8340586: JdkJfrEvent::get_all_klasses stores non-strong oops in JNI handles

Backport-of: d620426835a15fe75c42758567a7dab4ef0eb772

.github/workflows/build-cross-compile.yml

@@ -84,7 +84,7 @@ jobs:
           - target-cpu: riscv64
             gnu-arch: riscv64
             debian-arch: riscv64
-            debian-repository: https://httpredir.debian.org/debian/
+            debian-repository: https://snapshot.debian.org/archive/debian/20240228T034848Z/
             debian-version: sid
             tolerate-sysroot-errors: true
 

.github/workflows/main.yml

@@ -36,7 +36,7 @@ on:
       platforms:
         description: 'Platform(s) to execute on (comma separated, e.g. "linux-x64, macos, aarch64")'
         required: true
-        default: 'linux-x64, linux-x86, linux-x64-variants, linux-cross-compile, macos-x64, macos-aarch64, windows-x64, windows-aarch64, docs'
+        default: 'linux-x64, linux-x86-hs, linux-x64-variants, linux-cross-compile, macos-x64, macos-aarch64, windows-x64, windows-aarch64, docs'
       configure-arguments:
         description: 'Additional configure arguments'
         required: false
@@ -59,7 +59,7 @@ jobs:
     runs-on: ubuntu-22.04
     outputs:
       linux-x64: ${{ steps.include.outputs.linux-x64 }}
-      linux-x86: ${{ steps.include.outputs.linux-x86 }}
+      linux-x86-hs: ${{ steps.include.outputs.linux-x86-hs }}
       linux-x64-variants: ${{ steps.include.outputs.linux-x64-variants }}
       linux-cross-compile: ${{ steps.include.outputs.linux-cross-compile }}
       macos-x64: ${{ steps.include.outputs.macos-x64 }}
@@ -111,7 +111,7 @@ jobs:
           }
 
           echo "linux-x64=$(check_platform linux-x64 linux x64)" >> $GITHUB_OUTPUT
-          echo "linux-x86=$(check_platform linux-x86 linux x86)" >> $GITHUB_OUTPUT
+          echo "linux-x86-hs=$(check_platform linux-x86-hs linux x86)" >> $GITHUB_OUTPUT
           echo "linux-x64-variants=$(check_platform linux-x64-variants variants)" >> $GITHUB_OUTPUT
           echo "linux-cross-compile=$(check_platform linux-cross-compile cross-compile)" >> $GITHUB_OUTPUT
           echo "macos-x64=$(check_platform macos-x64 macos x64)" >> $GITHUB_OUTPUT
@@ -135,12 +135,13 @@ jobs:
       make-arguments: ${{ github.event.inputs.make-arguments }}
     if: needs.select.outputs.linux-x64 == 'true'
 
-  build-linux-x86:
-    name: linux-x86
+  build-linux-x86-hs:
+    name: linux-x86-hs
     needs: select
     uses: ./.github/workflows/build-linux.yml
     with:
       platform: linux-x86
+      make-target: 'hotspot'
       gcc-major-version: '10'
       gcc-package-suffix: '-multilib'
       apt-architecture: 'i386'
@@ -150,7 +151,7 @@ jobs:
       extra-conf-options: '--with-target-bits=32 --enable-fallback-linker --enable-libffi-bundling'
       configure-arguments: ${{ github.event.inputs.configure-arguments }}
       make-arguments: ${{ github.event.inputs.make-arguments }}
-    if: needs.select.outputs.linux-x86 == 'true'
+    if: needs.select.outputs.linux-x86-hs == 'true'
 
   build-linux-x64-hs-nopch:
     name: linux-x64-hs-nopch
@@ -300,16 +301,6 @@ jobs:
       bootjdk-platform: linux-x64
       runs-on: ubuntu-22.04
 
-  test-linux-x86:
-    name: linux-x86
-    needs:
-      - build-linux-x86
-    uses: ./.github/workflows/test.yml
-    with:
-      platform: linux-x86
-      bootjdk-platform: linux-x64
-      runs-on: ubuntu-22.04
-
   test-macos-x64:
     name: macos-x64
     needs:
@@ -347,7 +338,7 @@ jobs:
     if: always()
     needs:
       - build-linux-x64
-      - build-linux-x86
+      - build-linux-x86-hs
       - build-linux-x64-hs-nopch
       - build-linux-x64-hs-zero
       - build-linux-x64-hs-minimal
@@ -358,31 +349,28 @@ jobs:
       - build-windows-x64
       - build-windows-aarch64
       - test-linux-x64
-      - test-linux-x86
       - test-macos-x64
+      - test-macos-aarch64
       - test-windows-x64
 
     steps:
-        # Hack to get hold of the api environment variables that are only defined for actions
-      - name: 'Get API configuration'
-        id: api
-        uses: actions/github-script@v7
-        with:
-          script: 'return { url: process.env["ACTIONS_RUNTIME_URL"], token: process.env["ACTIONS_RUNTIME_TOKEN"] }'
-
       - name: 'Remove bundle artifacts'
         run: |
           # Find and remove all bundle artifacts
-          ALL_ARTIFACT_URLS="$(curl -s \
-              -H 'Accept: application/json;api-version=6.0-preview' \
-              -H 'Authorization: Bearer ${{ fromJson(steps.api.outputs.result).token }}' \
-              '${{ fromJson(steps.api.outputs.result).url }}_apis/pipelines/workflows/${{ github.run_id }}/artifacts?api-version=6.0-preview')"
-          BUNDLE_ARTIFACT_URLS="$(echo "$ALL_ARTIFACT_URLS" | jq -r -c '.value | map(select(.name|startswith("bundles-"))) | .[].url')"
-          for url in $BUNDLE_ARTIFACT_URLS; do
-            echo "Removing $url"
-            curl -s \
-                -H 'Accept: application/json;api-version=6.0-preview' \
-                -H 'Authorization: Bearer ${{ fromJson(steps.api.outputs.result).token }}' \
-                -X DELETE "$url" \
+          # See: https://docs.github.com/en/rest/actions/artifacts?apiVersion=2022-11-28
+          ALL_ARTIFACT_IDS="$(curl -sL \
+              -H 'Accept: application/vnd.github+json' \
+              -H 'Authorization: Bearer ${{ github.token }}' \
+              -H 'X-GitHub-Api-Version: 2022-11-28' \
+              '${{ github.api_url }}/repos/${{ github.repository }}/actions/runs/${{ github.run_id }}/artifacts?per_page=100')"
+          BUNDLE_ARTIFACT_IDS="$(echo "$ALL_ARTIFACT_IDS" | jq -r -c '.artifacts | map(select(.name|startswith("bundles-"))) | .[].id')"
+          for id in $BUNDLE_ARTIFACT_IDS; do
+            echo "Removing $id"
+            curl -sL \
+                -X DELETE \
+                -H 'Accept: application/vnd.github+json' \
+                -H 'Authorization: Bearer ${{ github.token }}' \
+                -H 'X-GitHub-Api-Version: 2022-11-28' \
+                "${{ github.api_url }}/repos/${{ github.repository }}/actions/artifacts/$id" \
             || echo "Failed to remove bundle"
           done

.jcheck/conf

@@ -1,15 +1,15 @@
 [general]
-project=jdk
+project=jdk-updates
 jbs=JDK
-version=23
+version=23.0.2
 
 [checks]
 error=author,committer,reviewers,merge,issues,executable,symlink,message,hg-tag,whitespace,problemlists
-warning=issuestitle
+warning=issuestitle,binary
 
 [repository]
 tags=(?:jdk-(?:[1-9]([0-9]*)(?:\.(?:0|[1-9][0-9]*)){0,4})(?:\+(?:(?:[0-9]+))|(?:-ga)))|(?:jdk[4-9](?:u\d{1,3})?-(?:(?:b\d{2,3})|(?:ga)))|(?:hs\d\d(?:\.\d{1,2})?-b\d\d)
-branches=
+branches=.*
 
 [census]
 version=0

SECURITY.md

@@ -0,0 +1,3 @@
+# JDK Vulnerabilities
+
+Please follow the process outlined in the [OpenJDK Vulnerability Policy](https://openjdk.org/groups/vulnerability/report) to disclose vulnerabilities in the JDK.

doc/building.html

@@ -614,10 +614,9 @@ be accepted by <code>configure</code>.</p>
 <code>--with-toolchain-type=clang</code>.</p>
 <h3 id="apple-xcode">Apple Xcode</h3>
 <p>The oldest supported version of Xcode is 13.0.</p>
-<p>You will need the Xcode command line developer tools to be able to
-build the JDK. (Actually, <em>only</em> the command line tools are
-needed, not the IDE.) The simplest way to install these is to run:</p>
-<pre><code>xcode-select --install</code></pre>
+<p>You will need to download Xcode either from the App Store or specific
+versions can be easily located via the <a
+href="https://xcodereleases.com">Xcode Releases</a> website.</p>
 <p>When updating Xcode, it is advisable to keep an older version for
 building the JDK. To use a specific version of Xcode you have multiple
 options:</p>

doc/building.md

@@ -422,13 +422,9 @@ To use clang instead of gcc on Linux, use `--with-toolchain-type=clang`.
 
 The oldest supported version of Xcode is 13.0.
 
-You will need the Xcode command line developer tools to be able to build the
-JDK. (Actually, *only* the command line tools are needed, not the IDE.) The
-simplest way to install these is to run:
-
-```
-xcode-select --install
-```
+You will need to download Xcode either from the App Store or specific versions
+can be easily located via the [Xcode Releases](https://xcodereleases.com)
+website.
 
 When updating Xcode, it is advisable to keep an older version for building the
 JDK. To use a specific version of Xcode you have multiple options:

make/autoconf/configure.ac

@@ -313,9 +313,11 @@ AC_OUTPUT
 
 # After AC_OUTPUT, we need to do final work
 CUSTOM_CONFIG_OUTPUT_GENERATED_HOOK
-BASIC_POST_CONFIG_OUTPUT
 
 # Finally output some useful information to the user
 HELP_PRINT_SUMMARY_AND_WARNINGS
 CUSTOM_SUMMARY_AND_WARNINGS_HOOK
 HELP_REPEAT_WARNINGS
+
+# All output is done. Do the post-config output management.
+BASIC_POST_CONFIG_OUTPUT

make/autoconf/jdk-options.m4

@@ -496,9 +496,15 @@ AC_DEFUN_ONCE([JDKOPT_SETUP_LEAK_SANITIZER],
 #
 AC_DEFUN_ONCE([JDKOPT_SETUP_UNDEFINED_BEHAVIOR_SANITIZER],
 [
+  UTIL_ARG_WITH(NAME: additional-ubsan-checks, TYPE: string,
+      DEFAULT: [],
+      DESC: [Customizes the ubsan checks],
+      OPTIONAL: true)
+
   # GCC reports lots of likely false positives for stringop-truncation and format-overflow.
   # Silence them for now.
-  UBSAN_CHECKS="-fsanitize=undefined -fsanitize=float-divide-by-zero -fno-sanitize=shift-base -fno-sanitize=alignment"
+  UBSAN_CHECKS="-fsanitize=undefined -fsanitize=float-divide-by-zero -fno-sanitize=shift-base -fno-sanitize=alignment \
+      $ADDITIONAL_UBSAN_CHECKS"
   UBSAN_CFLAGS="$UBSAN_CHECKS -Wno-stringop-truncation -Wno-format-overflow -fno-omit-frame-pointer -DUNDEFINED_BEHAVIOR_SANITIZER"
   UBSAN_LDFLAGS="$UBSAN_CHECKS"
   UTIL_ARG_ENABLE(NAME: ubsan, DEFAULT: false, RESULT: UBSAN_ENABLED,

make/autoconf/lib-alsa.m4

@@ -70,6 +70,25 @@ AC_DEFUN_ONCE([LIB_SETUP_ALSA],
         PKG_CHECK_MODULES(ALSA, alsa, [ALSA_FOUND=yes], [ALSA_FOUND=no])
       fi
     fi
+    if test "x$ALSA_FOUND" = xno; then
+      # If we have sysroot set, and no explicit library location is set,
+      # look at known locations in sysroot.
+      if test "x$SYSROOT" != "x" && test "x${with_alsa_lib}" == x; then
+        if test -f "$SYSROOT/usr/lib64/libasound.so" && test "x$OPENJDK_TARGET_CPU_BITS" = x64; then
+          ALSA_LIBS="-L$SYSROOT/usr/lib64 -lasound"
+          ALSA_FOUND=yes
+        elif test -f "$SYSROOT/usr/lib/libasound.so"; then
+          ALSA_LIBS="-L$SYSROOT/usr/lib -lasound"
+          ALSA_FOUND=yes
+        elif test -f "$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI/libasound.so"; then
+          ALSA_LIBS="-L$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI -lasound"
+          ALSA_FOUND=yes
+        elif test -f "$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU_AUTOCONF-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI/libasound.so"; then
+          ALSA_LIBS="-L$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU_AUTOCONF-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI -lasound"
+          ALSA_FOUND=yes
+        fi
+      fi
+    fi
     if test "x$ALSA_FOUND" = xno; then
       AC_CHECK_HEADERS([alsa/asoundlib.h],
           [

make/autoconf/lib-x11.m4

@@ -71,9 +71,9 @@ AC_DEFUN_ONCE([LIB_SETUP_X11],
           elif test -f "$SYSROOT/usr/lib/libX11.so"; then
             x_libraries="$SYSROOT/usr/lib"
           elif test -f "$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI/libX11.so"; then
-            x_libraries="$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI/libX11.so"
+            x_libraries="$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI"
           elif test -f "$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU_AUTOCONF-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI/libX11.so"; then
-            x_libraries="$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU_AUTOCONF-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI/libX11.so"
+            x_libraries="$SYSROOT/usr/lib/$OPENJDK_TARGET_CPU_AUTOCONF-$OPENJDK_TARGET_OS-$OPENJDK_TARGET_ABI"
           fi
         fi
       fi

make/autoconf/toolchain.m4

@@ -358,6 +358,11 @@ AC_DEFUN([TOOLCHAIN_EXTRACT_COMPILER_VERSION],
     #     Copyright (C) 2013 Free Software Foundation, Inc.
     #     This is free software; see the source for copying conditions.  There is NO
     #     warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+    # or look like
+    #     gcc (GCC) 10.2.1 20200825 (Alibaba 10.2.1-3.8 2.32)
+    #     Copyright (C) 2020 Free Software Foundation, Inc.
+    #     This is free software; see the source for copying conditions.  There is NO
+    #     warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
     COMPILER_VERSION_OUTPUT=`$COMPILER --version 2>&1`
     # Check that this is likely to be GCC.
     $ECHO "$COMPILER_VERSION_OUTPUT" | $GREP "Free Software Foundation" > /dev/null
@@ -371,7 +376,8 @@ AC_DEFUN([TOOLCHAIN_EXTRACT_COMPILER_VERSION],
     COMPILER_VERSION_STRING=`$ECHO $COMPILER_VERSION_OUTPUT | \
         $SED -e 's/ *Copyright .*//'`
     COMPILER_VERSION_NUMBER=`$ECHO $COMPILER_VERSION_OUTPUT | \
-        $SED -e 's/^.* \(@<:@1-9@:>@<:@0-9@:>@*\.@<:@0-9.@:>@*\)@<:@^0-9.@:>@.*$/\1/'`
+        $AWK -F ')' '{print [$]2}' | \
+        $AWK '{print [$]1}'`
   elif test  "x$TOOLCHAIN_TYPE" = xclang; then
     # clang --version output typically looks like
     #    Apple clang version 15.0.0 (clang-1500.3.9.4)

make/conf/github-actions.conf

@@ -29,17 +29,17 @@ GTEST_VERSION=1.14.0
 JTREG_VERSION=7.3.1+1
 
 LINUX_X64_BOOT_JDK_EXT=tar.gz
-LINUX_X64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22/830ec9fcccef480bb3e73fb7ecafe059/36/GPL/openjdk-22_linux-x64_bin.tar.gz
-LINUX_X64_BOOT_JDK_SHA256=4d65cc6ed28711768fd72c2043a7925f7c83f5f51bb64970bd9d52f7791fc6ac
-
-MACOS_X64_BOOT_JDK_EXT=tar.gz
-MACOS_X64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22/830ec9fcccef480bb3e73fb7ecafe059/36/GPL/openjdk-22_macos-x64_bin.tar.gz
-MACOS_X64_BOOT_JDK_SHA256=ae31fe10916429e3fe284266095067a5ce9fecbdc03ff1a079d20459f731ca36
+LINUX_X64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22.0.2/c9ecb94cd31b495da20a27d4581645e8/9/GPL/openjdk-22.0.2_linux-x64_bin.tar.gz
+LINUX_X64_BOOT_JDK_SHA256=41536f115668308ecf4eba92aaf6acaeb0936225828b741efd83b6173ba82963
 
 MACOS_AARCH64_BOOT_JDK_EXT=tar.gz
-MACOS_AARCH64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22/830ec9fcccef480bb3e73fb7ecafe059/36/GPL/openjdk-22_macos-aarch64_bin.tar.gz
-MACOS_AARCH64_BOOT_JDK_SHA256=d10f82429d01047968c52c7975c326388cb5d212791e14c1de21c987463a4b53
+MACOS_AARCH64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22.0.2/c9ecb94cd31b495da20a27d4581645e8/9/GPL/openjdk-22.0.2_macos-aarch64_bin.tar.gz
+MACOS_AARCH64_BOOT_JDK_SHA256=3dab98730234e1a87aec14bcb8171d2cae101e96ff4eed1dab96abbb08e843fd
+
+MACOS_X64_BOOT_JDK_EXT=tar.gz
+MACOS_X64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22.0.2/c9ecb94cd31b495da20a27d4581645e8/9/GPL/openjdk-22.0.2_macos-x64_bin.tar.gz
+MACOS_X64_BOOT_JDK_SHA256=e8b3ec7a7077711223d31156e771f11723cd7af31c2017f1bd2eda20855940fb
 
 WINDOWS_X64_BOOT_JDK_EXT=zip
-WINDOWS_X64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22/830ec9fcccef480bb3e73fb7ecafe059/36/GPL/openjdk-22_windows-x64_bin.zip
-WINDOWS_X64_BOOT_JDK_SHA256=8f5138fecb53c08c20abd4fa6812f9400051f3852582a2142ffda0dff73a5824
+WINDOWS_X64_BOOT_JDK_URL=https://download.java.net/java/GA/jdk22.0.2/c9ecb94cd31b495da20a27d4581645e8/9/GPL/openjdk-22.0.2_windows-x64_bin.zip
+WINDOWS_X64_BOOT_JDK_SHA256=f2a9b9ab944e71a64637fcdc6b13a1188cf02d4eb9ecf71dc927e98b3e45f5dc

make/conf/version-numbers.conf

@@ -28,15 +28,15 @@
 
 DEFAULT_VERSION_FEATURE=23
 DEFAULT_VERSION_INTERIM=0
-DEFAULT_VERSION_UPDATE=0
+DEFAULT_VERSION_UPDATE=2
 DEFAULT_VERSION_PATCH=0
 DEFAULT_VERSION_EXTRA1=0
 DEFAULT_VERSION_EXTRA2=0
 DEFAULT_VERSION_EXTRA3=0
-DEFAULT_VERSION_DATE=2024-09-17
+DEFAULT_VERSION_DATE=2025-01-21
 DEFAULT_VERSION_CLASSFILE_MAJOR=67  # "`$EXPR $DEFAULT_VERSION_FEATURE + 44`"
 DEFAULT_VERSION_CLASSFILE_MINOR=0
 DEFAULT_VERSION_DOCS_API_SINCE=11
 DEFAULT_ACCEPTABLE_BOOT_VERSIONS="22 23"
 DEFAULT_JDK_SOURCE_TARGET_VERSION=23
-DEFAULT_PROMOTED_VERSION_PRE=ea
+DEFAULT_PROMOTED_VERSION_PRE=

make/jdk/src/classes/build/tools/cldrconverter/CLDRConverter.java

@@ -786,7 +786,10 @@ public class CLDRConverter {
             String tzKey = Optional.ofNullable((String)handlerSupplMeta.get(tzid))
                                    .orElse(tzid);
             // Follow link, if needed
-            var tzLink = tzdbLinks.get(tzKey);
+            String tzLink = null;
+            for (var k = tzKey; tzdbLinks.containsKey(k);) {
+                k = tzLink = tzdbLinks.get(k);
+            }
             if (tzLink == null && tzdbLinks.containsValue(tzKey)) {
                 // reverse link search
                 // this is needed as in tzdb, "America/Buenos_Aires" links to
@@ -1214,7 +1217,7 @@ public class CLDRConverter {
     private static Set<String> getAvailableZoneIds() {
         assert handlerMetaZones != null;
         if (AVAILABLE_TZIDS == null) {
-            AVAILABLE_TZIDS = new HashSet<>(ZoneId.getAvailableZoneIds());
+            AVAILABLE_TZIDS = new HashSet<>(Arrays.asList(TimeZone.getAvailableIDs()));
             AVAILABLE_TZIDS.addAll(handlerMetaZones.keySet());
             AVAILABLE_TZIDS.remove(MetaZonesParseHandler.NO_METAZONE_KEY);
         }
@@ -1372,6 +1375,7 @@ public class CLDRConverter {
     private static void generateTZDBShortNamesMap() throws IOException {
         Files.walk(Path.of(tzDataDir), 1, FileVisitOption.FOLLOW_LINKS)
             .filter(p -> p.toFile().isFile())
+            .filter(p -> p.getFileName().toString().matches("africa|antarctica|asia|australasia|backward|etcetera|europe|northamerica|southamerica"))
             .forEach(p -> {
                 try {
                     String zone = null;
@@ -1394,43 +1398,41 @@ public class CLDRConverter {
                         }
                         // remove comments in-line
                         line = line.replaceAll("[ \t]*#.*", "");
-
+                        var tokens = line.split("[ \t]+", -1);
+                        var token0len = tokens.length > 0 ? tokens[0].length() : 0;
                         // Zone line
-                        if (line.startsWith("Zone")) {
+                        if (token0len > 0 && tokens[0].regionMatches(true, 0, "Zone", 0, token0len)) {
                             if (zone != null) {
                                 tzdbShortNamesMap.put(zone, format + NBSP + rule);
                             }
-                            var zl = line.split("[ \t]+", -1);
-                            zone = zl[1];
-                            rule = zl[3];
-                            format = flipIfNeeded(inVanguard, zl[4]);
+                            zone = tokens[1];
+                            rule = tokens[3];
+                            format = flipIfNeeded(inVanguard, tokens[4]);
                         } else {
                             if (zone != null) {
-                                if (line.startsWith("Rule") ||
-                                    line.startsWith("Link")) {
+                                if (token0len > 0 &&
+                                   (tokens[0].regionMatches(true, 0, "Rule", 0, token0len) ||
+                                    tokens[0].regionMatches(true, 0, "Link", 0, token0len))) {
                                     tzdbShortNamesMap.put(zone, format + NBSP + rule);
                                     zone = null;
                                     rule = null;
                                     format = null;
                                 } else {
-                                    var s = line.split("[ \t]+", -1);
-                                    rule = s[2];
-                                    format = flipIfNeeded(inVanguard, s[3]);
+                                    rule = tokens[2];
+                                    format = flipIfNeeded(inVanguard, tokens[3]);
                                 }
                             }
                         }
 
