/********************************************************************** Copyright(c) 2020 Arm Corporation All rights reserved. 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 Arm 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. **********************************************************************/ .arch armv8-a+crypto .text .align 2 .p2align 3,,7 /* Macros */ .macro declare_var_vector_reg name:req,reg:req \name\()_q .req q\reg \name\()_v .req v\reg \name\()_s .req s\reg .endm /* Variable list */ declare_var_vector_reg lane0_msg_0, 0 declare_var_vector_reg lane1_msg_0, 1 declare_var_vector_reg lane2_msg_0, 2 declare_var_vector_reg lane3_msg_0, 3 declare_var_vector_reg lane0_msg_1, 4 declare_var_vector_reg lane1_msg_1, 5 declare_var_vector_reg lane2_msg_1, 6 declare_var_vector_reg lane3_msg_1, 7 declare_var_vector_reg lane0_msg_2, 8 declare_var_vector_reg lane1_msg_2, 9 declare_var_vector_reg lane2_msg_2,10 declare_var_vector_reg lane3_msg_2,11 declare_var_vector_reg lane0_msg_3,12 declare_var_vector_reg lane1_msg_3,13 declare_var_vector_reg lane2_msg_3,14 declare_var_vector_reg lane3_msg_3,15 declare_var_vector_reg lane0_abcd ,16 declare_var_vector_reg lane1_abcd ,17 declare_var_vector_reg lane2_abcd ,18 declare_var_vector_reg lane3_abcd ,19 declare_var_vector_reg lane0_tmp0 ,20 declare_var_vector_reg lane1_tmp0 ,21 declare_var_vector_reg lane2_tmp0 ,22 declare_var_vector_reg lane3_tmp0 ,23 declare_var_vector_reg lane0_tmp1 ,24 declare_var_vector_reg lane1_tmp1 ,25 declare_var_vector_reg lane2_tmp1 ,26 declare_var_vector_reg lane3_tmp1 ,27 declare_var_vector_reg e0 ,28 declare_var_vector_reg e1 ,29 declare_var_vector_reg key ,30 declare_var_vector_reg tmp ,31 key_adr .req x4 msg_adr .req x5 block_cnt .req x6 offs .req x7 digest_adr .req x16 tmp0_adr .req x17 tmp1_adr .req x18 /** maros for round 4-67 */ .macro sha1_4_rounds inst:req,msg0:req,msg1:req,msg2:req,msg3:req,abcd:req,e0:req,tmp0:req,e1:req,tmp1:req sha1h lane0_\tmp0\()_s, lane0_\abcd\()_s sha1h lane1_\tmp0\()_s, lane1_\abcd\()_s sha1h lane2_\tmp0\()_s, lane2_\abcd\()_s sha1h lane3_\tmp0\()_s, lane3_\abcd\()_s mov \e0\()_v.S[0],lane0_\tmp0\()_v.S[0] mov \e0\()_v.S[1],lane1_\tmp0\()_v.S[0] mov \e0\()_v.S[2],lane2_\tmp0\()_v.S[0] mov \e0\()_v.S[3],lane3_\tmp0\()_v.S[0] mov lane0_\tmp0\()_v.S[0],\e1\()_v.S[0] mov lane1_\tmp0\()_v.S[0],\e1\()_v.S[1] mov lane2_\tmp0\()_v.S[0],\e1\()_v.S[2] mov lane3_\tmp0\()_v.S[0],\e1\()_v.S[3] \inst