8004124: Handle and/or warn about SI_KERNEL

Summary: Detect this crash in the signal handler and give a fatal error message instead of making us chase down bugs that don't reproduce
Reviewed-by: kvn, mgerdin, dholmes

src/os_cpu/linux_x86/vm/os_linux_x86.cpp

@@ -289,6 +289,16 @@ JVM_handle_linux_signal(int sig,
     }
 #endif // AMD64
 
+#ifndef AMD64
+    // Halt if SI_KERNEL before more crashes get misdiagnosed as Java bugs
+    // This can happen in any running code (currently more frequently in
+    // interpreter code but has been seen in compiled code)
+    if (sig == SIGSEGV && info->si_addr == 0 && info->si_code == SI_KERNEL) {
+      fatal("An irrecoverable SI_KERNEL SIGSEGV has occurred due "
+            "to unstable signal handling in this distribution.");
+    }
+#endif // AMD64
+
     // Handle ALL stack overflow variations here
     if (sig == SIGSEGV) {
       address addr = (address) info->si_addr;

src/share/vm/runtime/sharedRuntime.cpp

@@ -813,8 +813,11 @@ address SharedRuntime::continuation_for_implicit_exception(JavaThread* thread,
           // 3. Implict null exception in nmethod
 
           if (!cb->is_nmethod()) {
-            guarantee(cb->is_adapter_blob() || cb->is_method_handles_adapter_blob(),
-                      "exception happened outside interpreter, nmethods and vtable stubs (1)");
+            bool is_in_blob = cb->is_adapter_blob() || cb->is_method_handles_adapter_blob();
+            if (!is_in_blob) {
+              cb->print();
+              fatal(err_msg("exception happened outside interpreter, nmethods and vtable stubs at pc " INTPTR_FORMAT, pc));
+            }
             Events::log_exception(thread, "NullPointerException in code blob at " INTPTR_FORMAT, pc);
             // There is no handler here, so we will simply unwind.
             return StubRoutines::throw_NullPointerException_at_call_entry();