review request: 7184394: add intrinsics to use AES instructions

Roland Westrelin roland.westrelin at oracle.com
Thu Aug 9 11:00:58 PDT 2012


Hi Tom,

> http://cr.openjdk.java.net/~tdeneau/aes-intrinsics/webrev.04
> 
> has been posted which addresses the issues that Vladimir raised.

In callGenerator.cpp, this should go away:

966   // Vladimir said to remove this check
967   if (false && kit.stopped()) {
968     // Receiver's null check or other exception.
969     return kit.transfer_exceptions_into_jvms();
970   }

In library_call.cpp, typos:

5784 // the psuedo code we want to emulate with this predicate is:

5811   // we want to an instanceof comparison against the AESCrypt class

Also, it looks like this should go away as well:

5840   if (false) {
5841     //  src and dest are arrays.
5842     const Type* src_type = src->Value(&_gvn);
5843     const Type* dest_type = dest->Value(&_gvn);
5844     const TypeAryPtr* top_src = src_type->isa_aryptr();
5845     const TypeAryPtr* top_dest = dest_type->isa_aryptr();
5846     assert (top_src  != NULL && top_src->klass()  != NULL
5847             &&  top_dest != NULL && top_dest->klass() != NULL, "args are strange");
5848   }

Otherwise, I think it'ok to me.

Roland.


More information about the hotspot-compiler-dev mailing list