                         // Rule line
-                        if (line.startsWith("Rule")) {
-                            var rl = line.split("[ \t]+", -1);
-                            tzdbSubstLetters.put(rl[1] + NBSP + (rl[8].equals("0") ? STD : DST),
-                                    rl[9].replace(NO_SUBST, ""));
+                        if (token0len > 0 && tokens[0].regionMatches(true, 0, "Rule", 0, token0len)) {
+                            tzdbSubstLetters.put(tokens[1] + NBSP + (tokens[8].equals("0") ? STD : DST),
+                                    tokens[9].replace(NO_SUBST, ""));
                         }
 
                         // Link line
-                        if (line.startsWith("Link")) {
-                            var ll = line.split("[ \t]+", -1);
-                            tzdbLinks.put(ll[2], ll[1]);
+                        if (token0len > 0 && tokens[0].regionMatches(true, 0, "Link", 0, token0len)) {
+                            tzdbLinks.put(tokens[2], tokens[1]);
                         }
                     }
 
@@ -1491,13 +1493,14 @@ public class CLDRConverter {
     /*
      * Convert TZDB offsets to JDK's offsets, eg, "-08" to "GMT-08:00".
      * If it cannot recognize the pattern, return the argument as is.
+     * Returning null results in generating the GMT format at runtime.
      */
     private static String convertGMTName(String f) {
         try {
-            // Should pre-fill GMT format once COMPAT is gone.
-            // Till then, fall back to GMT format at runtime, after COMPAT short
-            // names are populated
-            ZoneOffset.of(f);
+            if (!f.equals("%z")) {
+                // Validate if the format is an offset
+                ZoneOffset.of(f);
+            }
             return null;
         } catch (DateTimeException dte) {
             // textual representation. return as is

make/jdk/src/classes/build/tools/intpoly/FieldGen.java

@@ -778,7 +778,7 @@ public class FieldGen {
         result.appendLine("}");
 
         result.appendLine("@Override");
-        result.appendLine("protected int mult(long[] a, long[] b, long[] r) {");
+        result.appendLine("protected void mult(long[] a, long[] b, long[] r) {");
         result.incrIndent();
         for (int i = 0; i < 2 * params.getNumLimbs() - 1; i++) {
             result.appendIndent();
@@ -804,9 +804,6 @@ public class FieldGen {
             }
         }
         result.append(");\n");
-        result.appendIndent();
-        result.append("return 0;");
-        result.appendLine();
         result.decrIndent();
         result.appendLine("}");
 
@@ -836,7 +833,7 @@ public class FieldGen {
         //      }
         //  }
         result.appendLine("@Override");
-        result.appendLine("protected int square(long[] a, long[] r) {");
+        result.appendLine("protected void square(long[] a, long[] r) {");
         result.incrIndent();
         for (int i = 0; i < 2 * params.getNumLimbs() - 1; i++) {
             result.appendIndent();
@@ -877,9 +874,6 @@ public class FieldGen {
             }
         }
         result.append(");\n");
-        result.appendIndent();
-        result.append("return 0;");
-        result.appendLine();
         result.decrIndent();
         result.appendLine("}");
 

make/jdk/src/classes/build/tools/taglet/SealedGraph.java

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2022, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2022, 2024, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -107,7 +107,7 @@ public final class SealedGraph implements Taglet {
             throw new RuntimeException(e);
         }
 
-        String simpleTypeName = element.getSimpleName().toString();
+        String simpleTypeName = packagelessCanonicalName(typeElement).replace('.', '/');
         String imageFile = simpleTypeName + "-sealed-graph.svg";
         int thumbnailHeight = 100; // also appears in the stylesheet
         String hoverImage = "<span>"
@@ -315,14 +315,14 @@ public final class SealedGraph implements Taglet {
                 case MEMBER -> packageName((TypeElement) element.getEnclosingElement());
             };
         }
+    }
 
-        private static String packagelessCanonicalName(TypeElement element) {
-            String result = element.getSimpleName().toString();
-            while (element.getNestingKind() == NestingKind.MEMBER) {
-                element = (TypeElement) element.getEnclosingElement();
-                result = element.getSimpleName().toString() + '.' + result;
-            }
-            return result;
+    private static String packagelessCanonicalName(TypeElement element) {
+        String result = element.getSimpleName().toString();
+        while (element.getNestingKind() == NestingKind.MEMBER) {
+            element = (TypeElement) element.getEnclosingElement();
+            result = element.getSimpleName().toString() + '.' + result;
         }
+        return result;
     }
 }

make/jdk/src/classes/build/tools/tzdb/TzdbZoneRulesProvider.java

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2014, 2020, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2014, 2024, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -164,7 +164,8 @@ class TzdbZoneRulesProvider {
                         }
                         continue;
                     }
-                    if (line.startsWith("Zone")) {        // parse Zone line
+                    int token0len = tokens.length > 0 ? tokens[0].length() : line.length();
+                    if (line.regionMatches(true, 0, "Zone", 0, token0len)) {        // parse Zone line
                         String name = tokens[1];
                         if (excludedZones.contains(name)){
                             continue;
@@ -182,13 +183,13 @@ class TzdbZoneRulesProvider {
                         if (zLine.parse(tokens, 2)) {
                             openZone = null;
                         }
-                    } else if (line.startsWith("Rule")) { // parse Rule line
+                    } else if (line.regionMatches(true, 0, "Rule", 0, token0len)) { // parse Rule line
                         String name = tokens[1];
                         if (!rules.containsKey(name)) {
                             rules.put(name, new ArrayList<RuleLine>(10));
                         }
                         rules.get(name).add(new RuleLine().parse(tokens));
-                    } else if (line.startsWith("Link")) { // parse link line
+                    } else if (line.regionMatches(true, 0, "Link", 0, token0len)) { // parse link line
                         if (tokens.length >= 3) {
                             String realId = tokens[1];
                             String aliasId = tokens[2];
@@ -304,7 +305,7 @@ class TzdbZoneRulesProvider {
             month = parseMonth(tokens[off++]);
             if (off < tokens.length) {
                 String dayRule = tokens[off++];
-                if (dayRule.startsWith("last")) {
+                if (dayRule.regionMatches(true, 0, "last", 0, 4)) {
                     dayOfMonth = -1;
                     dayOfWeek = parseDayOfWeek(dayRule.substring(4));
                     adjustForwards = false;
@@ -355,42 +356,45 @@ class TzdbZoneRulesProvider {
         }
 
         int parseYear(String year, int defaultYear) {
-            switch (year.toLowerCase()) {
-            case "min":  return 1900;
-            case "max":  return Year.MAX_VALUE;
-            case "only": return defaultYear;
-            }
+            int len = year.length();
+
+            if (year.regionMatches(true, 0, "minimum", 0, len)) return 1900;
+            if (year.regionMatches(true, 0, "maximum", 0, len)) return Year.MAX_VALUE;
+            if (year.regionMatches(true, 0, "only", 0, len)) return defaultYear;
+
             return Integer.parseInt(year);
         }
 
         Month parseMonth(String mon) {
-            switch (mon) {
-            case "Jan": return Month.JANUARY;
-            case "Feb": return Month.FEBRUARY;
-            case "Mar": return Month.MARCH;
-            case "Apr": return Month.APRIL;
-            case "May": return Month.MAY;
-            case "Jun": return Month.JUNE;
-            case "Jul": return Month.JULY;
-            case "Aug": return Month.AUGUST;
-            case "Sep": return Month.SEPTEMBER;
-            case "Oct": return Month.OCTOBER;
-            case "Nov": return Month.NOVEMBER;
-            case "Dec": return Month.DECEMBER;
-            }
+            int len = mon.length();
+
+            if (mon.regionMatches(true, 0, "January", 0, len)) return Month.JANUARY;
+            if (mon.regionMatches(true, 0, "February", 0, len)) return Month.FEBRUARY;
+            if (mon.regionMatches(true, 0, "March", 0, len)) return Month.MARCH;
+            if (mon.regionMatches(true, 0, "April", 0, len)) return Month.APRIL;
+            if (mon.regionMatches(true, 0, "May", 0, len)) return Month.MAY;
+            if (mon.regionMatches(true, 0, "June", 0, len)) return Month.JUNE;
+            if (mon.regionMatches(true, 0, "July", 0, len)) return Month.JULY;
+            if (mon.regionMatches(true, 0, "August", 0, len)) return Month.AUGUST;
+            if (mon.regionMatches(true, 0, "September", 0, len)) return Month.SEPTEMBER;
+            if (mon.regionMatches(true, 0, "October", 0, len)) return Month.OCTOBER;
+            if (mon.regionMatches(true, 0, "November", 0, len)) return Month.NOVEMBER;
+            if (mon.regionMatches(true, 0, "December", 0, len)) return Month.DECEMBER;
+
             throw new IllegalArgumentException("Unknown month: " + mon);
         }
 
         DayOfWeek parseDayOfWeek(String dow) {
-            switch (dow) {
-            case "Mon": return DayOfWeek.MONDAY;
-            case "Tue": return DayOfWeek.TUESDAY;
-            case "Wed": return DayOfWeek.WEDNESDAY;
-            case "Thu": return DayOfWeek.THURSDAY;
-            case "Fri": return DayOfWeek.FRIDAY;
-            case "Sat": return DayOfWeek.SATURDAY;
-            case "Sun": return DayOfWeek.SUNDAY;
-            }
+            int len = dow.length();
+
+            if (dow.regionMatches(true, 0, "Monday", 0, len)) return DayOfWeek.MONDAY;
+            if (dow.regionMatches(true, 0, "Tuesday", 0, len)) return DayOfWeek.TUESDAY;
+            if (dow.regionMatches(true, 0, "Wednesday", 0, len)) return DayOfWeek.WEDNESDAY;
+            if (dow.regionMatches(true, 0, "Thursday", 0, len)) return DayOfWeek.THURSDAY;
+            if (dow.regionMatches(true, 0, "Friday", 0, len)) return DayOfWeek.FRIDAY;
+            if (dow.regionMatches(true, 0, "Saturday", 0, len)) return DayOfWeek.SATURDAY;
+            if (dow.regionMatches(true, 0, "Sunday", 0, len)) return DayOfWeek.SUNDAY;
+
             throw new IllegalArgumentException("Unknown day-of-week: " + dow);
         }
 

make/src/classes/build/tools/jfr/GenerateJfrFiles.java

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2018, 2024, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -663,7 +663,7 @@ public class GenerateJfrFiles {
             out.write("");
             out.write("union JfrNativeSettings {");
             out.write("  // Array version.");
-            out.write("  jfrNativeEventSetting bits[NUMBER_OF_EVENTS];");
+            out.write("  jfrNativeEventSetting bits[NUMBER_OF_EVENTS + NUMBER_OF_RESERVED_EVENTS];");
             out.write("  // Then, to make it easy to debug,");
             out.write("  // add named struct members also.");
             out.write("  struct {");

make/test/JtregNativeHotspot.gmk

@@ -869,7 +869,7 @@ BUILD_HOTSPOT_JTREG_EXECUTABLES_JDK_LIBS_exedaemonDestroy := java.base:libjvm
 
 ifeq ($(call isTargetOs, windows), true)
   BUILD_HOTSPOT_JTREG_EXECUTABLES_CFLAGS_exeFPRegs := -MT
-  BUILD_HOTSPOT_JTREG_EXCLUDE += exesigtest.c libterminatedThread.c libTestJNI.c libCompleteExit.c libMonitorWithDeadObjectTest.c libTestPsig.c exeGetCreatedJavaVMs.c
+  BUILD_HOTSPOT_JTREG_EXCLUDE += exesigtest.c libterminatedThread.c libTestJNI.c libCompleteExit.c libMonitorWithDeadObjectTest.c libTestPsig.c exeGetCreatedJavaVMs.c libTestUnloadedClass.cpp
   BUILD_HOTSPOT_JTREG_LIBRARIES_JDK_LIBS_libnativeStack := java.base:libjvm
 else
   BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libbootclssearch_agent += -lpthread
@@ -1509,6 +1509,7 @@ else
   BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libCompleteExit += -lpthread
   BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libMonitorWithDeadObjectTest += -lpthread
   BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libnativeStack += -lpthread
+  BUILD_HOTSPOT_JTREG_LIBRARIES_LIBS_libTestUnloadedClass += -lpthread
   BUILD_HOTSPOT_JTREG_EXECUTABLES_LIBS_exeGetCreatedJavaVMs := -lpthread
   BUILD_HOTSPOT_JTREG_EXECUTABLES_JDK_LIBS_exeGetCreatedJavaVMs := java.base:libjvm
 

make/test/JtregNativeJdk.gmk

@@ -114,6 +114,8 @@ ifeq ($(call isTargetOs, linux), true)
   # stripping during the test libraries' build.
   BUILD_JDK_JTREG_LIBRARIES_CFLAGS_libFib := -g
   BUILD_JDK_JTREG_LIBRARIES_STRIP_SYMBOLS_libFib := false
+  # nio tests' libCreationTimeHelper native needs -ldl linker flag
+  BUILD_JDK_JTREG_LIBRARIES_LDFLAGS_libCreationTimeHelper := -ldl
 endif
 
 ifeq ($(ASAN_ENABLED), true)

src/hotspot/cpu/aarch64/globalDefinitions_aarch64.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1999, 2021, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1999, 2024, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2014, 2015, Red Hat Inc. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -76,4 +76,6 @@ const bool CCallingConventionRequiresIntsAsLongs = false;
 
 #define USE_POINTERS_TO_REGISTER_IMPL_ARRAY
 
+#define USE_TRAMPOLINE_STUB_FIX_OWNER
+
 #endif // CPU_AARCH64_GLOBALDEFINITIONS_AARCH64_HPP

src/hotspot/cpu/aarch64/macroAssembler_aarch64.cpp

@@ -1730,8 +1730,8 @@ void MacroAssembler::lookup_secondary_supers_table_slow_path(Register r_super_kl
   // The bitmap is full to bursting.
   // Implicit invariant: BITMAP_FULL implies (length > 0)
   assert(Klass::SECONDARY_SUPERS_BITMAP_FULL == ~uintx(0), "");
-  cmn(r_bitmap, (u1)1);
-  br(EQ, L_huge);
+  cmpw(r_array_length, (u1)(Klass::SECONDARY_SUPERS_TABLE_SIZE - 2));
+  br(GT, L_huge);
 
   // NB! Our caller has checked bits 0 and 1 in the bitmap. The
   // current slot (at secondary_supers[r_array_index]) has not yet

src/hotspot/cpu/aarch64/methodHandles_aarch64.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2024, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2014, Red Hat Inc. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -27,6 +27,7 @@
 #include "asm/macroAssembler.hpp"
 #include "classfile/javaClasses.inline.hpp"
 #include "classfile/vmClasses.hpp"
+#include "compiler/disassembler.hpp"
 #include "interpreter/interpreter.hpp"
 #include "interpreter/interpreterRuntime.hpp"
 #include "memory/allocation.inline.hpp"
@@ -36,7 +37,7 @@
 #include "runtime/frame.inline.hpp"
 #include "runtime/stubRoutines.hpp"
 
-#define __ _masm->
+#define __ Disassembler::hook<MacroAssembler>(__FILE__, __LINE__, _masm)->
 
 #ifdef PRODUCT
 #define BLOCK_COMMENT(str) /* nothing */

src/hotspot/cpu/aarch64/nativeInst_aarch64.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2024, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2014, 2020, Red Hat Inc. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -51,13 +51,18 @@ void NativeInstruction::wrote(int offset) {
 }
 
 address NativeCall::destination() const {
-  address addr = (address)this;
-  address destination = instruction_address() + displacement();
+  address addr = instruction_address();
+  address destination = addr + displacement();
+
+  // Performance optimization: no need to call find_blob() if it is a self-call
+  if (destination == addr) {
+    return destination;
+  }
 
   // Do we use a trampoline stub for this call?
   CodeBlob* cb = CodeCache::find_blob(addr);
-  assert(cb && cb->is_nmethod(), "sanity");
-  nmethod *nm = (nmethod *)cb;
+  assert(cb != nullptr && cb->is_nmethod(), "nmethod expected");
+  nmethod *nm = cb->as_nmethod();
   if (nm->stub_contains(destination) && is_NativeCallTrampolineStub_at(destination)) {
     // Yes we do, so get the destination from the trampoline stub.
     const address trampoline_stub_addr = destination;
@@ -72,12 +77,8 @@ address NativeCall::destination() const {
 // call instruction at all times.
 //
 // Used in the runtime linkage of calls; see class CompiledIC.
-//
-// Add parameter assert_lock to switch off assertion
-// during code generation, where no patching lock is needed.
-void NativeCall::set_destination_mt_safe(address dest, bool assert_lock) {
-  assert(!assert_lock ||
-         (Patching_lock->is_locked() || SafepointSynchronize::is_at_safepoint()) ||
+void NativeCall::set_destination_mt_safe(address dest) {
+  assert((CodeCache_lock->is_locked() || SafepointSynchronize::is_at_safepoint()) ||
          CompiledICLocker::is_safe(addr_at(0)),
          "concurrent code patching");
 
@@ -104,22 +105,18 @@ void NativeCall::set_destination_mt_safe(address dest, bool assert_lock) {
 }
 
 address NativeCall::get_trampoline() {
-  address call_addr = addr_at(0);
+  address call_addr = instruction_address();
 
   CodeBlob *code = CodeCache::find_blob(call_addr);
-  assert(code != nullptr, "Could not find the containing code blob");
+  assert(code != nullptr && code->is_nmethod(), "nmethod expected");
+  nmethod* nm = code->as_nmethod();
 
-  address bl_destination
-    = MacroAssembler::pd_call_destination(call_addr);
-  if (code->contains(bl_destination) &&
+  address bl_destination = call_addr + displacement();
+  if (nm->stub_contains(bl_destination) &&
       is_NativeCallTrampolineStub_at(bl_destination))
     return bl_destination;
 
-  if (code->is_nmethod()) {
-    return trampoline_stub_Relocation::get_trampoline_for(call_addr, (nmethod*)code);
-  }
-
-  return nullptr;
+  return trampoline_stub_Relocation::get_trampoline_for(call_addr, nm);
 }
 
