Mercurial > jdk8-mips64-public > hotspot / file comparison
comparison: src/cpu/sparc/vm/stubGenerator_sparc.cpp
src/cpu/sparc/vm/stubGenerator_sparc.cpp
- changeset 9806
- 758c07667682
- parent 9703
- 2fdf635bcf28
- parent 9788
- 44ef77ad417c
equal
deleted
inserted
replaced
| 9762:c97db0855565 | 9806:758c07667682 |
|---|---|
| 4786 __ delayed()->stf(FloatRegisterImpl::D, F14, state, 0x38); | 4786 __ delayed()->stf(FloatRegisterImpl::D, F14, state, 0x38); |
| 4787 | 4787 |
| 4788 return start; | 4788 return start; |
| 4789 } | 4789 } |
| 4790 | 4790 |
| 4791 /* Single and multi-block ghash operations */ | |
| 4792 address generate_ghash_processBlocks() { | |
| 4793 __ align(CodeEntryAlignment); | |
| 4794 Label L_ghash_loop, L_aligned, L_main; | |
| 4795 StubCodeMark mark(this, "StubRoutines", "ghash_processBlocks"); | |
| 4796 address start = __ pc(); | |
| 4797 | |
| 4798 Register state = I0; | |
| 4799 Register subkeyH = I1; | |
| 4800 Register data = I2; | |
| 4801 Register len = I3; | |
| 4802 | |
| 4803 __ save_frame(0); | |
| 4804 | |
| 4805 __ ldx(state, 0, O0); | |
| 4806 __ ldx(state, 8, O1); | |
| 4807 | |
| 4808 // Loop label for multiblock operations | |
| 4809 __ BIND(L_ghash_loop); | |
| 4810 | |
| 4811 // Check if 'data' is unaligned | |
| 4812 __ andcc(data, 7, G1); | |
| 4813 __ br(Assembler::zero, false, Assembler::pt, L_aligned); | |
| 4814 __ delayed()->nop(); | |
| 4815 | |
| 4816 Register left_shift = L1; | |
| 4817 Register right_shift = L2; | |
| 4818 Register data_ptr = L3; | |
| 4819 | |
| 4820 // Get left and right shift values in bits | |
| 4821 __ sll(G1, LogBitsPerByte, left_shift); | |
| 4822 __ mov(64, right_shift); | |
| 4823 __ sub(right_shift, left_shift, right_shift); | |
| 4824 | |
| 4825 // Align to read 'data' | |
| 4826 __ sub(data, G1, data_ptr); | |
| 4827 | |
| 4828 // Load first 8 bytes of 'data' | |
| 4829 __ ldx(data_ptr, 0, O4); | |
| 4830 __ sllx(O4, left_shift, O4); | |
| 4831 __ ldx(data_ptr, 8, O5); | |
| 4832 __ srlx(O5, right_shift, G4); | |
| 4833 __ bset(G4, O4); | |
| 4834 | |
| 4835 // Load second 8 bytes of 'data' | |
| 4836 __ sllx(O5, left_shift, O5); | |
| 4837 __ ldx(data_ptr, 16, G4); | |
| 4838 __ srlx(G4, right_shift, G4); | |
| 4839 __ ba(L_main); | |
| 4840 __ delayed()->bset(G4, O5); | |
| 4841 | |
| 4842 // If 'data' is aligned, load normally | |
| 4843 __ BIND(L_aligned); | |
| 4844 __ ldx(data, 0, O4); | |
| 4845 __ ldx(data, 8, O5); | |
| 4846 | |
| 4847 __ BIND(L_main); | |
| 4848 __ ldx(subkeyH, 0, O2); | |
| 4849 __ ldx(subkeyH, 8, O3); | |
| 4850 | |
| 4851 __ xor3(O0, O4, O0); | |
| 4852 __ xor3(O1, O5, O1); | |
| 4853 | |
| 4854 __ xmulxhi(O0, O3, G3); | |
| 4855 __ xmulx(O0, O2, O5); | |
| 4856 __ xmulxhi(O1, O2, G4); | |
| 4857 __ xmulxhi(O1, O3, G5); | |
| 4858 __ xmulx(O0, O3, G1); | |
| 4859 __ xmulx(O1, O3, G2); | |
| 4860 __ xmulx(O1, O2, O3); | |
| 4861 __ xmulxhi(O0, O2, O4); | |
| 4862 | |
| 4863 __ mov(0xE1, O0); | |
| 4864 __ sllx(O0, 56, O0); | |
| 4865 | |
| 4866 __ xor3(O5, G3, O5); | |
| 4867 __ xor3(O5, G4, O5); | |
| 4868 __ xor3(G5, G1, G1); | |
| 4869 __ xor3(G1, O3, G1); | |
| 4870 __ srlx(G2, 63, O1); | |
