Only Windows platform is affected.
The problem can be easily reproduced on non-English Windows (e.g., russian Windows 7 when Cp1251 character page is in use) using following snippet:

try {
    Runtime.getRuntime().exec("ttttttttt");
} catch (IOException ex) {
    ex.printStackTrace();
}

As the result you can see following output with invalid characters:

java.io.IOException: Cannot run program "ttttttttt": CreateProcess error=2, ?? ??????? ????? ????????? ????
        at java.lang.ProcessBuilder.start(Unknown Source)
        at java.lang.Runtime.exec(Unknown Source)
        at java.lang.Runtime.exec(Unknown Source)
        at java.lang.Runtime.exec(Unknown Source)
        ...
Caused by: java.io.IOException: CreateProcess error=2, ?? ??????? ????? ????????? ????
        at java.lang.ProcessImpl.create(Native Method)
        at java.lang.ProcessImpl.<init>(Unknown Source)
        at java.lang.ProcessImpl.start(Unknown Source)
        ... 6 more


Note, that detailed message with correct encoding can be obtained using the following trick:

new String(ex.getMessage().getBytes("ISO-8859-1"),Charset.defaultCharset().displayName());

The bug is reside in native part of JDK at jdk\src\windows\native\java\lang\ProcessImpl_md.c
See "static void win32Error(JNIEnv *env, const char *functionName)" method there.
This method does not consider character page of underlying system.

As an example how to correctly process encoding of the error message you can look at
jdk\src\share\native\common\jni_util.c, method JNIEXPORT void JNICALL JNU_ThrowByNameWithLastError(JNIEnv *env, const char *name, const char *defaultDetail)