[Bug 281382] java/openjdk21 segfaults with www/tomcat101 on aarch64

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=281382

--- Comment #5 from Harald Eilertsen <haraldei-freebsdbugz@anduin.net> ---
I'm able to reproduce, also on OpenJDK23, but I can't find any coredumps, for
some reason.

> I don't understand this crash, there is no code in JVSC which calls:
> Stack: [0x00000000411a4000,0x00000000811a4000],  sp=0x000000008119f9a0,  free space=1048558k
> Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
> J 994 c2 java.lang.StringBuilder.append(Ljava/lang/String;)Ljava/lang/StringBuilder; > java.base@21.0.4 (8 bytes) @ 0x000000009b20293c [0x000000009b202900+0x000000000000003c]
> j  sun.net.www.MimeEntry.setImageFileName(Ljava/lang/String;)V+71 java.base@21.0.4

That's right, this is called from the tomcat startup code, which is run by java
in the jsvc process. As seen from the bottom of the stack trace:

> j  org.apache.catalina.startup.Bootstrap.start()V+38
> j  java.lang.invoke.DirectMethodHandle$Holder.invokeSpecial(Ljava/lang/Object;Ljava/lang/Object;)V+10 java.base@21.0.4
> j  java.lang.invoke.LambdaForm$MH+0x0000000133468be0.invoke(Ljava/lang/Object;Ljava/lang/Object;)Ljava/lang/Object;+31 java.base@21.0.4
> J 1188 c1 jdk.internal.reflect.DirectMethodHandleAccessor.invoke(Ljava/lang/Object;[Ljava/lang/Object;)Ljava/lang/Object; java.base@21.0.4 (92 bytes) @ 0x0000000093901b54 [0x00000000939001c0+0x0000000000001994]
> J 1182 c1 java.lang.reflect.Method.invoke(Ljava/lang/Object;[Ljava/lang/Object;)Ljava/lang/Object; java.base@21.0.4 (108 bytes) @ 0x00000000938fe2e0 [0x00000000938fdd80+0x0000000000000560]
> j  org.apache.commons.daemon.support.DaemonLoader.start()Z+10
> v  ~StubRoutines::call_stub 0x000000009ac2113c
> V  [libjvm.so+0xa87b50]  AsyncGetCallTrace+0x115918
> V  [libjvm.so+0xafa038]  AsyncGetCallTrace+0x187e00
> V  [libjvm.so+0xaf9e20]  AsyncGetCallTrace+0x187be8
> C  [jsvc+0x19afc]
> C  [jsvc+0x16a4c]
> C  [jsvc+0x15a2c]

Not sure why this happens when run via jsvc, though. Tried to run it with a
debug version of the jdk, but that only made it crash earlier. Didn't really
give me any more information.

-- 
You are receiving this mail because:
You are the assignee for the bug.