wolfSSL/wolfssl
1
0
Fork 1
mirror of https://github.com/wolfSSL/wolfssl.git synced 2026-07-07 22:10:48 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #3510 from SparkiDev/sp_modinv_nct

Browse Source 
SP modinv: add non-constant time modinv
This commit is contained in:
toddouska
2020-12-10 16:06:12 -08:00
committed by GitHub
parent cb61dc7d2f 2862a9ce56
commit 2c652151ac
14 changed files with 8191 additions and 3075 deletions
Download Patch File Download Diff File Expand all files Collapse all files
wolfcrypt/src/aes_gcm_asm.S
+12 -12
View File
@@ -162,12 +162,12 @@ L_aes_gcm_mod2_128:
.text
.globl AES_GCM_encrypt
.type AES_GCM_encrypt,@function
.align 4
.align 16
AES_GCM_encrypt:
#else
.section __TEXT,__text
.globl _AES_GCM_encrypt
.p2align 2
.p2align 4
_AES_GCM_encrypt:
#endif /* __APPLE__ */
pushq %r13
@@ -2016,12 +2016,12 @@ L_AES_GCM_encrypt_store_tag_done:
.text
.globl AES_GCM_decrypt
.type AES_GCM_decrypt,@function
.align 4
.align 16
AES_GCM_decrypt:
#else
.section __TEXT,__text
.globl _AES_GCM_decrypt
.p2align 2
.p2align 4
_AES_GCM_decrypt:
#endif /* __APPLE__ */
pushq %r13
@@ -3555,12 +3555,12 @@ L_avx1_aes_gcm_mod2_128:
.text
.globl AES_GCM_encrypt_avx1
.type AES_GCM_encrypt_avx1,@function
.align 4
.align 16
AES_GCM_encrypt_avx1:
#else
.section __TEXT,__text
.globl _AES_GCM_encrypt_avx1
.p2align 2
.p2align 4
_AES_GCM_encrypt_avx1:
#endif /* __APPLE__ */
pushq %r13
@@ -5135,12 +5135,12 @@ L_AES_GCM_encrypt_avx1_store_tag_done:
.text
.globl AES_GCM_decrypt_avx1
.type AES_GCM_decrypt_avx1,@function
.align 4
.align 16
AES_GCM_decrypt_avx1:
#else
.section __TEXT,__text
.globl _AES_GCM_decrypt_avx1
.p2align 2
.p2align 4
_AES_GCM_decrypt_avx1:
#endif /* __APPLE__ */
pushq %r13
@@ -6455,12 +6455,12 @@ L_avx2_aes_gcm_mod2_128:
.text
.globl AES_GCM_encrypt_avx2
.type AES_GCM_encrypt_avx2,@function
.align 4
.align 16
AES_GCM_encrypt_avx2:
#else
.section __TEXT,__text
.globl _AES_GCM_encrypt_avx2
.p2align 2
.p2align 4
_AES_GCM_encrypt_avx2:
#endif /* __APPLE__ */
pushq %r13
@@ -7763,12 +7763,12 @@ L_AES_GCM_encrypt_avx2_store_tag_done:
.text
.globl AES_GCM_decrypt_avx2
.type AES_GCM_decrypt_avx2,@function
.align 4
.align 16
AES_GCM_decrypt_avx2:
#else
.section __TEXT,__text
.globl _AES_GCM_decrypt_avx2
.p2align 2
.p2align 4
_AES_GCM_decrypt_avx2:
#endif /* __APPLE__ */
pushq %r13
wolfcrypt/src/chacha_asm.S
+6 -6
View File
@@ -30,12 +30,12 @@
.text
.globl chacha_encrypt_x64
.type chacha_encrypt_x64,@function
.align 4
.align 16
chacha_encrypt_x64:
#else
.section __TEXT,__text
.globl _chacha_encrypt_x64
.p2align 2
.p2align 4
_chacha_encrypt_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -508,12 +508,12 @@ L_chacha20_avx1_four:
.text
.globl chacha_encrypt_avx1
.type chacha_encrypt_avx1,@function
.align 4
.align 16
chacha_encrypt_avx1:
#else
.section __TEXT,__text
.globl _chacha_encrypt_avx1
.p2align 2
.p2align 4
_chacha_encrypt_avx1:
#endif /* __APPLE__ */
subq $0x190, %rsp
@@ -1066,12 +1066,12 @@ L_chacha20_avx2_eight:
.text
.globl chacha_encrypt_avx2
.type chacha_encrypt_avx2,@function
.align 4
.align 16
chacha_encrypt_avx2:
#else
.section __TEXT,__text
.globl _chacha_encrypt_avx2
.p2align 2
.p2align 4
_chacha_encrypt_avx2:
#endif /* __APPLE__ */
subq $0x310, %rsp
wolfcrypt/src/fe_x25519_asm.S
+114 -114
View File
@@ -30,12 +30,12 @@
.text
.globl fe_init
.type fe_init,@function
.align 4
.align 16
fe_init:
#else
.section __TEXT,__text
.globl _fe_init
.p2align 2
.p2align 4
_fe_init:
#endif /* __APPLE__ */
#ifdef HAVE_INTEL_AVX2
@@ -233,12 +233,12 @@ L_fe_init_flags_done:
.text
.globl fe_frombytes
.type fe_frombytes,@function
.align 4
.align 16
fe_frombytes:
#else
.section __TEXT,__text
.globl _fe_frombytes
.p2align 2
.p2align 4
_fe_frombytes:
#endif /* __APPLE__ */
movq $0x7fffffffffffffff, %r9
@@ -259,12 +259,12 @@ _fe_frombytes:
.text
.globl fe_tobytes
.type fe_tobytes,@function
.align 4
.align 16
fe_tobytes:
#else
.section __TEXT,__text
.globl _fe_tobytes
.p2align 2
.p2align 4
_fe_tobytes:
#endif /* __APPLE__ */
movq $0x7fffffffffffffff, %r10
@@ -299,12 +299,12 @@ _fe_tobytes:
.text
.globl fe_1
.type fe_1,@function
.align 4
.align 16
fe_1:
#else
.section __TEXT,__text
.globl _fe_1
.p2align 2
.p2align 4
_fe_1:
#endif /* __APPLE__ */
# Set one
@@ -320,12 +320,12 @@ _fe_1:
.text
.globl fe_0
.type fe_0,@function
.align 4
.align 16
fe_0:
#else
.section __TEXT,__text
.globl _fe_0
.p2align 2
.p2align 4
_fe_0:
#endif /* __APPLE__ */
# Set zero
@@ -341,12 +341,12 @@ _fe_0:
.text
.globl fe_copy
.type fe_copy,@function
.align 4
.align 16
fe_copy:
#else
.section __TEXT,__text
.globl _fe_copy
.p2align 2
.p2align 4
_fe_copy:
#endif /* __APPLE__ */
# Copy
@@ -366,12 +366,12 @@ _fe_copy:
.text
.globl fe_sub
.type fe_sub,@function
.align 4
.align 16
fe_sub:
#else
.section __TEXT,__text
.globl _fe_sub
.p2align 2
.p2align 4
_fe_sub:
#endif /* __APPLE__ */
pushq %r12
@@ -409,12 +409,12 @@ _fe_sub:
.text
.globl fe_add
.type fe_add,@function
.align 4
.align 16
fe_add:
#else
.section __TEXT,__text
.globl _fe_add
.p2align 2
.p2align 4
_fe_add:
#endif /* __APPLE__ */
pushq %r12
@@ -452,12 +452,12 @@ _fe_add:
.text
.globl fe_neg
.type fe_neg,@function
.align 4
.align 16
fe_neg:
#else
.section __TEXT,__text
.globl _fe_neg
.p2align 2
.p2align 4
_fe_neg:
#endif /* __APPLE__ */
movq $-19, %rdx
@@ -480,12 +480,12 @@ _fe_neg:
.text
.globl fe_cmov
.type fe_cmov,@function
.align 4
.align 16
fe_cmov:
#else
.section __TEXT,__text
.globl _fe_cmov
.p2align 2
.p2align 4
_fe_cmov:
#endif /* __APPLE__ */
cmpl $0x01, %edx
@@ -509,12 +509,12 @@ _fe_cmov:
.text
.globl fe_isnonzero
.type fe_isnonzero,@function
.align 4
.align 16
fe_isnonzero:
#else
.section __TEXT,__text
.globl _fe_isnonzero
.p2align 2
.p2align 4
_fe_isnonzero:
#endif /* __APPLE__ */
movq $0x7fffffffffffffff, %r10
@@ -548,12 +548,12 @@ _fe_isnonzero:
.text
.globl fe_isnegative
.type fe_isnegative,@function
.align 4
.align 16
fe_isnegative:
#else
.section __TEXT,__text
.globl _fe_isnegative
.p2align 2
.p2align 4
_fe_isnegative:
#endif /* __APPLE__ */
movq $0x7fffffffffffffff, %r11
@@ -578,12 +578,12 @@ _fe_isnegative:
.text
.globl fe_cmov_table
.type fe_cmov_table,@function
.align 4
.align 16
fe_cmov_table:
#else
.section __TEXT,__text
.globl _fe_cmov_table
.p2align 2
.p2align 4
_fe_cmov_table:
#endif /* __APPLE__ */
pushq %r12
@@ -866,12 +866,12 @@ _fe_cmov_table:
.text
.globl fe_mul
.type fe_mul,@function
.align 4
.align 16
fe_mul:
#else
.section __TEXT,__text
.globl _fe_mul
.p2align 2
.p2align 4
_fe_mul:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -886,12 +886,12 @@ _fe_mul:
.text
.globl fe_sq
.type fe_sq,@function
.align 4
.align 16
fe_sq:
#else
.section __TEXT,__text
.globl _fe_sq
.p2align 2
.p2align 4
_fe_sq:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -906,12 +906,12 @@ _fe_sq:
.text
.globl fe_mul121666
.type fe_mul121666,@function
.align 4
.align 16
fe_mul121666:
#else
.section __TEXT,__text
.globl _fe_mul121666
.p2align 2
.p2align 4
_fe_mul121666:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -926,12 +926,12 @@ _fe_mul121666:
.text
.globl fe_sq2
.type fe_sq2,@function
.align 4
.align 16
fe_sq2:
#else
.section __TEXT,__text
.globl _fe_sq2
.p2align 2
.p2align 4
_fe_sq2:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -946,12 +946,12 @@ _fe_sq2:
.text
.globl fe_invert
.type fe_invert,@function
.align 4
.align 16
fe_invert:
#else
.section __TEXT,__text
.globl _fe_invert
.p2align 2
.p2align 4
_fe_invert:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -966,12 +966,12 @@ _fe_invert:
.text
.globl curve25519
.type curve25519,@function
.align 4
.align 16
curve25519:
#else
.section __TEXT,__text
.globl _curve25519
.p2align 2
.p2align 4
_curve25519:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -986,12 +986,12 @@ _curve25519:
.text
.globl fe_pow22523
.type fe_pow22523,@function
.align 4
.align 16
fe_pow22523:
#else
.section __TEXT,__text
.globl _fe_pow22523
.p2align 2
.p2align 4
_fe_pow22523:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1006,12 +1006,12 @@ _fe_pow22523:
.text
.globl fe_ge_to_p2
.type fe_ge_to_p2,@function
.align 4
.align 16
fe_ge_to_p2:
#else
.section __TEXT,__text
.globl _fe_ge_to_p2
.p2align 2
.p2align 4
_fe_ge_to_p2:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1026,12 +1026,12 @@ _fe_ge_to_p2:
.text
.globl fe_ge_to_p3
.type fe_ge_to_p3,@function
.align 4
.align 16
fe_ge_to_p3:
#else
.section __TEXT,__text
.globl _fe_ge_to_p3
.p2align 2
.p2align 4
_fe_ge_to_p3:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1046,12 +1046,12 @@ _fe_ge_to_p3:
.text
.globl fe_ge_dbl
.type fe_ge_dbl,@function
.align 4
.align 16
fe_ge_dbl:
#else
.section __TEXT,__text
.globl _fe_ge_dbl
.p2align 2
.p2align 4
_fe_ge_dbl:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1066,12 +1066,12 @@ _fe_ge_dbl:
.text
.globl fe_ge_madd
.type fe_ge_madd,@function
.align 4
.align 16
fe_ge_madd:
#else
.section __TEXT,__text
.globl _fe_ge_madd
.p2align 2
.p2align 4
_fe_ge_madd:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1086,12 +1086,12 @@ _fe_ge_madd:
.text
.globl fe_ge_msub
.type fe_ge_msub,@function
.align 4
.align 16
fe_ge_msub:
#else
.section __TEXT,__text
.globl _fe_ge_msub
.p2align 2
.p2align 4
_fe_ge_msub:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1106,12 +1106,12 @@ _fe_ge_msub:
.text
.globl fe_ge_add
.type fe_ge_add,@function
.align 4
.align 16
fe_ge_add:
#else
.section __TEXT,__text
.globl _fe_ge_add
.p2align 2
.p2align 4
_fe_ge_add:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1126,12 +1126,12 @@ _fe_ge_add:
.text
.globl fe_ge_sub
.type fe_ge_sub,@function
.align 4
.align 16
fe_ge_sub:
#else
.section __TEXT,__text
.globl _fe_ge_sub
.p2align 2
.p2align 4
_fe_ge_sub:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -1338,12 +1338,12 @@ _fe_ge_sub_p:
.text
.globl fe_mul_x64
.type fe_mul_x64,@function
.align 4
.align 16
fe_mul_x64:
#else
.section __TEXT,__text
.globl _fe_mul_x64
.p2align 2
.p2align 4
_fe_mul_x64:
#endif /* __APPLE__ */
pushq %r12
@@ -1518,12 +1518,12 @@ _fe_mul_x64:
.text
.globl fe_sq_x64
.type fe_sq_x64,@function
.align 4
.align 16
fe_sq_x64:
#else
.section __TEXT,__text
.globl _fe_sq_x64
.p2align 2
.p2align 4
_fe_sq_x64:
#endif /* __APPLE__ */
pushq %r12
@@ -1668,12 +1668,12 @@ _fe_sq_x64:
.text
.globl fe_sq_n_x64
.type fe_sq_n_x64,@function
.align 4
.align 16
fe_sq_n_x64:
#else
.section __TEXT,__text
.globl _fe_sq_n_x64
.p2align 2
.p2align 4
_fe_sq_n_x64:
#endif /* __APPLE__ */
pushq %r12
@@ -1824,12 +1824,12 @@ L_fe_sq_n_x64:
.text
.globl fe_mul121666_x64
.type fe_mul121666_x64,@function
.align 4
.align 16
fe_mul121666_x64:
#else
.section __TEXT,__text
.globl _fe_mul121666_x64
.p2align 2
.p2align 4
_fe_mul121666_x64:
#endif /* __APPLE__ */
pushq %r12
@@ -1875,12 +1875,12 @@ _fe_mul121666_x64:
.text
.globl fe_sq2_x64
.type fe_sq2_x64,@function
.align 4
.align 16
fe_sq2_x64:
#else
.section __TEXT,__text
.globl _fe_sq2_x64
.p2align 2
.p2align 4
_fe_sq2_x64:
#endif /* __APPLE__ */
pushq %r12
@@ -2038,12 +2038,12 @@ _fe_sq2_x64:
.text
.globl fe_invert_x64
.type fe_invert_x64,@function
.align 4
.align 16
fe_invert_x64:
#else
.section __TEXT,__text
.globl _fe_invert_x64
.p2align 2
.p2align 4
_fe_invert_x64:
#endif /* __APPLE__ */
subq $0x90, %rsp