 // Inserts a native call instruction at a given pc

src/hotspot/cpu/aarch64/nativeInst_aarch64.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2024, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2014, 2108, Red Hat Inc. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -173,6 +173,7 @@ public:
   int displacement() const { return (int_at(displacement_offset) << 6) >> 4; }
   address displacement_address() const { return addr_at(displacement_offset); }
   address return_address() const { return addr_at(return_address_offset); }
+  address raw_destination() const { return instruction_address() + displacement(); }
   address destination() const;
 
   void set_destination(address dest) {
@@ -212,9 +213,7 @@ public:
   //
   // Used in the runtime linkage of calls; see class CompiledIC.
   // (Cf. 4506997 and 4479829, where threads witnessed garbage displacements.)
-
-  // The parameter assert_lock disables the assertion during code generation.
-  void set_destination_mt_safe(address dest, bool assert_lock = true);
+  void set_destination_mt_safe(address dest);
 
   address get_trampoline();
 #if INCLUDE_JVMCI

src/hotspot/cpu/aarch64/relocInfo_aarch64.cpp

@@ -60,13 +60,12 @@ void Relocation::pd_set_data_value(address x, bool verify_only) {
 
 address Relocation::pd_call_destination(address orig_addr) {
   assert(is_call(), "should be a call here");
-  if (NativeCall::is_call_at(addr())) {
-    address trampoline = nativeCall_at(addr())->get_trampoline();
-    if (trampoline) {
-      return nativeCallTrampolineStub_at(trampoline)->destination();
+  if (orig_addr == nullptr) {
+    if (NativeCall::is_call_at(addr())) {
+      NativeCall* call = nativeCall_at(addr());
+      return call->destination();
     }
-  }
-  if (orig_addr != nullptr) {
+  } else {
     address new_addr = MacroAssembler::pd_call_destination(orig_addr);
     // If call is branch to self, don't try to relocate it, just leave it
     // as branch to self. This happens during code generation if the code
@@ -82,16 +81,26 @@ address Relocation::pd_call_destination(address orig_addr) {
 void Relocation::pd_set_call_destination(address x) {
   assert(is_call(), "should be a call here");
   if (NativeCall::is_call_at(addr())) {
-    address trampoline = nativeCall_at(addr())->get_trampoline();
-    if (trampoline) {
-      nativeCall_at(addr())->set_destination_mt_safe(x, /* assert_lock */false);
-      return;
-    }
+    NativeCall* call = nativeCall_at(addr());
+    call->set_destination(x);
+  } else {
+    MacroAssembler::pd_patch_instruction(addr(), x);
   }
-  MacroAssembler::pd_patch_instruction(addr(), x);
   assert(pd_call_destination(addr()) == x, "fail in reloc");
 }
 
+void trampoline_stub_Relocation::pd_fix_owner_after_move() {
+  NativeCall* call = nativeCall_at(owner());
+  assert(call->raw_destination() == owner(), "destination should be empty");
+  address trampoline = addr();
+  address dest = nativeCallTrampolineStub_at(trampoline)->destination();
+  if (!Assembler::reachable_from_branch_at(owner(), dest)) {
+    dest = trampoline;
+  }
+  call->set_destination(dest);
+}
+
+
 address* Relocation::pd_address_in_code() {
   return (address*)(addr() + 8);
 }

src/hotspot/cpu/aarch64/smallRegisterMap_aarch64.inline.hpp

@@ -30,8 +30,15 @@
 
 // Java frames don't have callee saved registers (except for rfp), so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
 private:
   static void assert_is_rfp(VMReg r) NOT_DEBUG_RETURN
                                      DEBUG_ONLY({ assert (r == rfp->as_VMReg() || r == rfp->as_VMReg()->next(), "Reg: %s", r->name()); })
@@ -48,17 +55,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-  #ifdef ASSERT
-    for(int i = 0; i < RegisterMap::reg_count; i++) {
-      VMReg r = VMRegImpl::as_VMReg(i);
-      if (map->location(r, (intptr_t*)nullptr) != nullptr) assert_is_rfp(r);
-    }
-  #endif
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     assert_is_rfp(reg);
     return (address)(sp - frame::sender_sp_offset);

src/hotspot/cpu/aarch64/stubGenerator_aarch64.cpp

@@ -7412,6 +7412,28 @@ class StubGenerator: public StubCodeGenerator {
     return start;
   }
 
+  // load Method* target of MethodHandle
+  // j_rarg0 = jobject receiver
+  // rmethod = result
+  address generate_upcall_stub_load_target() {
+    StubCodeMark mark(this, "StubRoutines", "upcall_stub_load_target");
+    address start = __ pc();
+
+    __ resolve_global_jobject(j_rarg0, rscratch1, rscratch2);
+      // Load target method from receiver
+    __ load_heap_oop(rmethod, Address(j_rarg0, java_lang_invoke_MethodHandle::form_offset()), rscratch1, rscratch2);
+    __ load_heap_oop(rmethod, Address(rmethod, java_lang_invoke_LambdaForm::vmentry_offset()), rscratch1, rscratch2);
+    __ load_heap_oop(rmethod, Address(rmethod, java_lang_invoke_MemberName::method_offset()), rscratch1, rscratch2);
+    __ access_load_at(T_ADDRESS, IN_HEAP, rmethod,
+                      Address(rmethod, java_lang_invoke_ResolvedMethodName::vmtarget_offset()),
+                      noreg, noreg);
+    __ str(rmethod, Address(rthread, JavaThread::callee_target_offset())); // just in case callee is deoptimized
+
+    __ ret(lr);
+
+    return start;
+  }
+
   // Continuation point for throwing of implicit exceptions that are
   // not handled in the current activation. Fabricates an exception
   // oop and initiates normal exception dispatching in this
@@ -8477,6 +8499,7 @@ class StubGenerator: public StubCodeGenerator {
 #endif
 
     StubRoutines::_upcall_stub_exception_handler = generate_upcall_stub_exception_handler();
+    StubRoutines::_upcall_stub_load_target = generate_upcall_stub_load_target();
 
     StubRoutines::aarch64::set_completed(); // Inidicate that arraycopy and zero_blocks stubs are generated
   }

src/hotspot/cpu/aarch64/templateInterpreterGenerator_aarch64.cpp

@@ -26,6 +26,7 @@
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
 #include "classfile/javaClasses.hpp"
+#include "compiler/disassembler.hpp"
 #include "compiler/compiler_globals.hpp"
 #include "gc/shared/barrierSetAssembler.hpp"
 #include "interpreter/bytecodeHistogram.hpp"
@@ -67,7 +68,7 @@
 // Max size with JVMTI
 int TemplateInterpreter::InterpreterCodeSize = 200 * 1024;
 
-#define __ _masm->
+#define __ Disassembler::hook<InterpreterMacroAssembler>(__FILE__, __LINE__, _masm)->
 
 //-----------------------------------------------------------------------------
 
@@ -2011,13 +2012,21 @@ void TemplateInterpreterGenerator::set_vtos_entry_points(Template* t,
                                                          address& vep) {
   assert(t->is_valid() && t->tos_in() == vtos, "illegal template");
   Label L;
-  aep = __ pc();  __ push_ptr();  __ b(L);
-  fep = __ pc();  __ push_f();    __ b(L);
-  dep = __ pc();  __ push_d();    __ b(L);
-  lep = __ pc();  __ push_l();    __ b(L);
-  bep = cep = sep =
-  iep = __ pc();  __ push_i();
-  vep = __ pc();
+  aep = __ pc();     // atos entry point
+      __ push_ptr();
+      __ b(L);
+  fep = __ pc();     // ftos entry point
+      __ push_f();
+      __ b(L);
+  dep = __ pc();     // dtos entry point
+      __ push_d();
+      __ b(L);
+  lep = __ pc();     // ltos entry point
+      __ push_l();
+      __ b(L);
+  bep = cep = sep = iep = __ pc();     // [bcsi]tos entry point
+      __ push_i();
+  vep = __ pc();     // vtos entry point
   __ bind(L);
   generate_and_dispatch(t);
 }

src/hotspot/cpu/aarch64/templateTable_aarch64.cpp

@@ -25,6 +25,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
+#include "compiler/disassembler.hpp"
 #include "compiler/compilerDefinitions.inline.hpp"
 #include "gc/shared/barrierSetAssembler.hpp"
 #include "gc/shared/collectedHeap.hpp"
@@ -49,7 +50,7 @@
 #include "runtime/synchronizer.hpp"
 #include "utilities/powerOfTwo.hpp"
 
-#define __ _masm->
+#define __ Disassembler::hook<InterpreterMacroAssembler>(__FILE__, __LINE__, _masm)->
 
 // Address computation: local variables
 

src/hotspot/cpu/aarch64/upcallLinker_aarch64.cpp

@@ -24,6 +24,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.hpp"
+#include "classfile/javaClasses.hpp"
 #include "logging/logStream.hpp"
 #include "memory/resourceArea.hpp"
 #include "prims/upcallLinker.hpp"
@@ -117,7 +118,7 @@ static void restore_callee_saved_registers(MacroAssembler* _masm, const ABIDescr
 static const int upcall_stub_code_base_size = 1024;
 static const int upcall_stub_size_per_arg = 16;
 
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,
@@ -222,7 +223,6 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
   __ block_comment("{ on_entry");
   __ lea(c_rarg0, Address(sp, frame_data_offset));
-  __ movptr(c_rarg1, (intptr_t)receiver);
   __ movptr(rscratch1, CAST_FROM_FN_PTR(uint64_t, UpcallLinker::on_entry));
   __ blr(rscratch1);
   __ mov(rthread, r0);
@@ -238,12 +238,10 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   arg_shuffle.generate(_masm, as_VMStorage(shuffle_reg), abi._shadow_space_bytes, 0);
   __ block_comment("} argument shuffle");
 
-  __ block_comment("{ receiver ");
-  __ get_vm_result(j_rarg0, rthread);
-  __ block_comment("} receiver ");
-
-  __ mov_metadata(rmethod, entry);
-  __ str(rmethod, Address(rthread, JavaThread::callee_target_offset())); // just in case callee is deoptimized
+  __ block_comment("{ load target ");
+  __ movptr(j_rarg0, (intptr_t)receiver);
+  __ far_call(RuntimeAddress(StubRoutines::upcall_stub_load_target()), rscratch1); // puts target Method* in rmethod
+  __ block_comment("} load target ");
 
   __ push_cont_fastpath(rthread);
 
@@ -318,7 +316,7 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
 #ifndef PRODUCT
   stringStream ss;
-  ss.print("upcall_stub_%s", entry->signature()->as_C_string());
+  ss.print("upcall_stub_%s", signature->as_C_string());
   const char* name = _masm->code_string(ss.as_string());
 #else // PRODUCT
   const char* name = "upcall_stub";

src/hotspot/cpu/arm/smallRegisterMap_arm.inline.hpp

@@ -30,8 +30,15 @@
 
 // Java frames don't have callee saved registers (except for rfp), so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
 private:
   static void assert_is_rfp(VMReg r) NOT_DEBUG_RETURN
                                      DEBUG_ONLY({ Unimplemented(); })
@@ -46,12 +53,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-    Unimplemented();
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     Unimplemented();
     return nullptr;

src/hotspot/cpu/arm/upcallLinker_arm.cpp

@@ -25,7 +25,7 @@
 #include "prims/upcallLinker.hpp"
 #include "utilities/debug.hpp"
 
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,

src/hotspot/cpu/ppc/c1_LIRAssembler_ppc.cpp

@@ -133,9 +133,20 @@ void LIR_Assembler::osr_entry() {
   // copied into place by code emitted in the IR.
 
   Register OSR_buf = osrBufferPointer()->as_register();
-  { assert(frame::interpreter_frame_monitor_size() == BasicObjectLock::size(), "adjust code below");
-    int monitor_offset = BytesPerWord * method()->max_locals() +
-      (2 * BytesPerWord) * (number_of_locks - 1);
+  {
+    assert(frame::interpreter_frame_monitor_size() == BasicObjectLock::size(), "adjust code below");
+
+    const int locals_space = BytesPerWord * method()->max_locals();
+    int monitor_offset = locals_space + (2 * BytesPerWord) * (number_of_locks - 1);
+    bool use_OSR_bias = false;
+
+    if (!Assembler::is_simm16(monitor_offset + BytesPerWord) && number_of_locks > 0) {
+      // Offsets too large for ld instructions. Use bias.
+      __ add_const_optimized(OSR_buf, OSR_buf, locals_space);
+      monitor_offset -= locals_space;
+      use_OSR_bias = true;
+    }
+
     // SharedRuntime::OSR_migration_begin() packs BasicObjectLocks in
     // the OSR buffer using 2 word entries: first the lock and then
     // the oop.
@@ -161,6 +172,11 @@ void LIR_Assembler::osr_entry() {
       __ ld(R0, slot_offset + 1*BytesPerWord, OSR_buf);
       __ std(R0, mo.disp(), mo.base());
     }
+
+    if (use_OSR_bias) {
+      // Restore.
+      __ sub_const_optimized(OSR_buf, OSR_buf, locals_space);
+    }
   }
 }
 

src/hotspot/cpu/ppc/frame_ppc.cpp

@@ -117,9 +117,9 @@ bool frame::safe_for_sender(JavaThread *thread) {
       return false;
     }
 
-    common_abi* sender_abi = (common_abi*) fp;
+    volatile common_abi* sender_abi = (common_abi*) fp; // May get updated concurrently by deoptimization!
     intptr_t* sender_sp = (intptr_t*) fp;
-    address   sender_pc = (address) sender_abi->lr;;
+    address   sender_pc = (address) sender_abi->lr;
 
     if (Continuation::is_return_barrier_entry(sender_pc)) {
       // If our sender_pc is the return barrier, then our "real" sender is the continuation entry
@@ -134,9 +134,18 @@ bool frame::safe_for_sender(JavaThread *thread) {
       return false;
     }
 
+    intptr_t* unextended_sender_sp = is_interpreted_frame() ? interpreter_frame_sender_sp() : sender_sp;
+
+    // If the sender is a deoptimized nmethod we need to check if the original pc is valid.
+    nmethod* sender_nm = sender_blob->as_nmethod_or_null();
+    if (sender_nm != nullptr && sender_nm->is_deopt_pc(sender_pc)) {
+      address orig_pc = *(address*)((address)unextended_sender_sp + sender_nm->orig_pc_offset());
+      if (!sender_nm->insts_contains_inclusive(orig_pc)) return false;
+    }
+
     // It should be safe to construct the sender though it might not be valid.
 
-    frame sender(sender_sp, sender_pc, nullptr /* unextended_sp */, nullptr /* fp */, sender_blob);
+    frame sender(sender_sp, sender_pc, unextended_sender_sp, nullptr /* fp */, sender_blob);
 
     // Do we have a valid fp?
     address sender_fp = (address) sender.fp();

src/hotspot/cpu/ppc/nativeInst_ppc.cpp

@@ -92,10 +92,10 @@ address NativeCall::destination() const {
 // Used in the runtime linkage of calls; see class CompiledIC.
 //
 // Add parameter assert_lock to switch off assertion
-// during code generation, where no patching lock is needed.
+// during code generation, where no lock is needed.
 void NativeCall::set_destination_mt_safe(address dest, bool assert_lock) {
   assert(!assert_lock ||
-         (Patching_lock->is_locked() || SafepointSynchronize::is_at_safepoint()) ||
+         (CodeCache_lock->is_locked() || SafepointSynchronize::is_at_safepoint()) ||
          CompiledICLocker::is_safe(addr_at(0)),
          "concurrent code patching");
 

src/hotspot/cpu/ppc/ppc.ad

@@ -3427,9 +3427,11 @@ encode %{
     call->_oop_map           = _oop_map;
     call->_jvms              = _jvms;
     call->_jvmadj            = _jvmadj;
+    call->_has_ea_local_in_scope = _has_ea_local_in_scope;
     call->_in_rms            = _in_rms;
     call->_nesting           = _nesting;
     call->_override_symbolic_info = _override_symbolic_info;
+    call->_arg_escape        = _arg_escape;
 
     // New call needs all inputs of old call.
     // Req...

src/hotspot/cpu/ppc/smallRegisterMap_ppc.inline.hpp

@@ -30,9 +30,16 @@
 
 // Java frames don't have callee saved registers, so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
-public:
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
+
   // as_RegisterMap is used when we didn't want to templatize and abstract over RegisterMap type to support SmallRegisterMap
   // Consider enhancing SmallRegisterMap to support those cases
   const RegisterMap* as_RegisterMap() const { return nullptr; }
@@ -44,19 +51,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-#ifdef ASSERT
-  for(int i = 0; i < RegisterMap::reg_count; i++) {
-    VMReg r = VMRegImpl::as_VMReg(i);
-    if (map->location(r, (intptr_t*)nullptr) != nullptr) {
-      assert(false, "Reg: %s", r->name()); // Should not reach here
-    }
-  }
-#endif
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     assert(false, "Reg: %s", reg->name());
     return nullptr;

src/hotspot/cpu/ppc/stubGenerator_ppc.cpp

@@ -4728,6 +4728,30 @@ class StubGenerator: public StubCodeGenerator {
     return start;
   }
 
+  // load Method* target of MethodHandle
+  // R3_ARG1 = jobject receiver
+  // R19_method = result Method*
+  address generate_upcall_stub_load_target() {
+
+    StubCodeMark mark(this, "StubRoutines", "upcall_stub_load_target");
+    address start = __ pc();
+
+    __ resolve_global_jobject(R3_ARG1, R22_tmp2, R23_tmp3, MacroAssembler::PRESERVATION_FRAME_LR_GP_FP_REGS);
+    // Load target method from receiver
+    __ load_heap_oop(R19_method, java_lang_invoke_MethodHandle::form_offset(), R3_ARG1,
+                     R22_tmp2, R23_tmp3, MacroAssembler::PRESERVATION_FRAME_LR_GP_FP_REGS, IS_NOT_NULL);
+    __ load_heap_oop(R19_method, java_lang_invoke_LambdaForm::vmentry_offset(), R19_method,
+                     R22_tmp2, R23_tmp3, MacroAssembler::PRESERVATION_FRAME_LR_GP_FP_REGS, IS_NOT_NULL);
+    __ load_heap_oop(R19_method, java_lang_invoke_MemberName::method_offset(), R19_method,
+                     R22_tmp2, R23_tmp3, MacroAssembler::PRESERVATION_FRAME_LR_GP_FP_REGS, IS_NOT_NULL);
+    __ ld(R19_method, java_lang_invoke_ResolvedMethodName::vmtarget_offset(), R19_method);
+    __ std(R19_method, in_bytes(JavaThread::callee_target_offset()), R16_thread); // just in case callee is deoptimized
+
+    __ blr();
+
+    return start;
+  }
+
   // Initialization
   void generate_initial_stubs() {
     // Generates all stubs and initializes the entry points
@@ -4808,6 +4832,7 @@ class StubGenerator: public StubCodeGenerator {
     generate_arraycopy_stubs();
 
     StubRoutines::_upcall_stub_exception_handler = generate_upcall_stub_exception_handler();
+    StubRoutines::_upcall_stub_load_target = generate_upcall_stub_load_target();
   }
 
   void generate_compiler_stubs() {

src/hotspot/cpu/ppc/upcallLinker_ppc.cpp

@@ -24,6 +24,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
+#include "classfile/javaClasses.hpp"
 #include "logging/logStream.hpp"
 #include "memory/resourceArea.hpp"
 #include "prims/upcallLinker.hpp"
@@ -118,7 +119,7 @@ static void restore_callee_saved_registers(MacroAssembler* _masm, const ABIDescr
 static const int upcall_stub_code_base_size = 1024;
 static const int upcall_stub_size_per_arg = 16; // arg save & restore + move
 
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,
@@ -221,7 +222,6 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   __ block_comment("{ on_entry");
   __ load_const_optimized(call_target_address, CAST_FROM_FN_PTR(uint64_t, UpcallLinker::on_entry), R0);
   __ addi(R3_ARG1, R1_SP, frame_data_offset);
-  __ load_const_optimized(R4_ARG2, (intptr_t)receiver, R0);
   __ call_c(call_target_address);
   __ mr(R16_thread, R3_RET);
   __ block_comment("} on_entry");
@@ -236,12 +236,12 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   arg_shuffle.generate(_masm, as_VMStorage(callerSP), frame::native_abi_minframe_size, frame::jit_out_preserve_size);
   __ block_comment("} argument shuffle");
 
-  __ block_comment("{ receiver ");
-  __ get_vm_result(R3_ARG1);
-  __ block_comment("} receiver ");
-
-  __ load_const_optimized(R19_method, (intptr_t)entry);
-  __ std(R19_method, in_bytes(JavaThread::callee_target_offset()), R16_thread);
+  __ block_comment("{ load target ");
+  __ load_const_optimized(call_target_address, StubRoutines::upcall_stub_load_target(), R0);
+  __ load_const_optimized(R3_ARG1, (intptr_t)receiver, R0);
+  __ mtctr(call_target_address);
+  __ bctrl(); // loads target Method* into R19_method
+  __ block_comment("} load target ");
 
   __ push_cont_fastpath();
 
@@ -326,7 +326,7 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
 #ifndef PRODUCT
   stringStream ss;
-  ss.print("upcall_stub_%s", entry->signature()->as_C_string());
+  ss.print("upcall_stub_%s", signature->as_C_string());
   const char* name = _masm->code_string(ss.as_string());
 #else // PRODUCT
   const char* name = "upcall_stub";

src/hotspot/cpu/riscv/assembler_riscv.hpp

@@ -1828,10 +1828,12 @@ enum Nf {
   // Vector unordered indexed load instructions
   INSN( vluxei8_v, 0b0000111, 0b000, 0b01, 0b0);
   INSN(vluxei32_v, 0b0000111, 0b110, 0b01, 0b0);
+  INSN(vluxei64_v, 0b0000111, 0b111, 0b01, 0b0);
 
   // Vector unordered indexed store instructions
   INSN( vsuxei8_v, 0b0100111, 0b000, 0b01, 0b0);
   INSN(vsuxei32_v, 0b0100111, 0b110, 0b01, 0b0);
+  INSN(vsuxei64_v, 0b0100111, 0b111, 0b01, 0b0);
 
