Details
-
Bug
-
Resolution: Fixed
-
P2
-
16-pool, 17
-
None
-
b13
Description
Two problems:
1) The input bytebuffer during decryption was not moved to the limit, The code uses bytebuffer duplicate() to do operations that do not effect the original input buffer.
2) Buffered data in decryption isn't cleared after the first operation is done. This can cause the buffered data to be used if the op is not re-initialized. Buffers are created during update() operations. Note while the fix does include an encryption fix too, GCM must be reinitialized per the GCM spec for key & IV.
3) Test fix for a manual test which an internal initialization NPE
There appear to be no jck or regression tests that handle the top two situations.
1) The input bytebuffer during decryption was not moved to the limit, The code uses bytebuffer duplicate() to do operations that do not effect the original input buffer.
2) Buffered data in decryption isn't cleared after the first operation is done. This can cause the buffered data to be used if the op is not re-initialized. Buffers are created during update() operations. Note while the fix does include an encryption fix too, GCM must be reinitialized per the GCM spec for key & IV.
3) Test fix for a manual test which an internal initialization NPE
There appear to be no jck or regression tests that handle the top two situations.
Attachments
Issue Links
- relates to
-
JDK-8253821 Improve ByteBuffer performance with GCM
- Resolved