@@ -2294,12 +2294,12 @@ _fe_invert_x64:
.text
.globl curve25519_x64
.type curve25519_x64,@function
.align 4
.align 16
curve25519_x64:
#else
.section __TEXT,__text
.globl _curve25519_x64
.p2align 2
.p2align 4
_curve25519_x64:
#endif /* __APPLE__ */
pushq %r12
@@ -4304,12 +4304,12 @@ L_curve25519_x64_bits:
.text
.globl fe_pow22523_x64
.type fe_pow22523_x64,@function
.align 4
.align 16
fe_pow22523_x64:
#else
.section __TEXT,__text
.globl _fe_pow22523_x64
.p2align 2
.p2align 4
_fe_pow22523_x64:
#endif /* __APPLE__ */
subq $0x70, %rsp
@@ -4559,12 +4559,12 @@ _fe_pow22523_x64:
.text
.globl fe_ge_to_p2_x64
.type fe_ge_to_p2_x64,@function
.align 4
.align 16
fe_ge_to_p2_x64:
#else
.section __TEXT,__text
.globl _fe_ge_to_p2_x64
.p2align 2
.p2align 4
_fe_ge_to_p2_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -5059,12 +5059,12 @@ _fe_ge_to_p2_x64:
.text
.globl fe_ge_to_p3_x64
.type fe_ge_to_p3_x64,@function
.align 4
.align 16
fe_ge_to_p3_x64:
#else
.section __TEXT,__text
.globl _fe_ge_to_p3_x64
.p2align 2
.p2align 4
_fe_ge_to_p3_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -5715,12 +5715,12 @@ _fe_ge_to_p3_x64:
.text
.globl fe_ge_dbl_x64
.type fe_ge_dbl_x64,@function
.align 4
.align 16
fe_ge_dbl_x64:
#else
.section __TEXT,__text
.globl _fe_ge_dbl_x64
.p2align 2
.p2align 4
_fe_ge_dbl_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -6412,12 +6412,12 @@ _fe_ge_dbl_x64:
.text
.globl fe_ge_madd_x64
.type fe_ge_madd_x64,@function
.align 4
.align 16
fe_ge_madd_x64:
#else
.section __TEXT,__text
.globl _fe_ge_madd_x64
.p2align 2
.p2align 4
_fe_ge_madd_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -7110,12 +7110,12 @@ _fe_ge_madd_x64:
.text
.globl fe_ge_msub_x64
.type fe_ge_msub_x64,@function
.align 4
.align 16
fe_ge_msub_x64:
#else
.section __TEXT,__text
.globl _fe_ge_msub_x64
.p2align 2
.p2align 4
_fe_ge_msub_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -7808,12 +7808,12 @@ _fe_ge_msub_x64:
.text
.globl fe_ge_add_x64
.type fe_ge_add_x64,@function
.align 4
.align 16
fe_ge_add_x64:
#else
.section __TEXT,__text
.globl _fe_ge_add_x64
.p2align 2
.p2align 4
_fe_ge_add_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -8662,12 +8662,12 @@ _fe_ge_add_x64:
.text
.globl fe_ge_sub_x64
.type fe_ge_sub_x64,@function
.align 4
.align 16
fe_ge_sub_x64:
#else
.section __TEXT,__text
.globl _fe_ge_sub_x64
.p2align 2
.p2align 4
_fe_ge_sub_x64:
#endif /* __APPLE__ */
pushq %rbx
@@ -9517,12 +9517,12 @@ _fe_ge_sub_x64:
.text
.globl fe_mul_avx2
.type fe_mul_avx2,@function
.align 4
.align 16
fe_mul_avx2:
#else
.section __TEXT,__text
.globl _fe_mul_avx2
.p2align 2
.p2align 4
_fe_mul_avx2:
#endif /* __APPLE__ */
pushq %r12
@@ -9669,12 +9669,12 @@ _fe_mul_avx2:
.text
.globl fe_sq_avx2
.type fe_sq_avx2,@function
.align 4
.align 16
fe_sq_avx2:
#else
.section __TEXT,__text
.globl _fe_sq_avx2
.p2align 2
.p2align 4
_fe_sq_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -9796,12 +9796,12 @@ _fe_sq_avx2:
.text
.globl fe_sq_n_avx2
.type fe_sq_n_avx2,@function
.align 4
.align 16
fe_sq_n_avx2:
#else
.section __TEXT,__text
.globl _fe_sq_n_avx2
.p2align 2
.p2align 4
_fe_sq_n_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -9929,12 +9929,12 @@ L_fe_sq_n_avx2:
.text
.globl fe_mul121666_avx2
.type fe_mul121666_avx2,@function
.align 4
.align 16
fe_mul121666_avx2:
#else
.section __TEXT,__text
.globl _fe_mul121666_avx2
.p2align 2
.p2align 4
_fe_mul121666_avx2:
#endif /* __APPLE__ */
pushq %r12
@@ -9970,12 +9970,12 @@ _fe_mul121666_avx2:
.text
.globl fe_sq2_avx2
.type fe_sq2_avx2,@function
.align 4
.align 16
fe_sq2_avx2:
#else
.section __TEXT,__text
.globl _fe_sq2_avx2
.p2align 2
.p2align 4
_fe_sq2_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -10108,12 +10108,12 @@ _fe_sq2_avx2:
.text
.globl fe_invert_avx2
.type fe_invert_avx2,@function
.align 4
.align 16
fe_invert_avx2:
#else
.section __TEXT,__text
.globl _fe_invert_avx2
.p2align 2
.p2align 4
_fe_invert_avx2:
#endif /* __APPLE__ */
subq $0x90, %rsp
@@ -10364,12 +10364,12 @@ _fe_invert_avx2:
.text
.globl curve25519_avx2
.type curve25519_avx2,@function
.align 4
.align 16
curve25519_avx2:
#else
.section __TEXT,__text
.globl _curve25519_avx2
.p2align 2
.p2align 4
_curve25519_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -12095,12 +12095,12 @@ L_curve25519_avx2_bits:
.text
.globl fe_pow22523_avx2
.type fe_pow22523_avx2,@function
.align 4
.align 16
fe_pow22523_avx2:
#else
.section __TEXT,__text
.globl _fe_pow22523_avx2
.p2align 2
.p2align 4
_fe_pow22523_avx2:
#endif /* __APPLE__ */
subq $0x70, %rsp
@@ -12350,12 +12350,12 @@ _fe_pow22523_avx2:
.text
.globl fe_ge_to_p2_avx2
.type fe_ge_to_p2_avx2,@function
.align 4
.align 16
fe_ge_to_p2_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_to_p2_avx2
.p2align 2
.p2align 4
_fe_ge_to_p2_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -12765,12 +12765,12 @@ _fe_ge_to_p2_avx2:
.text
.globl fe_ge_to_p3_avx2
.type fe_ge_to_p3_avx2,@function
.align 4
.align 16
fe_ge_to_p3_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_to_p3_avx2
.p2align 2
.p2align 4
_fe_ge_to_p3_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -13308,12 +13308,12 @@ _fe_ge_to_p3_avx2:
.text
.globl fe_ge_dbl_avx2
.type fe_ge_dbl_avx2,@function
.align 4
.align 16
fe_ge_dbl_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_dbl_avx2
.p2align 2
.p2align 4
_fe_ge_dbl_avx2:
#endif /* __APPLE__ */
pushq %rbp
@@ -13894,12 +13894,12 @@ _fe_ge_dbl_avx2:
.text
.globl fe_ge_madd_avx2
.type fe_ge_madd_avx2,@function
.align 4
.align 16
fe_ge_madd_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_madd_avx2
.p2align 2
.p2align 4
_fe_ge_madd_avx2:
#endif /* __APPLE__ */
pushq %rbp
@@ -14493,12 +14493,12 @@ _fe_ge_madd_avx2:
.text
.globl fe_ge_msub_avx2
.type fe_ge_msub_avx2,@function
.align 4
.align 16
fe_ge_msub_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_msub_avx2
.p2align 2
.p2align 4
_fe_ge_msub_avx2:
#endif /* __APPLE__ */
pushq %rbp
@@ -15091,12 +15091,12 @@ _fe_ge_msub_avx2:
.text
.globl fe_ge_add_avx2
.type fe_ge_add_avx2,@function
.align 4
.align 16
fe_ge_add_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_add_avx2
.p2align 2
.p2align 4
_fe_ge_add_avx2:
#endif /* __APPLE__ */
pushq %rbx
@@ -15817,12 +15817,12 @@ _fe_ge_add_avx2:
.text
.globl fe_ge_sub_avx2
.type fe_ge_sub_avx2,@function
.align 4
.align 16
fe_ge_sub_avx2:
#else
.section __TEXT,__text
.globl _fe_ge_sub_avx2
.p2align 2
.p2align 4
_fe_ge_sub_avx2:
#endif /* __APPLE__ */
pushq %rbx
wolfcrypt/src/poly1305_asm.S
+16 -16
View File
@@ -31,12 +31,12 @@
.text
.globl poly1305_setkey_avx
.type poly1305_setkey_avx,@function
.align 4
.align 16
poly1305_setkey_avx:
#else
.section __TEXT,__text
.globl _poly1305_setkey_avx
.p2align 2
.p2align 4
_poly1305_setkey_avx:
#endif /* __APPLE__ */
movabsq $0xffffffc0fffffff, %r10
@@ -91,12 +91,12 @@ _poly1305_setkey_avx:
.text
.globl poly1305_block_avx
.type poly1305_block_avx,@function
.align 4
.align 16
poly1305_block_avx:
#else
.section __TEXT,__text
.globl _poly1305_block_avx
.p2align 2
.p2align 4
_poly1305_block_avx:
#endif /* __APPLE__ */
pushq %r15
@@ -175,12 +175,12 @@ _poly1305_block_avx:
.text
.globl poly1305_blocks_avx
.type poly1305_blocks_avx,@function
.align 4
.align 16
poly1305_blocks_avx:
#else
.section __TEXT,__text
.globl _poly1305_blocks_avx
.p2align 2
.p2align 4
_poly1305_blocks_avx:
#endif /* __APPLE__ */
pushq %r15
@@ -263,12 +263,12 @@ L_poly1305_avx_blocks_start:
.text
.globl poly1305_final_avx
.type poly1305_final_avx,@function
.align 4
.align 16
poly1305_final_avx:
#else
.section __TEXT,__text
.globl _poly1305_final_avx
.p2align 2
.p2align 4
_poly1305_final_avx:
#endif /* __APPLE__ */
pushq %rbx
@@ -346,12 +346,12 @@ L_poly1305_avx_final_no_more:
.text
.globl poly1305_calc_powers_avx2
.type poly1305_calc_powers_avx2,@function
.align 4
.align 16
poly1305_calc_powers_avx2:
#else
.section __TEXT,__text
.globl _poly1305_calc_powers_avx2
.p2align 2
.p2align 4
_poly1305_calc_powers_avx2:
#endif /* __APPLE__ */
pushq %r12
@@ -614,12 +614,12 @@ _poly1305_calc_powers_avx2:
.text
.globl poly1305_setkey_avx2
.type poly1305_setkey_avx2,@function
.align 4
.align 16
poly1305_setkey_avx2:
#else
.section __TEXT,__text
.globl _poly1305_setkey_avx2
.p2align 2
.p2align 4
_poly1305_setkey_avx2:
#endif /* __APPLE__ */
#ifndef __APPLE__
@@ -669,12 +669,12 @@ L_poly1305_avx2_blocks_hibit:
.text
.globl poly1305_blocks_avx2
.type poly1305_blocks_avx2,@function
.align 4
.align 16
poly1305_blocks_avx2:
#else
.section __TEXT,__text
.globl _poly1305_blocks_avx2
.p2align 2
.p2align 4
_poly1305_blocks_avx2:
#endif /* __APPLE__ */
pushq %r12
@@ -1029,12 +1029,12 @@ L_poly1305_avx2_blocks_complete:
.text
.globl poly1305_final_avx2
.type poly1305_final_avx2,@function
.align 4
.align 16
poly1305_final_avx2:
#else
.section __TEXT,__text
.globl _poly1305_final_avx2
.p2align 2
.p2align 4
_poly1305_final_avx2:
#endif /* __APPLE__ */
movb $0x01, 616(%rdi)
wolfcrypt/src/sha256_asm.S
+16 -16
View File
@@ -89,12 +89,12 @@ L_avx1_sha256_flip_mask:
.text
.globl Transform_Sha256_AVX1
.type Transform_Sha256_AVX1,@function
.align 4
.align 16
Transform_Sha256_AVX1:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX1
.p2align 2
.p2align 4
_Transform_Sha256_AVX1:
#endif /* __APPLE__ */
pushq %rbx
@@ -2479,12 +2479,12 @@ _Transform_Sha256_AVX1:
.text
.globl Transform_Sha256_AVX1_Len
.type Transform_Sha256_AVX1_Len,@function
.align 4
.align 16
Transform_Sha256_AVX1_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX1_Len
.p2align 2
.p2align 4
_Transform_Sha256_AVX1_Len:
#endif /* __APPLE__ */
pushq %rbx
@@ -4943,12 +4943,12 @@ L_avx1_rorx_sha256_flip_mask:
.text
.globl Transform_Sha256_AVX1_RORX
.type Transform_Sha256_AVX1_RORX,@function
.align 4
.align 16
Transform_Sha256_AVX1_RORX:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX1_RORX
.p2align 2
.p2align 4
_Transform_Sha256_AVX1_RORX:
#endif /* __APPLE__ */
pushq %rbx
@@ -7291,12 +7291,12 @@ _Transform_Sha256_AVX1_RORX:
.text
.globl Transform_Sha256_AVX1_RORX_Len
.type Transform_Sha256_AVX1_RORX_Len,@function
.align 4
.align 16
Transform_Sha256_AVX1_RORX_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX1_RORX_Len
.p2align 2
.p2align 4
_Transform_Sha256_AVX1_RORX_Len:
#endif /* __APPLE__ */
pushq %rbx
@@ -9735,12 +9735,12 @@ L_avx2_sha256_flip_mask:
.text
.globl Transform_Sha256_AVX2
.type Transform_Sha256_AVX2,@function
.align 4
.align 16
Transform_Sha256_AVX2:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX2
.p2align 2
.p2align 4
_Transform_Sha256_AVX2:
#endif /* __APPLE__ */
pushq %rbx
@@ -12125,12 +12125,12 @@ _Transform_Sha256_AVX2:
.text
.globl Transform_Sha256_AVX2_Len
.type Transform_Sha256_AVX2_Len,@function
.align 4
.align 16
Transform_Sha256_AVX2_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX2_Len
.p2align 2
.p2align 4
_Transform_Sha256_AVX2_Len:
#endif /* __APPLE__ */
pushq %rbx
@@ -16331,12 +16331,12 @@ L_avx2_rorx_sha256_shuf_DC00:
.text
.globl Transform_Sha256_AVX2_RORX
.type Transform_Sha256_AVX2_RORX,@function
.align 4
.align 16
Transform_Sha256_AVX2_RORX:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX2_RORX
.p2align 2
.p2align 4
_Transform_Sha256_AVX2_RORX:
#endif /* __APPLE__ */
pushq %rbx
@@ -18696,12 +18696,12 @@ _Transform_Sha256_AVX2_RORX:
.text
.globl Transform_Sha256_AVX2_RORX_Len
.type Transform_Sha256_AVX2_RORX_Len,@function
.align 4
.align 16
Transform_Sha256_AVX2_RORX_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha256_AVX2_RORX_Len
.p2align 2
.p2align 4
_Transform_Sha256_AVX2_RORX_Len:
#endif /* __APPLE__ */
pushq %rbx
wolfcrypt/src/sha512_asm.S
+16 -16
View File
@@ -94,12 +94,12 @@ L_avx1_sha512_flip_mask:
.text
.globl Transform_Sha512_AVX1
.type Transform_Sha512_AVX1,@function
.align 4
.align 16
Transform_Sha512_AVX1:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX1
.p2align 2
.p2align 4
_Transform_Sha512_AVX1:
#endif /* __APPLE__ */
pushq %rbx
@@ -1368,12 +1368,12 @@ L_sha256_len_avx1_start:
.text
.globl Transform_Sha512_AVX1_Len
.type Transform_Sha512_AVX1_Len,@function
.align 4
.align 16
Transform_Sha512_AVX1_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX1_Len
.p2align 2
.p2align 4
_Transform_Sha512_AVX1_Len:
#endif /* __APPLE__ */
pushq %rbx
@@ -2724,12 +2724,12 @@ L_avx1_rorx_sha512_flip_mask:
.text
.globl Transform_Sha512_AVX1_RORX
.type Transform_Sha512_AVX1_RORX,@function
.align 4
.align 16
Transform_Sha512_AVX1_RORX:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX1_RORX
.p2align 2
.p2align 4
_Transform_Sha512_AVX1_RORX:
#endif /* __APPLE__ */
pushq %rbx
@@ -3927,12 +3927,12 @@ L_sha256_len_avx1_rorx_start:
.text
.globl Transform_Sha512_AVX1_RORX_Len
.type Transform_Sha512_AVX1_RORX_Len,@function
.align 4
.align 16
Transform_Sha512_AVX1_RORX_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX1_RORX_Len
.p2align 2
.p2align 4
_Transform_Sha512_AVX1_RORX_Len:
#endif /* __APPLE__ */
pushq %rbx
@@ -5334,12 +5334,12 @@ L_avx2_sha512_flip_mask:
.text
.globl Transform_Sha512_AVX2
.type Transform_Sha512_AVX2,@function
.align 4
.align 16
Transform_Sha512_AVX2:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX2
.p2align 2
.p2align 4
_Transform_Sha512_AVX2:
#endif /* __APPLE__ */
pushq %rbx
@@ -6396,12 +6396,12 @@ L_sha256_avx2_start:
.text
.globl Transform_Sha512_AVX2_Len
.type Transform_Sha512_AVX2_Len,@function
.align 4
.align 16
Transform_Sha512_AVX2_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX2_Len
.p2align 2
.p2align 4
_Transform_Sha512_AVX2_Len:
#endif /* __APPLE__ */
pushq %rbx
@@ -8191,12 +8191,12 @@ L_avx2_rorx_sha512_flip_mask:
.text
.globl Transform_Sha512_AVX2_RORX
.type Transform_Sha512_AVX2_RORX,@function
.align 4
.align 16
Transform_Sha512_AVX2_RORX:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX2_RORX
.p2align 2
.p2align 4
_Transform_Sha512_AVX2_RORX:
#endif /* __APPLE__ */
pushq %rbx
@@ -9195,12 +9195,12 @@ L_sha256_len_avx2_rorx_start:
.text
.globl Transform_Sha512_AVX2_RORX_Len
.type Transform_Sha512_AVX2_RORX_Len,@function
.align 4
.align 16
Transform_Sha512_AVX2_RORX_Len:
#else
.section __TEXT,__text
.globl _Transform_Sha512_AVX2_RORX_Len
.p2align 2
.p2align 4
_Transform_Sha512_AVX2_RORX_Len:
#endif /* __APPLE__ */
pushq %rbx
wolfcrypt/src/sp_arm32.c
+653 -10
View File
@@ -31807,7 +31807,7 @@ typedef struct sp_256_proj_point_add_8_ctx {
sp_digit* z;
} sp_256_proj_point_add_8_ctx;
static int sp_256_proj_point_add_8_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
static int sp_256_proj_point_add_8_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
const sp_point_256* p, const sp_point_256* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -36466,7 +36466,7 @@ static int sp_256_mont_inv_order_8_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
{
int err = FP_WOULDBLOCK;
sp_256_mont_inv_order_8_ctx* ctx = (sp_256_mont_inv_order_8_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_8_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -36665,9 +36665,9 @@ int sp_ecc_sign_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_256_ecc_mulmod_8_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_256_ecc_mulmod_8_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p256_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -36920,6 +36920,293 @@ int sp_ecc_sign_256(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
static void sp_256_rshift1_8(sp_digit* r, sp_digit* a)
{
__asm__ __volatile__ (
"mov r10, #0\n\t"
"mov r9, #0\n\t"
"ldr r3, [%[a], #16]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #24]\n\t"
"ldr r6, [%[a], #28]\n\t"
"lsr r7, r3, #1\n\t"
"and r3, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"mov r9, r3\n\t"
"str r7, [%[r], #16]\n\t"
"str r8, [%[r], #20]\n\t"
"str r10, [%[r], #24]\n\t"
"str r14, [%[r], #28]\n\t"
"ldr r3, [%[r], #0]\n\t"
"ldr r4, [%[r], #4]\n\t"
"ldr r5, [%[r], #8]\n\t"
"ldr r6, [%[r], #12]\n\t"
"lsr r7, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"str r7, [%[r], #0]\n\t"
"str r8, [%[r], #4]\n\t"
"str r10, [%[r], #8]\n\t"
"str r14, [%[r], #12]\n\t"
:
: [r] "r" (r), [a] "r" (a)
: "memory", "r3", "r4", "r5", "r6", "r7", "r8", "r10", "r14", "r9"
);
}
/* Divide the number by 2 mod the modulus. (r = a / 2 % m)
*
* r Result of division by 2.
* a Number to divide.
* m Modulus.
*/
static void sp_256_div2_mod_8(sp_digit* r, const sp_digit* a,
const sp_digit* m)
{
__asm__ __volatile__ (
"mov r10, #0\n\t"
"ldr r3, [%[a], #0]\n\t"
"ands r9, r3, #1\n\t"
"beq 1f\n\t"
"ldr r4, [%[a], #4]\n\t"
"ldr r5, [%[a], #8]\n\t"
"ldr r6, [%[a], #12]\n\t"
"ldr r7, [%[m], #0]\n\t"
"ldr r8, [%[m], #4]\n\t"
"ldr r10, [%[m], #8]\n\t"
"ldr r14, [%[m], #12]\n\t"
"adds r3, r3, r7\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r10\n\t"
"adcs r6, r6, r14\n\t"
"str r3, [%[r], #0]\n\t"
"str r4, [%[r], #4]\n\t"
"str r5, [%[r], #8]\n\t"
"str r6, [%[r], #12]\n\t"
"ldr r3, [%[a], #16]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #24]\n\t"
"ldr r6, [%[a], #28]\n\t"
"ldr r7, [%[m], #16]\n\t"
"ldr r8, [%[m], #20]\n\t"
"ldr r10, [%[m], #24]\n\t"
"ldr r14, [%[m], #28]\n\t"
"adcs r3, r3, r7\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r10\n\t"
"adcs r6, r6, r14\n\t"
"adc r9, r10, r10\n\t"
"b 2f\n\t"
"\n1:\n\t"
"ldr r3, [%[a], #16]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #24]\n\t"
"ldr r6, [%[a], #28]\n\t"
"\n2:\n\t"
"lsr r7, r3, #1\n\t"
"and r3, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"mov r9, r3\n\t"
"str r7, [%[r], #16]\n\t"
"str r8, [%[r], #20]\n\t"
"str r10, [%[r], #24]\n\t"
"str r14, [%[r], #28]\n\t"
"ldr r3, [%[r], #0]\n\t"
"ldr r4, [%[r], #4]\n\t"
"ldr r5, [%[r], #8]\n\t"
"ldr r6, [%[r], #12]\n\t"
"lsr r7, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"str r7, [%[r], #0]\n\t"
"str r8, [%[r], #4]\n\t"
"str r10, [%[r], #8]\n\t"
"str r14, [%[r], #12]\n\t"
:
: [r] "r" (r), [a] "r" (a), [m] "r" (m)
: "memory", "r3", "r4", "r5", "r6", "r7", "r8", "r10", "r14", "r9"
);
}
static int sp_256_num_bits_8(sp_digit* a)
{
int r = 0;
__asm__ __volatile__ (
"ldr r2, [%[a], #28]\n\t"
"cmp r2, #0\n\t"
"beq 7f\n\t"
"mov r3, #256\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n7:\n\t"
"ldr r2, [%[a], #24]\n\t"
"cmp r2, #0\n\t"
"beq 6f\n\t"
"mov r3, #224\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n6:\n\t"
"ldr r2, [%[a], #20]\n\t"
"cmp r2, #0\n\t"
"beq 5f\n\t"
"mov r3, #192\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n5:\n\t"
"ldr r2, [%[a], #16]\n\t"
"cmp r2, #0\n\t"
"beq 4f\n\t"
"mov r3, #160\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n4:\n\t"
"ldr r2, [%[a], #12]\n\t"
"cmp r2, #0\n\t"
"beq 3f\n\t"
"mov r3, #128\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n3:\n\t"
"ldr r2, [%[a], #8]\n\t"
"cmp r2, #0\n\t"
"beq 2f\n\t"
"mov r3, #96\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n2:\n\t"
"ldr r2, [%[a], #4]\n\t"
"cmp r2, #0\n\t"
"beq 1f\n\t"
"mov r3, #64\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n1:\n\t"
"ldr r2, [%[a], #0]\n\t"
"mov r3, #32\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"\n9:\n\t"
: [r] "+r" (r)
: [a] "r" (a)
: "r2", "r3"
);
return r;
}
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_256_mod_inv_8(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
sp_digit u[8];
sp_digit v[8];
sp_digit b[8];
sp_digit d[8];
int ut, vt;
sp_digit o;
XMEMCPY(u, m, sizeof(u));
XMEMCPY(v, a, sizeof(v));
ut = sp_256_num_bits_8(u);
vt = sp_256_num_bits_8(v);
XMEMSET(b, 0, sizeof(b));
if ((v[0] & 1) == 0) {
sp_256_rshift1_8(v, v);
XMEMCPY(d, m, sizeof(u));
d[0] += 1;
sp_256_rshift1_8(d, d);
vt--;
while ((v[0] & 1) == 0) {
sp_256_rshift1_8(v, v);
sp_256_div2_mod_8(d, d, m);
vt--;
}
}
else {
XMEMSET(d+1, 0, sizeof(d)-sizeof(sp_digit));
d[0] = 1;
}
while (ut > 1 && vt > 1) {
if (ut > vt || (ut == vt && sp_256_cmp_8(u, v) >= 0)) {
sp_256_sub_8(u, u, v);
o = sp_256_sub_8(b, b, d);
if (o != 0)
sp_256_add_8(b, b, m);
ut = sp_256_num_bits_8(u);
do {
sp_256_rshift1_8(u, u);
sp_256_div2_mod_8(b, b, m);
ut--;
}
while (ut > 0 && (u[0] & 1) == 0);
}
else {
sp_256_sub_8(v, v, u);
o = sp_256_sub_8(d, d, b);
if (o != 0)
sp_256_add_8(d, d, m);
vt = sp_256_num_bits_8(v);
do {
sp_256_rshift1_8(v, v);
sp_256_div2_mod_8(d, d, m);
vt--;
}
while (vt > 0 && (v[0] & 1) == 0);
}
}
if (ut == 1)
XMEMCPY(r, b, sizeof(b));
else
XMEMCPY(r, d, sizeof(d));
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 256)
@@ -37045,7 +37332,7 @@ int sp_ecc_verify_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_256_proj_point_dbl_8_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_256_proj_point_dbl_8_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -37168,6 +37455,11 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
sp_256_from_mp(p2->y, 8, pY);
sp_256_from_mp(p2->z, 8, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_256_mod_inv_8(s, s, p256_order);
}
#endif /* !WOLFSSL_SP_SMALL */
{
sp_256_mul_8(s, s, p256_norm_order);
}
@@ -37175,12 +37467,20 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_256_norm_8(s);
#ifdef WOLFSSL_SP_SMALL
{
sp_256_mont_inv_order_8(s, s, tmp);
sp_256_mont_mul_order_8(u1, u1, s);
sp_256_mont_mul_order_8(u2, u2, s);
}
#else
{
sp_256_mont_mul_order_8(u1, u1, s);
sp_256_mont_mul_order_8(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
err = sp_256_ecc_mulmod_base_8(p1, u1, 0, 0, heap);
}
if (err == MP_OKAY) {
@@ -41455,7 +41755,7 @@ typedef struct sp_384_proj_point_add_12_ctx {
sp_digit* z;
} sp_384_proj_point_add_12_ctx;
static int sp_384_proj_point_add_12_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
static int sp_384_proj_point_add_12_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
const sp_point_384* p, const sp_point_384* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -45249,7 +45549,7 @@ static int sp_384_mont_inv_order_12_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const
{
int err = FP_WOULDBLOCK;
sp_384_mont_inv_order_12_ctx* ctx = (sp_384_mont_inv_order_12_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_384_mont_inv_order_12_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -45419,9 +45719,9 @@ int sp_ecc_sign_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_384_ecc_mulmod_12_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_384_ecc_mulmod_12_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p384_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -45674,6 +45974,336 @@ int sp_ecc_sign_384(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
/* Divide the number by 2 mod the modulus. (r = a / 2 % m)
*
* r Result of division by 2.