 #undef INSN
 

src/hotspot/cpu/riscv/gc/x/x_riscv.ad

@@ -52,11 +52,11 @@ static void x_load_barrier_slow_path(MacroAssembler* masm, const MachNode* node,
 %}
 
 // Load Pointer
-instruct xLoadP(iRegPNoSp dst, memory mem, iRegPNoSp tmp)
+instruct xLoadP(iRegPNoSp dst, memory mem, iRegPNoSp tmp, rFlagsReg cr)
 %{
   match(Set dst (LoadP mem));
   predicate(UseZGC && !ZGenerational && (n->as_Load()->barrier_data() != 0));
-  effect(TEMP dst, TEMP tmp);
+  effect(TEMP dst, TEMP tmp, KILL cr);
 
   ins_cost(4 * DEFAULT_COST);
 
@@ -71,11 +71,11 @@ instruct xLoadP(iRegPNoSp dst, memory mem, iRegPNoSp tmp)
   ins_pipe(iload_reg_mem);
 %}
 
-instruct xCompareAndSwapP(iRegINoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp) %{
+instruct xCompareAndSwapP(iRegINoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set res (CompareAndSwapP mem (Binary oldval newval)));
   match(Set res (WeakCompareAndSwapP mem (Binary oldval newval)));
   predicate(UseZGC && !ZGenerational && !needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() == XLoadBarrierStrong);
-  effect(TEMP_DEF res, TEMP tmp);
+  effect(TEMP_DEF res, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -105,11 +105,11 @@ instruct xCompareAndSwapP(iRegINoSp res, indirect mem, iRegP oldval, iRegP newva
   ins_pipe(pipe_slow);
 %}
 
-instruct xCompareAndSwapPAcq(iRegINoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp) %{
+instruct xCompareAndSwapPAcq(iRegINoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set res (CompareAndSwapP mem (Binary oldval newval)));
   match(Set res (WeakCompareAndSwapP mem (Binary oldval newval)));
   predicate(UseZGC && !ZGenerational && needs_acquiring_load_reserved(n) && (n->as_LoadStore()->barrier_data() == XLoadBarrierStrong));
-  effect(TEMP_DEF res, TEMP tmp);
+  effect(TEMP_DEF res, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -139,10 +139,10 @@ instruct xCompareAndSwapPAcq(iRegINoSp res, indirect mem, iRegP oldval, iRegP ne
   ins_pipe(pipe_slow);
 %}
 
-instruct xCompareAndExchangeP(iRegPNoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp) %{
+instruct xCompareAndExchangeP(iRegPNoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set res (CompareAndExchangeP mem (Binary oldval newval)));
   predicate(UseZGC && !ZGenerational && !needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() == XLoadBarrierStrong);
-  effect(TEMP_DEF res, TEMP tmp);
+  effect(TEMP_DEF res, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -167,10 +167,10 @@ instruct xCompareAndExchangeP(iRegPNoSp res, indirect mem, iRegP oldval, iRegP n
   ins_pipe(pipe_slow);
 %}
 
-instruct xCompareAndExchangePAcq(iRegPNoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp) %{
+instruct xCompareAndExchangePAcq(iRegPNoSp res, indirect mem, iRegP oldval, iRegP newval, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set res (CompareAndExchangeP mem (Binary oldval newval)));
   predicate(UseZGC && !ZGenerational && needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() == XLoadBarrierStrong);
-  effect(TEMP_DEF res, TEMP tmp);
+  effect(TEMP_DEF res, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -195,10 +195,10 @@ instruct xCompareAndExchangePAcq(iRegPNoSp res, indirect mem, iRegP oldval, iReg
   ins_pipe(pipe_slow);
 %}
 
-instruct xGetAndSetP(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp) %{
+instruct xGetAndSetP(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set prev (GetAndSetP mem newv));
   predicate(UseZGC && !ZGenerational && !needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP_DEF prev, TEMP tmp);
+  effect(TEMP_DEF prev, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -212,10 +212,10 @@ instruct xGetAndSetP(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp) %{
   ins_pipe(pipe_serial);
 %}
 
-instruct xGetAndSetPAcq(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp) %{
+instruct xGetAndSetPAcq(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set prev (GetAndSetP mem newv));
   predicate(UseZGC && !ZGenerational && needs_acquiring_load_reserved(n) && (n->as_LoadStore()->barrier_data() != 0));
-  effect(TEMP_DEF prev, TEMP tmp);
+  effect(TEMP_DEF prev, TEMP tmp, KILL cr);
 
   ins_cost(VOLATILE_REF_COST);
 

src/hotspot/cpu/riscv/gc/z/z_riscv.ad

@@ -90,11 +90,11 @@ static void z_store_barrier(MacroAssembler* masm, const MachNode* node, Address
 %}
 
 // Load Pointer
-instruct zLoadP(iRegPNoSp dst, memory mem, iRegPNoSp tmp)
+instruct zLoadP(iRegPNoSp dst, memory mem, iRegPNoSp tmp, rFlagsReg cr)
 %{
   match(Set dst (LoadP mem));
   predicate(UseZGC && ZGenerational && n->as_Load()->barrier_data() != 0);
-  effect(TEMP dst, TEMP tmp);
+  effect(TEMP dst, TEMP tmp, KILL cr);
 
   ins_cost(4 * DEFAULT_COST);
 
@@ -110,11 +110,11 @@ instruct zLoadP(iRegPNoSp dst, memory mem, iRegPNoSp tmp)
 %}
 
 // Store Pointer
-instruct zStoreP(memory mem, iRegP src, iRegPNoSp tmp1, iRegPNoSp tmp2)
+instruct zStoreP(memory mem, iRegP src, iRegPNoSp tmp1, iRegPNoSp tmp2, rFlagsReg cr)
 %{
   predicate(UseZGC && ZGenerational && n->as_Store()->barrier_data() != 0);
   match(Set mem (StoreP mem src));
-  effect(TEMP tmp1, TEMP tmp2);
+  effect(TEMP tmp1, TEMP tmp2, KILL cr);
 
   ins_cost(125); // XXX
   format %{ "sd    $mem, $src\t# ptr" %}
@@ -127,11 +127,11 @@ instruct zStoreP(memory mem, iRegP src, iRegPNoSp tmp1, iRegPNoSp tmp2)
 %}
 
 instruct zCompareAndSwapP(iRegINoSp res, indirect mem, iRegP oldval, iRegP newval,
-                          iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1) %{
+                          iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1, rFlagsReg cr) %{
   match(Set res (CompareAndSwapP mem (Binary oldval newval)));
   match(Set res (WeakCompareAndSwapP mem (Binary oldval newval)));
   predicate(UseZGC && ZGenerational && !needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res);
+  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -150,11 +150,11 @@ instruct zCompareAndSwapP(iRegINoSp res, indirect mem, iRegP oldval, iRegP newva
 %}
 
 instruct zCompareAndSwapPAcq(iRegINoSp res, indirect mem, iRegP oldval, iRegP newval,
-                             iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1) %{
+                             iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1, rFlagsReg cr) %{
   match(Set res (CompareAndSwapP mem (Binary oldval newval)));
   match(Set res (WeakCompareAndSwapP mem (Binary oldval newval)));
   predicate(UseZGC && ZGenerational && needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res);
+  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -173,10 +173,10 @@ instruct zCompareAndSwapPAcq(iRegINoSp res, indirect mem, iRegP oldval, iRegP ne
 %}
 
 instruct zCompareAndExchangeP(iRegPNoSp res, indirect mem, iRegP oldval, iRegP newval,
-                              iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1) %{
+                              iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1, rFlagsReg cr) %{
   match(Set res (CompareAndExchangeP mem (Binary oldval newval)));
   predicate(UseZGC && ZGenerational && !needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res);
+  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -195,10 +195,10 @@ instruct zCompareAndExchangeP(iRegPNoSp res, indirect mem, iRegP oldval, iRegP n
 %}
 
 instruct zCompareAndExchangePAcq(iRegPNoSp res, indirect mem, iRegP oldval, iRegP newval,
-                                 iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1) %{
+                                 iRegPNoSp oldval_tmp, iRegPNoSp newval_tmp, iRegPNoSp tmp1, rFlagsReg cr) %{
   match(Set res (CompareAndExchangeP mem (Binary oldval newval)));
   predicate(UseZGC && ZGenerational && needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res);
+  effect(TEMP oldval_tmp, TEMP newval_tmp, TEMP tmp1, TEMP_DEF res, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -216,10 +216,10 @@ instruct zCompareAndExchangePAcq(iRegPNoSp res, indirect mem, iRegP oldval, iReg
   ins_pipe(pipe_slow);
 %}
 
-instruct zGetAndSetP(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp) %{
+instruct zGetAndSetP(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set prev (GetAndSetP mem newv));
   predicate(UseZGC && ZGenerational && !needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP_DEF prev, TEMP tmp);
+  effect(TEMP_DEF prev, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 
@@ -234,10 +234,10 @@ instruct zGetAndSetP(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp) %{
   ins_pipe(pipe_serial);
 %}
 
-instruct zGetAndSetPAcq(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp) %{
+instruct zGetAndSetPAcq(indirect mem, iRegP newv, iRegPNoSp prev, iRegPNoSp tmp, rFlagsReg cr) %{
   match(Set prev (GetAndSetP mem newv));
   predicate(UseZGC && ZGenerational && needs_acquiring_load_reserved(n) && n->as_LoadStore()->barrier_data() != 0);
-  effect(TEMP_DEF prev, TEMP tmp);
+  effect(TEMP_DEF prev, TEMP tmp, KILL cr);
 
   ins_cost(2 * VOLATILE_REF_COST);
 

src/hotspot/cpu/riscv/interpreterRT_riscv.cpp

@@ -74,6 +74,16 @@ InterpreterRuntime::SignatureHandlerGenerator::SignatureHandlerGenerator(
   _stack_offset = 0;
 }
 
+// The C ABI specifies:
+// "integer scalars narrower than XLEN bits are widened according to the sign
+// of their type up to 32 bits, then sign-extended to XLEN bits."
+// Applies for both passed in register and stack.
+//
+// Java uses 32-bit stack slots; jint, jshort, jchar, jbyte uses one slot.
+// Native uses 64-bit stack slots for all integer scalar types.
+//
+// lw loads the Java stack slot, sign-extends and
+// sd store this widened integer into a 64 bit native stack slot.
 void InterpreterRuntime::SignatureHandlerGenerator::pass_int() {
   const Address src(from(), Interpreter::local_offset_in_bytes(offset()));
 
@@ -82,7 +92,7 @@ void InterpreterRuntime::SignatureHandlerGenerator::pass_int() {
     __ lw(reg, src);
   } else {
     __ lw(x10, src);
-    __ sw(x10, Address(to(), next_stack_offset()));
+    __ sd(x10, Address(to(), next_stack_offset()));
   }
 }
 

src/hotspot/cpu/riscv/macroAssembler_riscv.cpp

@@ -146,18 +146,6 @@ bool MacroAssembler::is_li32_at(address instr) {
          check_li32_data_dependency(instr);
 }
 
-bool MacroAssembler::is_li64_at(address instr) {
-  return is_lui_at(instr) && // lui
-         is_addi_at(instr + instruction_size) && // addi
-         is_slli_shift_at(instr + instruction_size * 2, 12) &&  // Slli Rd, Rs, 12
-         is_addi_at(instr + instruction_size * 3) && // addi
-         is_slli_shift_at(instr + instruction_size * 4, 12) &&  // Slli Rd, Rs, 12
-         is_addi_at(instr + instruction_size * 5) && // addi
-         is_slli_shift_at(instr + instruction_size * 6, 8) &&   // Slli Rd, Rs, 8
-         is_addi_at(instr + instruction_size * 7) && // addi
-         check_li64_data_dependency(instr);
-}
-
 bool MacroAssembler::is_lwu_to_zr(address instr) {
   assert_cond(instr != nullptr);
   return (extract_opcode(instr) == 0b0000011 &&
@@ -909,37 +897,9 @@ void MacroAssembler::li32(Register Rd, int32_t imm) {
   upper = (int32_t)upper;
   // lui Rd, imm[31:12] + imm[11]
   lui(Rd, upper);
-  // use addiw to distinguish li32 to li64
   addiw(Rd, Rd, lower);
 }
 
-void MacroAssembler::li64(Register Rd, int64_t imm) {
-  // Load upper 32 bits. upper = imm[63:32], but if imm[31] == 1 or
-  // (imm[31:20] == 0x7ff && imm[19] == 1), upper = imm[63:32] + 1.
-  int64_t lower = imm & 0xffffffff;
-  lower -= ((lower << 44) >> 44);
-  int64_t tmp_imm = ((uint64_t)(imm & 0xffffffff00000000)) + (uint64_t)lower;
-  int32_t upper = (tmp_imm - (int32_t)lower) >> 32;
-
-  // Load upper 32 bits
-  int64_t up = upper, lo = upper;
-  lo = (lo << 52) >> 52;
-  up -= lo;
-  up = (int32_t)up;
-  lui(Rd, up);
-  addi(Rd, Rd, lo);
-
-  // Load the rest 32 bits.
-  slli(Rd, Rd, 12);
-  addi(Rd, Rd, (int32_t)lower >> 20);
-  slli(Rd, Rd, 12);
-  lower = ((int32_t)imm << 12) >> 20;
-  addi(Rd, Rd, lower);
-  slli(Rd, Rd, 8);
-  lower = imm & 0xff;
-  addi(Rd, Rd, lower);
-}
-
 void MacroAssembler::li(Register Rd, int64_t imm) {
   // int64_t is in range 0x8000 0000 0000 0000 ~ 0x7fff ffff ffff ffff
   // li -> c.li
@@ -1586,27 +1546,6 @@ static int patch_addr_in_movptr2(address instruction_address, address target) {
   return MacroAssembler::movptr2_instruction_size;
 }
 
-static int patch_imm_in_li64(address branch, address target) {
-  const int LI64_INSTRUCTIONS_NUM = 8;                                          // lui + addi + slli + addi + slli + addi + slli + addi
-  int64_t lower = (intptr_t)target & 0xffffffff;
-  lower = lower - ((lower << 44) >> 44);
-  int64_t tmp_imm = ((uint64_t)((intptr_t)target & 0xffffffff00000000)) + (uint64_t)lower;
-  int32_t upper =  (tmp_imm - (int32_t)lower) >> 32;
-  int64_t tmp_upper = upper, tmp_lower = upper;
-  tmp_lower = (tmp_lower << 52) >> 52;
-  tmp_upper -= tmp_lower;
-  tmp_upper >>= 12;
-  // Load upper 32 bits. Upper = target[63:32], but if target[31] = 1 or (target[31:20] == 0x7ff && target[19] == 1),
-  // upper = target[63:32] + 1.
-  Assembler::patch(branch + 0,  31, 12, tmp_upper & 0xfffff);                       // Lui.
-  Assembler::patch(branch + 4,  31, 20, tmp_lower & 0xfff);                         // Addi.
-  // Load the rest 32 bits.
-  Assembler::patch(branch + 12, 31, 20, ((int32_t)lower >> 20) & 0xfff);            // Addi.
-  Assembler::patch(branch + 20, 31, 20, (((intptr_t)target << 44) >> 52) & 0xfff);  // Addi.
-  Assembler::patch(branch + 28, 31, 20, (intptr_t)target & 0xff);                   // Addi.
-  return LI64_INSTRUCTIONS_NUM * MacroAssembler::instruction_size;
-}
-
 static int patch_imm_in_li16u(address branch, uint16_t target) {
   Assembler::patch(branch, 31, 12, target); // patch lui only
   return MacroAssembler::instruction_size;
@@ -1677,16 +1616,6 @@ static address get_target_of_movptr2(address insn_addr) {
   return ret;
 }
 
-static address get_target_of_li64(address insn_addr) {
-  assert_cond(insn_addr != nullptr);
-  intptr_t target_address = (((int64_t)Assembler::sextract(Assembler::ld_instr(insn_addr), 31, 12)) & 0xfffff) << 44; // Lui.
-  target_address += ((int64_t)Assembler::sextract(Assembler::ld_instr(insn_addr + 4), 31, 20)) << 32;                 // Addi.
-  target_address += ((int64_t)Assembler::sextract(Assembler::ld_instr(insn_addr + 12), 31, 20)) << 20;                // Addi.
-  target_address += ((int64_t)Assembler::sextract(Assembler::ld_instr(insn_addr + 20), 31, 20)) << 8;                 // Addi.
-  target_address += ((int64_t)Assembler::sextract(Assembler::ld_instr(insn_addr + 28), 31, 20));                      // Addi.
-  return (address)target_address;
-}
-
 address MacroAssembler::get_target_of_li32(address insn_addr) {
   assert_cond(insn_addr != nullptr);
   intptr_t target_address = (((int64_t)Assembler::sextract(Assembler::ld_instr(insn_addr), 31, 12)) & 0xfffff) << 12; // Lui.
@@ -1709,8 +1638,6 @@ int MacroAssembler::pd_patch_instruction_size(address instruction_address, addre
     return patch_addr_in_movptr1(instruction_address, target);
   } else if (MacroAssembler::is_movptr2_at(instruction_address)) {              // movptr2
     return patch_addr_in_movptr2(instruction_address, target);
-  } else if (MacroAssembler::is_li64_at(instruction_address)) {                 // li64
-    return patch_imm_in_li64(instruction_address, target);
   } else if (MacroAssembler::is_li32_at(instruction_address)) {                 // li32
     int64_t imm = (intptr_t)target;
     return patch_imm_in_li32(instruction_address, (int32_t)imm);
@@ -1741,8 +1668,6 @@ address MacroAssembler::target_addr_for_insn(address insn_addr) {
     return get_target_of_movptr1(insn_addr);
   } else if (MacroAssembler::is_movptr2_at(insn_addr)) {          // movptr2
     return get_target_of_movptr2(insn_addr);
-  } else if (MacroAssembler::is_li64_at(insn_addr)) {             // li64
-    return get_target_of_li64(insn_addr);
   } else if (MacroAssembler::is_li32_at(insn_addr)) {             // li32
     return get_target_of_li32(insn_addr);
   } else {
@@ -5072,15 +4997,21 @@ static int reg2offset_out(VMReg r) {
   return (r->reg2stack() + SharedRuntime::out_preserve_stack_slots()) * VMRegImpl::stack_slot_size;
 }
 
-// On 64 bit we will store integer like items to the stack as
-// 64 bits items (riscv64 abi) even though java would only store
-// 32bits for a parameter. On 32bit it will simply be 32 bits
-// So this routine will do 32->32 on 32bit and 32->64 on 64bit
+// The C ABI specifies:
+// "integer scalars narrower than XLEN bits are widened according to the sign
+// of their type up to 32 bits, then sign-extended to XLEN bits."
+// Applies for both passed in register and stack.
+//
+// Java uses 32-bit stack slots; jint, jshort, jchar, jbyte uses one slot.
+// Native uses 64-bit stack slots for all integer scalar types.
+//
+// lw loads the Java stack slot, sign-extends and
+// sd store this widened integer into a 64 bit native stack slot.
 void MacroAssembler::move32_64(VMRegPair src, VMRegPair dst, Register tmp) {
   if (src.first()->is_stack()) {
     if (dst.first()->is_stack()) {
       // stack to stack
-      ld(tmp, Address(fp, reg2offset_in(src.first())));
+      lw(tmp, Address(fp, reg2offset_in(src.first())));
       sd(tmp, Address(sp, reg2offset_out(dst.first())));
     } else {
       // stack to reg

src/hotspot/cpu/riscv/macroAssembler_riscv.hpp

@@ -781,7 +781,6 @@ public:
 
   void li16u(Register Rd, uint16_t imm);
   void li32(Register Rd, int32_t imm);
-  void li64(Register Rd, int64_t imm);
   void li  (Register Rd, int64_t imm);  // optimized load immediate
 
   // mv
@@ -1676,40 +1675,6 @@ private:
            extract_rs1(last_instr) == extract_rd(add);
   }
 
-  // the instruction sequence of li64 is as below:
-  //     lui
-  //     addi
-  //     slli
-  //     addi
-  //     slli
-  //     addi
-  //     slli
-  //     addi
-  static bool check_li64_data_dependency(address instr) {
-    address lui = instr;
-    address addi1 = lui + instruction_size;
-    address slli1 = addi1 + instruction_size;
-    address addi2 = slli1 + instruction_size;
-    address slli2 = addi2 + instruction_size;
-    address addi3 = slli2 + instruction_size;
-    address slli3 = addi3 + instruction_size;
-    address addi4 = slli3 + instruction_size;
-    return extract_rs1(addi1) == extract_rd(lui) &&
-           extract_rs1(addi1) == extract_rd(addi1) &&
-           extract_rs1(slli1) == extract_rd(addi1) &&
-           extract_rs1(slli1) == extract_rd(slli1) &&
-           extract_rs1(addi2) == extract_rd(slli1) &&
-           extract_rs1(addi2) == extract_rd(addi2) &&
-           extract_rs1(slli2) == extract_rd(addi2) &&
-           extract_rs1(slli2) == extract_rd(slli2) &&
-           extract_rs1(addi3) == extract_rd(slli2) &&
-           extract_rs1(addi3) == extract_rd(addi3) &&
-           extract_rs1(slli3) == extract_rd(addi3) &&
-           extract_rs1(slli3) == extract_rd(slli3) &&
-           extract_rs1(addi4) == extract_rd(slli3) &&
-           extract_rs1(addi4) == extract_rd(addi4);
-  }
-
   // the instruction sequence of li16u is as below:
   //     lui
   //     srli
@@ -1754,7 +1719,6 @@ private:
   }
 
   static bool is_li32_at(address instr);
-  static bool is_li64_at(address instr);
   static bool is_pc_relative_at(address branch);
 
   static bool is_membar(address addr) {

src/hotspot/cpu/riscv/methodHandles_riscv.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2024, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2014, Red Hat Inc. All rights reserved.
  * Copyright (c) 2020, 2023, Huawei Technologies Co., Ltd. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
@@ -28,6 +28,7 @@
 #include "asm/macroAssembler.hpp"
 #include "classfile/javaClasses.inline.hpp"
 #include "classfile/vmClasses.hpp"
+#include "compiler/disassembler.hpp"
 #include "interpreter/interpreter.hpp"
 #include "interpreter/interpreterRuntime.hpp"
 #include "memory/allocation.inline.hpp"
@@ -37,7 +38,7 @@
 #include "runtime/frame.inline.hpp"
 #include "runtime/stubRoutines.hpp"
 
