<div dir="ltr"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"># Internal Error<br>(/opt/mach5/mesos/work_dir/slaves/3fe4557b-066a-4392-9b05-ef874140c35c-S62/frameworks/1735e8a2-a1db-478c-8104-60c8b0af87dd-0196/executors/cfba4e4e-51a1-47f1-9d82-c6f09de94dd2/runs/e5810681-4ef9-46d2-9b14-5ef945fda550/workspace/open/src/hotspot/share/code/SCCache.cpp:4205),<br>pid=1235, tid=1253<br># fatal error: Address 0x00007fe4817c5d30 for <unknown> is missing in<br>SCA table</blockquote><div><br></div><div>This error on Linux/x64 can be fixed by increasing MAX_STR_COUNT from 200 to 300.</div><div>Here is the patch for that:</div><div><br></div><div>diff --git a/src/hotspot/share/code/SCCache.cpp b/src/hotspot/share/code/SCCache.cpp<br>index f42fab991a5..5df7a84f5d6 100644<br>--- a/src/hotspot/share/code/SCCache.cpp<br>+++ b/src/hotspot/share/code/SCCache.cpp<br>@@ -3956,7 +3956,7 @@ SCAddressTable::~SCAddressTable() {<br> }<br> }<br> <br>-#define MAX_STR_COUNT 200<br>+#define MAX_STR_COUNT 300<br> static const char* _C_strings[MAX_STR_COUNT] = {nullptr};<br> static int _C_strings_count = 0;<br> static int _C_strings_s[MAX_STR_COUNT] = {0};<br></div><div><br></div><div> With this change test/hotspot/jtreg/runtime/cds/appcds/applications/QuarkusGettingStarted.java#leyden works fine.</div><div><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On aarch64 (Linux or MacOS), this failure seems to affect many more test<br>cases. Also, instead of <unknown>, we see<br><br># Internal Error (SCCache.cpp:4159), pid=87789, tid=28419<br># fatal error: Address 0x0000000114d4aa80 for<br>Stub:_large_arrays_hashcode_boolean is missing in SCA table</blockquote><div><br></div><div>I don't have an aarch64 machine, but the error message indicates some stub(s) in cpu/aarch64/stubRoutines_aarch64.hpp are missing in the SCA table.</div><div>I think adding them should resolve this issue. Patch for adding those stubs is:</div><div><br></div><div>diff --git a/src/hotspot/share/code/SCCache.cpp b/src/hotspot/share/code/SCCache.cpp<br>index f42fab991a5..89b5a8bf0e2 100644<br>--- a/src/hotspot/share/code/SCCache.cpp<br>+++ b/src/hotspot/share/code/SCCache.cpp<br>@@ -3813,6 +3813,11 @@ void SCAddressTable::init() {<br> SET_ADDRESS(_stubs, StubRoutines::aarch64::count_positives());<br> SET_ADDRESS(_stubs, StubRoutines::aarch64::count_positives_long());<br> SET_ADDRESS(_stubs, StubRoutines::aarch64::large_array_equals());<br>+ SET_ADDRESS(_stubs, StubRoutines::aarch64::large_arrays_hashcode(T_BOOLEAN));<br>+ SET_ADDRESS(_stubs, StubRoutines::aarch64::large_arrays_hashcode(T_BYTE));<br>+ SET_ADDRESS(_stubs, StubRoutines::aarch64::large_arrays_hashcode(T_CHAR));<br>+ SET_ADDRESS(_stubs, StubRoutines::aarch64::large_arrays_hashcode(T_INT));<br>+ SET_ADDRESS(_stubs, StubRoutines::aarch64::large_arrays_hashcode(T_SHORT));<br> SET_ADDRESS(_stubs, StubRoutines::aarch64::compare_long_string_LL());<br> SET_ADDRESS(_stubs, StubRoutines::aarch64::compare_long_string_UU());<br> SET_ADDRESS(_stubs, StubRoutines::aarch64::compare_long_string_LU());<br></div><div><br></div><div>@Ioi, can you please try these patches.<br></div><div><br></div><div>Thanks,</div><div><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr">- Ashutosh Mehra</div></div></div><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Nov 22, 2024 at 12:27 AM <<a href="mailto:ioi.lam@oracle.com">ioi.lam@oracle.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">As mentioned in today's Leyden meeting, I have merged JEP 483 into <br>
premain in my local repo. Since there are some new failures, I've put <br>
the merged version in a temporary branch:<br>
<br>
<a href="https://github.com/openjdk/leyden/tree/premain-temp-merge-after-jep-483" rel="noreferrer" target="_blank">https://github.com/openjdk/leyden/tree/premain-temp-merge-after-jep-483</a><br>
<br>
[1] I have disabled some of the new CDS optimizations. See cdsConfig.cpp<br>
<br>
// FIXME -- leyden+JEP483 merge {<br>