* a Number to divide.
* m Modulus.
*/
static void sp_384_div2_mod_12(sp_digit* r, const sp_digit* a,
const sp_digit* m)
{
__asm__ __volatile__ (
"ldr r4, [%[a]]\n\t"
"ands r8, r4, #1\n\t"
"beq 1f\n\t"
"mov r12, #0\n\t"
"ldr r5, [%[a], #4]\n\t"
"ldr r6, [%[a], #8]\n\t"
"ldr r7, [%[a], #12]\n\t"
"ldr r8, [%[m], #0]\n\t"
"ldr r9, [%[m], #4]\n\t"
"ldr r10, [%[m], #8]\n\t"
"ldr r14, [%[m], #12]\n\t"
"adds r4, r4, r8\n\t"
"adcs r5, r5, r9\n\t"
"adcs r6, r6, r10\n\t"
"adcs r7, r7, r14\n\t"
"str r4, [%[r], #0]\n\t"
"str r5, [%[r], #4]\n\t"
"str r6, [%[r], #8]\n\t"
"str r7, [%[r], #12]\n\t"
"ldr r4, [%[a], #16]\n\t"
"ldr r5, [%[a], #20]\n\t"
"ldr r6, [%[a], #24]\n\t"
"ldr r7, [%[a], #28]\n\t"
"ldr r8, [%[m], #16]\n\t"
"ldr r9, [%[m], #20]\n\t"
"ldr r10, [%[m], #24]\n\t"
"ldr r14, [%[m], #28]\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r9\n\t"
"adcs r6, r6, r10\n\t"
"adcs r7, r7, r14\n\t"
"str r4, [%[r], #16]\n\t"
"str r5, [%[r], #20]\n\t"
"str r6, [%[r], #24]\n\t"
"str r7, [%[r], #28]\n\t"
"ldr r4, [%[a], #32]\n\t"
"ldr r5, [%[a], #36]\n\t"
"ldr r6, [%[a], #40]\n\t"
"ldr r7, [%[a], #44]\n\t"
"ldr r8, [%[m], #32]\n\t"
"ldr r9, [%[m], #36]\n\t"
"ldr r10, [%[m], #40]\n\t"
"ldr r14, [%[m], #44]\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r9\n\t"
"adcs r6, r6, r10\n\t"
"adcs r7, r7, r14\n\t"
"str r4, [%[r], #32]\n\t"
"str r5, [%[r], #36]\n\t"
"str r6, [%[r], #40]\n\t"
"str r7, [%[r], #44]\n\t"
"adc r8, r12, r12\n\t"
"b 2f\n\t"
"\n1:\n\t"
"ldr r5, [%[a], #2]\n\t"
"str r4, [%[r], #0]\n\t"
"str r5, [%[r], #2]\n\t"
"ldr r4, [%[a], #4]\n\t"
"ldr r5, [%[a], #6]\n\t"
"str r4, [%[r], #4]\n\t"
"str r5, [%[r], #6]\n\t"
"ldr r4, [%[a], #8]\n\t"
"ldr r5, [%[a], #10]\n\t"
"str r4, [%[r], #8]\n\t"
"str r5, [%[r], #10]\n\t"
"ldr r4, [%[a], #12]\n\t"
"ldr r5, [%[a], #14]\n\t"
"str r4, [%[r], #12]\n\t"
"str r5, [%[r], #14]\n\t"
"ldr r4, [%[a], #16]\n\t"
"ldr r5, [%[a], #18]\n\t"
"str r4, [%[r], #16]\n\t"
"str r5, [%[r], #18]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #22]\n\t"
"str r4, [%[r], #20]\n\t"
"str r5, [%[r], #22]\n\t"
"\n2:\n\t"
"ldr r3, [%[r]]\n\t"
"ldr r4, [%[r], #4]\n\t"
"lsr r3, r3, #1\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #8]\n\t"
"str r3, [%[r], #0]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"ldr r3, [%[a], #12]\n\t"
"str r4, [%[r], #4]\n\t"
"orr r5, r5, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"ldr r4, [%[a], #16]\n\t"
"str r5, [%[r], #8]\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #20]\n\t"
"str r3, [%[r], #12]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"ldr r3, [%[a], #24]\n\t"
"str r4, [%[r], #16]\n\t"
"orr r5, r5, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"ldr r4, [%[a], #28]\n\t"
"str r5, [%[r], #20]\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #32]\n\t"
"str r3, [%[r], #24]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"ldr r3, [%[a], #36]\n\t"
"str r4, [%[r], #28]\n\t"
"orr r5, r5, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"ldr r4, [%[a], #40]\n\t"
"str r5, [%[r], #32]\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #44]\n\t"
"str r3, [%[r], #36]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"orr r5, r5, r8, lsl #31\n\t"
"str r4, [%[r], #40]\n\t"
"str r5, [%[r], #44]\n\t"
:
: [r] "r" (r), [a] "r" (a), [m] "r" (m)
: "memory", "r3", "r4", "r5", "r6", "r7", "r8", "r9", "r10", "r14"
);
}
static int sp_384_num_bits_12(sp_digit* a)
{
int r = 0;
__asm__ __volatile__ (
"ldr r2, [%[a], #44]\n\t"
"cmp r2, #0\n\t"
"beq 11f\n\t"
"mov r3, #384\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n11:\n\t"
"ldr r2, [%[a], #40]\n\t"
"cmp r2, #0\n\t"
"beq 10f\n\t"
"mov r3, #352\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n10:\n\t"
"ldr r2, [%[a], #36]\n\t"
"cmp r2, #0\n\t"
"beq 9f\n\t"
"mov r3, #320\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n9:\n\t"
"ldr r2, [%[a], #32]\n\t"
"cmp r2, #0\n\t"
"beq 8f\n\t"
"mov r3, #288\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n8:\n\t"
"ldr r2, [%[a], #28]\n\t"
"cmp r2, #0\n\t"
"beq 7f\n\t"
"mov r3, #256\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n7:\n\t"
"ldr r2, [%[a], #24]\n\t"
"cmp r2, #0\n\t"
"beq 6f\n\t"
"mov r3, #224\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n6:\n\t"
"ldr r2, [%[a], #20]\n\t"
"cmp r2, #0\n\t"
"beq 5f\n\t"
"mov r3, #192\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n5:\n\t"
"ldr r2, [%[a], #16]\n\t"
"cmp r2, #0\n\t"
"beq 4f\n\t"
"mov r3, #160\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n4:\n\t"
"ldr r2, [%[a], #12]\n\t"
"cmp r2, #0\n\t"
"beq 3f\n\t"
"mov r3, #128\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n3:\n\t"
"ldr r2, [%[a], #8]\n\t"
"cmp r2, #0\n\t"
"beq 2f\n\t"
"mov r3, #96\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n2:\n\t"
"ldr r2, [%[a], #4]\n\t"
"cmp r2, #0\n\t"
"beq 1f\n\t"
"mov r3, #64\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n1:\n\t"
"ldr r2, [%[a], #0]\n\t"
"mov r3, #32\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"\n13:\n\t"
: [r] "+r" (r)
: [a] "r" (a)
: "r2", "r3"
);
return r;
}
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_384_mod_inv_12(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
sp_digit u[12];
sp_digit v[12];
sp_digit b[12];
sp_digit d[12];
int ut, vt;
sp_digit o;
XMEMCPY(u, m, sizeof(u));
XMEMCPY(v, a, sizeof(v));
ut = sp_384_num_bits_12(u);
vt = sp_384_num_bits_12(v);
XMEMSET(b, 0, sizeof(b));
if ((v[0] & 1) == 0) {
sp_384_rshift1_12(v, v);
XMEMCPY(d, m, sizeof(u));
d[0] += 1;
sp_384_rshift1_12(d, d);
vt--;
while ((v[0] & 1) == 0) {
sp_384_rshift1_12(v, v);
sp_384_div2_mod_12(d, d, m);
vt--;
}
}
else {
XMEMSET(d+1, 0, sizeof(d)-sizeof(sp_digit));
d[0] = 1;
}
while (ut > 1 && vt > 1) {
if (ut > vt || (ut == vt && sp_384_cmp_12(u, v) >= 0)) {
sp_384_sub_12(u, u, v);
o = sp_384_sub_12(b, b, d);
if (o != 0)
sp_384_add_12(b, b, m);
ut = sp_384_num_bits_12(u);
do {
sp_384_rshift1_12(u, u);
sp_384_div2_mod_12(b, b, m);
ut--;
}
while (ut > 0 && (u[0] & 1) == 0);
}
else {
sp_384_sub_12(v, v, u);
o = sp_384_sub_12(d, d, b);
if (o != 0)
sp_384_add_12(d, d, m);
vt = sp_384_num_bits_12(v);
do {
sp_384_rshift1_12(v, v);
sp_384_div2_mod_12(d, d, m);
vt--;
}
while (vt > 0 && (v[0] & 1) == 0);
}
}
if (ut == 1)
XMEMCPY(r, b, sizeof(b));
else
XMEMCPY(r, d, sizeof(d));
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 384)
@@ -45799,7 +46429,7 @@ int sp_ecc_verify_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_384_proj_point_dbl_12_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_384_proj_point_dbl_12_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -45922,6 +46552,11 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
sp_384_from_mp(p2->y, 12, pY);
sp_384_from_mp(p2->z, 12, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_384_mod_inv_12(s, s, p384_order);
}
#endif /* !WOLFSSL_SP_SMALL */
{
sp_384_mul_12(s, s, p384_norm_order);
}
@@ -45929,12 +46564,20 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_384_norm_12(s);
#ifdef WOLFSSL_SP_SMALL
{
sp_384_mont_inv_order_12(s, s, tmp);
sp_384_mont_mul_order_12(u1, u1, s);
sp_384_mont_mul_order_12(u2, u2, s);
}
#else
{
sp_384_mont_mul_order_12(u1, u1, s);
sp_384_mont_mul_order_12(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
err = sp_384_ecc_mulmod_base_12(p1, u1, 0, 0, heap);
}
if (err == MP_OKAY) {
wolfcrypt/src/sp_arm64.c
+498 -15
View File
@@ -1208,7 +1208,7 @@ SP_NOINLINE static void sp_2048_sqr_16(sp_digit* r, const sp_digit* a)
u += sp_2048_add_16(r + 8, r + 8, z1);
u += sp_2048_add_8(r + 16, r + 16, z2);
sp_2048_add_zero_8(r + 24, z2 + 8, u);
}
/* Sub b from a into a. (a -= b)
@@ -1629,7 +1629,7 @@ SP_NOINLINE static void sp_2048_sqr_32(sp_digit* r, const sp_digit* a)
u += sp_2048_add_32(r + 16, r + 16, z1);
u += sp_2048_add_16(r + 32, r + 32, z2);
sp_2048_add_zero_16(r + 48, z2 + 16, u);
}
#endif /* !WOLFSSL_SP_SMALL */
@@ -8126,7 +8126,7 @@ SP_NOINLINE static void sp_3072_sqr_24(sp_digit* r, const sp_digit* a)
u += sp_3072_add_24(r + 12, r + 12, z1);
u += sp_3072_add_12(r + 24, r + 24, z2);
sp_3072_add_zero_12(r + 36, z2 + 12, u);
}
/* Sub b from a into a. (a -= b)
@@ -8661,7 +8661,7 @@ SP_NOINLINE static void sp_3072_sqr_48(sp_digit* r, const sp_digit* a)
u += sp_3072_add_48(r + 24, r + 24, z1);
u += sp_3072_add_24(r + 48, r + 48, z2);
sp_3072_add_zero_24(r + 72, z2 + 24, u);
}
#endif /* !WOLFSSL_SP_SMALL */
@@ -14651,7 +14651,7 @@ SP_NOINLINE static void sp_4096_sqr_64(sp_digit* r, const sp_digit* a)
u += sp_4096_add_64(r + 32, r + 32, z1);
u += sp_4096_add_32(r + 64, r + 64, z2);
sp_4096_add_zero_32(r + 96, z2 + 32, u);
}
#endif /* !WOLFSSL_SP_SMALL */
@@ -20611,7 +20611,7 @@ typedef struct sp_256_proj_point_add_4_ctx {
sp_digit* z;
} sp_256_proj_point_add_4_ctx;