-#define __ _masm->
+#define __ Disassembler::hook<MacroAssembler>(__FILE__, __LINE__, _masm)->
 
 #ifdef PRODUCT
 #define BLOCK_COMMENT(str) /* nothing */

src/hotspot/cpu/riscv/nativeInst_riscv.cpp

@@ -68,10 +68,10 @@ address NativeCall::destination() const {
 // Used in the runtime linkage of calls; see class CompiledIC.
 //
 // Add parameter assert_lock to switch off assertion
-// during code generation, where no patching lock is needed.
+// during code generation, where no lock is needed.
 void NativeCall::set_destination_mt_safe(address dest, bool assert_lock) {
   assert(!assert_lock ||
-         (Patching_lock->is_locked() || SafepointSynchronize::is_at_safepoint()) ||
+         (CodeCache_lock->is_locked() || SafepointSynchronize::is_at_safepoint()) ||
          CompiledICLocker::is_safe(addr_at(0)),
          "concurrent code patching");
 

src/hotspot/cpu/riscv/riscv.ad

@@ -84,8 +84,8 @@ reg_def R0      ( NS,  NS,  Op_RegI, 0,  x0->as_VMReg()         ); // zr
 reg_def R0_H    ( NS,  NS,  Op_RegI, 0,  x0->as_VMReg()->next() );
 reg_def R1      ( NS,  SOC, Op_RegI, 1,  x1->as_VMReg()         ); // ra
 reg_def R1_H    ( NS,  SOC, Op_RegI, 1,  x1->as_VMReg()->next() );
-reg_def R2      ( NS,  SOE, Op_RegI, 2,  x2->as_VMReg()         ); // sp
-reg_def R2_H    ( NS,  SOE, Op_RegI, 2,  x2->as_VMReg()->next() );
+reg_def R2      ( NS,  NS,  Op_RegI, 2,  x2->as_VMReg()         ); // sp
+reg_def R2_H    ( NS,  NS,  Op_RegI, 2,  x2->as_VMReg()->next() );
 reg_def R3      ( NS,  NS,  Op_RegI, 3,  x3->as_VMReg()         ); // gp
 reg_def R3_H    ( NS,  NS,  Op_RegI, 3,  x3->as_VMReg()->next() );
 reg_def R4      ( NS,  NS,  Op_RegI, 4,  x4->as_VMReg()         ); // tp

src/hotspot/cpu/riscv/riscv_v.ad

@@ -4795,36 +4795,49 @@ instruct vcountTrailingZeros(vReg dst, vReg src) %{
 
 // ------------------------------ Vector Load Gather ---------------------------
 
-instruct gather_load(vReg dst, indirect mem, vReg idx) %{
-  predicate(type2aelembytes(Matcher::vector_element_basic_type(n)) == 4 ||
-            type2aelembytes(Matcher::vector_element_basic_type(n)) == 8);
+instruct gather_loadS(vReg dst, indirect mem, vReg idx) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n)) == 4);
   match(Set dst (LoadVectorGather mem idx));
   effect(TEMP_DEF dst);
-  format %{ "gather_load $dst, $mem, $idx" %}
+  format %{ "gather_loadS $dst, $mem, $idx" %}
   ins_encode %{
-    __ vmv1r_v(as_VectorRegister($dst$$reg), as_VectorRegister($idx$$reg));
     BasicType bt = Matcher::vector_element_basic_type(this);
     Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
     __ vsetvli_helper(bt, Matcher::vector_length(this));
-    __ vsll_vi(as_VectorRegister($dst$$reg), as_VectorRegister($dst$$reg), (int)sew);
+    __ vsll_vi(as_VectorRegister($dst$$reg), as_VectorRegister($idx$$reg), (int)sew);
     __ vluxei32_v(as_VectorRegister($dst$$reg), as_Register($mem$$base),
                   as_VectorRegister($dst$$reg));
  %}
   ins_pipe(pipe_slow);
 %}
 
-instruct gather_load_masked(vReg dst, indirect mem, vReg idx, vRegMask_V0 v0, vReg tmp) %{
-  predicate(type2aelembytes(Matcher::vector_element_basic_type(n)) == 4 ||
-            type2aelembytes(Matcher::vector_element_basic_type(n)) == 8);
-  match(Set dst (LoadVectorGatherMasked mem (Binary idx v0)));
-  effect(TEMP_DEF dst, TEMP tmp);
-  format %{ "gather_load_masked $dst, $mem, $idx, $v0\t# KILL $tmp" %}
+instruct gather_loadD(vReg dst, indirect mem, vReg idx) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n)) == 8);
+  match(Set dst (LoadVectorGather mem idx));
+  effect(TEMP_DEF dst);
+  format %{ "gather_loadD $dst, $mem, $idx" %}
   ins_encode %{
-    __ vmv1r_v(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg));
     BasicType bt = Matcher::vector_element_basic_type(this);
     Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
     __ vsetvli_helper(bt, Matcher::vector_length(this));
-    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($tmp$$reg), (int)sew);
+    __ vzext_vf2(as_VectorRegister($dst$$reg), as_VectorRegister($idx$$reg));
+    __ vsll_vi(as_VectorRegister($dst$$reg), as_VectorRegister($dst$$reg), (int)sew);
+    __ vluxei64_v(as_VectorRegister($dst$$reg), as_Register($mem$$base),
+                  as_VectorRegister($dst$$reg));
+ %}
+  ins_pipe(pipe_slow);
+%}
+
+instruct gather_loadS_masked(vReg dst, indirect mem, vReg idx, vRegMask_V0 v0, vReg tmp) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n)) == 4);
+  match(Set dst (LoadVectorGatherMasked mem (Binary idx v0)));
+  effect(TEMP_DEF dst, TEMP tmp);
+  format %{ "gather_loadS_masked $dst, $mem, $idx, $v0\t# KILL $tmp" %}
+  ins_encode %{
+    BasicType bt = Matcher::vector_element_basic_type(this);
+    Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
+    __ vsetvli_helper(bt, Matcher::vector_length(this));
+    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg), (int)sew);
     __ vxor_vv(as_VectorRegister($dst$$reg), as_VectorRegister($dst$$reg),
                as_VectorRegister($dst$$reg));
     __ vluxei32_v(as_VectorRegister($dst$$reg), as_Register($mem$$base),
@@ -4833,44 +4846,93 @@ instruct gather_load_masked(vReg dst, indirect mem, vReg idx, vRegMask_V0 v0, vR
   ins_pipe(pipe_slow);
 %}
 
+instruct gather_loadD_masked(vReg dst, indirect mem, vReg idx, vRegMask_V0 v0, vReg tmp) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n)) == 8);
+  match(Set dst (LoadVectorGatherMasked mem (Binary idx v0)));
+  effect(TEMP_DEF dst, TEMP tmp);
+  format %{ "gather_loadD_masked $dst, $mem, $idx, $v0\t# KILL $tmp" %}
+  ins_encode %{
+    BasicType bt = Matcher::vector_element_basic_type(this);
+    Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
+    __ vsetvli_helper(bt, Matcher::vector_length(this));
+    __ vzext_vf2(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg));
+    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($tmp$$reg), (int)sew);
+    __ vxor_vv(as_VectorRegister($dst$$reg), as_VectorRegister($dst$$reg),
+               as_VectorRegister($dst$$reg));
+    __ vluxei64_v(as_VectorRegister($dst$$reg), as_Register($mem$$base),
+                  as_VectorRegister($tmp$$reg), Assembler::v0_t);
+ %}
+  ins_pipe(pipe_slow);
+%}
+
 // ------------------------------ Vector Store Scatter -------------------------
 
-instruct scatter_store(indirect mem, vReg src, vReg idx, vReg tmp) %{
-  predicate(type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 4 ||
-            type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 8);
+instruct scatter_storeS(indirect mem, vReg src, vReg idx, vReg tmp) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 4);
   match(Set mem (StoreVectorScatter mem (Binary src idx)));
   effect(TEMP tmp);
-  format %{ "scatter_store $mem, $idx, $src\t# KILL $tmp" %}
+  format %{ "scatter_storeS $mem, $idx, $src\t# KILL $tmp" %}
   ins_encode %{
-    __ vmv1r_v(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg));
     BasicType bt = Matcher::vector_element_basic_type(this, $src);
     Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
     __ vsetvli_helper(bt, Matcher::vector_length(this, $src));
-    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($tmp$$reg), (int)sew);
+    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg), (int)sew);
     __ vsuxei32_v(as_VectorRegister($src$$reg), as_Register($mem$$base),
                   as_VectorRegister($tmp$$reg));
   %}
   ins_pipe(pipe_slow);
 %}
 
-instruct scatter_store_masked(indirect mem, vReg src, vReg idx, vRegMask_V0 v0, vReg tmp) %{
-  predicate(type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 4 ||
-            type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 8);
-  match(Set mem (StoreVectorScatterMasked mem (Binary src (Binary idx v0))));
+instruct scatter_storeD(indirect mem, vReg src, vReg idx, vReg tmp) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 8);
+  match(Set mem (StoreVectorScatter mem (Binary src idx)));
   effect(TEMP tmp);
-  format %{ "scatter_store_masked $mem, $idx, $src, $v0\t# KILL $tmp" %}
+  format %{ "scatter_storeD $mem, $idx, $src\t# KILL $tmp" %}
   ins_encode %{
-    __ vmv1r_v(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg));
     BasicType bt = Matcher::vector_element_basic_type(this, $src);
     Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
     __ vsetvli_helper(bt, Matcher::vector_length(this, $src));
+    __ vzext_vf2(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg));
     __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($tmp$$reg), (int)sew);
+    __ vsuxei64_v(as_VectorRegister($src$$reg), as_Register($mem$$base),
+                  as_VectorRegister($tmp$$reg));
+  %}
+  ins_pipe(pipe_slow);
+%}
+
+instruct scatter_storeS_masked(indirect mem, vReg src, vReg idx, vRegMask_V0 v0, vReg tmp) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 4);
+  match(Set mem (StoreVectorScatterMasked mem (Binary src (Binary idx v0))));
+  effect(TEMP tmp);
+  format %{ "scatter_storeS_masked $mem, $idx, $src, $v0\t# KILL $tmp" %}
+  ins_encode %{
+    BasicType bt = Matcher::vector_element_basic_type(this, $src);
+    Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
+    __ vsetvli_helper(bt, Matcher::vector_length(this, $src));
+    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg), (int)sew);
     __ vsuxei32_v(as_VectorRegister($src$$reg), as_Register($mem$$base),
                   as_VectorRegister($tmp$$reg), Assembler::v0_t);
   %}
   ins_pipe(pipe_slow);
 %}
 
+instruct scatter_storeD_masked(indirect mem, vReg src, vReg idx, vRegMask_V0 v0, vReg tmp) %{
+  predicate(type2aelembytes(Matcher::vector_element_basic_type(n->in(3)->in(1))) == 8);
+  match(Set mem (StoreVectorScatterMasked mem (Binary src (Binary idx v0))));
+  effect(TEMP tmp);
+  format %{ "scatter_storeD_masked $mem, $idx, $src, $v0\t# KILL $tmp" %}
+  ins_encode %{
+    BasicType bt = Matcher::vector_element_basic_type(this, $src);
+    Assembler::SEW sew = Assembler::elemtype_to_sew(bt);
+    __ vsetvli_helper(bt, Matcher::vector_length(this, $src));
+    __ vzext_vf2(as_VectorRegister($tmp$$reg), as_VectorRegister($idx$$reg));
+    __ vsll_vi(as_VectorRegister($tmp$$reg), as_VectorRegister($tmp$$reg), (int)sew);
+    __ vsuxei64_v(as_VectorRegister($src$$reg), as_Register($mem$$base),
+                  as_VectorRegister($tmp$$reg), Assembler::v0_t);
+  %}
+  ins_pipe(pipe_slow);
+%}
+
 // ------------------------------ Populate Index to a Vector -------------------
 
 instruct populateindex(vReg dst, iRegIorL2I src1, iRegIorL2I src2, vReg tmp) %{

src/hotspot/cpu/riscv/smallRegisterMap_riscv.inline.hpp

@@ -30,8 +30,15 @@
 
 // Java frames don't have callee saved registers (except for fp), so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
 private:
   static void assert_is_fp(VMReg r) NOT_DEBUG_RETURN
                                     DEBUG_ONLY({ assert (r == fp->as_VMReg() || r == fp->as_VMReg()->next(), "Reg: %s", r->name()); })
@@ -48,17 +55,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-  #ifdef ASSERT
-    for(int i = 0; i < RegisterMap::reg_count; i++) {
-      VMReg r = VMRegImpl::as_VMReg(i);
-      if (map->location(r, (intptr_t*)nullptr) != nullptr) assert_is_fp(r);
-    }
-  #endif
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     assert_is_fp(reg);
     return (address)(sp - 2);

src/hotspot/cpu/riscv/stubGenerator_riscv.cpp

@@ -5374,6 +5374,29 @@ static const int64_t right_3_bits = right_n_bits(3);
     return start;
   }
 
+  // load Method* target of MethodHandle
+  // j_rarg0 = jobject receiver
+  // xmethod = Method* result
+  address generate_upcall_stub_load_target() {
+
+    StubCodeMark mark(this, "StubRoutines", "upcall_stub_load_target");
+    address start = __ pc();
+
+    __ resolve_global_jobject(j_rarg0, t0, t1);
+      // Load target method from receiver
+    __ load_heap_oop(xmethod, Address(j_rarg0, java_lang_invoke_MethodHandle::form_offset()), t0, t1);
+    __ load_heap_oop(xmethod, Address(xmethod, java_lang_invoke_LambdaForm::vmentry_offset()), t0, t1);
+    __ load_heap_oop(xmethod, Address(xmethod, java_lang_invoke_MemberName::method_offset()), t0, t1);
+    __ access_load_at(T_ADDRESS, IN_HEAP, xmethod,
+                      Address(xmethod, java_lang_invoke_ResolvedMethodName::vmtarget_offset()),
+                      noreg, noreg);
+    __ sd(xmethod, Address(xthread, JavaThread::callee_target_offset())); // just in case callee is deoptimized
+
+    __ ret();
+
+    return start;
+  }
+
   // Continuation point for throwing of implicit exceptions that are
   // not handled in the current activation. Fabricates an exception
   // oop and initiates normal exception dispatching in this
@@ -5567,6 +5590,7 @@ static const int64_t right_3_bits = right_n_bits(3);
     }
 
     StubRoutines::_upcall_stub_exception_handler = generate_upcall_stub_exception_handler();
+    StubRoutines::_upcall_stub_load_target = generate_upcall_stub_load_target();
 
     StubRoutines::riscv::set_completed();
   }

src/hotspot/cpu/riscv/templateInterpreterGenerator_riscv.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 2024, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2014, 2020, Red Hat Inc. All rights reserved.
  * Copyright (c) 2020, 2022, Huawei Technologies Co., Ltd. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
@@ -27,6 +27,7 @@
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
 #include "classfile/javaClasses.hpp"
+#include "compiler/disassembler.hpp"
 #include "gc/shared/barrierSetAssembler.hpp"
 #include "interpreter/bytecodeHistogram.hpp"
 #include "interpreter/bytecodeTracer.hpp"
@@ -70,7 +71,7 @@
 // Max size with JVMTI
 int TemplateInterpreter::InterpreterCodeSize = 256 * 1024;
 
-#define __ _masm->
+#define __ Disassembler::hook<InterpreterMacroAssembler>(__FILE__, __LINE__, _masm)->
 
 //-----------------------------------------------------------------------------
 
@@ -1748,13 +1749,21 @@ void TemplateInterpreterGenerator::set_vtos_entry_points(Template* t,
                                                          address& vep) {
   assert(t != nullptr && t->is_valid() && t->tos_in() == vtos, "illegal template");
   Label L;
-  aep = __ pc();  __ push_ptr();  __ j(L);
-  fep = __ pc();  __ push_f();    __ j(L);
-  dep = __ pc();  __ push_d();    __ j(L);
-  lep = __ pc();  __ push_l();    __ j(L);
-  bep = cep = sep =
-  iep = __ pc();  __ push_i();
-  vep = __ pc();
+  aep = __ pc();     // atos entry point
+      __ push_ptr();
+      __ j(L);
+  fep = __ pc();     // ftos entry point
+      __ push_f();
+      __ j(L);
+  dep = __ pc();     // dtos entry point
+      __ push_d();
+      __ j(L);
+  lep = __ pc();     // ltos entry point
+      __ push_l();
+      __ j(L);
+  bep = cep = sep = iep = __ pc();     // [bcsi]tos entry point
+      __ push_i();
+  vep = __ pc();     // vtos entry point
   __ bind(L);
   generate_and_dispatch(t);
 }

src/hotspot/cpu/riscv/templateTable_riscv.cpp

@@ -26,6 +26,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
+#include "compiler/disassembler.hpp"
 #include "gc/shared/barrierSetAssembler.hpp"
 #include "gc/shared/collectedHeap.hpp"
 #include "gc/shared/tlab_globals.hpp"
@@ -49,7 +50,7 @@
 #include "runtime/synchronizer.hpp"
 #include "utilities/powerOfTwo.hpp"
 
-#define __ _masm->
+#define __ Disassembler::hook<InterpreterMacroAssembler>(__FILE__, __LINE__, _masm)->
 
 // Address computation: local variables
 
@@ -178,7 +179,6 @@ void TemplateTable::patch_bytecode(Bytecodes::Code bc, Register bc_reg,
         __ la(temp_reg, Address(temp_reg, in_bytes(ResolvedFieldEntry::put_code_offset())));
       }
       // Load-acquire the bytecode to match store-release in ResolvedFieldEntry::fill_in()
-      __ membar(MacroAssembler::AnyAny);
       __ lbu(temp_reg, Address(temp_reg, 0));
       __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
       __ mv(bc_reg, bc);
@@ -320,7 +320,6 @@ void TemplateTable::ldc(LdcType type) {
   // get type
   __ addi(x13, x11, tags_offset);
   __ add(x13, x10, x13);
-  __ membar(MacroAssembler::AnyAny);
   __ lbu(x13, Address(x13, 0));
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
 
@@ -2189,7 +2188,6 @@ void TemplateTable::resolve_cache_and_index_for_method(int byte_no,
       break;
   }
   // Load-acquire the bytecode to match store-release in InterpreterRuntime
-  __ membar(MacroAssembler::AnyAny);
   __ lbu(temp, Address(temp, 0));
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
 
@@ -2241,7 +2239,6 @@ void TemplateTable::resolve_cache_and_index_for_field(int byte_no,
     __ la(temp, Address(Rcache, in_bytes(ResolvedFieldEntry::put_code_offset())));
   }
   // Load-acquire the bytecode to match store-release in ResolvedFieldEntry::fill_in()
-  __ membar(MacroAssembler::AnyAny);
   __ lbu(temp, Address(temp, 0));
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
   __ mv(t0, (int) code);  // have we resolved this bytecode?
@@ -2403,7 +2400,6 @@ void TemplateTable::load_invokedynamic_entry(Register method) {
   Label resolved;
 
   __ load_resolved_indy_entry(cache, index);
-  __ membar(MacroAssembler::AnyAny);
   __ ld(method, Address(cache, in_bytes(ResolvedIndyEntry::method_offset())));
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
 
@@ -2418,7 +2414,6 @@ void TemplateTable::load_invokedynamic_entry(Register method) {
   __ call_VM(noreg, entry, method);
   // Update registers with resolved info
   __ load_resolved_indy_entry(cache, index);
-  __ membar(MacroAssembler::AnyAny);
   __ ld(method, Address(cache, in_bytes(ResolvedIndyEntry::method_offset())));
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
 
@@ -3533,7 +3528,6 @@ void TemplateTable::_new() {
   const int tags_offset = Array<u1>::base_offset_in_bytes();
   __ add(t0, x10, x13);
   __ la(t0, Address(t0, tags_offset));
-  __ membar(MacroAssembler::AnyAny);
   __ lbu(t0, t0);
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
   __ sub(t1, t0, (u1)JVM_CONSTANT_Class);
@@ -3652,7 +3646,6 @@ void TemplateTable::checkcast() {
   // See if bytecode has already been quicked
   __ add(t0, x13, Array<u1>::base_offset_in_bytes());
   __ add(x11, t0, x9);
-  __ membar(MacroAssembler::AnyAny);
   __ lbu(x11, x11);
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
   __ sub(t0, x11, (u1)JVM_CONSTANT_Class);
@@ -3708,7 +3701,6 @@ void TemplateTable::instanceof() {
   // See if bytecode has already been quicked
   __ add(t0, x13, Array<u1>::base_offset_in_bytes());
   __ add(x11, t0, x9);
-  __ membar(MacroAssembler::AnyAny);
   __ lbu(x11, x11);
   __ membar(MacroAssembler::LoadLoad | MacroAssembler::LoadStore);
   __ sub(t0, x11, (u1)JVM_CONSTANT_Class);

src/hotspot/cpu/riscv/upcallLinker_riscv.cpp

@@ -25,6 +25,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.hpp"
+#include "classfile/javaClasses.hpp"
 #include "logging/logStream.hpp"
 #include "memory/resourceArea.hpp"
 #include "prims/upcallLinker.hpp"
@@ -117,7 +118,7 @@ static void restore_callee_saved_registers(MacroAssembler* _masm, const ABIDescr
 static const int upcall_stub_code_base_size = 1024;
 static const int upcall_stub_size_per_arg = 16;
 
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,
@@ -223,7 +224,6 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
   __ block_comment("{ on_entry");
   __ la(c_rarg0, Address(sp, frame_data_offset));
-  __ movptr(c_rarg1, (address) receiver);
   __ rt_call(CAST_FROM_FN_PTR(address, UpcallLinker::on_entry));
   __ mv(xthread, x10);
   __ reinit_heapbase();
@@ -260,12 +260,10 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   arg_shuffle.generate(_masm, as_VMStorage(shuffle_reg), abi._shadow_space_bytes, 0);
   __ block_comment("} argument shuffle");
 
-  __ block_comment("{ receiver ");
-  __ get_vm_result(j_rarg0, xthread);
-  __ block_comment("} receiver ");
-
-  __ mov_metadata(xmethod, entry);
-  __ sd(xmethod, Address(xthread, JavaThread::callee_target_offset())); // just in case callee is deoptimized
+  __ block_comment("{ load target ");
+  __ movptr(j_rarg0, (address) receiver);
+  __ far_call(RuntimeAddress(StubRoutines::upcall_stub_load_target())); // loads Method* into xmethod
+  __ block_comment("} load target ");
 
   __ push_cont_fastpath(xthread);
 
@@ -338,7 +336,7 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
 #ifndef PRODUCT
   stringStream ss;
-  ss.print("upcall_stub_%s", entry->signature()->as_C_string());
+  ss.print("upcall_stub_%s", signature->as_C_string());
   const char *name = _masm->code_string(ss.as_string());
 #else // PRODUCT
   const char* name = "upcall_stub";

src/hotspot/cpu/s390/c1_LIRAssembler_s390.cpp

@@ -131,9 +131,19 @@ void LIR_Assembler::osr_entry() {
   // copied into place by code emitted in the IR.
 
