so not create new ByteBuf instances that need to be released

Result:

No more leaks
2020-10-18 14:10:12 +02:00
Norman Maurer
f9414d129f
Update java patch versions (#10703)
Motivation:

We should use the latest patch releases when building via docker

Modifications:

Update all java versions to the latest patch release

Result:

Use latest releases
2020-10-17 20:24:18 +02:00
Aayush Atharva
e051b5f338
Add null rule check in rules array of RuleBasedIpFilter (#10527)
Motivation:

We can filter out `null` rules while initializing the instance of `RuleBasedIpFilter` so we don't have to keep checking for `null` rules while iterating through `rules` array in `for loop` which is just a waste of CPU cycles.

Modification:
Added `null` rule check inside the constructor.

Result:
No more wasting CPU cycles on check the `null` rule each time in `for loop` and makes the overall operation more faster.
2020-10-17 20:22:43 +02:00
Artem Smotrakov
1ca7d5db81
Fix or suppress LGTM findings (#10689)
Motivation:

LGTM reports multiple issues. They need to be triaged,
and real ones should be fixed.

Modifications:
- Fixed multiple issues reported by LGTM, such as redundant conditions,
  resource leaks, typos, possible integer overflows.
- Suppressed false-positives.
- Added a few testcases.

Result:

Fixed several possible issues, get rid of false alarms in the LGTM report.
2020-10-17 09:49:44 +02:00
Norman Maurer
5e808eb9a5
Add missing -XX:+AllowRedefinitionToAddDeleteMethods for blockhound testsuite (#10700)
Motivation:

We need to also add -XX:+AllowRedefinitionToAddDeleteMethods for JDK15 and 16 as otherwise blockhound will not work

Modifications:

Add profiles for JDK15 and JDK16

Result:

Blockhound tests pass again
2020-10-17 09:15:49 +02:00
Norman Maurer
cc79f5f4ff
Fix unit tests that sometimes failed due timeouts (#10698)
Motivation:

We had two unit tests that sometimes failed due timeouts. After insepecting these I noticed these can be improved to run faster while still do the right validation

Modifications:

- Only submit one task for execution per execute
- Cleanup

Result:

No test failures due timeout
2020-10-16 21:00:26 +02:00
Norman Maurer
fbb42a292f
Add docker-compose files to compile with OpenJDK15 (#10697)
Motivation:

OpenJDK15 was released, we should compile with it on the CI

Modifications:

Add docker-compose files to be able to compile with OpenJDK15

Result:

Compile with latest major JDK version
2020-10-16 11:33:06 +02:00
Norman Maurer
9563d2bd61
Cleanup PoolChunk / PoolSubpage and add a few more asserts (#10690)
Motivation:

As the PooledByteBufAllocator is a critical part of netty we should ensure it works as expected.

Modifications:

- Add a few more asserts to ensure we not see any corrupted state
- Null out slot in the subpage array once the subpage was freed and removed from the pool
- Merge methods into constructor as it was only called from the constructor anyway.

Result:

Code cleanup
2020-10-15 21:01:57 +02:00
Norman Maurer
0ed788c15e
Fire SniCompletionEvent after onLookupComplete(...) was called (#10688)
Motivation:

Users may want to do special actions when onComplete(...) was called and depend on these once they receive the SniCompletionEvent

Modifications:

Switch order and so call onLookupComplete(...) before we fire the event

Result:

Fixes https://github.com/netty/netty/issues/10655
2020-10-15 21:01:03 +02:00
Norman Maurer
6b613682ba
Ensure we don't leak the ClassLoader in the backtrace (#10691)
Motivation:

We have a few classes in which we store and reuse static instances of various exceptions. When doing so it is important to also override fillInStacktrace() and so prevent the leak of the ClassLoader in the internal backtrace field.

Modifications:

- Add overrides of fillInStracktrace when needed
- Move ThrowableUtil usage in the static methods

Result:

Fixes https://github.com/netty/netty/pull/10686
2020-10-15 20:41:29 +02:00
Artem Smotrakov
a3c154f267
Added a security policy (#10692)
Motivation:

The process of reporting security issues should be documented
and easy to find.

Modification:

Added a SECURITY.md file that describes how to report a security issue.

Result:

It's a bit easier to find the docs that describe how security issues
should be reported. Also, when someone creates an issue the repository,
they will see a link to the security policy.
2020-10-15 20:39:37 +02:00
Aayush Atharva
08ff3442ec
Update OWASP Links in Cookie class (#10677)
Motivation:
Fix Broken Link of OWASP HttpOnly Cookie in Cookie class.

