Adopt to JDK22

.github/workflows/ci.yml

@@ -24,7 +24,7 @@ jobs:
         uses: actions/setup-java@v3
         with:
           distribution: oracle
-          java-version: 21
+          java-version: 22
           cache: maven
 
       - name: Check CPU features for Maven tests

.github/workflows/codeql-analysis.yml

@@ -28,7 +28,7 @@ jobs:
         uses: actions/setup-java@v3
         with:
           distribution: oracle
-          java-version: 21
+          java-version: 22
           cache: maven
 
       - name: Initialize CodeQL

.github/workflows/release.yml

@@ -18,7 +18,7 @@ jobs:
         uses: actions/setup-java@v3
         with:
           distribution: oracle
-          java-version: 21
+          java-version: 22
           cache: maven
       - name: Publish to GitHub Packages via Apache Maven
         run: mvn -B -DskipTests deploy javadoc:javadoc

README.md

@@ -5,15 +5,15 @@ ffmasm
 ![CodeQL](../../actions/workflows/codeql-analysis.yml/badge.svg)
 
 ffmasm is an assembler for hand-assembling from Java.  
-It uses Foreign Function & Memory API, so the application can call assembled code via [MethodHandle](https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/lang/invoke/MethodHandle.html).
+It uses Foreign Function & Memory API, so the application can call assembled code via [MethodHandle](https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/invoke/MethodHandle.html).
 
 * Javadoc: https://yasuenag.github.io/ffmasm/
 * Maven package: https://github.com/YaSuenag/ffmasm/packages/
 * Supported instructions: See builder classes in [com.yasuenag.ffmasm.amd64](https://yasuenag.github.io/ffmasm/com.yasuenag.ffmasm/com/yasuenag/ffmasm/amd64/package-summary.html).
 
 # Requirements
 
-Java 21
+Java 22
 
 # Supported platform
 
@@ -45,7 +45,7 @@ See [Javadoc](https://yasuenag.github.io/ffmasm/) and [cpumodel](examples/cpumod
 ## 1. Create `CodeSegment`
 
 `CodeSegment` is a storage for assembled code. In Linux, it would be allocated by `mmap(2)` with executable bit.  
-It implements [AutoCloseable](https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/lang/AutoCloseable.html), so you can use try-with-resources in below:
+It implements [AutoCloseable](https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/AutoCloseable.html), so you can use try-with-resources in below:
 
 ```java
 try(var seg = new CodeSegment()){
@@ -55,7 +55,7 @@ try(var seg = new CodeSegment()){
 
 ## 2. Create `MethodHandle` via `AMD64AsmBuilder`
 
-You can assemble the code via `AMD64AsmBuilder`. It would be instanciated via `create()`, and it should be passed both `CodeSegment` and [FunctionDescriptor](https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/lang/foreign/FunctionDescriptor.html).
+You can assemble the code via `AMD64AsmBuilder`. It would be instanciated via `create()`, and it should be passed both `CodeSegment` and [FunctionDescriptor](https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/foreign/FunctionDescriptor.html).
 
 In following example, the method is defined as `(I)I` (JNI signature) in `FunctionDescriptor`.  
 `AMD64AsmBuilder` is builder pattern, so you can add instruction in below. Following example shows method argument (`int`) would be returned straightly.
@@ -74,10 +74,10 @@ var method = AMD64AsmBuilder.create(seg, desc)
     /* mov %rdi, %rax    */ .movRM(Register.RDI, Register.RAX, OptionalInt.empty())
     /* leave             */ .leave()
     /* ret               */ .ret()
-                            .build(Linker.Option.isTrivial());
+                            .build(Linker.Option.critical(false));
 ```
 
-NOTE: [Linker.Option.isTrivial()](https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/lang/foreign/Linker.Option.html#isTrivial()) is recommended to pass `build()` method due to performance, but it might be cause of some issues in JVM (time to synchronize safepoint, memory corruption, etc). See Javadoc of `isTrivial()`.
+NOTE: [Linker.Option.critical()](https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/foreign/Linker.Option.html#critical(boolean)) is recommended to pass `build()` method due to performance, but it might be cause of some issues in JVM (time to synchronize safepoint, memory corruption, etc). See Javadoc of `critical()`.
 