lane0_\abcd\()_q,lane0_\tmp0\()_s,lane0_\tmp1\()_v.4s \inst lane1_\abcd\()_q,lane1_\tmp0\()_s,lane1_\tmp1\()_v.4s \inst lane2_\abcd\()_q,lane2_\tmp0\()_s,lane2_\tmp1\()_v.4s \inst lane3_\abcd\()_q,lane3_\tmp0\()_s,lane3_\tmp1\()_v.4s ld1 {lane0_\tmp0\()_v.4s-lane3_\tmp0\()_v.4s},[\tmp0\()_adr] add lane0_\tmp1\()_v.4s,lane0_\msg3\()_v.4s,key_v.4s add lane1_\tmp1\()_v.4s,lane1_\msg3\()_v.4s,key_v.4s add lane2_\tmp1\()_v.4s,lane2_\msg3\()_v.4s,key_v.4s add lane3_\tmp1\()_v.4s,lane3_\msg3\()_v.4s,key_v.4s st1 {lane0_\tmp1\()_v.4s-lane3_\tmp1\()_v.4s},[\tmp1\()_adr] sha1su1 lane0_\msg0\()_v.4s,lane0_\msg3\()_v.4s sha1su1 lane1_\msg0\()_v.4s,lane1_\msg3\()_v.4s sha1su1 lane2_\msg0\()_v.4s,lane2_\msg3\()_v.4s sha1su1 lane3_\msg0\()_v.4s,lane3_\msg3\()_v.4s sha1su0 lane0_\msg1\()_v.4s,lane0_\msg2\()_v.4s,lane0_\msg3\()_v.4s sha1su0 lane1_\msg1\()_v.4s,lane1_\msg2\()_v.4s,lane1_\msg3\()_v.4s sha1su0 lane2_\msg1\()_v.4s,lane2_\msg2\()_v.4s,lane2_\msg3\()_v.4s sha1su0 lane3_\msg1\()_v.4s,lane3_\msg2\()_v.4s,lane3_\msg3\()_v.4s .endm /* void mh_sha1_block_ce(const uint8_t * input_data, uint32_t digests[SHA1_DIGEST_WORDS][HASH_SEGS], uint8_t frame_buffer[MH_SHA1_BLOCK_SIZE], uint32_t num_blocks) */ /* Arguements list */ input_data .req x0 digests .req x1 frame_buffer .req x2 num_blocks .req w3 .global mh_sha1_block_ce .type mh_sha1_block_ce, %function mh_sha1_block_ce: //save temp vector registers stp d8, d9, [sp, -128]! stp d10, d11, [sp, 16] stp d12, d13, [sp, 32] stp d14, d15, [sp, 48] mov tmp0_adr,frame_buffer add tmp1_adr,tmp0_adr,128 start_loop: mov block_cnt,0 mov msg_adr,input_data lane_loop: mov offs,64 adr key_adr,KEY_0 //load msg 0 ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[0],[msg_adr],offs ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[1],[msg_adr],offs ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[2],[msg_adr],offs ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[3],[msg_adr],offs ld4 {lane0_msg_1_v.S-lane3_msg_1_v.S}[0],[msg_adr],offs ld4 {lane0_msg_1_v.S-lane3_msg_1_v.S}[1],[msg_adr],offs ld4 {lane0_msg_1_v.S-lane3_msg_1_v.S}[2],[msg_adr],offs ld4 {lane0_msg_1_v.S-lane3_msg_1_v.S}[3],[msg_adr],offs ld4 {lane0_msg_2_v.S-lane3_msg_2_v.S}[0],[msg_adr],offs ld4 {lane0_msg_2_v.S-lane3_msg_2_v.S}[1],[msg_adr],offs ld4 {lane0_msg_2_v.S-lane3_msg_2_v.S}[2],[msg_adr],offs ld4 {lane0_msg_2_v.S-lane3_msg_2_v.S}[3],[msg_adr],offs ld4 {lane0_msg_3_v.S-lane3_msg_3_v.S}[0],[msg_adr],offs