static int sp_256_proj_point_add_4_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
static int sp_256_proj_point_add_4_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
const sp_point_256* p, const sp_point_256* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -36877,7 +36877,7 @@ static int sp_256_mont_inv_order_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
{
int err = FP_WOULDBLOCK;
sp_256_mont_inv_order_4_ctx* ctx = (sp_256_mont_inv_order_4_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_4_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -37076,9 +37076,9 @@ int sp_ecc_sign_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_256_ecc_mulmod_4_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_256_ecc_mulmod_4_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p256_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -37331,6 +37331,293 @@ int sp_ecc_sign_256(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_256_mod_inv_4(sp_digit* r, const sp_digit* a,
const sp_digit* m)
{
__asm__ __volatile__ (
"ldr x20, [%[m], 0]\n\t"
"ldr x21, [%[m], 8]\n\t"
"ldr x22, [%[m], 16]\n\t"
"ldr x23, [%[m], 24]\n\t"
"ldr x7, [%[a], 0]\n\t"
"ldr x8, [%[a], 8]\n\t"
"ldr x9, [%[a], 16]\n\t"
"ldr x10, [%[a], 24]\n\t"
"mov x3, x20\n\t"
"mov x4, x21\n\t"
"mov x5, x22\n\t"
"mov x6, x23\n\t"
"mov x11, xzr\n\t"
"mov x12, xzr\n\t"
"mov x13, xzr\n\t"
"mov x14, xzr\n\t"
"mov x15, 1\n\t"
"mov x16, xzr\n\t"
"mov x17, xzr\n\t"
"mov x19, xzr\n\t"
"cmp x6, 0\n\t"
"b.eq 10f\n\t"
"mov x26, 256\n\t"
"clz x24, x6\n\t"
"sub x24, x26, x24\n\t"
"b 13f\n\t"
"\n10:\n\t"
"cmp x5, 0\n\t"
"b.eq 11f\n\t"
"mov x26, 192\n\t"
"clz x24, x5\n\t"
"sub x24, x26, x24\n\t"
"b 13f\n\t"
"\n11:\n\t"
"cmp x4, 0\n\t"
"b.eq 12f\n\t"
"mov x26, 128\n\t"
"clz x24, x4\n\t"
"sub x24, x26, x24\n\t"
"b 13f\n\t"
"\n12:\n\t"
"mov x26, 64\n\t"
"clz x24, x3\n\t"
"sub x24, x26, x24\n\t"
"\n13:\n\t"
"cmp x10, 0\n\t"
"b.eq 20f\n\t"
"mov x26, 256\n\t"
"clz x25, x10\n\t"
"sub x25, x26, x25\n\t"
"b 23f\n\t"
"\n20:\n\t"
"cmp x9, 0\n\t"
"b.eq 21f\n\t"
"mov x26, 192\n\t"
"clz x25, x9\n\t"
"sub x25, x26, x25\n\t"
"b 23f\n\t"
"\n21:\n\t"
"cmp x8, 0\n\t"
"b.eq 22f\n\t"
"mov x26, 128\n\t"
"clz x25, x8\n\t"
"sub x25, x26, x25\n\t"
"b 23f\n\t"
"\n22:\n\t"
"mov x26, 64\n\t"
"clz x25, x7\n\t"
"sub x25, x26, x25\n\t"
"\n23:\n\t"
"tst x7, 1\n\t"
"b.ne 90f\n\t"
"\n1:\n\t"
"lsr x7, x7, 1\n\t"
"lsr x27, x8, 1\n\t"
"lsr x28, x9, 1\n\t"
"orr x7, x7, x8, lsl 63\n\t"
"orr x8, x27, x9, lsl 63\n\t"
"orr x9, x28, x10, lsl 63\n\t"
"lsr x10, x10, 1\n\t"
"sub x25, x25, 1\n\t"
"ands x26, x15, 1\n\t"
"b.eq 2f\n\t"
"adds x15, x15, x20\n\t"
"adcs x16, x16, x21\n\t"
"adcs x17, x17, x22\n\t"
"adcs x19, x19, x23\n\t"
"cset x26, cs\n\t"
"\n2:\n\t"
"lsr x15, x15, 1\n\t"
"lsr x27, x16, 1\n\t"
"lsr x28, x17, 1\n\t"
"lsr x29, x19, 1\n\t"
"orr x15, x15, x16, lsl 63\n\t"
"orr x16, x27, x17, lsl 63\n\t"
"orr x17, x28, x19, lsl 63\n\t"
"orr x19, x29, x26, lsl 63\n\t"
"tst x7, 1\n\t"
"b.eq 1b\n\t"
"\n90:\n\t"
"cmp x24, 1\n\t"
"b.eq 100f\n\t"
"cmp x25, 1\n\t"
"b.eq 101f\n\t"
"cmp x24, x25\n\t"
"b.hi 91f\n\t"
"b.cc 92f\n\t"
"cmp x6, x10\n\t"
"b.hi 91f\n\t"
"b.cc 92f\n\t"
"cmp x5, x9\n\t"
"b.hi 91f\n\t"
"b.cc 92f\n\t"
"cmp x4, x8\n\t"
"b.hi 91f\n\t"
"b.cc 92f\n\t"
"cmp x3, x7\n\t"
"b.cc 92f\n\t"
"\n91:\n\t"
"subs x3, x3, x7\n\t"
"sbcs x4, x4, x8\n\t"
"sbcs x5, x5, x9\n\t"
"sbc x6, x6, x10\n\t"
"subs x11, x11, x15\n\t"
"sbcs x12, x12, x16\n\t"
"sbcs x13, x13, x17\n\t"
"sbcs x14, x14, x19\n\t"
"b.cs 30f\n\t"
"adds x11, x11, x20\n\t"
"adcs x12, x12, x21\n\t"
"adcs x13, x13, x22\n\t"
"adc x14, x14, x23\n\t"
"\n30:\n\t"
"cmp x6, 0\n\t"
"b.eq 40f\n\t"
"mov x26, 256\n\t"
"clz x24, x6\n\t"
"sub x24, x26, x24\n\t"
"b 43f\n\t"
"\n40:\n\t"
"cmp x5, 0\n\t"
"b.eq 41f\n\t"
"mov x26, 192\n\t"
"clz x24, x5\n\t"
"sub x24, x26, x24\n\t"
"b 43f\n\t"
"\n41:\n\t"
"cmp x4, 0\n\t"
"b.eq 42f\n\t"
"mov x26, 128\n\t"
"clz x24, x4\n\t"
"sub x24, x26, x24\n\t"
"b 43f\n\t"
"\n42:\n\t"
"mov x26, 64\n\t"
"clz x24, x3\n\t"
"sub x24, x26, x24\n\t"
"\n43:\n\t"
"\n50:\n\t"
"lsr x3, x3, 1\n\t"
"lsr x27, x4, 1\n\t"
"lsr x28, x5, 1\n\t"
"orr x3, x3, x4, lsl 63\n\t"
"orr x4, x27, x5, lsl 63\n\t"
"orr x5, x28, x6, lsl 63\n\t"
"lsr x6, x6, 1\n\t"
"sub x24, x24, 1\n\t"
"ands x26, x11, 1\n\t"
"b.eq 51f\n\t"
"adds x11, x11, x20\n\t"
"adcs x12, x12, x21\n\t"
"adcs x13, x13, x22\n\t"
"adcs x14, x14, x23\n\t"
"cset x26, cs\n\t"
"\n51:\n\t"
"lsr x11, x11, 1\n\t"
"lsr x27, x12, 1\n\t"
"lsr x28, x13, 1\n\t"
"lsr x29, x14, 1\n\t"
"orr x11, x11, x12, lsl 63\n\t"
"orr x12, x27, x13, lsl 63\n\t"
"orr x13, x28, x14, lsl 63\n\t"
"orr x14, x29, x26, lsl 63\n\t"
"tst x3, 1\n\t"
"b.eq 50b\n\t"
"b 90b\n\t"
"\n92:\n\t"
"subs x7, x7, x3\n\t"
"sbcs x8, x8, x4\n\t"
"sbcs x9, x9, x5\n\t"
"sbc x10, x10, x6\n\t"
"subs x15, x15, x11\n\t"
"sbcs x16, x16, x12\n\t"
"sbcs x17, x17, x13\n\t"
"sbcs x19, x19, x14\n\t"
"b.cs 60f\n\t"
"adds x15, x15, x20\n\t"
"adcs x16, x16, x21\n\t"
"adcs x17, x17, x22\n\t"
"adc x19, x19, x23\n\t"
"\n60:\n\t"
"cmp x10, 0\n\t"
"b.eq 70f\n\t"
"mov x26, 256\n\t"
"clz x25, x10\n\t"
"sub x25, x26, x25\n\t"
"b 73f\n\t"
"\n70:\n\t"
"cmp x9, 0\n\t"
"b.eq 71f\n\t"
"mov x26, 192\n\t"
"clz x25, x9\n\t"
"sub x25, x26, x25\n\t"
"b 73f\n\t"
"\n71:\n\t"
"cmp x8, 0\n\t"
"b.eq 72f\n\t"
"mov x26, 128\n\t"
"clz x25, x8\n\t"
"sub x25, x26, x25\n\t"
"b 73f\n\t"
"\n72:\n\t"
"mov x26, 64\n\t"
"clz x25, x7\n\t"
"sub x25, x26, x25\n\t"
"\n73:\n\t"
"\n80:\n\t"
"lsr x7, x7, 1\n\t"
"lsr x27, x8, 1\n\t"
"lsr x28, x9, 1\n\t"
"orr x7, x7, x8, lsl 63\n\t"
"orr x8, x27, x9, lsl 63\n\t"
"orr x9, x28, x10, lsl 63\n\t"
"lsr x10, x10, 1\n\t"
"sub x25, x25, 1\n\t"
"ands x26, x15, 1\n\t"
"b.eq 81f\n\t"
"adds x15, x15, x20\n\t"
"adcs x16, x16, x21\n\t"
"adcs x17, x17, x22\n\t"
"adcs x19, x19, x23\n\t"
"cset x26, cs\n\t"
"\n81:\n\t"
"lsr x15, x15, 1\n\t"
"lsr x27, x16, 1\n\t"
"lsr x28, x17, 1\n\t"
"lsr x29, x19, 1\n\t"
"orr x15, x15, x16, lsl 63\n\t"
"orr x16, x27, x17, lsl 63\n\t"
"orr x17, x28, x19, lsl 63\n\t"
"orr x19, x29, x26, lsl 63\n\t"
"tst x7, 1\n\t"
"b.eq 80b\n\t"
"b 90b\n\t"
"\n100:\n\t"
"str x11, [%[r], 0]\n\t"
"str x12, [%[r], 8]\n\t"
"str x13, [%[r], 16]\n\t"
"str x14, [%[r], 24]\n\t"
"b 102f\n\t"
"\n101:\n\t"
"str x15, [%[r], 0]\n\t"
"str x16, [%[r], 8]\n\t"
"str x17, [%[r], 16]\n\t"
"str x19, [%[r], 24]\n\t"
"\n102:\n\t"
:
: [r] "r" (r), [a] "r" (a), [m] "r" (m)
: "memory", "x3", "x4", "x5", "x6", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14", "x15", "x16", "x17", "x19", "x20", "x21", "x22", "x23", "x24", "x25", "x26", "x27", "x28", "x29"
);
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 256)
@@ -37456,7 +37743,7 @@ int sp_ecc_verify_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_256_proj_point_dbl_4_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_256_proj_point_dbl_4_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -37579,6 +37866,11 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
sp_256_from_mp(p2->y, 4, pY);
sp_256_from_mp(p2->z, 4, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_256_mod_inv_4(s, s, p256_order);
}
#endif /* !WOLFSSL_SP_SMALL */
{
sp_256_mul_4(s, s, p256_norm_order);
}
@@ -37586,12 +37878,20 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_256_norm_4(s);
#ifdef WOLFSSL_SP_SMALL
{
sp_256_mont_inv_order_4(s, s, tmp);
sp_256_mont_mul_order_4(u1, u1, s);
sp_256_mont_mul_order_4(u2, u2, s);
}
#else
{
sp_256_mont_mul_order_4(u1, u1, s);
sp_256_mont_mul_order_4(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
err = sp_256_ecc_mulmod_base_4(p1, u1, 0, 0, heap);
}
if (err == MP_OKAY) {
@@ -40180,7 +40480,7 @@ typedef struct sp_384_proj_point_add_6_ctx {
sp_digit* z;
} sp_384_proj_point_add_6_ctx;
static int sp_384_proj_point_add_6_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
static int sp_384_proj_point_add_6_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
const sp_point_384* p, const sp_point_384* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -43366,7 +43666,7 @@ static int sp_384_mont_inv_order_6_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
{
int err = FP_WOULDBLOCK;
sp_384_mont_inv_order_6_ctx* ctx = (sp_384_mont_inv_order_6_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_384_mont_inv_order_6_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -43536,9 +43836,9 @@ int sp_ecc_sign_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_384_ecc_mulmod_6_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_384_ecc_mulmod_6_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p384_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -43791,6 +44091,176 @@ int sp_ecc_sign_384(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
/* Divide the number by 2 mod the modulus. (r = a / 2 % m)
*
* r Result of division by 2.
* a Number to divide.
* m Modulus.