 ## 3. Method call
 
@@ -126,26 +126,6 @@ try(var seg = new CodeSegment()){
 }
 ```
 
-# Memory pinning
-
-You can pin Java primitive array via `Pinning`. It is same semantics of `GetPrimitiveArrayCritical()` / `ReleasePrimitiveArrayCritical()` in JNI. It expects performance improvement when you refer Java array in ffmasm code. However it might be serious problem (e.g. preventing GC) if you keep pinned memory long time. So you should use it carefully.
-
-Following example shows pin `array` at first, then we modify value via `MemorySegment`.
-
-```java
-int[] array = new int[]{1, 2, 3, 4};
-
-var pinnedMem = Pinning.getInstance()
-                       .pin(array)
-                       .reinterpret(ValueLayout.JAVA_INT.byteSize() * array.length);
-for(int idx = 0; idx < expected.length; idx++){
-  pinnedMem.setAtIndex(ValueLayout.JAVA_INT, idx, idx * 2);
-}
-Pinning.getInstance().unpin(pinnedMem);
-
-// array is {0, 2, 4, 6} at this point
-```
-
 # License
 
 The GNU Lesser General Public License, version 3.0

benchmarks/funccall/README.md

@@ -7,7 +7,7 @@ This benchmark runs RDTSC instruction. JNI function is written in assembly code
 
 # Requirements
 
-* Java 21
+* Java 22
 * Maven
 * GNU Make
 * GNU assembler

benchmarks/funccall/pom.xml

@@ -11,9 +11,9 @@
 
     <properties>
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
-        <maven.compiler.source>21</maven.compiler.source>
-        <maven.compiler.target>21</maven.compiler.target>
-        <jmh.version>1.35</jmh.version>
+        <maven.compiler.source>22</maven.compiler.source>
+        <maven.compiler.target>22</maven.compiler.target>
+        <jmh.version>1.37</jmh.version>
     </properties>
 
     <dependencies>
@@ -22,12 +22,6 @@
             <artifactId>jmh-core</artifactId>
             <version>${jmh.version}</version>
         </dependency>
-        <dependency>
-            <groupId>org.openjdk.jmh</groupId>
-            <artifactId>jmh-generator-annprocess</artifactId>
-            <version>${jmh.version}</version>
-            <scope>provided</scope>
-        </dependency>
         <dependency>
             <groupId>com.yasuenag</groupId>
             <artifactId>ffmasm</artifactId>
@@ -46,6 +40,13 @@
                         <arg>-Xlint:all</arg>
                     </compilerArgs>
                     <enablePreview>true</enablePreview>
+                    <annotationProcessorPaths>
+                        <path>
+                            <groupId>org.openjdk.jmh</groupId>
+                            <artifactId>jmh-generator-annprocess</artifactId>
+                            <version>${jmh.version}</version>
+                        </path>
+                    </annotationProcessorPaths>
                 </configuration>
             </plugin>
             <plugin>
@@ -76,7 +77,6 @@
                                 <argument>-classpath</argument>
                                 <classpath/>
                                 <argument>-Djava.library.path=.</argument>
-                                <argument>--enable-preview</argument>
                                 <argument>--enable-native-access=ALL-UNNAMED</argument>
                                 <argument>com.yasuenag.ffmasm.benchmark.funccall.FuncCallComparison</argument>
                             </arguments>
@@ -86,7 +86,7 @@
             </plugin>
             <plugin>
                 <artifactId>maven-shade-plugin</artifactId>
-                <version>3.4.0</version>
+                <version>3.5.1</version>
                 <configuration>
                     <createDependencyReducedPom>false</createDependencyReducedPom>
                 </configuration>

benchmarks/funccall/src/main/java/com/yasuenag/ffmasm/benchmark/funccall/FuncCallComparison.java

@@ -13,7 +13,7 @@
 
 @State(Scope.Benchmark)
 @BenchmarkMode(Mode.Throughput)
-@Fork(value = 1, jvmArgsAppend = {"--enable-preview", "--enable-native-access=ALL-UNNAMED", "-Djava.library.path=.", "-Xms4g", "-Xmx4g", "-XX:+UnlockExperimentalVMOptions", "-XX:+UseEpsilonGC", "-XX:+AlwaysPreTouch"})
+@Fork(value = 1, jvmArgsAppend = {"--enable-native-access=ALL-UNNAMED", "-Djava.library.path=.", "-Xms4g", "-Xmx4g", "-XX:+UnlockExperimentalVMOptions", "-XX:+UseEpsilonGC", "-XX:+AlwaysPreTouch"})
 @Warmup(iterations = 1, time = 3, timeUnit = TimeUnit.SECONDS)
 @Measurement(iterations = 3, time = 10, timeUnit = TimeUnit.SECONDS)
 public class FuncCallComparison{
@@ -39,7 +39,7 @@ public void setup(){
            /* ret            */ .ret()
                                 .getMemorySegment();
 
-      ffmRDTSC = Linker.nativeLinker().downcallHandle(mem, desc, Linker.Option.isTrivial());
+      ffmRDTSC = Linker.nativeLinker().downcallHandle(mem, desc, Linker.Option.critical(false));
 
       var register = NativeRegister.create(this.getClass());
       register.registerNatives(Map.of(this.getClass().getMethod("rdtscFFMDirect"), mem));

benchmarks/vectorapi/README.md

@@ -7,7 +7,7 @@ This benchmark adds with 256bit packed int. Pure Java and Vector API are expecte
 
 # Requirements
 
-* Java 21
+* Java 22
 * Maven
 
 # How to build
@@ -22,5 +22,5 @@ $ mvn package
 # Run benchmark
 