| 4871 __ srlx(G1, 63, G3); | |
| 4872 __ sllx(G2, 63, O3); | |
| 4873 __ sllx(G2, 58, O2); | |
| 4874 __ xor3(O3, O2, O2); | |
| 4875 | |
| 4876 __ sllx(G1, 1, G1); | |
| 4877 __ or3(G1, O1, G1); | |
| 4878 | |
| 4879 __ xor3(G1, O2, G1); | |
| 4880 | |
| 4881 __ sllx(G2, 1, G2); | |
| 4882 | |
| 4883 __ xmulxhi(G1, O0, O1); | |
| 4884 __ xmulx(G1, O0, O2); | |
| 4885 __ xmulxhi(G2, O0, O3); | |
| 4886 __ xmulx(G2, O0, G1); | |
| 4887 | |
| 4888 __ xor3(O4, O1, O4); | |
| 4889 __ xor3(O5, O2, O5); | |
| 4890 __ xor3(O5, O3, O5); | |
| 4891 | |
| 4892 __ sllx(O4, 1, O2); | |
| 4893 __ srlx(O5, 63, O3); | |
| 4894 | |
| 4895 __ or3(O2, O3, O0); | |
| 4896 | |
| 4897 __ sllx(O5, 1, O1); | |
| 4898 __ srlx(G1, 63, O2); | |
| 4899 __ or3(O1, O2, O1); | |
| 4900 __ xor3(O1, G3, O1); | |
| 4901 | |
| 4902 __ deccc(len); | |
| 4903 __ br(Assembler::notZero, true, Assembler::pt, L_ghash_loop); | |
| 4904 __ delayed()->add(data, 16, data); | |
| 4905 | |
| 4906 __ stx(O0, I0, 0); | |
| 4907 __ stx(O1, I0, 8); | |
| 4908 | |
| 4909 __ ret(); | |
| 4910 __ delayed()->restore(); | |
| 4911 | |
| 4912 return start; | |
| 4913 } | |
| 4914 | |
| 4791 void generate_initial() { | 4915 void generate_initial() { |
| 4792 // Generates all stubs and initializes the entry points | 4916 // Generates all stubs and initializes the entry points |
| 4793 | 4917 |
| 4794 //------------------------------------------------------------------------------------------------------------------------ | 4918 //------------------------------------------------------------------------------------------------------------------------ |
| 4795 // entry points that exist in all platforms | 4919 // entry points that exist in all platforms |
| 4857 if (UseAESIntrinsics) { | 4981 if (UseAESIntrinsics) { |
| 4858 StubRoutines::_aescrypt_encryptBlock = generate_aescrypt_encryptBlock(); | 4982 StubRoutines::_aescrypt_encryptBlock = generate_aescrypt_encryptBlock(); |
| 4859 StubRoutines::_aescrypt_decryptBlock = generate_aescrypt_decryptBlock(); | 4983 StubRoutines::_aescrypt_decryptBlock = generate_aescrypt_decryptBlock(); |
| 4860 StubRoutines::_cipherBlockChaining_encryptAESCrypt = generate_cipherBlockChaining_encryptAESCrypt(); | 4984 StubRoutines::_cipherBlockChaining_encryptAESCrypt = generate_cipherBlockChaining_encryptAESCrypt(); |
| 4861 StubRoutines::_cipherBlockChaining_decryptAESCrypt = generate_cipherBlockChaining_decryptAESCrypt_Parallel(); | 4985 StubRoutines::_cipherBlockChaining_decryptAESCrypt = generate_cipherBlockChaining_decryptAESCrypt_Parallel(); |
| 4986 } | |
| 4987 // generate GHASH intrinsics code | |
| 4988 if (UseGHASHIntrinsics) { | |
| 4989 StubRoutines::_ghash_processBlocks = generate_ghash_processBlocks(); | |
| 4862 } | 4990 } |
| 4863 | 4991 |
| 4864 // generate SHA1/SHA256/SHA512 intrinsics code | 4992 // generate SHA1/SHA256/SHA512 intrinsics code |
| 4865 if (UseSHA1Intrinsics) { | 4993 if (UseSHA1Intrinsics) { |
| 4866 StubRoutines::_sha1_implCompress = generate_sha1_implCompress(false, "sha1_implCompress"); | 4994 StubRoutines::_sha1_implCompress = generate_sha1_implCompress(false, "sha1_implCompress"); |