*/
static void sp_384_div2_mod_6(sp_digit* r, const sp_digit* a,
const sp_digit* m)
{
__asm__ __volatile__ (
"ldr x3, [%[a], 0]\n\t"
"ldr x4, [%[a], 8]\n\t"
"ldr x5, [%[a], 16]\n\t"
"ldr x6, [%[a], 24]\n\t"
"ldr x7, [%[a], 32]\n\t"
"ldr x8, [%[a], 40]\n\t"
"ldr x9, [%[m], 0]\n\t"
"ldr x10, [%[m], 8]\n\t"
"ldr x11, [%[m], 16]\n\t"
"ldr x12, [%[m], 24]\n\t"
"ldr x13, [%[m], 32]\n\t"
"ldr x14, [%[m], 40]\n\t"
"ands x15, x3, 1\n\t"
"b.eq 1f\n\t"
"adds x3, x3, x9\n\t"
"adcs x4, x4, x10\n\t"
"adcs x5, x5, x11\n\t"
"adcs x6, x6, x12\n\t"
"adcs x7, x7, x13\n\t"
"adcs x8, x8, x14\n\t"
"cset x15, cs\n\t"
"\n1:\n\t"
"lsr x3, x3, 1\n\t"
"lsr x10, x4, 1\n\t"
"lsr x11, x5, 1\n\t"
"lsr x12, x6, 1\n\t"
"lsr x13, x7, 1\n\t"
"lsr x14, x8, 1\n\t"
"orr x3, x3, x4, lsl 63\n\t"
"orr x4, x10, x5, lsl 63\n\t"
"orr x5, x11, x6, lsl 63\n\t"
"orr x6, x12, x7, lsl 63\n\t"
"orr x7, x13, x8, lsl 63\n\t"
"orr x8, x14, x15, lsl 63\n\t"
"str x3, [%[r], 0]\n\t"
"str x4, [%[r], 8]\n\t"
"str x5, [%[r], 16]\n\t"
"str x6, [%[r], 24]\n\t"
"str x7, [%[r], 32]\n\t"
"str x8, [%[r], 40]\n\t"
:
: [r] "r" (r), [a] "r" (a), [m] "r" (m)
: "memory", "x3", "x4", "x5", "x6", "x7", "x8", "x9", "x10", "x11", "x12", "x13", "x14", "x15"
);
}
static int sp_384_num_bits_64_6(sp_digit n)
{
int64_t r = -1;
__asm__ __volatile__ (
"mov x1, 64\n\t"
"clz %[r], %[n]\n\t"
"sub %[r], x1, %[r]"
: [r] "+r" (r)
: [n] "r" (n)
: "x1"
);
return r + 1;
}
static int sp_384_num_bits_6(const sp_digit* a)
{
int i;
int r = 0;
for (i=5; i>=0; i--) {
if (a[i] != 0) {
r = sp_384_num_bits_64_6(a[i]);
r += i * 64;
break;
}
}
return r;
}
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_384_mod_inv_6(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
sp_digit u[6];
sp_digit v[6];
sp_digit b[6];
sp_digit d[6];
int ut, vt;
sp_digit o;
XMEMCPY(u, m, sizeof(u));
XMEMCPY(v, a, sizeof(v));
ut = sp_384_num_bits_6(u);
vt = sp_384_num_bits_6(v);
XMEMSET(b, 0, sizeof(b));
if ((v[0] & 1) == 0) {
sp_384_rshift1_6(v, v);
XMEMCPY(d, m, sizeof(u));
d[0] += 1;
sp_384_rshift1_6(d, d);
vt--;
while ((v[0] & 1) == 0) {
sp_384_rshift1_6(v, v);
sp_384_div2_mod_6(d, d, m);
vt--;
}
}
else {
XMEMSET(d+1, 0, sizeof(d)-sizeof(sp_digit));
d[0] = 1;
}
while (ut > 1 && vt > 1) {
if (ut > vt || (ut == vt && sp_384_cmp_6(u, v) >= 0)) {
sp_384_sub_6(u, u, v);
o = sp_384_sub_6(b, b, d);
if (o != 0)
sp_384_add_6(b, b, m);
ut = sp_384_num_bits_6(u);
do {
sp_384_rshift1_6(u, u);
sp_384_div2_mod_6(b, b, m);
ut--;
}
while (ut > 0 && (u[0] & 1) == 0);
}
else {
sp_384_sub_6(v, v, u);
o = sp_384_sub_6(d, d, b);
if (o != 0)
sp_384_add_6(d, d, m);
vt = sp_384_num_bits_6(v);
do {
sp_384_rshift1_6(v, v);
sp_384_div2_mod_6(d, d, m);
vt--;
}
while (vt > 0 && (v[0] & 1) == 0);
}
}
if (ut == 1)
XMEMCPY(r, b, sizeof(b));
else
XMEMCPY(r, d, sizeof(d));
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 384)
@@ -43916,7 +44386,7 @@ int sp_ecc_verify_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_384_proj_point_dbl_6_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_384_proj_point_dbl_6_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -44039,6 +44509,11 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
sp_384_from_mp(p2->y, 6, pY);
sp_384_from_mp(p2->z, 6, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_384_mod_inv_6(s, s, p384_order);
}
#endif /* !WOLFSSL_SP_SMALL */
{
sp_384_mul_6(s, s, p384_norm_order);
}
@@ -44046,12 +44521,20 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_384_norm_6(s);
#ifdef WOLFSSL_SP_SMALL
{
sp_384_mont_inv_order_6(s, s, tmp);
sp_384_mont_mul_order_6(u1, u1, s);
sp_384_mont_mul_order_6(u2, u2, s);
}
#else
{
sp_384_mont_mul_order_6(u1, u1, s);
sp_384_mont_mul_order_6(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
err = sp_384_ecc_mulmod_base_6(p1, u1, 0, 0, heap);
}
if (err == MP_OKAY) {
wolfcrypt/src/sp_armthumb.c
+1395 -10
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/sp_c32.c
+457 -141
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/sp_c64.c
+450 -130
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/sp_cortexm.c
+662 -33
View File
@@ -16798,7 +16798,7 @@ typedef struct sp_256_proj_point_add_8_ctx {
sp_digit* z;
} sp_256_proj_point_add_8_ctx;
static int sp_256_proj_point_add_8_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
static int sp_256_proj_point_add_8_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
const sp_point_256* p, const sp_point_256* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -21245,7 +21245,7 @@ static int sp_256_mont_inv_order_8_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
{
int err = FP_WOULDBLOCK;
sp_256_mont_inv_order_8_ctx* ctx = (sp_256_mont_inv_order_8_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_8_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -21444,9 +21444,9 @@ int sp_ecc_sign_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_256_ecc_mulmod_8_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_256_ecc_mulmod_8_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p256_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -21699,6 +21699,291 @@ int sp_ecc_sign_256(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
static void sp_256_rshift1_8(sp_digit* r, sp_digit* a)
{
__asm__ __volatile__ (
"mov r10, #0\n\t"
"mov r9, #0\n\t"
"ldr r3, [%[a], #16]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #24]\n\t"
"ldr r6, [%[a], #28]\n\t"
"lsr r7, r3, #1\n\t"
"and r3, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"mov r9, r3\n\t"
"str r7, [%[r], #16]\n\t"
"str r8, [%[r], #20]\n\t"
"str r10, [%[r], #24]\n\t"
"str r14, [%[r], #28]\n\t"
"ldr r3, [%[r], #0]\n\t"
"ldr r4, [%[r], #4]\n\t"
"ldr r5, [%[r], #8]\n\t"
"ldr r6, [%[r], #12]\n\t"
"lsr r7, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"str r7, [%[r], #0]\n\t"
"str r8, [%[r], #4]\n\t"
"str r10, [%[r], #8]\n\t"
"str r14, [%[r], #12]\n\t"
:
: [r] "r" (r), [a] "r" (a)
: "memory", "r3", "r4", "r5", "r6", "r7", "r8", "r10", "r14", "r9"
);
}
/* Divide the number by 2 mod the modulus. (r = a / 2 % m)
*
* r Result of division by 2.
* a Number to divide.
* m Modulus.
*/
static void sp_256_div2_mod_8(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
__asm__ __volatile__ (
"mov r10, #0\n\t"
"ldr r3, [%[a], #0]\n\t"
"ands r9, r3, #1\n\t"
"beq 1f\n\t"
"ldr r4, [%[a], #4]\n\t"
"ldr r5, [%[a], #8]\n\t"
"ldr r6, [%[a], #12]\n\t"
"ldr r7, [%[m], #0]\n\t"
"ldr r8, [%[m], #4]\n\t"
"ldr r10, [%[m], #8]\n\t"
"ldr r14, [%[m], #12]\n\t"
"adds r3, r3, r7\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r10\n\t"
"adcs r6, r6, r14\n\t"
"str r3, [%[r], #0]\n\t"
"str r4, [%[r], #4]\n\t"
"str r5, [%[r], #8]\n\t"
"str r6, [%[r], #12]\n\t"
"ldr r3, [%[a], #16]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #24]\n\t"
"ldr r6, [%[a], #28]\n\t"
"ldr r7, [%[m], #16]\n\t"
"ldr r8, [%[m], #20]\n\t"
"ldr r10, [%[m], #24]\n\t"
"ldr r14, [%[m], #28]\n\t"
"adcs r3, r3, r7\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r10\n\t"
"adcs r6, r6, r14\n\t"
"adc r9, r10, r10\n\t"
"b 2f\n\t"
"\n1:\n\t"
"ldr r3, [%[a], #16]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #24]\n\t"
"ldr r6, [%[a], #28]\n\t"
"\n2:\n\t"
"lsr r7, r3, #1\n\t"
"and r3, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"mov r9, r3\n\t"
"str r7, [%[r], #16]\n\t"
"str r8, [%[r], #20]\n\t"
"str r10, [%[r], #24]\n\t"
"str r14, [%[r], #28]\n\t"
"ldr r3, [%[r], #0]\n\t"
"ldr r4, [%[r], #4]\n\t"
"ldr r5, [%[r], #8]\n\t"
"ldr r6, [%[r], #12]\n\t"
"lsr r7, r3, #1\n\t"
"lsr r8, r4, #1\n\t"
"lsr r10, r5, #1\n\t"
"lsr r14, r6, #1\n\t"
"orr r7, r7, r4, lsl #31\n\t"
"orr r8, r8, r5, lsl #31\n\t"
"orr r10, r10, r6, lsl #31\n\t"
"orr r14, r14, r9, lsl #31\n\t"
"str r7, [%[r], #0]\n\t"
"str r8, [%[r], #4]\n\t"
"str r10, [%[r], #8]\n\t"
"str r14, [%[r], #12]\n\t"
:
: [r] "r" (r), [a] "r" (a), [m] "r" (m)
: "memory", "r3", "r4", "r5", "r6", "r7", "r8", "r10", "r14", "r9"
);
}
static int sp_256_num_bits_8(sp_digit* a)
{
int r = 0;
__asm__ __volatile__ (
"ldr r2, [%[a], #28]\n\t"
"cmp r2, #0\n\t"
"beq 7f\n\t"
"mov r3, #256\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n7:\n\t"
"ldr r2, [%[a], #24]\n\t"
"cmp r2, #0\n\t"
"beq 6f\n\t"
"mov r3, #224\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n6:\n\t"
"ldr r2, [%[a], #20]\n\t"
"cmp r2, #0\n\t"
"beq 5f\n\t"
"mov r3, #192\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n5:\n\t"
"ldr r2, [%[a], #16]\n\t"
"cmp r2, #0\n\t"
"beq 4f\n\t"
"mov r3, #160\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n4:\n\t"
"ldr r2, [%[a], #12]\n\t"
"cmp r2, #0\n\t"
"beq 3f\n\t"
"mov r3, #128\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n3:\n\t"
"ldr r2, [%[a], #8]\n\t"
"cmp r2, #0\n\t"
"beq 2f\n\t"
"mov r3, #96\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n2:\n\t"
"ldr r2, [%[a], #4]\n\t"
"cmp r2, #0\n\t"
"beq 1f\n\t"
"mov r3, #64\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 9f\n\t"
"\n1:\n\t"
"ldr r2, [%[a], #0]\n\t"
"mov r3, #32\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"\n9:\n\t"
: [r] "+r" (r)
: [a] "r" (a)
: "r2", "r3"
);
return r;
}
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_256_mod_inv_8(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
sp_digit u[8];
sp_digit v[8];
sp_digit b[8];
sp_digit d[8];
int ut, vt;
sp_digit o;
XMEMCPY(u, m, sizeof(u));
XMEMCPY(v, a, sizeof(v));
ut = sp_256_num_bits_8(u);
vt = sp_256_num_bits_8(v);
XMEMSET(b, 0, sizeof(b));
if ((v[0] & 1) == 0) {
sp_256_rshift1_8(v, v);
XMEMCPY(d, m, sizeof(u));
d[0] += 1;
sp_256_rshift1_8(d, d);
vt--;
while ((v[0] & 1) == 0) {
sp_256_rshift1_8(v, v);
sp_256_div2_mod_8(d, d, m);
vt--;
}
}
else {
XMEMSET(d+1, 0, sizeof(d)-sizeof(sp_digit));
d[0] = 1;
}
while (ut > 1 && vt > 1) {
if (ut > vt || (ut == vt && sp_256_cmp_8(u, v) >= 0)) {
sp_256_sub_8(u, u, v);
o = sp_256_sub_8(b, b, d);
if (o != 0)
sp_256_add_8(b, b, m);
ut = sp_256_num_bits_8(u);
do {
sp_256_rshift1_8(u, u);
sp_256_div2_mod_8(b, b, m);
ut--;
}
while (ut > 0 && (u[0] & 1) == 0);
}
else {
sp_256_sub_8(v, v, u);
o = sp_256_sub_8(d, d, b);
if (o != 0)
sp_256_add_8(d, d, m);
vt = sp_256_num_bits_8(v);
do {
sp_256_rshift1_8(v, v);
sp_256_div2_mod_8(d, d, m);
vt--;
}
while (vt > 0 && (v[0] & 1) == 0);
}
}
if (ut == 1)
XMEMCPY(r, b, sizeof(b));
else
XMEMCPY(r, d, sizeof(d));
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 256)
@@ -21824,7 +22109,7 @@ int sp_ecc_verify_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_256_proj_point_dbl_8_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_256_proj_point_dbl_8_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -21947,6 +22232,11 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
sp_256_from_mp(p2->y, 8, pY);
sp_256_from_mp(p2->z, 8, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_256_mod_inv_8(s, s, p256_order);
}
#endif /* !WOLFSSL_SP_SMALL */
{
sp_256_mul_8(s, s, p256_norm_order);
}
@@ -21954,12 +22244,20 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_256_norm_8(s);
#ifdef WOLFSSL_SP_SMALL
{
sp_256_mont_inv_order_8(s, s, tmp);
sp_256_mont_mul_order_8(u1, u1, s);
sp_256_mont_mul_order_8(u2, u2, s);
}
#else
{
sp_256_mont_mul_order_8(u1, u1, s);
sp_256_mont_mul_order_8(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
err = sp_256_ecc_mulmod_base_8(p1, u1, 0, 0, heap);
}
if (err == MP_OKAY) {
@@ -23907,64 +24205,53 @@ static void sp_384_rshift1_12(sp_digit* r, sp_digit* a)
"ldr r2, [%[a]]\n\t"
"ldr r3, [%[a], #4]\n\t"
"lsr r2, r2, #1\n\t"
"lsl r5, r3, #31\n\t"
"orr r2, r2, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"orr r2, r2, r5\n\t"
"ldr r4, [%[a], #8]\n\t"
"str r2, [%[r], #0]\n\t"
"lsl r5, r4, #31\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"orr r3, r3, r5\n\t"
"ldr r2, [%[a], #12]\n\t"
"str r3, [%[r], #4]\n\t"
"lsl r5, r2, #31\n\t"
"orr r4, r4, r2, lsl #31\n\t"
"lsr r2, r2, #1\n\t"
"orr r4, r4, r5\n\t"
"ldr r3, [%[a], #16]\n\t"
"str r4, [%[r], #8]\n\t"
"lsl r5, r3, #31\n\t"
"orr r2, r2, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"orr r2, r2, r5\n\t"
"ldr r4, [%[a], #20]\n\t"
"str r2, [%[r], #12]\n\t"
"lsl r5, r4, #31\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"orr r3, r3, r5\n\t"
"ldr r2, [%[a], #24]\n\t"
"str r3, [%[r], #16]\n\t"
"lsl r5, r2, #31\n\t"
"orr r4, r4, r2, lsl #31\n\t"
"lsr r2, r2, #1\n\t"
"orr r4, r4, r5\n\t"
"ldr r3, [%[a], #28]\n\t"
"str r4, [%[r], #20]\n\t"
"lsl r5, r3, #31\n\t"
"orr r2, r2, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"orr r2, r2, r5\n\t"
"ldr r4, [%[a], #32]\n\t"
"str r2, [%[r], #24]\n\t"
"lsl r5, r4, #31\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"orr r3, r3, r5\n\t"
"ldr r2, [%[a], #36]\n\t"
"str r3, [%[r], #28]\n\t"
"lsl r5, r2, #31\n\t"
"orr r4, r4, r2, lsl #31\n\t"
"lsr r2, r2, #1\n\t"
"orr r4, r4, r5\n\t"
"ldr r3, [%[a], #40]\n\t"
"str r4, [%[r], #32]\n\t"
"lsl r5, r3, #31\n\t"
"orr r2, r2, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"orr r2, r2, r5\n\t"
"ldr r4, [%[a], #44]\n\t"
"str r2, [%[r], #36]\n\t"
"lsl r5, r4, #31\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"orr r3, r3, r5\n\t"
"str r3, [%[r], #40]\n\t"
"str r4, [%[r], #44]\n\t"
:
: [r] "r" (r), [a] "r" (a)
: "memory", "r2", "r3", "r4", "r5"
: "memory", "r2", "r3", "r4"
);
}
@@ -24216,7 +24503,7 @@ typedef struct sp_384_proj_point_add_12_ctx {
sp_digit* z;
} sp_384_proj_point_add_12_ctx;
static int sp_384_proj_point_add_12_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
static int sp_384_proj_point_add_12_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
const sp_point_384* p, const sp_point_384* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -27887,7 +28174,7 @@ static int sp_384_mont_inv_order_12_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const
{
int err = FP_WOULDBLOCK;
sp_384_mont_inv_order_12_ctx* ctx = (sp_384_mont_inv_order_12_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_384_mont_inv_order_12_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -28057,9 +28344,9 @@ int sp_ecc_sign_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_384_ecc_mulmod_12_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_384_ecc_mulmod_12_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p384_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -28312,6 +28599,335 @@ int sp_ecc_sign_384(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
/* Divide the number by 2 mod the modulus. (r = a / 2 % m)
*
* r Result of division by 2.