Modification:
Updated the broken link.

Result:
Broken Link Fix for better Documentation.
2020-10-15 14:08:36 +02:00
Aayush Atharva
6cb72be2dd
Use ObjectUtil for multiple operations (#10679)
Motivation:
We should use ObjectUtil for checking if Compression parameters are in range. This will reduce LOC and make code more readable.

Modification:
Used ObjectUtil

Result:
More readable code
2020-10-14 12:01:42 +02:00
dependabot[bot]
36f682141f
Bump junit from 4.12 to 4.13.1 (#10681)
Bumps [junit](https://github.com/junit-team/junit4) from 4.12 to 4.13.1.
- [Release notes](https://github.com/junit-team/junit4/releases)
- [Changelog](https://github.com/junit-team/junit4/blob/main/doc/ReleaseNotes4.12.md)
- [Commits](https://github.com/junit-team/junit4/compare/r4.12...r4.13.1)

Signed-off-by: dependabot[bot] <support@github.com>

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2020-10-14 12:00:12 +02:00
Chris Vest
fd8c1874b4
Fix #10614 by making UnorderedTPEExecutor.scheduleAtFixedRate run tasks more than once (#10659)
Motivation:
All scheduled executors should behave in accordance to their API.
The bug here is that scheduled tasks were not run more than once because we executed the runnables directly, instead of through the provided runnable future.

Modification:
We now run tasks through the provided future, so that when each run completes, the internal state of the task is reset and the ScheduledThreadPoolExecutor is informed of the completion.
This allows the executor to prepare the next run.

Result:
The UnorderedThreadPoolEventExecutor is now able to run scheduled tasks more than once.
Which is what one would expect from the API.
2020-10-14 11:09:16 +02:00
Chris Vest
cd1581faad
Add a build profile for JDK 16 (#10676)
Motivation:
Java 16 will come around eventually anyway, and this makes it easier for people to experiment with Early Access builds.

Modification:
Added Maven profiles for JDK 16 to relevant pom files.

Result:
Netty now builds on JDK 16 pre-releases (provided they've not broken compatibility in some way).
2020-10-13 14:24:58 +02:00
root
01b7e18632 [maven-release-plugin] prepare for next development iteration 2020-10-13 06:29:26 +00:00
root
d4a0050ef3 [maven-release-plugin] prepare release netty-4.1.53.Final 2020-10-13 06:29:02 +00:00
Aayush Atharva
9d457c3f0f
Add checkInRange in ObjectUtil (#10668)
Motivation:
We check lots of numbers if it lies in a range. So it's better to add a method in `ObjectUtil` to check if a number lies inside a range.

Modification:
Added Range check method.

Result:
A faster and better way to check if a number lies inside a range.
2020-10-12 18:25:58 +02:00
Aayush Atharva
4c231a2e77
Simplify and Remove useless Bit operations from HpackHuffmanDecoder (#10656)
Motivation:
We should simplify and remove useless bit operations to make code more efficient, faster, and easier to understand.

Modification:
Simplified and Removed Useless Bit Operations

Result:
Simpler Code.
2020-10-12 11:44:44 +02:00
Artem Smotrakov
5b55ee434b
Avoid casting numbers to narrower types (#10645)
Motivation:

Avoid implicit conversions to narrower types in
AbstractMemoryHttpData and Bzip2HuffmanStageEncoder classes
reported by LGTM.

Modifications:

Updated the classes to avoid implicit casting to narrower types.
It doesn't look like that an integer overflow is possible there,
therefore no checks for overflows were added.

Result:

No warnings about implicit conversions to narrower types.
2020-10-12 09:33:47 +02:00
Artem Smotrakov
00f21845f8
Suppress warnings about weak hash algorithms (#10647)
Motivation:

LGTM reported that WebSocketUtil uses MD5 and SHA-1
that are considered weak. Although those algorithms
are insecure, they are required by draft-ietf-hybi-thewebsocketprotocol-00
specification that is implemented in the corresponding WebSocket
handshakers. Once the handshakers are removed, WebSocketUtil can be
updated to stop using those weak hash functions.

Modifications:

Added SuppressWarnings annotations.

Result:

Suppressed warnings.
2020-10-12 09:24:17 +02:00