diff options
Diffstat (limited to 'arch/x86/crypto/ghash-clmulni-intel_asm.S')
| -rw-r--r-- | arch/x86/crypto/ghash-clmulni-intel_asm.S | 36 |
1 files changed, 18 insertions, 18 deletions
diff --git a/arch/x86/crypto/ghash-clmulni-intel_asm.S b/arch/x86/crypto/ghash-clmulni-intel_asm.S index bb9735fbb865..c4fbaa82ed7a 100644 --- a/arch/x86/crypto/ghash-clmulni-intel_asm.S +++ b/arch/x86/crypto/ghash-clmulni-intel_asm.S @@ -4,7 +4,7 @@ * instructions. This file contains accelerated part of ghash * implementation. More information about PCLMULQDQ can be found at: * - * http://software.intel.com/en-us/articles/carry-less-multiplication-and-its-usage-for-computing-the-gcm-mode/ + * https://www.intel.com/content/dam/develop/external/us/en/documents/clmul-wp-rev-2-02-2014-04-20.pdf * * Copyright (c) 2009 Intel Corp. * Author: Huang Ying <ying.huang@intel.com> @@ -14,7 +14,6 @@ */ #include <linux/linkage.h> -#include <asm/inst.h> #include <asm/frame.h> .section .rodata.cst16.bswap_mask, "aM", @progbits, 16 @@ -51,9 +50,9 @@ SYM_FUNC_START_LOCAL(__clmul_gf128mul_ble) pxor DATA, T2 pxor SHASH, T3 - PCLMULQDQ 0x00 SHASH DATA # DATA = a0 * b0 - PCLMULQDQ 0x11 SHASH T1 # T1 = a1 * b1 - PCLMULQDQ 0x00 T3 T2 # T2 = (a1 + a0) * (b1 + b0) + pclmulqdq $0x00, SHASH, DATA # DATA = a0 * b0 + pclmulqdq $0x11, SHASH, T1 # T1 = a1 * b1 + pclmulqdq $0x00, T3, T2 # T2 = (a1 + a0) * (b1 + b0) pxor DATA, T2 pxor T1, T2 # T2 = a0 * b1 + a1 * b0 @@ -86,48 +85,49 @@ SYM_FUNC_START_LOCAL(__clmul_gf128mul_ble) psrlq $1, T2 pxor T2, T1 pxor T1, DATA - ret + RET SYM_FUNC_END(__clmul_gf128mul_ble) -/* void clmul_ghash_mul(char *dst, const u128 *shash) */ +/* void clmul_ghash_mul(char *dst, const le128 *shash) */ SYM_FUNC_START(clmul_ghash_mul) FRAME_BEGIN movups (%rdi), DATA movups (%rsi), SHASH - movaps .Lbswap_mask, BSWAP - PSHUFB_XMM BSWAP DATA + movaps .Lbswap_mask(%rip), BSWAP + pshufb BSWAP, DATA call __clmul_gf128mul_ble - PSHUFB_XMM BSWAP DATA + pshufb BSWAP, DATA movups DATA, (%rdi) FRAME_END - ret + RET SYM_FUNC_END(clmul_ghash_mul) /* - * void clmul_ghash_update(char *dst, const char *src, unsigned int srclen, - * const u128 *shash); + * int clmul_ghash_update(char *dst, const char *src, unsigned int srclen, + * const le128 *shash); */ SYM_FUNC_START(clmul_ghash_update) FRAME_BEGIN cmp $16, %rdx jb .Lupdate_just_ret # check length - movaps .Lbswap_mask, BSWAP + movaps .Lbswap_mask(%rip), BSWAP movups (%rdi), DATA movups (%rcx), SHASH - PSHUFB_XMM BSWAP DATA + pshufb BSWAP, DATA .align 4 .Lupdate_loop: movups (%rsi), IN1 - PSHUFB_XMM BSWAP IN1 + pshufb BSWAP, IN1 pxor IN1, DATA call __clmul_gf128mul_ble sub $16, %rdx add $16, %rsi cmp $16, %rdx jge .Lupdate_loop - PSHUFB_XMM BSWAP DATA + pshufb BSWAP, DATA movups DATA, (%rdi) .Lupdate_just_ret: + mov %rdx, %rax FRAME_END - ret + RET SYM_FUNC_END(clmul_ghash_update) |