   Register OSR_buf = osrBufferPointer()->as_register();
-  { assert(frame::interpreter_frame_monitor_size() == BasicObjectLock::size(), "adjust code below");
-    int monitor_offset = BytesPerWord * method()->max_locals() +
-      (2 * BytesPerWord) * (number_of_locks - 1);
+  {
+    assert(frame::interpreter_frame_monitor_size() == BasicObjectLock::size(), "adjust code below");
+
+    const int locals_space = BytesPerWord * method() -> max_locals();
+    int monitor_offset = locals_space + (2 * BytesPerWord) * (number_of_locks - 1);
+    bool large_offset = !Immediate::is_simm20(monitor_offset + BytesPerWord) && number_of_locks > 0;
+
+    if (large_offset) {
+      // z_lg can only handle displacement upto 20bit signed binary integer
+      __ z_algfi(OSR_buf, locals_space);
+      monitor_offset -= locals_space;
+    }
+
     // SharedRuntime::OSR_migration_begin() packs BasicObjectLocks in
     // the OSR buffer using 2 word entries: first the lock and then
     // the oop.
@@ -147,6 +157,10 @@ void LIR_Assembler::osr_entry() {
       __ z_lg(Z_R1_scratch, slot_offset + 1*BytesPerWord, OSR_buf);
       __ z_stg(Z_R1_scratch, frame_map()->address_for_monitor_object(i));
     }
+
+    if (large_offset) {
+      __ z_slgfi(OSR_buf, locals_space);
+    }
   }
 }
 

src/hotspot/cpu/s390/gc/shared/barrierSetAssembler_s390.cpp

@@ -105,16 +105,60 @@ void BarrierSetAssembler::store_at(MacroAssembler* masm, DecoratorSet decorators
   }
 }
 
+// Generic implementation. GCs can provide an optimized one.
 void BarrierSetAssembler::resolve_jobject(MacroAssembler* masm, Register value, Register tmp1, Register tmp2) {
-  NearLabel Ldone;
-  __ z_ltgr(tmp1, value);
-  __ z_bre(Ldone);          // Use null result as-is.
 
-  __ z_nill(value, ~JNIHandles::tag_mask);
-  __ z_lg(value, 0, value); // Resolve (untagged) jobject.
+  assert_different_registers(value, tmp1, tmp2);
+  NearLabel done, weak_tag, verify, tagged;
+  __ z_ltgr(value, value);
+  __ z_bre(done);          // Use null result as-is.
 
+  __ z_tmll(value, JNIHandles::tag_mask);
+  __ z_btrue(tagged); // not zero
+
+  // Resolve Local handle
+  __ access_load_at(T_OBJECT, IN_NATIVE | AS_RAW, Address(value, 0), value, tmp1, tmp2);
+  __ z_bru(verify);
+
+  __ bind(tagged);
+  __ testbit(value, exact_log2(JNIHandles::TypeTag::weak_global)); // test for weak tag
+  __ z_btrue(weak_tag);
+
+  // resolve global handle
+  __ access_load_at(T_OBJECT, IN_NATIVE, Address(value, -JNIHandles::TypeTag::global), value, tmp1, tmp2);
+  __ z_bru(verify);
+
+  __ bind(weak_tag);
+  // resolve jweak.
+  __ access_load_at(T_OBJECT, IN_NATIVE | ON_PHANTOM_OOP_REF,
+                    Address(value, -JNIHandles::TypeTag::weak_global), value, tmp1, tmp2);
+  __ bind(verify);
   __ verify_oop(value, FILE_AND_LINE);
-  __ bind(Ldone);
+  __ bind(done);
+}
+
+// Generic implementation. GCs can provide an optimized one.
+void BarrierSetAssembler::resolve_global_jobject(MacroAssembler* masm, Register value, Register tmp1, Register tmp2) {
+  assert_different_registers(value, tmp1, tmp2);
+  NearLabel done;
+
+  __ z_ltgr(value, value);
+  __ z_bre(done); // use null as-is.
+
+#ifdef ASSERT
+  {
+    NearLabel valid_global_tag;
+    __ testbit(value, exact_log2(JNIHandles::TypeTag::global)); // test for global tag
+    __ z_btrue(valid_global_tag);
+    __ stop("non global jobject using resolve_global_jobject");
+    __ bind(valid_global_tag);
+  }
+#endif // ASSERT
+
+  // Resolve global handle
+  __ access_load_at(T_OBJECT, IN_NATIVE, Address(value, -JNIHandles::TypeTag::global), value, tmp1, tmp2);
+  __ verify_oop(value, FILE_AND_LINE);
+  __ bind(done);
 }
 
 void BarrierSetAssembler::try_resolve_jobject_in_native(MacroAssembler* masm, Register jni_env,

src/hotspot/cpu/s390/gc/shared/barrierSetAssembler_s390.hpp

@@ -51,6 +51,7 @@ public:
                         const Address& addr, Register val, Register tmp1, Register tmp2, Register tmp3);
 
   virtual void resolve_jobject(MacroAssembler* masm, Register value, Register tmp1, Register tmp2);
+  virtual void resolve_global_jobject(MacroAssembler* masm, Register value, Register tmp1, Register tmp2);
 
   virtual void try_resolve_jobject_in_native(MacroAssembler* masm, Register jni_env,
                                              Register obj, Register tmp, Label& slowpath);

src/hotspot/cpu/s390/gc/shared/modRefBarrierSetAssembler_s390.cpp

@@ -1,6 +1,6 @@
 /*
- * Copyright (c) 2018, 2019, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2018, 2019 SAP SE. All rights reserved.
+ * Copyright (c) 2018, 2024, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2018, 2024 SAP SE. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -26,6 +26,7 @@
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
 #include "gc/shared/modRefBarrierSetAssembler.hpp"
+#include "runtime/jniHandles.hpp"
 
 #define __ masm->
 
@@ -58,3 +59,16 @@ void ModRefBarrierSetAssembler::store_at(MacroAssembler* masm, DecoratorSet deco
     BarrierSetAssembler::store_at(masm, decorators, type, dst, val, tmp1, tmp2, tmp3);
   }
 }
+
+void ModRefBarrierSetAssembler::resolve_jobject(MacroAssembler* masm, Register value, Register tmp1, Register tmp2) {
+  NearLabel done;
+
+  __ z_ltgr(value, value);
+  __ z_bre(done);  // use null as-is.
+
+  __ z_nill(value, ~JNIHandles::tag_mask);
+  __ z_lg(value, 0, value); // Resolve (untagged) jobject.
+
+  __ verify_oop(value, FILE_AND_LINE);
+  __ bind(done);
+}

src/hotspot/cpu/s390/gc/shared/modRefBarrierSetAssembler_s390.hpp

@@ -1,6 +1,6 @@
 /*
- * Copyright (c) 2018, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2018 SAP SE. All rights reserved.
+ * Copyright (c) 2018, 2024, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2018, 2024 SAP SE. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -48,6 +48,8 @@ public:
 
   virtual void store_at(MacroAssembler* masm, DecoratorSet decorators, BasicType type,
                         const Address& dst, Register val, Register tmp1, Register tmp2, Register tmp3);
+
+  virtual void resolve_jobject(MacroAssembler* masm, Register value, Register tmp1, Register tmp2);
 };
 
 #endif // CPU_S390_GC_SHARED_MODREFBARRIERSETASSEMBLER_S390_HPP

src/hotspot/cpu/s390/macroAssembler_s390.cpp

@@ -3380,6 +3380,11 @@ void MacroAssembler::resolve_jobject(Register value, Register tmp1, Register tmp
   bs->resolve_jobject(this, value, tmp1, tmp2);
 }
 
+void MacroAssembler::resolve_global_jobject(Register value, Register tmp1, Register tmp2) {
+  BarrierSetAssembler* bs = BarrierSet::barrier_set()->barrier_set_assembler();
+  bs->resolve_global_jobject(this, value, tmp1, tmp2);
+}
+
 // Last_Java_sp must comply to the rules in frame_s390.hpp.
 void MacroAssembler::set_last_Java_frame(Register last_Java_sp, Register last_Java_pc, bool allow_relocation) {
   BLOCK_COMMENT("set_last_Java_frame {");

src/hotspot/cpu/s390/macroAssembler_s390.hpp

@@ -730,6 +730,7 @@ class MacroAssembler: public Assembler {
   void lightweight_unlock(Register obj, Register hdr, Register tmp, Label& slow);
 
   void resolve_jobject(Register value, Register tmp1, Register tmp2);
+  void resolve_global_jobject(Register value, Register tmp1, Register tmp2);
 
   // Support for last Java frame (but use call_VM instead where possible).
  private:
@@ -791,7 +792,6 @@ class MacroAssembler: public Assembler {
   void compare_klass_ptr(Register Rop1, int64_t disp, Register Rbase, bool maybenull);
 
   // Access heap oop, handle encoding and GC barriers.
- private:
   void access_store_at(BasicType type, DecoratorSet decorators,
                        const Address& addr, Register val,
                        Register tmp1, Register tmp2, Register tmp3);

src/hotspot/cpu/s390/nativeInst_s390.cpp

@@ -658,8 +658,8 @@ void NativeGeneralJump::insert_unconditional(address code_pos, address entry) {
 
 void NativeGeneralJump::replace_mt_safe(address instr_addr, address code_buffer) {
   assert(((intptr_t)instr_addr & (BytesPerWord-1)) == 0, "requirement for mt safe patching");
-  // Bytes_after_jump cannot change, because we own the Patching_lock.
-  assert(Patching_lock->owned_by_self(), "must hold lock to patch instruction");
+  // Bytes_after_jump cannot change, because we own the CodeCache_lock.
+  assert(CodeCache_lock->owned_by_self(), "must hold lock to patch instruction");
   intptr_t bytes_after_jump = (*(intptr_t*)instr_addr)  & 0x000000000000ffffL; // 2 bytes after jump.
   intptr_t load_const_bytes = (*(intptr_t*)code_buffer) & 0xffffffffffff0000L;
   *(intptr_t*)instr_addr = load_const_bytes | bytes_after_jump;

src/hotspot/cpu/s390/smallRegisterMap_s390.inline.hpp

@@ -30,8 +30,15 @@
 
 // Java frames don't have callee saved registers (except for rfp), so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
 private:
   static void assert_is_rfp(VMReg r) NOT_DEBUG_RETURN
                                      DEBUG_ONLY({ Unimplemented(); })
@@ -46,12 +53,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-    Unimplemented();
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     Unimplemented();
     return nullptr;

src/hotspot/cpu/s390/stubGenerator_s390.cpp

@@ -3107,6 +3107,29 @@ class StubGenerator: public StubCodeGenerator {
     return start;
   }
 
+  // load Method* target of MethodHandle
+  // Z_ARG1 = jobject receiver
+  // Z_method = Method* result
+  address generate_upcall_stub_load_target() {
+    StubCodeMark mark(this, "StubRoutines", "upcall_stub_load_target");
+    address start = __ pc();
+
+    __ resolve_global_jobject(Z_ARG1, Z_tmp_1, Z_tmp_2);
+      // Load target method from receiver
+    __ load_heap_oop(Z_method, Address(Z_ARG1, java_lang_invoke_MethodHandle::form_offset()),
+                    noreg, noreg, IS_NOT_NULL);
+    __ load_heap_oop(Z_method, Address(Z_method, java_lang_invoke_LambdaForm::vmentry_offset()),
+                    noreg, noreg, IS_NOT_NULL);
+    __ load_heap_oop(Z_method, Address(Z_method, java_lang_invoke_MemberName::method_offset()),
+                    noreg, noreg, IS_NOT_NULL);
+    __ z_lg(Z_method, Address(Z_method, java_lang_invoke_ResolvedMethodName::vmtarget_offset()));
+    __ z_stg(Z_method, Address(Z_thread, JavaThread::callee_target_offset())); // just in case callee is deoptimized
+
+    __ z_br(Z_R14);
+
+    return start;
+  }
+
   void generate_initial_stubs() {
     // Generates all stubs and initializes the entry points.
 
@@ -3188,6 +3211,7 @@ class StubGenerator: public StubCodeGenerator {
     }
 
     StubRoutines::_upcall_stub_exception_handler = generate_upcall_stub_exception_handler();
+    StubRoutines::_upcall_stub_load_target = generate_upcall_stub_load_target();
   }
 
   void generate_compiler_stubs() {

src/hotspot/cpu/s390/upcallLinker_s390.cpp

@@ -23,6 +23,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.inline.hpp"
+#include "classfile/javaClasses.hpp"
 #include "logging/logStream.hpp"
 #include "memory/resourceArea.hpp"
 #include "prims/upcallLinker.hpp"
@@ -116,7 +117,7 @@ static void restore_callee_saved_registers(MacroAssembler* _masm, const ABIDescr
 
 static const int upcall_stub_code_base_size = 1024;
 static const int upcall_stub_size_per_arg = 16; // arg save & restore + move
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,
@@ -206,7 +207,6 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   __ block_comment("on_entry {");
   __ load_const_optimized(call_target_address, CAST_FROM_FN_PTR(uint64_t, UpcallLinker::on_entry));
   __ z_aghik(Z_ARG1, Z_SP, frame_data_offset);
-  __ load_const_optimized(Z_ARG2, (intptr_t)receiver);
   __ call(call_target_address);
   __ z_lgr(Z_thread, Z_RET);
   __ block_comment("} on_entry");
@@ -216,12 +216,11 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   arg_shuffle.generate(_masm, shuffle_reg, abi._shadow_space_bytes, frame::z_jit_out_preserve_size);
   __ block_comment("} argument_shuffle");
 
-  __ block_comment("receiver {");
-  __ get_vm_result(Z_ARG1);
-  __ block_comment("} receiver");
-
-  __ load_const_optimized(Z_method, (intptr_t)entry);
-  __ z_stg(Z_method, Address(Z_thread, in_bytes(JavaThread::callee_target_offset())));
+  __ block_comment("load_target {");
+  __ load_const_optimized(Z_ARG1, (intptr_t)receiver);
+  __ load_const_optimized(call_target_address, StubRoutines::upcall_stub_load_target());
+  __ call(call_target_address); // load taget Method* into Z_method
+  __ block_comment("} load_target");
 
   __ z_lg(call_target_address, Address(Z_method, in_bytes(Method::from_compiled_offset())));
   __ call(call_target_address);
@@ -274,7 +273,7 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
 #ifndef PRODUCT
   stringStream ss;
-  ss.print("upcall_stub_%s", entry->signature()->as_C_string());
+  ss.print("upcall_stub_%s", signature->as_C_string());
   const char* name = _masm->code_string(ss.as_string());
 #else // PRODUCT
   const char* name = "upcall_stub";

src/hotspot/cpu/x86/macroAssembler_x86.cpp

@@ -4940,9 +4940,8 @@ void MacroAssembler::lookup_secondary_supers_table_slow_path(Register r_super_kl
 
   // The bitmap is full to bursting.
   // Implicit invariant: BITMAP_FULL implies (length > 0)
-  assert(Klass::SECONDARY_SUPERS_BITMAP_FULL == ~uintx(0), "");
-  cmpq(r_bitmap, (int32_t)-1); // sign-extends immediate to 64-bit value
-  jcc(Assembler::equal, L_huge);
+  cmpl(r_array_length, (int32_t)Klass::SECONDARY_SUPERS_TABLE_SIZE - 2);
+  jcc(Assembler::greater, L_huge);
 
   // NB! Our caller has checked bits 0 and 1 in the bitmap. The
   // current slot (at secondary_supers[r_array_index]) has not yet

src/hotspot/cpu/x86/nativeInst_x86.cpp

@@ -218,7 +218,7 @@ void NativeCall::insert(address code_pos, address entry) {
 // (spinlock). Then patches the last byte, and then atomically replaces
 // the jmp's with the first 4 byte of the new instruction.
 void NativeCall::replace_mt_safe(address instr_addr, address code_buffer) {
-  assert(Patching_lock->is_locked() ||
+  assert(CodeCache_lock->is_locked() ||
          SafepointSynchronize::is_at_safepoint(), "concurrent code patching");
   assert (instr_addr != nullptr, "illegal address for code patching");
 
@@ -281,7 +281,7 @@ void NativeCall::set_destination_mt_safe(address dest) {
   debug_only(verify());
   // Make sure patching code is locked.  No two threads can patch at the same
   // time but one may be executing this code.
-  assert(Patching_lock->is_locked() || SafepointSynchronize::is_at_safepoint() ||
+  assert(CodeCache_lock->is_locked() || SafepointSynchronize::is_at_safepoint() ||
          CompiledICLocker::is_safe(instruction_address()), "concurrent code patching");
   // Both C1 and C2 should now be generating code which aligns the patched address
   // to be within a single cache line.

src/hotspot/cpu/x86/smallRegisterMap_x86.inline.hpp

@@ -30,8 +30,16 @@
 
 // Java frames don't have callee saved registers (except for rbp), so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
+
 private:
   static void assert_is_rbp(VMReg r) NOT_DEBUG_RETURN
                                      DEBUG_ONLY({ assert(r == rbp->as_VMReg() || r == rbp->as_VMReg()->next(), "Reg: %s", r->name()); })
@@ -48,17 +56,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-  #ifdef ASSERT
-    for(int i = 0; i < RegisterMap::reg_count; i++) {
-      VMReg r = VMRegImpl::as_VMReg(i);
-      if (map->location(r, (intptr_t*)nullptr) != nullptr) assert_is_rbp(r);
-    }
-  #endif
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     assert_is_rbp(reg);
     return (address)(sp - frame::sender_sp_offset);

src/hotspot/cpu/x86/stubGenerator_x86_64.cpp

@@ -24,6 +24,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.hpp"
+#include "classfile/javaClasses.hpp"
 #include "classfile/vmIntrinsics.hpp"
 #include "compiler/oopMap.hpp"
 #include "gc/shared/barrierSet.hpp"
@@ -3988,6 +3989,28 @@ address StubGenerator::generate_upcall_stub_exception_handler() {
   return start;
 }
 