ld4 {lane0_msg_3_v.S-lane3_msg_3_v.S}[1],[msg_adr],offs ld4 {lane0_msg_3_v.S-lane3_msg_3_v.S}[2],[msg_adr],offs ld4 {lane0_msg_3_v.S-lane3_msg_3_v.S}[3],[msg_adr],offs add digest_adr,digests,block_cnt ld4 {lane0_abcd_v.S-lane3_abcd_v.S}[0],[digest_adr],offs ld4 {lane0_abcd_v.S-lane3_abcd_v.S}[1],[digest_adr],offs ld4 {lane0_abcd_v.S-lane3_abcd_v.S}[2],[digest_adr],offs ld4 {lane0_abcd_v.S-lane3_abcd_v.S}[3],[digest_adr],offs ldr e0_q,[digest_adr] //load key_0 ldr key_q,[key_adr] rev32 lane0_msg_0_v.16b,lane0_msg_0_v.16b rev32 lane1_msg_0_v.16b,lane1_msg_0_v.16b rev32 lane2_msg_0_v.16b,lane2_msg_0_v.16b rev32 lane3_msg_0_v.16b,lane3_msg_0_v.16b rev32 lane0_msg_1_v.16b,lane0_msg_1_v.16b rev32 lane1_msg_1_v.16b,lane1_msg_1_v.16b rev32 lane2_msg_1_v.16b,lane2_msg_1_v.16b rev32 lane3_msg_1_v.16b,lane3_msg_1_v.16b rev32 lane0_msg_2_v.16b,lane0_msg_2_v.16b rev32 lane1_msg_2_v.16b,lane1_msg_2_v.16b rev32 lane2_msg_2_v.16b,lane2_msg_2_v.16b rev32 lane3_msg_2_v.16b,lane3_msg_2_v.16b rev32 lane0_msg_3_v.16b,lane0_msg_3_v.16b rev32 lane1_msg_3_v.16b,lane1_msg_3_v.16b rev32 lane2_msg_3_v.16b,lane2_msg_3_v.16b rev32 lane3_msg_3_v.16b,lane3_msg_3_v.16b add lane0_tmp1_v.4s,lane0_msg_1_v.4s,key_v.4s add lane1_tmp1_v.4s,lane1_msg_1_v.4s,key_v.4s add lane2_tmp1_v.4s,lane2_msg_1_v.4s,key_v.4s add lane3_tmp1_v.4s,lane3_msg_1_v.4s,key_v.4s st1 {lane0_tmp1_v.4s-lane3_tmp1_v.4s},[tmp1_adr] add lane0_tmp0_v.4s,lane0_msg_0_v.4s,key_v.4s add lane1_tmp0_v.4s,lane1_msg_0_v.4s,key_v.4s add lane2_tmp0_v.4s,lane2_msg_0_v.4s,key_v.4s add lane3_tmp0_v.4s,lane3_msg_0_v.4s,key_v.4s /* rounds 0-3 */ sha1h lane0_tmp1_s,lane0_abcd_s sha1h lane1_tmp1_s,lane1_abcd_s sha1h lane2_tmp1_s,lane2_abcd_s sha1h lane3_tmp1_s,lane3_abcd_s mov e1_v.S[0],lane0_tmp1_v.S[0] mov e1_v.S[1],lane1_tmp1_v.S[0] mov e1_v.S[2],lane2_tmp1_v.S[0] mov e1_v.S[3],lane3_tmp1_v.S[0] mov lane0_tmp1_v.S[0],e0_v.S[0] mov lane1_tmp1_v.S[0],e0_v.S[1] mov lane2_tmp1_v.S[0],e0_v.S[2] mov lane3_tmp1_v.S[0],e0_v.S[3] sha1c lane0_abcd_q,lane0_tmp1_s,lane0_tmp0_v.4s sha1c lane1_abcd_q,lane1_tmp1_s,lane1_tmp0_v.4s sha1c lane2_abcd_q,lane2_tmp1_s,lane2_tmp0_v.4s sha1c lane3_abcd_q,lane3_tmp1_s,lane3_tmp0_v.4s ld1 {lane0_tmp1_v.4s-lane3_tmp1_v.4s},[tmp1_adr] add lane0_tmp0_v.4s,lane0_msg_2_v.4s,key_v.4s sha1su0 lane0_msg_0_v.4s,lane0_msg_1_v.4s,lane0_msg_2_v.4s add lane1_tmp0_v.4s,lane1_msg_2_v.4s,key_v.4s sha1su0 lane1_msg_0_v.4s,lane1_msg_1_v.4s,lane1_msg_2_v.4s add