On Fri, 4 Jun 2021 23:49:31 GMT, Smita Kamath <svkam...@openjdk.org> wrote:

>> I would like to submit AES-GCM optimization for x86_64 architectures 
>> supporting AVX3+VAES (Evex encoded AES). This optimization interleaves AES 
>> and GHASH operations.
>> Performance gain of ~1.5x - 2x for message sizes 8k and above.
>
> Smita Kamath has updated the pull request incrementally with one additional 
> commit since the last revision:
> 
>   8267125:Updated intrinsic signature to remove copies of counter, state and 
> subkeyHtbl

Do you plan to implement `decrypt` intrinsic too?

src/hotspot/share/opto/library_call.cpp line 547:

> 545: 
> 546:   case vmIntrinsics::_galoisCounterMode_AESCrypt:
> 547:     return inline_galoisCounterMode_AESCrypt(intrinsic_id());

You don't need to pass `intrinsic_id()` for this implementation unless you plan 
to add decrypt intrinsic later.

src/hotspot/share/opto/library_call.cpp line 6545:

> 6543:          top_out != NULL && top_out->klass() != NULL, "args are 
> strange");
> 6544: 
> 6545:   // checks are the responsibility of the caller

Do you have all NULL for all objects and range checks in Java code for this 
intrinsic?

src/hotspot/share/opto/library_call.cpp line 6564:

> 6562:   Node* subkeyHtbl = load_field_from_object(ghash_object, "subkeyHtbl", 
> "[J");
> 6563:   Node* state = load_field_from_object(ghash_object, "state", "[J");
> 6564:   if (embeddedCipherObj == NULL || counter == NULL || subkeyHtbl == 
> NULL || state == NULL) return false;

Follow coding style for such long condition:

if () {
  return false;
}

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

PR: https://git.openjdk.java.net/jdk/pull/4019

Reply via email to