+// load Method* target of MethodHandle
+// j_rarg0 = jobject receiver
+// rbx = result
+address StubGenerator::generate_upcall_stub_load_target() {
+  StubCodeMark mark(this, "StubRoutines", "upcall_stub_load_target");
+  address start = __ pc();
+
+  __ resolve_global_jobject(j_rarg0, r15_thread, rscratch1);
+    // Load target method from receiver
+  __ load_heap_oop(rbx, Address(j_rarg0, java_lang_invoke_MethodHandle::form_offset()), rscratch1);
+  __ load_heap_oop(rbx, Address(rbx, java_lang_invoke_LambdaForm::vmentry_offset()), rscratch1);
+  __ load_heap_oop(rbx, Address(rbx, java_lang_invoke_MemberName::method_offset()), rscratch1);
+  __ access_load_at(T_ADDRESS, IN_HEAP, rbx,
+                    Address(rbx, java_lang_invoke_ResolvedMethodName::vmtarget_offset()),
+                    noreg, noreg);
+  __ movptr(Address(r15_thread, JavaThread::callee_target_offset()), rbx); // just in case callee is deoptimized
+
+  __ ret(0);
+
+  return start;
+}
+
 address StubGenerator::generate_lookup_secondary_supers_table_stub(u1 super_klass_index) {
   StubCodeMark mark(this, "StubRoutines", "lookup_secondary_supers_table");
 
@@ -4190,6 +4213,7 @@ void StubGenerator::generate_final_stubs() {
   }
 
   StubRoutines::_upcall_stub_exception_handler = generate_upcall_stub_exception_handler();
+  StubRoutines::_upcall_stub_load_target = generate_upcall_stub_load_target();
 }
 
 void StubGenerator::generate_compiler_stubs() {

src/hotspot/cpu/x86/stubGenerator_x86_64.hpp

@@ -615,6 +615,7 @@ class StubGenerator: public StubCodeGenerator {
 
   // shared exception handler for FFM upcall stubs
   address generate_upcall_stub_exception_handler();
+  address generate_upcall_stub_load_target();
 
   // Specialized stub implementations for UseSecondarySupersTable.
   address generate_lookup_secondary_supers_table_stub(u1 super_klass_index);

src/hotspot/cpu/x86/stubGenerator_x86_64_poly_mont.cpp

@@ -249,7 +249,6 @@ address StubGenerator::generate_intpoly_montgomeryMult_P256() {
   const Register tmp     = r9;
 
   montgomeryMultiply(aLimbs, bLimbs, rLimbs, tmp, _masm);
-  __ mov64(rax, 0x1); // Return 1 (Fig. 5, Step 6 [1] skipped in montgomeryMultiply)
 
   __ leave();
   __ ret(0);

src/hotspot/cpu/x86/upcallLinker_x86_32.cpp

@@ -24,7 +24,7 @@
 #include "precompiled.hpp"
 #include "prims/upcallLinker.hpp"
 
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,

src/hotspot/cpu/x86/upcallLinker_x86_64.cpp

@@ -23,7 +23,7 @@
 
 #include "precompiled.hpp"
 #include "asm/macroAssembler.hpp"
-#include "code/codeBlob.hpp"
+#include "classfile/javaClasses.hpp"
 #include "code/codeBlob.hpp"
 #include "code/vmreg.inline.hpp"
 #include "compiler/disassembler.hpp"
@@ -165,10 +165,10 @@ static void restore_callee_saved_registers(MacroAssembler* _masm, const ABIDescr
   __ block_comment("} restore_callee_saved_regs ");
 }
 
-static const int upcall_stub_code_base_size = 1024;
+static const int upcall_stub_code_base_size = 1200;
 static const int upcall_stub_size_per_arg = 16;
 
-address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject receiver, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,
@@ -277,7 +277,6 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   __ block_comment("{ on_entry");
   __ vzeroupper();
   __ lea(c_rarg0, Address(rsp, frame_data_offset));
-  __ movptr(c_rarg1, (intptr_t)receiver);
   // stack already aligned
   __ call(RuntimeAddress(CAST_FROM_FN_PTR(address, UpcallLinker::on_entry)));
   __ movptr(r15_thread, rax);
@@ -293,12 +292,10 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
   arg_shuffle.generate(_masm, shuffle_reg, abi._shadow_space_bytes, 0);
   __ block_comment("} argument shuffle");
 
-  __ block_comment("{ receiver ");
-  __ get_vm_result(j_rarg0, r15_thread);
-  __ block_comment("} receiver ");
-
-  __ mov_metadata(rbx, entry);
-  __ movptr(Address(r15_thread, JavaThread::callee_target_offset()), rbx); // just in case callee is deoptimized
+  __ block_comment("{ load target ");
+  __ movptr(j_rarg0, (intptr_t)receiver);
+  __ call(RuntimeAddress(StubRoutines::upcall_stub_load_target())); // puts target Method* in rbx
+  __ block_comment("} load target ");
 
   __ push_cont_fastpath();
 
@@ -373,7 +370,7 @@ address UpcallLinker::make_upcall_stub(jobject receiver, Method* entry,
 
 #ifndef PRODUCT
   stringStream ss;
-  ss.print("upcall_stub_%s", entry->signature()->as_C_string());
+  ss.print("upcall_stub_%s", signature->as_C_string());
   const char* name = _masm->code_string(ss.freeze());
 #else // PRODUCT
   const char* name = "upcall_stub";

src/hotspot/cpu/zero/smallRegisterMap_zero.inline.hpp

@@ -30,8 +30,15 @@
 
 // Java frames don't have callee saved registers (except for rfp), so we can use a smaller RegisterMap
 class SmallRegisterMap {
+  constexpr SmallRegisterMap() = default;
+  ~SmallRegisterMap() = default;
+  NONCOPYABLE(SmallRegisterMap);
+
 public:
-  static constexpr SmallRegisterMap* instance = nullptr;
+  static const SmallRegisterMap* instance() {
+    static constexpr SmallRegisterMap the_instance{};
+    return &the_instance;
+  }
 private:
   static void assert_is_rfp(VMReg r) NOT_DEBUG_RETURN
                                      DEBUG_ONLY({ Unimplemented(); })
@@ -46,12 +53,6 @@ public:
     return map;
   }
 
-  SmallRegisterMap() {}
-
-  SmallRegisterMap(const RegisterMap* map) {
-    Unimplemented();
-  }
-
   inline address location(VMReg reg, intptr_t* sp) const {
     Unimplemented();
     return nullptr;

src/hotspot/cpu/zero/upcallLinker_zero.cpp

@@ -24,7 +24,7 @@
 #include "precompiled.hpp"
 #include "prims/upcallLinker.hpp"
 
-address UpcallLinker::make_upcall_stub(jobject mh, Method* entry,
+address UpcallLinker::make_upcall_stub(jobject mh, Symbol* signature,
                                        BasicType* out_sig_bt, int total_out_args,
                                        BasicType ret_type,
                                        jobject jabi, jobject jconv,

src/hotspot/os/linux/os_linux.cpp

@@ -314,7 +314,10 @@ static jlong host_free_swap() {
 }
 
 jlong os::free_swap_space() {
-  jlong host_free_swap_val = host_free_swap();
+  // os::total_swap_space() might return the containerized limit which might be
+  // less than host_free_swap(). The upper bound of free swap needs to be the lower of the two.
+  jlong host_free_swap_val = MIN2(os::total_swap_space(), host_free_swap());
+  assert(host_free_swap_val >= 0, "sysinfo failed?");
   if (OSContainer::is_containerized()) {
     jlong mem_swap_limit = OSContainer::memory_and_swap_limit_in_bytes();
     jlong mem_limit = OSContainer::memory_limit_in_bytes();

src/hotspot/os/windows/os_windows.cpp

@@ -1934,7 +1934,10 @@ void os::win32::print_windows_version(outputStream* st) {
       // - 2016 GA 10/2016 build: 14393
       // - 2019 GA 11/2018 build: 17763
       // - 2022 GA 08/2021 build: 20348
-      if (build_number > 20347) {
+      // - 2025 Preview build   : 26040
+      if (build_number > 26039) {
+        st->print("Server 2025");
+      } else if (build_number > 20347) {
         st->print("Server 2022");
       } else if (build_number > 17762) {
         st->print("Server 2019");
@@ -3440,7 +3443,8 @@ static char* map_or_reserve_memory_aligned(size_t size, size_t alignment, int fi
                                      os::attempt_reserve_memory_at(aligned_base, size, false, flag);
   }
 
-  assert(aligned_base != nullptr, "Did not manage to re-map after %d attempts?", max_attempts);
+  assert(aligned_base != nullptr,
+      "Did not manage to re-map after %d attempts (size %zu, alignment %zu, file descriptor %d)", max_attempts, size, alignment, file_desc);
 
   return aligned_base;
 }
@@ -4067,6 +4071,39 @@ int    os::win32::_build_minor               = 0;
 bool   os::win32::_processor_group_warning_displayed = false;
 bool   os::win32::_job_object_processor_group_warning_displayed = false;
 
+void getWindowsInstallationType(char* buffer, int bufferSize) {
+  HKEY hKey;
+  const char* subKey = "SOFTWARE\\Microsoft\\Windows NT\\CurrentVersion";
+  const char* valueName = "InstallationType";
+
+  DWORD valueLength = bufferSize;
+
+  // Initialize buffer with empty string
+  buffer[0] = '\0';
+
+  // Open the registry key
+  if (RegOpenKeyExA(HKEY_LOCAL_MACHINE, subKey, 0, KEY_READ, &hKey) != ERROR_SUCCESS) {
+    // Return empty buffer if key cannot be opened
+    return;
+  }
+
+  // Query the value
+  if (RegQueryValueExA(hKey, valueName, NULL, NULL, (LPBYTE)buffer, &valueLength) != ERROR_SUCCESS) {
+    RegCloseKey(hKey);
+    buffer[0] = '\0';
+    return;
+  }
+
+  RegCloseKey(hKey);
+}
+
+bool isNanoServer() {
+  const int BUFFER_SIZE = 256;
+  char installationType[BUFFER_SIZE];
+  getWindowsInstallationType(installationType, BUFFER_SIZE);
+  return (strcmp(installationType, "Nano Server") == 0);
+}
+
 void os::win32::initialize_windows_version() {
   assert(_major_version == 0, "windows version already initialized.");
 
@@ -4084,7 +4121,13 @@ void os::win32::initialize_windows_version() {
     warning("Attempt to determine system directory failed: %s", buf_len != 0 ? error_msg_buffer : "<unknown error>");
     return;
   }
-  strncat(kernel32_path, "\\kernel32.dll", MAX_PATH - ret);
+
+  if (isNanoServer()) {
+    // On Windows Nanoserver the kernel32.dll is located in the forwarders subdirectory
+    strncat(kernel32_path, "\\forwarders\\kernel32.dll", MAX_PATH - ret);
+  } else {
+    strncat(kernel32_path, "\\kernel32.dll", MAX_PATH - ret);
+  }
 
   DWORD version_size = GetFileVersionInfoSize(kernel32_path, nullptr);
   if (version_size == 0) {

src/hotspot/share/asm/codeBuffer.cpp

@@ -135,10 +135,11 @@ CodeBuffer::~CodeBuffer() {
     // Previous incarnations of this buffer are held live, so that internal
     // addresses constructed before expansions will not be confused.
     cb->free_blob();
-    // free any overflow storage
-    delete cb->_overflow_arena;
   }
-
+  if (_overflow_arena != nullptr) {
+    // free any overflow storage
+    delete _overflow_arena;
+  }
   if (_shared_trampoline_requests != nullptr) {
     delete _shared_trampoline_requests;
   }
@@ -960,8 +961,6 @@ void CodeBuffer::take_over_code_from(CodeBuffer* cb) {
     CodeSection* this_sect = code_section(n);
     this_sect->take_over_code_from(cb_sect);
   }
-  _overflow_arena = cb->_overflow_arena;
-  cb->_overflow_arena = nullptr;
   // Make sure the old cb won't try to use it or free it.
   DEBUG_ONLY(cb->_blob = (BufferBlob*)badAddress);
 }

src/hotspot/share/c1/c1_GraphBuilder.cpp

@@ -1389,6 +1389,11 @@ void GraphBuilder::jsr(int dest) {
   // If the bytecodes are strange (jumping out of a jsr block) then we
   // might end up trying to re-parse a block containing a jsr which
   // has already been activated. Watch for this case and bail out.
+  if (next_bci() >= method()->code_size()) {
+    // This can happen if the subroutine does not terminate with a ret,
+    // effectively turning the jsr into a goto.
+    BAILOUT("too-complicated jsr/ret structure");
+  }
   for (ScopeData* cur_scope_data = scope_data();
        cur_scope_data != nullptr && cur_scope_data->parsing_jsr() && cur_scope_data->scope() == scope();
        cur_scope_data = cur_scope_data->parent()) {
@@ -1563,7 +1568,7 @@ void GraphBuilder::method_return(Value x, bool ignore_return) {
   // The conditions for a memory barrier are described in Parse::do_exits().
   bool need_mem_bar = false;
   if (method()->name() == ciSymbols::object_initializer_name() &&
-       (scope()->wrote_final() ||
+       (scope()->wrote_final() || scope()->wrote_stable() ||
          (AlwaysSafeConstructors && scope()->wrote_fields()) ||
          (support_IRIW_for_not_multiple_copy_atomic_cpu && scope()->wrote_volatile()))) {
     need_mem_bar = true;
@@ -1741,15 +1746,17 @@ void GraphBuilder::access_field(Bytecodes::Code code) {
     }
   }
 
-  if (field->is_final() && (code == Bytecodes::_putfield)) {
-    scope()->set_wrote_final();
-  }
-
   if (code == Bytecodes::_putfield) {
     scope()->set_wrote_fields();
     if (field->is_volatile()) {
       scope()->set_wrote_volatile();
     }
+    if (field->is_final()) {
+      scope()->set_wrote_final();
+    }
+    if (field->is_stable()) {
+      scope()->set_wrote_stable();
+    }
   }
 
   const int offset = !needs_patching ? field->offset_in_bytes() : -1;
@@ -3731,6 +3738,9 @@ bool GraphBuilder::try_inline_intrinsics(ciMethod* callee, bool ignore_return) {
 bool GraphBuilder::try_inline_jsr(int jsr_dest_bci) {
   // Introduce a new callee continuation point - all Ret instructions
   // will be replaced with Gotos to this point.
+  if (next_bci() >= method()->code_size()) {
+    return false;
+  }
   BlockBegin* cont = block_at(next_bci());
   assert(cont != nullptr, "continuation must exist (BlockListBuilder starts a new block after a jsr");
 

src/hotspot/share/c1/c1_IR.cpp

@@ -146,6 +146,7 @@ IRScope::IRScope(Compilation* compilation, IRScope* caller, int caller_bci, ciMe
   _wrote_final        = false;
   _wrote_fields       = false;
   _wrote_volatile     = false;
+  _wrote_stable       = false;
   _start              = nullptr;
 
   if (osr_bci != -1) {

src/hotspot/share/c1/c1_IR.hpp

@@ -149,6 +149,7 @@ class IRScope: public CompilationResourceObj {
   bool          _wrote_final;                    // has written final field
   bool          _wrote_fields;                   // has written fields
   bool          _wrote_volatile;                 // has written volatile field
+  bool          _wrote_stable;                   // has written @Stable field
   BlockBegin*   _start;                          // the start block, successsors are method entries
 
   ResourceBitMap _requires_phi_function;         // bit is set if phi functions at loop headers are necessary for a local variable
@@ -187,6 +188,8 @@ class IRScope: public CompilationResourceObj {
   bool          wrote_fields    () const         { return _wrote_fields; }
   void          set_wrote_volatile()             { _wrote_volatile = true; }
   bool          wrote_volatile    () const       { return _wrote_volatile; }
+  void          set_wrote_stable()               { _wrote_stable = true; }
+  bool          wrote_stable() const             { return _wrote_stable; }
 };
 
 

src/hotspot/share/c1/c1_RangeCheckElimination.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2012, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2012, 2024, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -483,14 +483,14 @@ void RangeCheckEliminator::in_block_motion(BlockBegin *block, AccessIndexedList
 
           if (c) {
             jint value = c->type()->as_IntConstant()->value();
-            if (value != min_jint) {
-              if (ao->op() == Bytecodes::_isub) {
-                value = -value;
-              }
+            if (ao->op() == Bytecodes::_iadd) {
               base = java_add(base, value);
-              last_integer = base;
-              last_instruction = other;
+            } else {
+              assert(ao->op() == Bytecodes::_isub, "unexpected bytecode");
+              base = java_subtract(base, value);
             }
+            last_integer = base;
+            last_instruction = other;
             index = other;
           } else {
             break;

src/hotspot/share/c1/c1_Runtime1.cpp

@@ -880,7 +880,7 @@ static Klass* resolve_field_return_klass(const methodHandle& caller, int bci, TR
 //             movl reg, [reg1 + <const>]  (for field offsets)
 //             jmp continue
 //             <being_init offset> <bytes to copy> <bytes to skip>
-// patch_stub: jmp Runtim1::patch_code (through a runtime stub)
+// patch_stub: jmp Runtime1::patch_code (through a runtime stub)
 //             jmp patch_site
 //
 // If the class is being initialized the patch body is rewritten and
@@ -1096,7 +1096,7 @@ JRT_ENTRY(void, Runtime1::patch_code(JavaThread* current, Runtime1::StubID stub_
   // Now copy code back
 
   {
-    MutexLocker ml_patch (current, Patching_lock, Mutex::_no_safepoint_check_flag);
+    MutexLocker ml_code (current, CodeCache_lock, Mutex::_no_safepoint_check_flag);
     //
     // Deoptimization may have happened while we waited for the lock.
     // In that case we don't bother to do any patching we just return
@@ -1261,12 +1261,8 @@ JRT_ENTRY(void, Runtime1::patch_code(JavaThread* current, Runtime1::StubID stub_
         }
       }
     }
-  }
-
-  // If we are patching in a non-perm oop, make sure the nmethod
-  // is on the right list.
-  {
-    MutexLocker ml_code (current, CodeCache_lock, Mutex::_no_safepoint_check_flag);
+    // If we are patching in a non-perm oop, make sure the nmethod
+    // is on the right list.
     nmethod* nm = CodeCache::find_nmethod(caller_frame.pc());
     guarantee(nm != nullptr, "only nmethods can contain non-perm oops");
 

src/hotspot/share/cds/archiveHeapLoader.cpp

@@ -34,6 +34,7 @@
 #include "memory/iterator.inline.hpp"
 #include "memory/resourceArea.hpp"
 #include "memory/universe.hpp"
+#include "sanitizers/ub.hpp"
 #include "utilities/bitMap.inline.hpp"
 #include "utilities/copy.hpp"
 
@@ -61,6 +62,7 @@ ptrdiff_t ArchiveHeapLoader::_mapped_heap_delta = 0;
 
 // Every mapped region is offset by _mapped_heap_delta from its requested address.
 // See FileMapInfo::heap_region_requested_address().
+ATTRIBUTE_NO_UBSAN
 void ArchiveHeapLoader::init_mapped_heap_info(address mapped_heap_bottom, ptrdiff_t delta, int dumptime_oop_shift) {
   assert(!_mapped_heap_relocation_initialized, "only once");
   if (!UseCompressedOops) {

src/hotspot/share/cds/cppVtables.cpp

@@ -66,19 +66,17 @@
 
 class CppVtableInfo {
   intptr_t _vtable_size;
-  intptr_t _cloned_vtable[1];
+  intptr_t _cloned_vtable[1]; // Pseudo flexible array member.
+  static size_t cloned_vtable_offset() { return offset_of(CppVtableInfo, _cloned_vtable); }
 public:
-  static int num_slots(int vtable_size) {
-    return 1 + vtable_size; // Need to add the space occupied by _vtable_size;
-  }
   int vtable_size()           { return int(uintx(_vtable_size)); }
   void set_vtable_size(int n) { _vtable_size = intptr_t(n); }
-  intptr_t* cloned_vtable()   { return &_cloned_vtable[0]; }
-  void zero()                 { memset(_cloned_vtable, 0, sizeof(intptr_t) * vtable_size()); }
+  // Using _cloned_vtable[i] for i > 0 causes undefined behavior. We use address calculation instead.
+  intptr_t* cloned_vtable()   { return (intptr_t*)((char*)this + cloned_vtable_offset()); }
+  void zero()                 { memset(cloned_vtable(), 0, sizeof(intptr_t) * vtable_size()); }
   // Returns the address of the next CppVtableInfo that can be placed immediately after this CppVtableInfo
   static size_t byte_size(int vtable_size) {
-    CppVtableInfo i;
-    return pointer_delta(&i._cloned_vtable[vtable_size], &i, sizeof(u1));
+    return cloned_vtable_offset() + (sizeof(intptr_t) * vtable_size);
   }
 };
 

src/hotspot/share/ci/ciEnv.cpp

@@ -1672,7 +1672,10 @@ void ciEnv::dump_replay_data_helper(outputStream* out) {
   for (int i = 0; i < objects->length(); i++) {
     objects->at(i)->dump_replay_data(out);
   }
-  dump_compile_data(out);
+
+  if (this->task() != nullptr) {
+    dump_compile_data(out);
+  }
   out->flush();
 }
 

src/hotspot/share/classfile/classLoaderDataGraph.cpp

@@ -241,9 +241,14 @@ LockedClassesDo::~LockedClassesDo() {
 
 
 // Iterating over the CLDG needs to be locked because
-// unloading can remove entries concurrently soon.
-template <bool keep_alive = true>
-class ClassLoaderDataGraphIteratorBase : public StackObj {
+// unloading can remove entries concurrently.
+// This iterator does not keep the CLD alive.
+// Any CLD OopHandles (modules, mirrors, resolved refs)
+// resolved must be treated as no keepalive. And requires
+// that its CLD's holder is kept alive if they escape the
+// caller's safepoint or ClassLoaderDataGraph_lock
+// critical section.
+class ClassLoaderDataGraph::ClassLoaderDataGraphIterator : public StackObj {
   ClassLoaderData* _next;
   Thread*          _thread;
   HandleMark       _hm;  // clean up handles when this is done.
@@ -251,12 +256,8 @@ class ClassLoaderDataGraphIteratorBase : public StackObj {
                             // unless verifying at a safepoint.
 