lane2_tmp0_v.4s,lane2_msg_2_v.4s,key_v.4s sha1su0 lane2_msg_0_v.4s,lane2_msg_1_v.4s,lane2_msg_2_v.4s add lane3_tmp0_v.4s,lane3_msg_2_v.4s,key_v.4s sha1su0 lane3_msg_0_v.4s,lane3_msg_1_v.4s,lane3_msg_2_v.4s st1 {lane0_tmp0_v.4s-lane3_tmp0_v.4s},[tmp0_adr] sha1_4_rounds sha1c,msg_0,msg_1,msg_2,msg_3,abcd,e0,tmp0,e1,tmp1 /* rounds 4-7 */ sha1_4_rounds sha1c,msg_1,msg_2,msg_3,msg_0,abcd,e1,tmp1,e0,tmp0 adr key_adr,KEY_1 ldr key_q,[key_adr] sha1_4_rounds sha1c,msg_2,msg_3,msg_0,msg_1,abcd,e0,tmp0,e1,tmp1 /* rounds 12-15 */ sha1_4_rounds sha1c,msg_3,msg_0,msg_1,msg_2,abcd,e1,tmp1,e0,tmp0 sha1_4_rounds sha1p,msg_0,msg_1,msg_2,msg_3,abcd,e0,tmp0,e1,tmp1 /* rounds 20-23 */ sha1_4_rounds sha1p,msg_1,msg_2,msg_3,msg_0,abcd,e1,tmp1,e0,tmp0 sha1_4_rounds sha1p,msg_2,msg_3,msg_0,msg_1,abcd,e0,tmp0,e1,tmp1 adr key_adr,KEY_2 ldr key_q,[key_adr] sha1_4_rounds sha1p,msg_3,msg_0,msg_1,msg_2,abcd,e1,tmp1,e0,tmp0 sha1_4_rounds sha1p,msg_0,msg_1,msg_2,msg_3,abcd,e0,tmp0,e1,tmp1 /* rounds 36-39 */ sha1_4_rounds sha1m,msg_1,msg_2,msg_3,msg_0,abcd,e1,tmp1,e0,tmp0 sha1_4_rounds sha1m,msg_2,msg_3,msg_0,msg_1,abcd,e0,tmp0,e1,tmp1 sha1_4_rounds sha1m,msg_3,msg_0,msg_1,msg_2,abcd,e1,tmp1,e0,tmp0 adr key_adr,KEY_3 ldr key_q,[key_adr] sha1_4_rounds sha1m,msg_0,msg_1,msg_2,msg_3,abcd,e0,tmp0,e1,tmp1 /* rounds 52-55 */ sha1_4_rounds sha1m,msg_1,msg_2,msg_3,msg_0,abcd,e1,tmp1,e0,tmp0 sha1_4_rounds sha1p,msg_2,msg_3,msg_0,msg_1,abcd,e0,tmp0,e1,tmp1 sha1_4_rounds sha1p,msg_3,msg_0,msg_1,msg_2,abcd,e1,tmp1,e0,tmp0 //msg2 and msg1 are free mov lane0_msg_2_v.S[0],e1_v.S[0] mov lane1_msg_2_v.S[0],e1_v.S[1] mov lane2_msg_2_v.S[0],e1_v.S[2] mov lane3_msg_2_v.S[0],e1_v.S[3] /* rounds 68-71 */ sha1h lane0_msg_1_s,lane0_abcd_s sha1h lane1_msg_1_s,lane1_abcd_s sha1h lane2_msg_1_s,lane2_abcd_s sha1h lane3_msg_1_s,lane3_abcd_s sha1p lane0_abcd_q,lane0_msg_2_s,lane0_tmp1_v.4s sha1p lane1_abcd_q,lane1_msg_2_s,lane1_tmp1_v.4s sha1p lane2_abcd_q,lane2_msg_2_s,lane2_tmp1_v.4s sha1p lane3_abcd_q,lane3_msg_2_s,lane3_tmp1_v.4s add lane0_tmp1_v.4s,lane0_msg_3_v.4s,key_v.4s add lane1_tmp1_v.4s,lane1_msg_3_v.4s,key_v.4s add lane2_tmp1_v.4s,lane2_msg_3_v.4s,key_v.4s add lane3_tmp1_v.4s,lane3_msg_3_v.4s,key_v.4s sha1su1 lane0_msg_0_v.4s,lane0_msg_3_v.4s sha1su1 lane1_msg_0_v.4s,lane1_msg_3_v.4s sha1su1 lane2_msg_0_v.4s,lane2_msg_3_v.4s sha1su1 lane3_msg_0_v.4s,lane3_msg_3_v.4s /* rounds 72-75 */ sha1h lane0_msg_2_s,lane0_abcd_s sha1h lane1_msg_2_s,lane1_abcd_s sha1h