 ```sh
-$ $JAVA_HOME/bin/java -jar target/ffmasm-benchmark-vectorapi-1.0.3.jar
+$ $JAVA_HOME/bin/java -jar target/ffmasm-benchmark-vectorapi-1.0.4.jar
 ```

benchmarks/vectorapi/pom.xml

@@ -7,13 +7,13 @@
     <groupId>com.yasuenag</groupId>
     <artifactId>ffmasm-benchmark-vectorapi</artifactId>
     <packaging>jar</packaging>
-    <version>1.0.3</version>
+    <version>1.0.4</version>
 
     <properties>
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
-        <maven.compiler.source>21</maven.compiler.source>
-        <maven.compiler.target>21</maven.compiler.target>
-        <jmh.version>1.35</jmh.version>
+        <maven.compiler.source>22</maven.compiler.source>
+        <maven.compiler.target>22</maven.compiler.target>
+        <jmh.version>1.37</jmh.version>
     </properties>
 
     <dependencies>
@@ -22,12 +22,6 @@
             <artifactId>jmh-core</artifactId>
             <version>${jmh.version}</version>
         </dependency>
-        <dependency>
-            <groupId>org.openjdk.jmh</groupId>
-            <artifactId>jmh-generator-annprocess</artifactId>
-            <version>${jmh.version}</version>
-            <scope>provided</scope>
-        </dependency>
         <dependency>
             <groupId>com.yasuenag</groupId>
             <artifactId>ffmasm</artifactId>
@@ -49,7 +43,13 @@
                         <arg>--add-exports</arg>
                         <arg>jdk.incubator.vector/jdk.incubator.vector=ALL-UNNAMED</arg>
                     </compilerArgs>
-                    <enablePreview>true</enablePreview>
+                    <annotationProcessorPaths>
+                        <path>
+                            <groupId>org.openjdk.jmh</groupId>
+                            <artifactId>jmh-generator-annprocess</artifactId>
+                            <version>${jmh.version}</version>
+                        </path>
+                    </annotationProcessorPaths>
                 </configuration>
             </plugin>
             <plugin>

benchmarks/vectorapi/src/main/java/com/yasuenag/ffmasm/benchmark/vectorapi/VectorOpComparison.java

@@ -15,7 +15,7 @@
 
 @State(Scope.Benchmark)
 @BenchmarkMode(Mode.Throughput)
-@Fork(value = 1, jvmArgsAppend = {"--enable-preview", "--enable-native-access=ALL-UNNAMED", "--add-modules", "jdk.incubator.vector", "--add-exports", "jdk.incubator.vector/jdk.incubator.vector=ALL-UNNAMED", "-Xms6g", "-Xmx6g"})
+@Fork(value = 1, jvmArgsAppend = {"--enable-native-access=ALL-UNNAMED", "--add-modules", "jdk.incubator.vector", "--add-exports", "jdk.incubator.vector/jdk.incubator.vector=ALL-UNNAMED", "-Xms6g", "-Xmx6g"})
 @Warmup(iterations = 1, time = 3, timeUnit = TimeUnit.SECONDS)
 @Measurement(iterations = 3, time = 10, timeUnit = TimeUnit.SECONDS)
 public class VectorOpComparison{
@@ -27,6 +27,8 @@ public class VectorOpComparison{
   private MethodHandle ffm;
   private MemorySegment srcSeg;
   private MemorySegment destSeg;
+  private MemorySegment pinnedSrcSeg;
+  private MemorySegment pinnedDestSeg;
   private MethodHandle ffmPinned;
 
   private IntVector vectorSrc;
