diff options
Diffstat (limited to 'src/spdk/intel-ipsec-mb/avx/mb_mgr_aes_xcbc_submit_avx.asm')
-rw-r--r-- | src/spdk/intel-ipsec-mb/avx/mb_mgr_aes_xcbc_submit_avx.asm | 272 |
1 files changed, 272 insertions, 0 deletions
diff --git a/src/spdk/intel-ipsec-mb/avx/mb_mgr_aes_xcbc_submit_avx.asm b/src/spdk/intel-ipsec-mb/avx/mb_mgr_aes_xcbc_submit_avx.asm new file mode 100644 index 000000000..38f6a6470 --- /dev/null +++ b/src/spdk/intel-ipsec-mb/avx/mb_mgr_aes_xcbc_submit_avx.asm @@ -0,0 +1,272 @@ +;; +;; Copyright (c) 2012-2018, Intel Corporation +;; +;; Redistribution and use in source and binary forms, with or without +;; modification, are permitted provided that the following conditions are met: +;; +;; * Redistributions of source code must retain the above copyright notice, +;; this list of conditions and the following disclaimer. +;; * Redistributions in binary form must reproduce the above copyright +;; notice, this list of conditions and the following disclaimer in the +;; documentation and/or other materials provided with the distribution. +;; * Neither the name of Intel Corporation nor the names of its contributors +;; may be used to endorse or promote products derived from this software +;; without specific prior written permission. +;; +;; THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +;; AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +;; IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +;; DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE +;; FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +;; DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +;; SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +;; CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +;; OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +;; OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +;; + +%include "include/os.asm" +%include "job_aes_hmac.asm" +%include "mb_mgr_datastruct.asm" + +%include "include/reg_sizes.asm" + +%include "include/memcpy.asm" +%include "include/const.inc" + +%ifndef AES_XCBC_X8 +%define AES_XCBC_X8 aes_xcbc_mac_128_x8 +%define SUBMIT_JOB_AES_XCBC submit_job_aes_xcbc_avx +%endif + +; void AES_XCBC_X8(AES_XCBC_ARGS_x8 *args, UINT64 len_in_bytes); +extern AES_XCBC_X8 + + +section .data +default rel + +align 16 +dupw: ;ddq 0x01000100010001000100010001000100 + dq 0x0100010001000100, 0x0100010001000100 +x80: ;ddq 0x00000000000000000000000000000080 + dq 0x0000000000000080, 0x0000000000000000 + +section .text + +%ifdef LINUX +%define arg1 rdi +%define arg2 rsi +%else +%define arg1 rcx +%define arg2 rdx +%endif + +%define state arg1 +%define job arg2 +%define len2 arg2 + +%define job_rax rax + +%if 1 +; idx needs to be in rbp +%define len r11 +%define idx rbp +%define tmp2 rbp +%define tmp r14 + +%define lane r8 +%define icv r9 +%define p2 r9 + +%define last_len r10 + +%define lane_data r12 +%define p r13 + +%define unused_lanes rbx +%endif + +; STACK_SPACE needs to be an odd multiple of 8 +; This routine and its callee clobbers all GPRs +struc STACK +_gpr_save: resq 8 +_rsp_save: resq 1 +endstruc + +; JOB* SUBMIT_JOB_AES_XCBC(MB_MGR_AES_XCBC_OOO *state, JOB_AES_HMAC *job) +; arg 1 : state +; arg 2 : job +MKGLOBAL(SUBMIT_JOB_AES_XCBC,function,internal) +SUBMIT_JOB_AES_XCBC: + + mov rax, rsp + sub rsp, STACK_size + and rsp, -16 + + mov [rsp + _gpr_save + 8*0], rbx + mov [rsp + _gpr_save + 8*1], rbp + mov [rsp + _gpr_save + 8*2], r12 + mov [rsp + _gpr_save + 8*3], r13 + mov [rsp + _gpr_save + 8*4], r14 + mov [rsp + _gpr_save + 8*5], r15 +%ifndef LINUX + mov [rsp + _gpr_save + 8*6], rsi + mov [rsp + _gpr_save + 8*7], rdi +%endif + mov [rsp + _rsp_save], rax ; original SP + + mov unused_lanes, [state + _aes_xcbc_unused_lanes] + mov lane, unused_lanes + and lane, 0xF + shr unused_lanes, 4 + imul lane_data, lane, _XCBC_LANE_DATA_size + lea lane_data, [state + _aes_xcbc_ldata + lane_data] + mov len, [job + _msg_len_to_hash_in_bytes] + mov [state + _aes_xcbc_unused_lanes], unused_lanes + mov [lane_data + _xcbc_job_in_lane], job + mov dword [lane_data + _xcbc_final_done], 0 + mov tmp, [job + _k1_expanded] + mov [state + _aes_xcbc_args_keys + lane*8], tmp + mov p, [job + _src] + add p, [job + _hash_start_src_offset_in_bytes] + + mov last_len, len + + cmp len, 16 + jle small_buffer + + mov [state + _aes_xcbc_args_in + lane*8], p + add p, len ; set point to end of data + + and last_len, 15 ; Check lsbs of msg len + jnz slow_copy ; if not 16B mult, do slow copy + +fast_copy: + vmovdqu xmm0, [p - 16] ; load last block M[n] + mov tmp, [job + _k2] ; load K2 address + vmovdqu xmm1, [tmp] ; load K2 + vpxor xmm0, xmm0, xmm1 ; M[n] XOR K2 + vmovdqa [lane_data + _xcbc_final_block], xmm0 + sub len, 16 ; take last block off length +end_fast_copy: + vpxor xmm0, xmm0, xmm0 + shl lane, 4 ; multiply by 16 + vmovdqa [state + _aes_xcbc_args_ICV + lane], xmm0 + + vmovdqa xmm0, [state + _aes_xcbc_lens] + XVPINSRW xmm0, xmm1, tmp, lane, len, no_scale + vmovdqa [state + _aes_xcbc_lens], xmm0 + + cmp unused_lanes, 0xf + jne return_null + +start_loop: + ; Find min length + vphminposuw xmm1, xmm0 + vpextrw DWORD(len2), xmm1, 0 ; min value + vpextrw DWORD(idx), xmm1, 1 ; min index (0...7) + cmp len2, 0 + je len_is_0 + + vpshufb xmm1, xmm1, [rel dupw] ; duplicate words across all lanes + vpsubw xmm0, xmm0, xmm1 + vmovdqa [state + _aes_xcbc_lens], xmm0 + + ; "state" and "args" are the same address, arg1 + ; len is arg2 + call AES_XCBC_X8 + ; state and idx are intact + +len_is_0: + ; process completed job "idx" + imul lane_data, idx, _XCBC_LANE_DATA_size + lea lane_data, [state + _aes_xcbc_ldata + lane_data] + cmp dword [lane_data + _xcbc_final_done], 0 + jne end_loop + + mov dword [lane_data + _xcbc_final_done], 1 + + vmovdqa xmm0, [state + _aes_xcbc_lens] + XVPINSRW xmm0, xmm1, tmp, idx, 16, scale_x16 + vmovdqa [state + _aes_xcbc_lens], xmm0 + + lea tmp, [lane_data + _xcbc_final_block] + mov [state + _aes_xcbc_args_in + 8*idx], tmp + jmp start_loop + +end_loop: + ; process completed job "idx" + mov job_rax, [lane_data + _xcbc_job_in_lane] + mov icv, [job_rax + _auth_tag_output] + mov unused_lanes, [state + _aes_xcbc_unused_lanes] + mov qword [lane_data + _xcbc_job_in_lane], 0 + or dword [job_rax + _status], STS_COMPLETED_HMAC + shl unused_lanes, 4 + or unused_lanes, idx + shl idx, 4 ; multiply by 16 + mov [state + _aes_xcbc_unused_lanes], unused_lanes + + ; copy 12 bytes + vmovdqa xmm0, [state + _aes_xcbc_args_ICV + idx] + vmovq [icv], xmm0 + vpextrd [icv + 8], xmm0, 2 + +%ifdef SAFE_DATA + ;; Clear ICV + vpxor xmm0, xmm0 + vmovdqa [state + _aes_xcbc_args_ICV + idx], xmm0 + + ;; Clear final block (32 bytes) + vmovdqa [lane_data + _xcbc_final_block], xmm0 + vmovdqa [lane_data + _xcbc_final_block + 16], xmm0 +%endif + +return: + + mov rbx, [rsp + _gpr_save + 8*0] + mov rbp, [rsp + _gpr_save + 8*1] + mov r12, [rsp + _gpr_save + 8*2] + mov r13, [rsp + _gpr_save + 8*3] + mov r14, [rsp + _gpr_save + 8*4] + mov r15, [rsp + _gpr_save + 8*5] +%ifndef LINUX + mov rsi, [rsp + _gpr_save + 8*6] + mov rdi, [rsp + _gpr_save + 8*7] +%endif + mov rsp, [rsp + _rsp_save] ; original SP + + ret + +small_buffer: + ; For buffers <= 16 Bytes + ; The input data is set to final block + lea tmp, [lane_data + _xcbc_final_block] ; final block + mov [state + _aes_xcbc_args_in + lane*8], tmp + add p, len ; set point to end of data + cmp len, 16 + je fast_copy + +slow_copy: + and len, ~15 ; take final block off len + sub p, last_len ; adjust data pointer + lea p2, [lane_data + _xcbc_final_block + 16] ; upper part of final + sub p2, last_len ; adjust data pointer backwards + memcpy_avx_16_1 p2, p, last_len, tmp, tmp2 + vmovdqa xmm0, [rel x80] ; fill reg with padding + vmovdqu [lane_data + _xcbc_final_block + 16], xmm0 ; add padding + vmovdqu xmm0, [p2] ; load final block to process + mov tmp, [job + _k3] ; load K3 address + vmovdqu xmm1, [tmp] ; load K3 + vpxor xmm0, xmm0, xmm1 ; M[n] XOR K3 + vmovdqu [lane_data + _xcbc_final_block], xmm0 ; write final block + jmp end_fast_copy + +return_null: + xor job_rax, job_rax + jmp return + + +%ifdef LINUX +section .note.GNU-stack noalloc noexec nowrite progbits +%endif |