FLAG_SET_ERGO(ArchiveDynamicProxies, false);<br>
FLAG_SET_ERGO(ArchiveLoaderLookupCache, false);<br>
FLAG_SET_ERGO(ArchivePackages, false);<br>
FLAG_SET_ERGO(ArchiveProtectionDomains, false);<br>
FLAG_SET_ERGO(ArchiveReflectionData, false);<br>
<br>
The reason is that JEP 483 is more aggressive in picking classes to be <br>
AOT-initialized. For example, with the -XX:+ArchiveProtectionDomains <br>
option, after the 483 merge, the jdk.internal.loader.NativeLibraries <br>
class is AOT-initialized. As a result, the static field <br>
nativeLibraryLockMap is archived, so it remembers all the native libs <br>
that were loaded during the assembly phase. When we attempt to load <br>
these libs again during production, we get an "is being loaded in <br>
another classloader" error.<br>
<br>
Therefore, we need more refactoring of the core libraries (such as <br>
adding runtimeSetup() methods) before some of these optimizations can be <br>
enabled again.<br>
<br>
[2] There's a new failure in the AOT compiler. Could someone familiar <br>
with the compiler take a look?<br>
<br>
# Internal Error <br>
(/opt/mach5/mesos/work_dir/slaves/3fe4557b-066a-4392-9b05-ef874140c35c-S62/frameworks/1735e8a2-a1db-478c-8104-60c8b0af87dd-0196/executors/cfba4e4e-51a1-47f1-9d82-c6f09de94dd2/runs/e5810681-4ef9-46d2-9b14-5ef945fda550/workspace/open/src/hotspot/share/code/SCCache.cpp:4205), <br>
pid=1235, tid=1253<br>
# fatal error: Address 0x00007fe4817c5d30 for <unknown> is missing in <br>
SCA table<br>
<br>
V [libjvm.so+0x366795] SCAddressTable::id_for_address(unsigned char*, <br>
RelocIterator, CodeBuffer*) [clone .part.0]+0x455 (SCCache.cpp:4205)<br>
V [libjvm.so+0x36df0a] SCCache::write_relocations(CodeBuffer*, <br>
unsigned int&)+0x68a (SCCache.cpp:4147)<br>
V [libjvm.so+0x3702d2] SCCache::write_nmethod(methodHandle const&, <br>
int, int, CodeOffsets*, int, DebugInformationRecorder*, Dependencies*, <br>
CodeBuffer*, int, OopMapSet*, ExceptionHandlerTable*, <br>
ImplicitExceptionTable*, AbstractCompiler*, CompLevel, bool, bool, bool, <br>
bool, bool, bool)+0xb92 (SCCache.cpp:3335)<br>
V [libjvm.so+0x3707b2] SCCache::store_nmethod(methodHandle const&, <br>
int, int, CodeOffsets*, int, DebugInformationRecorder*, Dependencies*, <br>
CodeBuffer*, int, OopMapSet*, ExceptionHandlerTable*, <br>
ImplicitExceptionTable*, AbstractCompiler*, CompLevel, bool, bool, bool, <br>
bool, bool, bool)+0x182 (SCCache.cpp:3102)<br>
V [libjvm.so+0x96ac99] ciEnv::register_method(ciMethod*, int, <br>
CodeOffsets*, int, CodeBuffer*, int, OopMapSet*, ExceptionHandlerTable*, <br>
ImplicitExceptionTable*, AbstractCompiler*, bool, bool, bool, bool, <br>
bool, bool, int, bool, SCCEntry*)+0xcf9 (ciEnv.cpp:1099)<br>
V [libjvm.so+0x161c22e] PhaseOutput::install_code(ciMethod*, int, <br>
AbstractCompiler*, bool, bool)+0x15e (output.cpp:3445)<br>
V [libjvm.so+0xa9f46b] Compile::Code_Gen()+0x5cb (compile.cpp:3036)<br>
V [libjvm.so+0xaa216f] Compile::Compile(ciEnv*, ciMethod*, int, <br>
Options, DirectiveSet*)+0x1c7f (compile.cpp:886)<br>
V [libjvm.so+0x8dfde1] C2Compiler::compile_method(ciEnv*, ciMethod*, <br>
int, bool, DirectiveSet*)+0x261 (c2compiler.cpp:172)<br>
V [libjvm.so+0xaaff07] <br>
CompileBroker::invoke_compiler_on_method(CompileTask*)+0xcb7 <br>
(compileBroker.cpp:2626)<br>
V [libjvm.so+0xab2448] CompileBroker::compiler_thread_loop()+0x758 <br>
(compileBroker.cpp:2272)<br>
V [libjvm.so+0xfac4ce] JavaThread::thread_main_inner()+0xee <br>
(javaThread.cpp:774)<br>
<br>
On Linux/x64, this can be reliably reproduced with <br>
test/hotspot/jtreg/runtime/cds/appcds/applications/QuarkusGettingStarted.java#leyden<br>
<br>
On aarch64 (Linux or MacOS), this failure seems to affect many more test <br>
cases. Also, instead of <unknown>, we see<br>
<br>
# Internal Error (SCCache.cpp:4159), pid=87789, tid=28419<br>
# fatal error: Address 0x0000000114d4aa80 for <br>
Stub:_large_arrays_hashcode_boolean is missing in SCA table<br>
<br>
Thanks<br>
<br>
- Ioi<br>
<br>
</blockquote></div>