@@ -47,7 +49,7 @@ public void setup(){
 /* vmovdqa %ymm0, (%rdi)       */ .vmovdqaRM(Register.YMM0, Register.RDI, OptionalInt.of(0))
 /* leave                       */ .leave()
 /* ret                         */ .ret()
-                                  .build(Linker.Option.isTrivial());
+                                  .build(Linker.Option.critical(false));
 
       var descPinned = FunctionDescriptor.ofVoid(ValueLayout.ADDRESS, ValueLayout.ADDRESS);
       ffmPinned = AMD64AsmBuilder.create(AVXAsmBuilder.class, seg, descPinned)
@@ -59,7 +61,7 @@ public void setup(){
 /* vmovdqu %ymm0, (%rdi)       */ .vmovdquRM(Register.YMM0, Register.RDI, OptionalInt.of(0))
 /* leave                       */ .leave()
 /* ret                         */ .ret()
-                                  .build(Linker.Option.isTrivial());
+                                  .build(Linker.Option.critical(true));
     }
     catch(PlatformException | UnsupportedPlatformException e){
       throw new RuntimeException(e);
@@ -80,49 +82,13 @@ public void paramSetup(){
     MemorySegment.copy(randArray, 0, srcSeg, ValueLayout.JAVA_INT, 0, 8);
     MemorySegment.copy(result, 0, destSeg, ValueLayout.JAVA_INT, 0, 8);
 
+    pinnedSrcSeg = MemorySegment.ofArray(randArray);
+    pinnedDestSeg = MemorySegment.ofArray(result);
+
     vectorSrc = IntVector.fromArray(IntVector.SPECIES_256, randArray, 0);
     vectorDest = IntVector.fromArray(IntVector.SPECIES_256, result, 0);
   }
 
-  @State(Scope.Benchmark)
-  public static class PinnedState{
-
-    private static final Pinning pinning;
-
-    private MemorySegment pinnedSrcSeg;
-    private MemorySegment pinnedDestSeg;
-
-    static{
-      try{
-        pinning = Pinning.getInstance();
-      }
-      catch(Throwable t){
-        throw new RuntimeException(t);
-      }
-    }
-
-    @Setup(Level.Iteration)
-    public void setup(){
-      pinnedSrcSeg = pinning.pin(randArray);
-      pinnedDestSeg = pinning.pin(result);
-    }
-
-    @TearDown(Level.Iteration)
-    public void tearDownInIterate(){
-      pinning.unpin(pinnedSrcSeg);
-      pinning.unpin(pinnedDestSeg);
-    }
-
-    public MemorySegment getPinnedSrcSeg(){
-      return pinnedSrcSeg;
-    }
-
-    public MemorySegment getPinnedDestSeg(){
-      return pinnedDestSeg;
-    }
-
-  }
-
   @Benchmark
   public int[] invokeJava(){
     for(int i = 0; i < 8; i++){
@@ -143,9 +109,9 @@ public int[] invokeFFM(){
   }
 
   @Benchmark
-  public int[] invokePinnedFFM(PinnedState state){
+  public int[] invokePinnedFFM(){
     try{
-      ffmPinned.invoke(state.getPinnedDestSeg(), state.getPinnedSrcSeg());
+      ffmPinned.invoke(pinnedDestSeg, pinnedSrcSeg);
       return result;
     }
     catch(Throwable t){

examples/cpumodel/README.md

@@ -5,7 +5,7 @@ This is an example of ffmasm to obtain CPU model from `CPUID` instruction in AMD
 
 # Requirements
 
-* Java 21
+* Java 22
 * AMD64 Linux or Windows
 * Maven
 
@@ -27,7 +27,7 @@ $ mvn clean package
 ## 3. Run via Maven
 
 ```
-$ mvn exec:exec
+$ mvn exec:java
 ```
 
 Press enter to exit.