* a Number to divide.
* m Modulus.
*/
static void sp_384_div2_mod_12(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
__asm__ __volatile__ (
"ldr r4, [%[a]]\n\t"
"ands r8, r4, #1\n\t"
"beq 1f\n\t"
"mov r12, #0\n\t"
"ldr r5, [%[a], #4]\n\t"
"ldr r6, [%[a], #8]\n\t"
"ldr r7, [%[a], #12]\n\t"
"ldr r8, [%[m], #0]\n\t"
"ldr r9, [%[m], #4]\n\t"
"ldr r10, [%[m], #8]\n\t"
"ldr r14, [%[m], #12]\n\t"
"adds r4, r4, r8\n\t"
"adcs r5, r5, r9\n\t"
"adcs r6, r6, r10\n\t"
"adcs r7, r7, r14\n\t"
"str r4, [%[r], #0]\n\t"
"str r5, [%[r], #4]\n\t"
"str r6, [%[r], #8]\n\t"
"str r7, [%[r], #12]\n\t"
"ldr r4, [%[a], #16]\n\t"
"ldr r5, [%[a], #20]\n\t"
"ldr r6, [%[a], #24]\n\t"
"ldr r7, [%[a], #28]\n\t"
"ldr r8, [%[m], #16]\n\t"
"ldr r9, [%[m], #20]\n\t"
"ldr r10, [%[m], #24]\n\t"
"ldr r14, [%[m], #28]\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r9\n\t"
"adcs r6, r6, r10\n\t"
"adcs r7, r7, r14\n\t"
"str r4, [%[r], #16]\n\t"
"str r5, [%[r], #20]\n\t"
"str r6, [%[r], #24]\n\t"
"str r7, [%[r], #28]\n\t"
"ldr r4, [%[a], #32]\n\t"
"ldr r5, [%[a], #36]\n\t"
"ldr r6, [%[a], #40]\n\t"
"ldr r7, [%[a], #44]\n\t"
"ldr r8, [%[m], #32]\n\t"
"ldr r9, [%[m], #36]\n\t"
"ldr r10, [%[m], #40]\n\t"
"ldr r14, [%[m], #44]\n\t"
"adcs r4, r4, r8\n\t"
"adcs r5, r5, r9\n\t"
"adcs r6, r6, r10\n\t"
"adcs r7, r7, r14\n\t"
"str r4, [%[r], #32]\n\t"
"str r5, [%[r], #36]\n\t"
"str r6, [%[r], #40]\n\t"
"str r7, [%[r], #44]\n\t"
"adc r8, r12, r12\n\t"
"b 2f\n\t"
"\n1:\n\t"
"ldr r5, [%[a], #2]\n\t"
"str r4, [%[r], #0]\n\t"
"str r5, [%[r], #2]\n\t"
"ldr r4, [%[a], #4]\n\t"
"ldr r5, [%[a], #6]\n\t"
"str r4, [%[r], #4]\n\t"
"str r5, [%[r], #6]\n\t"
"ldr r4, [%[a], #8]\n\t"
"ldr r5, [%[a], #10]\n\t"
"str r4, [%[r], #8]\n\t"
"str r5, [%[r], #10]\n\t"
"ldr r4, [%[a], #12]\n\t"
"ldr r5, [%[a], #14]\n\t"
"str r4, [%[r], #12]\n\t"
"str r5, [%[r], #14]\n\t"
"ldr r4, [%[a], #16]\n\t"
"ldr r5, [%[a], #18]\n\t"
"str r4, [%[r], #16]\n\t"
"str r5, [%[r], #18]\n\t"
"ldr r4, [%[a], #20]\n\t"
"ldr r5, [%[a], #22]\n\t"
"str r4, [%[r], #20]\n\t"
"str r5, [%[r], #22]\n\t"
"\n2:\n\t"
"ldr r3, [%[r]]\n\t"
"ldr r4, [%[r], #4]\n\t"
"lsr r3, r3, #1\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #8]\n\t"
"str r3, [%[r], #0]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"ldr r3, [%[a], #12]\n\t"
"str r4, [%[r], #4]\n\t"
"orr r5, r5, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"ldr r4, [%[a], #16]\n\t"
"str r5, [%[r], #8]\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #20]\n\t"
"str r3, [%[r], #12]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"ldr r3, [%[a], #24]\n\t"
"str r4, [%[r], #16]\n\t"
"orr r5, r5, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"ldr r4, [%[a], #28]\n\t"
"str r5, [%[r], #20]\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #32]\n\t"
"str r3, [%[r], #24]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"ldr r3, [%[a], #36]\n\t"
"str r4, [%[r], #28]\n\t"
"orr r5, r5, r3, lsl #31\n\t"
"lsr r3, r3, #1\n\t"
"ldr r4, [%[a], #40]\n\t"
"str r5, [%[r], #32]\n\t"
"orr r3, r3, r4, lsl #31\n\t"
"lsr r4, r4, #1\n\t"
"ldr r5, [%[a], #44]\n\t"
"str r3, [%[r], #36]\n\t"
"orr r4, r4, r5, lsl #31\n\t"
"lsr r5, r5, #1\n\t"
"orr r5, r5, r8, lsl #31\n\t"
"str r4, [%[r], #40]\n\t"
"str r5, [%[r], #44]\n\t"
:
: [r] "r" (r), [a] "r" (a), [m] "r" (m)
: "memory", "r3", "r4", "r5", "r6", "r7", "r8", "r9", "r10", "r14"
);
}
static int sp_384_num_bits_12(sp_digit* a)
{
int r = 0;
__asm__ __volatile__ (
"ldr r2, [%[a], #44]\n\t"
"cmp r2, #0\n\t"
"beq 11f\n\t"
"mov r3, #384\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n11:\n\t"
"ldr r2, [%[a], #40]\n\t"
"cmp r2, #0\n\t"
"beq 10f\n\t"
"mov r3, #352\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n10:\n\t"
"ldr r2, [%[a], #36]\n\t"
"cmp r2, #0\n\t"
"beq 9f\n\t"
"mov r3, #320\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n9:\n\t"
"ldr r2, [%[a], #32]\n\t"
"cmp r2, #0\n\t"
"beq 8f\n\t"
"mov r3, #288\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n8:\n\t"
"ldr r2, [%[a], #28]\n\t"
"cmp r2, #0\n\t"
"beq 7f\n\t"
"mov r3, #256\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n7:\n\t"
"ldr r2, [%[a], #24]\n\t"
"cmp r2, #0\n\t"
"beq 6f\n\t"
"mov r3, #224\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n6:\n\t"
"ldr r2, [%[a], #20]\n\t"
"cmp r2, #0\n\t"
"beq 5f\n\t"
"mov r3, #192\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n5:\n\t"
"ldr r2, [%[a], #16]\n\t"
"cmp r2, #0\n\t"
"beq 4f\n\t"
"mov r3, #160\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n4:\n\t"
"ldr r2, [%[a], #12]\n\t"
"cmp r2, #0\n\t"
"beq 3f\n\t"
"mov r3, #128\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n3:\n\t"
"ldr r2, [%[a], #8]\n\t"
"cmp r2, #0\n\t"
"beq 2f\n\t"
"mov r3, #96\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n2:\n\t"
"ldr r2, [%[a], #4]\n\t"
"cmp r2, #0\n\t"
"beq 1f\n\t"
"mov r3, #64\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"b 13f\n\t"
"\n1:\n\t"
"ldr r2, [%[a], #0]\n\t"
"mov r3, #32\n\t"
"clz %[r], r2\n\t"
"sub %[r], r3, %[r]\n\t"
"\n13:\n\t"
: [r] "+r" (r)
: [a] "r" (a)
: "r2", "r3"
);
return r;
}
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_384_mod_inv_12(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
sp_digit u[12];
sp_digit v[12];
sp_digit b[12];
sp_digit d[12];
int ut, vt;
sp_digit o;
XMEMCPY(u, m, sizeof(u));
XMEMCPY(v, a, sizeof(v));
ut = sp_384_num_bits_12(u);
vt = sp_384_num_bits_12(v);
XMEMSET(b, 0, sizeof(b));
if ((v[0] & 1) == 0) {
sp_384_rshift1_12(v, v);
XMEMCPY(d, m, sizeof(u));
d[0] += 1;
sp_384_rshift1_12(d, d);
vt--;
while ((v[0] & 1) == 0) {
sp_384_rshift1_12(v, v);
sp_384_div2_mod_12(d, d, m);
vt--;
}
}
else {
XMEMSET(d+1, 0, sizeof(d)-sizeof(sp_digit));
d[0] = 1;
}
while (ut > 1 && vt > 1) {
if (ut > vt || (ut == vt && sp_384_cmp_12(u, v) >= 0)) {
sp_384_sub_12(u, u, v);
o = sp_384_sub_12(b, b, d);
if (o != 0)
sp_384_add_12(b, b, m);
ut = sp_384_num_bits_12(u);
do {
sp_384_rshift1_12(u, u);
sp_384_div2_mod_12(b, b, m);
ut--;
}
while (ut > 0 && (u[0] & 1) == 0);
}
else {
sp_384_sub_12(v, v, u);
o = sp_384_sub_12(d, d, b);
if (o != 0)
sp_384_add_12(d, d, m);
vt = sp_384_num_bits_12(v);
do {
sp_384_rshift1_12(v, v);
sp_384_div2_mod_12(d, d, m);
vt--;
}
while (vt > 0 && (v[0] & 1) == 0);
}
}
if (ut == 1)
XMEMCPY(r, b, sizeof(b));
else
XMEMCPY(r, d, sizeof(d));
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 384)
@@ -28437,7 +29053,7 @@ int sp_ecc_verify_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_384_proj_point_dbl_12_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_384_proj_point_dbl_12_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -28560,6 +29176,11 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
sp_384_from_mp(p2->y, 12, pY);
sp_384_from_mp(p2->z, 12, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_384_mod_inv_12(s, s, p384_order);
}
#endif /* !WOLFSSL_SP_SMALL */
{
sp_384_mul_12(s, s, p384_norm_order);
}
@@ -28567,12 +29188,20 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_384_norm_12(s);
#ifdef WOLFSSL_SP_SMALL
{
sp_384_mont_inv_order_12(s, s, tmp);
sp_384_mont_mul_order_12(u1, u1, s);
sp_384_mont_mul_order_12(u2, u2, s);
}
#else
{
sp_384_mont_mul_order_12(u1, u1, s);
sp_384_mont_mul_order_12(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
err = sp_384_ecc_mulmod_base_12(p1, u1, 0, 0, heap);
}
if (err == MP_OKAY) {
wolfcrypt/src/sp_x86_64.c
+179 -52
View File
@@ -7091,7 +7091,7 @@ typedef struct sp_256_proj_point_add_4_ctx {
sp_digit* z;
} sp_256_proj_point_add_4_ctx;
static int sp_256_proj_point_add_4_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
static int sp_256_proj_point_add_4_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
const sp_point_256* p, const sp_point_256* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -8171,7 +8171,7 @@ typedef struct sp_256_proj_point_add_avx2_4_ctx {
sp_digit* z;
} sp_256_proj_point_add_avx2_4_ctx;
static int sp_256_proj_point_add_avx2_4_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
static int sp_256_proj_point_add_avx2_4_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
const sp_point_256* p, const sp_point_256* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -22716,6 +22716,13 @@ static void sp_256_mont_sqr_n_order_4(sp_digit* r, const sp_digit* a, int n)
}
#endif /* !WOLFSSL_SP_SMALL */
#ifdef WOLFSSL_SP_NONBLOCK
/* Context of non-blocking moduluar inversion with Montgomery form number. */
typedef struct sp_256_mont_inv_order_4_ctx {
int state; /* State of next operation. */
int i; /* Index of bit in order. */
} sp_256_mont_inv_order_4_ctx;
/* Invert the number, in Montgomery form, modulo the order of the P256 curve.
