RFR: 8266310: deadlock while loading the JNI code [v5]

[Aleksei Voitylov]

On Fri, 21 May 2021 15:49:09 GMT, Aleksei Voitylov <avoitylov at openjdk.org> wrote:

>> Please review this PR which fixes the deadlock in ClassLoader between the two lock objects - a lock object associated with the class being loaded, and the ClassLoader.loadedLibraryNames hash map, locked during the native library load operation.
>> 
>> Problem being fixed:
>> 
>> The initial reproducer demonstrated a deadlock between the JarFile/ZipFile and the hash map. That deadlock exists even when the ZipFile/JarFile lock is removed because there's another lock object in the class loader, associated with the name of the class being loaded. Such objects are stored in ClassLoader.parallelLockMap. The deadlock occurs when JNI_OnLoad() loads exactly the same class, whose signature is being verified in another thread.
>> 
>> Proposed fix:
>> 
>> The proposed patch suggests to get rid of locking loadedLibraryNames hash map and synchronize on each entry name, as it's done with class names in see ClassLoader.getClassLoadingLock(name) method.
>> 
>> The patch introduces nativeLibraryLockMap which holds the lock objects for each library name, and the getNativeLibraryLock() private method is used to lazily initialize the corresponding lock object. nativeLibraryContext was changed to ThreadLocal, so that in any concurrent thread it would have a NativeLibrary object on top of the stack, that's being currently loaded/unloaded in that thread. nativeLibraryLockMap accumulates the names of all native libraries loaded - in line with class loading code, it is not explicitly cleared.
>> 
>> Testing:  jtreg and jck testing with no regressions. A new regression test was developed.
>
> Aleksei Voitylov has updated the pull request incrementally with one additional commit since the last revision:
> 
>   fix whitespace

The PR was updated as follows:

- the fix resolves the deadlock problem when different libraries are loaded. 
- the lock on the library name is held just like in the first version of PR to avoid simultaneous JNI_OnLoad calls. 
- the test now checks that JNI_Onload and JNI_OnUnload are executed only once.

Within the current limitations it is not possible to resolve the deadlock when a library being loaded attempts, in JNI_OnLoad,  to load a class that depends on the same library. It is not the intent of this PR to do so.

-------------

PR: https://git.openjdk.java.net/jdk/pull/3976