lane2_msg_2_s,lane2_abcd_s sha1h lane3_msg_2_s,lane3_abcd_s sha1p lane0_abcd_q,lane0_msg_1_s,lane0_tmp0_v.4s sha1p lane1_abcd_q,lane1_msg_1_s,lane1_tmp0_v.4s sha1p lane2_abcd_q,lane2_msg_1_s,lane2_tmp0_v.4s sha1p lane3_abcd_q,lane3_msg_1_s,lane3_tmp0_v.4s /* rounds 76-79 */ sha1h lane0_msg_1_s,lane0_abcd_s sha1h lane1_msg_1_s,lane1_abcd_s sha1h lane2_msg_1_s,lane2_abcd_s sha1h lane3_msg_1_s,lane3_abcd_s sha1p lane0_abcd_q,lane0_msg_2_s,lane0_tmp1_v.4s sha1p lane1_abcd_q,lane1_msg_2_s,lane1_tmp1_v.4s sha1p lane2_abcd_q,lane2_msg_2_s,lane2_tmp1_v.4s sha1p lane3_abcd_q,lane3_msg_2_s,lane3_tmp1_v.4s add digest_adr,digests,block_cnt ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[0],[digest_adr],offs ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[1],[digest_adr],offs ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[2],[digest_adr],offs ld4 {lane0_msg_0_v.S-lane3_msg_0_v.S}[3],[digest_adr],offs ld4 {lane0_msg_3_v.S-lane3_msg_3_v.S}[0],[digest_adr] add lane0_abcd_v.4S,lane0_abcd_v.4S,lane0_msg_0_v.4S add lane1_abcd_v.4S,lane1_abcd_v.4S,lane1_msg_0_v.4S add lane2_abcd_v.4S,lane2_abcd_v.4S,lane2_msg_0_v.4S add lane3_abcd_v.4S,lane3_abcd_v.4S,lane3_msg_0_v.4S add lane0_msg_1_v.4S,lane0_msg_1_v.4S,lane0_msg_3_v.4S add lane1_msg_1_v.4S,lane1_msg_1_v.4S,lane1_msg_3_v.4S add lane2_msg_1_v.4S,lane2_msg_1_v.4S,lane2_msg_3_v.4S add lane3_msg_1_v.4S,lane3_msg_1_v.4S,lane3_msg_3_v.4S add digest_adr,digests,block_cnt st4 {lane0_abcd_v.S-lane3_abcd_v.S}[0],[digest_adr],offs st4 {lane0_abcd_v.S-lane3_abcd_v.S}[1],[digest_adr],offs st4 {lane0_abcd_v.S-lane3_abcd_v.S}[2],[digest_adr],offs st4 {lane0_abcd_v.S-lane3_abcd_v.S}[3],[digest_adr],offs st4 {lane0_msg_1_v.S-lane3_msg_1_v.S}[0],[digest_adr] add block_cnt,block_cnt,16 cmp block_cnt,64 add msg_adr,input_data,block_cnt add digest_adr,digests,block_cnt bcc lane_loop subs num_blocks,num_blocks,1 add input_data,input_data,1024 bhi start_loop exit_func: //restore temp register ldp d10, d11, [sp, 16] ldp d12, d13, [sp, 32] ldp d14, d15, [sp, 48] ldp d8, d9, [sp], 128 ret .size mh_sha1_block_ce, .-mh_sha1_block_ce .section .rodata.cst16,"aM",@progbits,16 .align 4 KEY_0: .word 0x5a827999 .word 0x5a827999 .word 0x5a827999 .word 0x5a827999 KEY_1: .word 0x6ed9eba1 .word 0x6ed9eba1 .word 0x6ed9eba1 .word 0x6ed9eba1 KEY_2: .word 0x8f1bbcdc .word 0x8f1bbcdc .word 0x8f1bbcdc .word 0x8f1bbcdc KEY_3: .word 0xca62c1d6 .word 0xca62c1d6 .word 0xca62c1d6 .word 0xca62c1d6