* (r = 1 / a mod order)
*
@@ -22723,18 +22730,12 @@ static void sp_256_mont_sqr_n_order_4(sp_digit* r, const sp_digit* a, int n)
* a Number to invert.
* td Temporary data.
*/
#ifdef WOLFSSL_SP_NONBLOCK
typedef struct sp_256_mont_inv_order_4_ctx {
int state;
int i;
} sp_256_mont_inv_order_4_ctx;
static int sp_256_mont_inv_order_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const sp_digit* a,
sp_digit* t)
{
int err = FP_WOULDBLOCK;
sp_256_mont_inv_order_4_ctx* ctx = (sp_256_mont_inv_order_4_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_4_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -22764,6 +22765,13 @@ static int sp_256_mont_inv_order_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
}
#endif /* WOLFSSL_SP_NONBLOCK */
/* Invert the number, in Montgomery form, modulo the order of the P256 curve.
* (r = 1 / a mod order)
*
* r Inverse result.
* a Number to invert.
* td Temporary data.
*/
static void sp_256_mont_inv_order_4(sp_digit* r, const sp_digit* a,
sp_digit* td)
{
@@ -22858,32 +22866,8 @@ static void sp_256_mont_inv_order_4(sp_digit* r, const sp_digit* a,
}
#ifdef HAVE_INTEL_AVX2
extern void sp_256_sqr_avx2_4(sp_digit* r, const sp_digit* a);
#define sp_256_mont_reduce_order_avx2_4 sp_256_mont_reduce_avx2_4
extern void sp_256_mont_reduce_avx2_4(sp_digit* a, const sp_digit* m, sp_digit mp);
/* Multiply two number mod the order of P256 curve. (r = a * b mod order)
*
* r Result of the multiplication.
* a First operand of the multiplication.
* b Second operand of the multiplication.
*/
static void sp_256_mont_mul_order_avx2_4(sp_digit* r, const sp_digit* a, const sp_digit* b)
{
sp_256_mul_avx2_4(r, a, b);
sp_256_mont_reduce_order_avx2_4(r, p256_order, p256_mp_order);
}
/* Square number mod the order of P256 curve. (r = a * a mod order)
*
* r Result of the squaring.
* a Number to square.
*/
static void sp_256_mont_sqr_order_avx2_4(sp_digit* r, const sp_digit* a)
{
sp_256_sqr_avx2_4(r, a);
sp_256_mont_reduce_order_avx2_4(r, p256_order, p256_mp_order);
}
extern void sp_256_mont_mul_order_avx2_4(sp_digit* r, const sp_digit* a, const sp_digit* b);
extern void sp_256_mont_sqr_order_avx2_4(sp_digit* r, const sp_digit* a);
#ifndef WOLFSSL_SP_SMALL
/* Square number mod the order of P256 curve a number of times.
@@ -22903,6 +22887,13 @@ static void sp_256_mont_sqr_n_order_avx2_4(sp_digit* r, const sp_digit* a, int n
}
#endif /* !WOLFSSL_SP_SMALL */
#ifdef WOLFSSL_SP_NONBLOCK
/* Context of non-blocking moduluar inversion with Montgomery form number. */
typedef struct sp_256_mont_inv_order_avx2_4_ctx {
int state; /* State of next operation. */
int i; /* Index of bit in order. */
} sp_256_mont_inv_order_avx2_4_ctx;
/* Invert the number, in Montgomery form, modulo the order of the P256 curve.
* (r = 1 / a mod order)
*
@@ -22910,18 +22901,12 @@ static void sp_256_mont_sqr_n_order_avx2_4(sp_digit* r, const sp_digit* a, int n
* a Number to invert.
* td Temporary data.
*/
#ifdef WOLFSSL_SP_NONBLOCK
typedef struct sp_256_mont_inv_order_avx2_4_ctx {
int state;
int i;
} sp_256_mont_inv_order_avx2_4_ctx;
static int sp_256_mont_inv_order_avx2_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const sp_digit* a,
sp_digit* t)
{
int err = FP_WOULDBLOCK;
sp_256_mont_inv_order_avx2_4_ctx* ctx = (sp_256_mont_inv_order_avx2_4_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_avx2_4_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -22951,6 +22936,13 @@ static int sp_256_mont_inv_order_avx2_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, co
}
#endif /* WOLFSSL_SP_NONBLOCK */
/* Invert the number, in Montgomery form, modulo the order of the P256 curve.
* (r = 1 / a mod order)
*
* r Inverse result.
* a Number to invert.
* td Temporary data.
*/
static void sp_256_mont_inv_order_avx2_4(sp_digit* r, const sp_digit* a,
sp_digit* td)
{
@@ -23121,9 +23113,9 @@ int sp_ecc_sign_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_256_ecc_mulmod_4_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_256_ecc_mulmod_4_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p256_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -23404,6 +23396,10 @@ int sp_ecc_sign_256(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
extern void sp_256_mod_inv_4(sp_digit* r, const sp_digit* a, const sp_digit* m);
extern void sp_256_mod_inv_avx2_4(sp_digit* r, const sp_digit* a, const sp_digit* m);
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 256)
@@ -23529,7 +23525,7 @@ int sp_ecc_verify_256_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_256_proj_point_dbl_4_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_256_proj_point_dbl_4_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -23655,6 +23651,17 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
sp_256_from_mp(p2->y, 4, pY);
sp_256_from_mp(p2->z, 4, pZ);
#ifndef WOLFSSL_SP_SMALL
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_256_mod_inv_avx2_4(s, s, p256_order);
}
else
#endif
{
sp_256_mod_inv_4(s, s, p256_order);
}
#endif /* !WOLFSSL_SP_SMALL */
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_256_mul_avx2_4(s, s, p256_norm_order);
@@ -23668,6 +23675,7 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_256_norm_4(s);
#ifdef WOLFSSL_SP_SMALL
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_256_mont_inv_order_avx2_4(s, s, tmp);
@@ -23682,6 +23690,20 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, mp_int* pX,
sp_256_mont_mul_order_4(u2, u2, s);
}
#else
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_256_mont_mul_order_avx2_4(u1, u1, s);
sp_256_mont_mul_order_avx2_4(u2, u2, s);
}
else
#endif
{
sp_256_mont_mul_order_4(u1, u1, s);
sp_256_mont_mul_order_4(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags))
err = sp_256_ecc_mulmod_base_avx2_4(p1, u1, 0, 0, heap);
@@ -25425,7 +25447,7 @@ typedef struct sp_384_proj_point_add_6_ctx {
sp_digit* z;
} sp_384_proj_point_add_6_ctx;
static int sp_384_proj_point_add_6_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
static int sp_384_proj_point_add_6_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
const sp_point_384* p, const sp_point_384* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -26557,7 +26579,7 @@ typedef struct sp_384_proj_point_add_avx2_6_ctx {
sp_digit* z;
} sp_384_proj_point_add_avx2_6_ctx;
static int sp_384_proj_point_add_avx2_6_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
static int sp_384_proj_point_add_avx2_6_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
const sp_point_384* p, const sp_point_384* q, sp_digit* t)
{
int err = FP_WOULDBLOCK;
@@ -29796,7 +29818,7 @@ static int sp_384_mont_inv_order_6_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
{
int err = FP_WOULDBLOCK;
sp_384_mont_inv_order_6_ctx* ctx = (sp_384_mont_inv_order_6_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_384_mont_inv_order_6_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -29950,7 +29972,7 @@ static int sp_384_mont_inv_order_avx2_6_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, co
{
int err = FP_WOULDBLOCK;
sp_384_mont_inv_order_avx2_6_ctx* ctx = (sp_384_mont_inv_order_avx2_6_ctx*)sp_ctx;
typedef char ctx_size_test[sizeof(sp_384_mont_inv_order_avx2_6_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
(void)sizeof(ctx_size_test);
@@ -30121,9 +30143,9 @@ int sp_ecc_sign_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen, W
}
XMEMSET(&ctx->mulmod_ctx, 0, sizeof(ctx->mulmod_ctx));
ctx->state = 2;
break;
break;
case 2: /* MULMOD */
err = sp_384_ecc_mulmod_6_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
err = sp_384_ecc_mulmod_6_nb((sp_ecc_ctx_t*)&ctx->mulmod_ctx,
&ctx->point, &p384_base, ctx->k, 1, 1, heap);
if (err == MP_OKAY) {
ctx->state = 3;
@@ -30404,6 +30426,91 @@ int sp_ecc_sign_384(const byte* hash, word32 hashLen, WC_RNG* rng, mp_int* priv,
}
#endif /* HAVE_ECC_SIGN */
#ifndef WOLFSSL_SP_SMALL
extern void sp_384_rshift1_6(sp_digit* r, const sp_digit* a);
extern void sp_384_div2_mod_6(sp_digit* r, const sp_digit* a, const sp_digit* m);
extern int sp_384_num_bits_6(const sp_digit * a);
/* Non-constant time modular inversion.
*
* @param [out] r Resulting number.
* @param [in] a Number to invert.
* @param [in] m Modulus.
* @return MP_OKAY on success.
*/
static int sp_384_mod_inv_6(sp_digit* r, const sp_digit* a, const sp_digit* m)
{
sp_digit u[6];
sp_digit v[6];
sp_digit b[6];
sp_digit d[6];
int ut, vt;
sp_digit o;
XMEMCPY(u, m, sizeof(u));
XMEMCPY(v, a, sizeof(v));
ut = sp_384_num_bits_6(u);
vt = sp_384_num_bits_6(v);
XMEMSET(b, 0, sizeof(b));
if ((v[0] & 1) == 0) {
sp_384_rshift1_6(v, v);
XMEMCPY(d, m, sizeof(u));
d[0] += 1;
sp_384_rshift1_6(d, d);
vt--;
while ((v[0] & 1) == 0) {
sp_384_rshift1_6(v, v);
sp_384_div2_mod_6(d, d, m);
vt--;
}
}
else {
XMEMSET(d+1, 0, sizeof(d)-sizeof(sp_digit));
d[0] = 1;
}
while (ut > 1 && vt > 1) {
if (ut > vt || (ut == vt && sp_384_cmp_6(u, v) >= 0)) {
sp_384_sub_6(u, u, v);
o = sp_384_sub_6(b, b, d);
if (o != 0)
sp_384_add_6(b, b, m);
ut = sp_384_num_bits_6(u);
do {
sp_384_rshift1_6(u, u);
sp_384_div2_mod_6(b, b, m);
ut--;
}
while (ut > 0 && (u[0] & 1) == 0);
}
else {
sp_384_sub_6(v, v, u);
o = sp_384_sub_6(d, d, b);
if (o != 0)
sp_384_add_6(d, d, m);
vt = sp_384_num_bits_6(v);
do {
sp_384_rshift1_6(v, v);
sp_384_div2_mod_6(d, d, m);
vt--;
}
while (vt > 0 && (v[0] & 1) == 0);
}
}
if (ut == 1)
XMEMCPY(r, b, sizeof(b));
else
XMEMCPY(r, d, sizeof(d));
return MP_OKAY;
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_ECC_VERIFY
/* Verify the signature values with the hash and public key.
* e = Truncate(hash, 384)
@@ -30529,7 +30636,7 @@ int sp_ecc_verify_384_nb(sp_ecc_ctx_t* sp_ctx, const byte* hash, word32 hashLen,
ctx->state = 11;
break;
case 10: /* DBL */
err = sp_384_proj_point_dbl_6_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
err = sp_384_proj_point_dbl_6_nb((sp_ecc_ctx_t*)&ctx->dbl_ctx, &ctx->p1,
&ctx->p2, ctx->tmp);
if (err == MP_OKAY) {
ctx->state = 11;
@@ -30655,6 +30762,11 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
sp_384_from_mp(p2->y, 6, pY);
sp_384_from_mp(p2->z, 6, pZ);
#ifndef WOLFSSL_SP_SMALL
{
sp_384_mod_inv_6(s, s, p384_order);
}
#endif /* !WOLFSSL_SP_SMALL */
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_384_mul_avx2_6(s, s, p384_norm_order);
@@ -30668,6 +30780,7 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
}
if (err == MP_OKAY) {
sp_384_norm_6(s);
#ifdef WOLFSSL_SP_SMALL
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_384_mont_inv_order_avx2_6(s, s, tmp);
@@ -30682,6 +30795,20 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, mp_int* pX,
sp_384_mont_mul_order_6(u2, u2, s);
}
#else
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
sp_384_mont_mul_order_avx2_6(u1, u1, s);
sp_384_mont_mul_order_avx2_6(u2, u2, s);
}
else
#endif
{
sp_384_mont_mul_order_6(u1, u1, s);
sp_384_mont_mul_order_6(u2, u2, s);
}
#endif /* WOLFSSL_SP_SMALL */
#ifdef HAVE_INTEL_AVX2
if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags))
err = sp_384_ecc_mulmod_base_avx2_6(p1, u1, 0, 0, heap);
wolfcrypt/src/sp_x86_64_asm.S
+3717 -2504
View File
File diff suppressed because it is too large Load Diff