 public:
-  ClassLoaderDataGraphIteratorBase() : _next(ClassLoaderDataGraph::_head), _thread(Thread::current()), _hm(_thread) {
-    if (keep_alive) {
-      assert_locked_or_safepoint(ClassLoaderDataGraph_lock);
-    } else {
-      assert_at_safepoint();
-    }
+  ClassLoaderDataGraphIterator() : _next(ClassLoaderDataGraph::_head), _thread(Thread::current()), _hm(_thread) {
+    assert_locked_or_safepoint(ClassLoaderDataGraph_lock);
   }
 
   ClassLoaderData* get_next() {
@@ -266,10 +267,6 @@ public:
       cld = cld->next();
     }
     if (cld != nullptr) {
-      if (keep_alive) {
-        // Keep cld that is being returned alive.
-        Handle(_thread, cld->holder());
-      }
       _next = cld->next();
     } else {
       _next = nullptr;
@@ -278,9 +275,6 @@ public:
   }
 };
 
-using ClassLoaderDataGraphIterator = ClassLoaderDataGraphIteratorBase<true /* keep_alive */>;
-using ClassLoaderDataGraphIteratorNoKeepAlive = ClassLoaderDataGraphIteratorBase<false /* keep_alive */>;
-
 void ClassLoaderDataGraph::loaded_cld_do(CLDClosure* cl) {
   ClassLoaderDataGraphIterator iter;
   while (ClassLoaderData* cld = iter.get_next()) {
@@ -288,13 +282,6 @@ void ClassLoaderDataGraph::loaded_cld_do(CLDClosure* cl) {
   }
 }
 
-void ClassLoaderDataGraph::loaded_cld_do_no_keepalive(CLDClosure* cl) {
-  ClassLoaderDataGraphIteratorNoKeepAlive iter;
-  while (ClassLoaderData* cld = iter.get_next()) {
-    cl->do_cld(cld);
-  }
-}
-
 // These functions assume that the caller has locked the ClassLoaderDataGraph_lock
 // if they are not calling the function from a safepoint.
 void ClassLoaderDataGraph::classes_do(KlassClosure* klass_closure) {
@@ -318,6 +305,16 @@ void ClassLoaderDataGraph::methods_do(void f(Method*)) {
   }
 }
 
+void ClassLoaderDataGraph::modules_do_keepalive(void f(ModuleEntry*)) {
+  assert_locked_or_safepoint(Module_lock);
+  ClassLoaderDataGraphIterator iter;
+  while (ClassLoaderData* cld = iter.get_next()) {
+    // Keep the holder alive.
+    (void)cld->holder();
+    cld->modules_do(f);
+  }
+}
+
 void ClassLoaderDataGraph::modules_do(void f(ModuleEntry*)) {
   assert_locked_or_safepoint(Module_lock);
   ClassLoaderDataGraphIterator iter;
@@ -334,9 +331,11 @@ void ClassLoaderDataGraph::packages_do(void f(PackageEntry*)) {
   }
 }
 
-void ClassLoaderDataGraph::loaded_classes_do(KlassClosure* klass_closure) {
+void ClassLoaderDataGraph::loaded_classes_do_keepalive(KlassClosure* klass_closure) {
   ClassLoaderDataGraphIterator iter;
   while (ClassLoaderData* cld = iter.get_next()) {
+    // Keep the holder alive.
+    (void)cld->holder();
     cld->loaded_classes_do(klass_closure);
   }
 }
@@ -346,7 +345,7 @@ void ClassLoaderDataGraph::classes_unloading_do(void f(Klass* const)) {
 }
 
 void ClassLoaderDataGraph::verify_dictionary() {
-  ClassLoaderDataGraphIteratorNoKeepAlive iter;
+  ClassLoaderDataGraphIterator iter;
   while (ClassLoaderData* cld = iter.get_next()) {
     if (cld->dictionary() != nullptr) {
       cld->dictionary()->verify();
@@ -354,26 +353,28 @@ void ClassLoaderDataGraph::verify_dictionary() {
   }
 }
 
-#define FOR_ALL_DICTIONARY(X)   ClassLoaderDataGraphIterator iter; \
-                                while (ClassLoaderData* X = iter.get_next()) \
-                                  if (X->dictionary() != nullptr)
-
 void ClassLoaderDataGraph::print_dictionary(outputStream* st) {
-  FOR_ALL_DICTIONARY(cld) {
-    st->print("Dictionary for ");
-    cld->print_value_on(st);
-    st->cr();
-    cld->dictionary()->print_on(st);
-    st->cr();
+  ClassLoaderDataGraphIterator iter;
+  while (ClassLoaderData *cld = iter.get_next()) {
+    if (cld->dictionary() != nullptr) {
+      st->print("Dictionary for ");
+      cld->print_value_on(st);
+      st->cr();
+      cld->dictionary()->print_on(st);
+      st->cr();
+    }
   }
 }
 
 void ClassLoaderDataGraph::print_table_statistics(outputStream* st) {
-  FOR_ALL_DICTIONARY(cld) {
-    ResourceMark rm; // loader_name_and_id
-    stringStream tempst;
-    tempst.print("System Dictionary for %s class loader", cld->loader_name_and_id());
-    cld->dictionary()->print_table_statistics(st, tempst.freeze());
+  ClassLoaderDataGraphIterator iter;
+  while (ClassLoaderData *cld = iter.get_next()) {
+    if (cld->dictionary() != nullptr) {
+      ResourceMark rm; // loader_name_and_id
+      stringStream tempst;
+      tempst.print("System Dictionary for %s class loader", cld->loader_name_and_id());
+      cld->dictionary()->print_table_statistics(st, tempst.freeze());
+    }
   }
 }
 
@@ -550,7 +551,7 @@ Klass* ClassLoaderDataGraphKlassIteratorAtomic::next_klass() {
 }
 
 void ClassLoaderDataGraph::verify() {
-  ClassLoaderDataGraphIteratorNoKeepAlive iter;
+  ClassLoaderDataGraphIterator iter;
   while (ClassLoaderData* cld = iter.get_next()) {
     cld->verify();
   }

src/hotspot/share/classfile/classLoaderDataGraph.hpp

@@ -37,10 +37,10 @@ class ClassLoaderDataGraph : public AllStatic {
   friend class ClassLoaderDataGraphMetaspaceIterator;
   friend class ClassLoaderDataGraphKlassIteratorAtomic;
   friend class ClassLoaderDataGraphKlassIteratorStatic;
-  template <bool keep_alive>
-  friend class ClassLoaderDataGraphIteratorBase;
   friend class VMStructs;
  private:
+  class ClassLoaderDataGraphIterator;
+
   // All CLDs (except unlinked CLDs) can be reached by walking _head->_next->...
   static ClassLoaderData* volatile _head;
 
@@ -71,8 +71,12 @@ class ClassLoaderDataGraph : public AllStatic {
   static void roots_cld_do(CLDClosure* strong, CLDClosure* weak);
   static void always_strong_cld_do(CLDClosure* cl);
   // Iteration through CLDG not by GC.
+  // All the do suffixed functions do not keep the CLD alive. Any CLD OopHandles
+  // (modules, mirrors, resolved refs) resolved must be treated as no keepalive.
+  // And requires that its CLD's holder is kept alive if they escape the
+  // caller's safepoint or ClassLoaderDataGraph_lock critical section.
+  // The do_keepalive suffixed functions will keep all CLDs alive.
   static void loaded_cld_do(CLDClosure* cl);
-  static void loaded_cld_do_no_keepalive(CLDClosure* cl);
   // klass do
   // Walking classes through the ClassLoaderDataGraph include array classes.  It also includes
   // classes that are allocated but not loaded, classes that have errors, and scratch classes
@@ -81,9 +85,10 @@ class ClassLoaderDataGraph : public AllStatic {
   static void classes_do(KlassClosure* klass_closure);
   static void classes_do(void f(Klass* const));
   static void methods_do(void f(Method*));
+  static void modules_do_keepalive(void f(ModuleEntry*));
   static void modules_do(void f(ModuleEntry*));
   static void packages_do(void f(PackageEntry*));
-  static void loaded_classes_do(KlassClosure* klass_closure);
+  static void loaded_classes_do_keepalive(KlassClosure* klass_closure);
   static void classes_unloading_do(void f(Klass* const));
   static bool do_unloading();
 

src/hotspot/share/classfile/classLoaderStats.cpp

@@ -165,7 +165,7 @@ void ClassLoaderStatsClosure::addEmptyParents(oop cl) {
 
 void ClassLoaderStatsVMOperation::doit() {
   ClassLoaderStatsClosure clsc (_out);
-  ClassLoaderDataGraph::loaded_cld_do_no_keepalive(&clsc);
+  ClassLoaderDataGraph::loaded_cld_do(&clsc);
   clsc.print();
 }
 

src/hotspot/share/classfile/javaClasses.cpp

@@ -788,6 +788,7 @@ int java_lang_Class::_class_loader_offset;
 int java_lang_Class::_module_offset;
 int java_lang_Class::_protection_domain_offset;
 int java_lang_Class::_component_mirror_offset;
+int java_lang_Class::_init_lock_offset;
 int java_lang_Class::_signers_offset;
 int java_lang_Class::_name_offset;
 int java_lang_Class::_source_file_offset;
@@ -911,6 +912,12 @@ void java_lang_Class::initialize_mirror_fields(Klass* k,
                                                Handle protection_domain,
                                                Handle classData,
                                                TRAPS) {
+  // Allocate a simple java object for a lock.
+  // This needs to be a java object because during class initialization
+  // it can be held across a java call.
+  typeArrayOop r = oopFactory::new_typeArray(T_INT, 0, CHECK);
+  set_init_lock(mirror(), r);
+
   // Set protection domain also
   set_protection_domain(mirror(), protection_domain());
 
@@ -1132,6 +1139,10 @@ bool java_lang_Class::restore_archived_mirror(Klass *k,
   if (!k->is_array_klass()) {
     // - local static final fields with initial values were initialized at dump time
 
+    // create the init_lock
+    typeArrayOop r = oopFactory::new_typeArray(T_INT, 0, CHECK_(false));
+    set_init_lock(mirror(), r);
+
     if (protection_domain.not_null()) {
       set_protection_domain(mirror(), protection_domain());
     }
@@ -1196,6 +1207,15 @@ oop java_lang_Class::component_mirror(oop java_class) {
   return java_class->obj_field(_component_mirror_offset);
 }
 
+oop java_lang_Class::init_lock(oop java_class) {
+  assert(_init_lock_offset != 0, "must be set");
+  return java_class->obj_field(_init_lock_offset);
+}
+void java_lang_Class::set_init_lock(oop java_class, oop init_lock) {
+  assert(_init_lock_offset != 0, "must be set");
+  java_class->obj_field_put(_init_lock_offset, init_lock);
+}
+
 objArrayOop java_lang_Class::signers(oop java_class) {
   assert(_signers_offset != 0, "must be set");
   return (objArrayOop)java_class->obj_field(_signers_offset);
@@ -1414,7 +1434,6 @@ void java_lang_Class::compute_offsets() {
 
   InstanceKlass* k = vmClasses::Class_klass();
   CLASS_FIELDS_DO(FIELD_COMPUTE_OFFSET);
-
   CLASS_INJECTED_FIELDS(INJECTED_FIELD_COMPUTE_OFFSET);
 }
 

src/hotspot/share/classfile/javaClasses.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2024, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -210,6 +210,7 @@ class java_lang_String : AllStatic {
   macro(java_lang_Class, protection_domain,      object_signature,  false) \
   macro(java_lang_Class, signers,                object_signature,  false) \
   macro(java_lang_Class, source_file,            object_signature,  false) \
+  macro(java_lang_Class, init_lock,              object_signature,  false)
 
 class java_lang_Class : AllStatic {
   friend class VMStructs;
@@ -226,6 +227,7 @@ class java_lang_Class : AllStatic {
   static int _static_oop_field_count_offset;
 
   static int _protection_domain_offset;
+  static int _init_lock_offset;
   static int _signers_offset;
   static int _class_loader_offset;
   static int _module_offset;
@@ -240,6 +242,7 @@ class java_lang_Class : AllStatic {
   static GrowableArray<Klass*>* _fixup_mirror_list;
   static GrowableArray<Klass*>* _fixup_module_field_list;
 
+  static void set_init_lock(oop java_class, oop init_lock);
   static void set_protection_domain(oop java_class, oop protection_domain);
   static void set_class_loader(oop java_class, oop class_loader);
   static void set_component_mirror(oop java_class, oop comp_mirror);
@@ -292,6 +295,10 @@ class java_lang_Class : AllStatic {
 
   // Support for embedded per-class oops
   static oop  protection_domain(oop java_class);
+  static oop  init_lock(oop java_class);
+  static void clear_init_lock(oop java_class) {
+    set_init_lock(java_class, nullptr);
+  }
   static oop  component_mirror(oop java_class);
   static objArrayOop  signers(oop java_class);
   static void set_signers(oop java_class, objArrayOop signers);

src/hotspot/share/classfile/symbolTable.cpp

@@ -344,8 +344,23 @@ Symbol* SymbolTable::lookup_common(const char* name,
   return sym;
 }
 
+// Symbols should represent entities from the constant pool that are
+// limited to <64K in length, but usage errors creep in allowing Symbols
+// to be used for arbitrary strings. For debug builds we will assert if
+// a string is too long, whereas product builds will truncate it.
+static int check_length(const char* name, int len) {
+  assert(len <= Symbol::max_length(),
+         "String length %d exceeds the maximum Symbol length of %d", len, Symbol::max_length());
+  if (len > Symbol::max_length()) {
+    warning("A string \"%.80s ... %.80s\" exceeds the maximum Symbol "
+            "length of %d and has been truncated", name, (name + len - 80), Symbol::max_length());
+    len = Symbol::max_length();
+  }
+  return len;
+}
+
 Symbol* SymbolTable::new_symbol(const char* name, int len) {
-  assert(len <= Symbol::max_length(), "sanity");
+  len = check_length(name, len);
   unsigned int hash = hash_symbol(name, len, _alt_hash);
   Symbol* sym = lookup_common(name, len, hash);
   if (sym == nullptr) {
@@ -485,6 +500,7 @@ void SymbolTable::new_symbols(ClassLoaderData* loader_data, const constantPoolHa
   for (int i = 0; i < names_count; i++) {
     const char *name = names[i];
     int len = lengths[i];
+    assert(len <= Symbol::max_length(), "must be - these come from the constant pool");
     unsigned int hash = hashValues[i];
     assert(lookup_shared(name, len, hash) == nullptr, "must have checked already");
     Symbol* sym = do_add_if_needed(name, len, hash, is_permanent);
@@ -494,6 +510,7 @@ void SymbolTable::new_symbols(ClassLoaderData* loader_data, const constantPoolHa
 }
 
 Symbol* SymbolTable::do_add_if_needed(const char* name, int len, uintx hash, bool is_permanent) {
+  assert(len <= Symbol::max_length(), "caller should have ensured this");
   SymbolTableLookup lookup(name, len, hash);
   SymbolTableGet stg;
   bool clean_hint = false;
@@ -542,7 +559,7 @@ Symbol* SymbolTable::do_add_if_needed(const char* name, int len, uintx hash, boo
 
 Symbol* SymbolTable::new_permanent_symbol(const char* name) {
   unsigned int hash = 0;
-  int len = (int)strlen(name);
+  int len = check_length(name, (int)strlen(name));
   Symbol* sym = SymbolTable::lookup_only(name, len, hash);
   if (sym == nullptr) {
     sym = do_add_if_needed(name, len, hash, /* is_permanent */ true);

src/hotspot/share/classfile/systemDictionary.hpp

@@ -177,7 +177,7 @@ class SystemDictionary : AllStatic {
 
   static void classes_do(MetaspaceClosure* it);
   // Iterate over all methods in all klasses
-
+  // Will not keep metadata alive. See ClassLoaderDataGraph::methods_do.
   static void methods_do(void f(Method*));
 
   // Garbage collection support

src/hotspot/share/classfile/verifier.cpp

@@ -32,6 +32,7 @@
 #include "classfile/stackMapTableFormat.hpp"
 #include "classfile/symbolTable.hpp"
 #include "classfile/systemDictionary.hpp"
+#include "classfile/systemDictionaryShared.hpp"
 #include "classfile/verifier.hpp"
 #include "classfile/vmClasses.hpp"
 #include "classfile/vmSymbols.hpp"
@@ -212,6 +213,11 @@ bool Verifier::verify(InstanceKlass* klass, bool should_verify_class, TRAPS) {
          exception_name == vmSymbols::java_lang_ClassFormatError())) {
       log_info(verification)("Fail over class verification to old verifier for: %s", klass->external_name());
       log_info(class, init)("Fail over class verification to old verifier for: %s", klass->external_name());
+      // Exclude any classes that fail over during dynamic dumping
+      if (CDSConfig::is_dumping_dynamic_archive()) {
+        SystemDictionaryShared::warn_excluded(klass, "Failed over class verification while dynamic dumping");
+        SystemDictionaryShared::set_excluded(klass);
+      }
       message_buffer = NEW_RESOURCE_ARRAY(char, message_buffer_len);
       exception_message = message_buffer;
       exception_name = inference_verify(

src/hotspot/share/classfile/vmIntrinsics.hpp

@@ -529,8 +529,8 @@ class methodHandle;
   /* support for sun.security.util.math.intpoly.MontgomeryIntegerPolynomialP256 */                                      \
   do_class(sun_security_util_math_intpoly_MontgomeryIntegerPolynomialP256, "sun/security/util/math/intpoly/MontgomeryIntegerPolynomialP256")  \
   do_intrinsic(_intpoly_montgomeryMult_P256, sun_security_util_math_intpoly_MontgomeryIntegerPolynomialP256, intPolyMult_name, intPolyMult_signature, F_R) \
-  do_name(intPolyMult_name, "mult")                                                                                     \
-  do_signature(intPolyMult_signature, "([J[J[J)I")                                                                      \
+  do_name(intPolyMult_name, "multImpl")                                                                                     \
+  do_signature(intPolyMult_signature, "([J[J[J)V")                                                                      \
                                                                                                                         \
   do_class(sun_security_util_math_intpoly_IntegerPolynomial, "sun/security/util/math/intpoly/IntegerPolynomial")        \
   do_intrinsic(_intpoly_assign, sun_security_util_math_intpoly_IntegerPolynomial, intPolyAssign_name, intPolyAssign_signature, F_S) \

src/hotspot/share/classfile/vmSymbols.hpp

@@ -557,6 +557,7 @@ class SerializeClosure;
   template(bool_array_signature,                      "[Z")                                       \
   template(byte_array_signature,                      "[B")                                       \
   template(char_array_signature,                      "[C")                                       \
+  template(int_array_signature,                       "[I")                                       \
   template(runnable_signature,                        "Ljava/lang/Runnable;")                     \
   template(continuation_signature,                    "Ljdk/internal/vm/Continuation;")           \
   template(continuationscope_signature,               "Ljdk/internal/vm/ContinuationScope;")      \

src/hotspot/share/code/codeCache.cpp

@@ -227,6 +227,11 @@ void CodeCache::initialize_heaps() {
 
   if (!non_nmethod.set) {
     non_nmethod.size += compiler_buffer_size;
+    // Further down, just before FLAG_SET_ERGO(), all segment sizes are
+    // aligned down to the next lower multiple of min_size. For large page
+    // sizes, this may result in (non_nmethod.size == 0) which is not acceptable.
+    // Therefore, force non_nmethod.size to at least min_size.
+    non_nmethod.size = MAX2(non_nmethod.size, min_size);
   }
 
   if (!profiled.set && !non_profiled.set) {

src/hotspot/share/code/dependencies.cpp

@@ -897,7 +897,7 @@ void Dependencies::DepStream::print_dependency(outputStream* st, Klass* witness,
 void Dependencies::DepStream::initial_asserts(size_t byte_limit) {
   assert(must_be_in_vm(), "raw oops here");
   _byte_limit = byte_limit;
-  _type       = (DepType)(end_marker-1);  // defeat "already at end" assert
+  _type       = undefined_dependency;  // defeat "already at end" assert
   assert((_code!=nullptr) + (_deps!=nullptr) == 1, "one or t'other");
 }
 #endif //ASSERT

src/hotspot/share/code/dependencies.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2005, 2023, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2005, 2024, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -103,6 +103,9 @@ class Dependencies: public ResourceObj {
   // type now includes N, that is, all super types of N.
   //
   enum DepType {
+    // _type is initially set to -1, to prevent "already at end" assert
+    undefined_dependency = -1,
+
     end_marker = 0,
 
     // An 'evol' dependency simply notes that the contents of the