diff options
Diffstat (limited to 'lib/accelerated/x86/elf/ghash-x86_64.s')
-rw-r--r-- | lib/accelerated/x86/elf/ghash-x86_64.s | 80 |
1 files changed, 69 insertions, 11 deletions
diff --git a/lib/accelerated/x86/elf/ghash-x86_64.s b/lib/accelerated/x86/elf/ghash-x86_64.s index e2568a6fd6..1e4d18b341 100644 --- a/lib/accelerated/x86/elf/ghash-x86_64.s +++ b/lib/accelerated/x86/elf/ghash-x86_64.s @@ -1,4 +1,4 @@ -# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org> +# Copyright (c) 2011-2016, Andy Polyakov <appro@openssl.org> # All rights reserved. # # Redistribution and use in source and binary forms, with or without @@ -44,9 +44,27 @@ .type gcm_gmult_4bit,@function .align 16 gcm_gmult_4bit: +.cfi_startproc pushq %rbx +.cfi_adjust_cfa_offset 8 +.cfi_offset %rbx,-16 pushq %rbp +.cfi_adjust_cfa_offset 8 +.cfi_offset %rbp,-24 pushq %r12 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r12,-32 + pushq %r13 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r13,-40 + pushq %r14 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r14,-48 + pushq %r15 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r15,-56 + subq $280,%rsp +.cfi_adjust_cfa_offset 280 .Lgmult_prologue: movzbq 15(%rdi),%r8 @@ -123,22 +141,41 @@ gcm_gmult_4bit: movq %r8,8(%rdi) movq %r9,(%rdi) - movq 16(%rsp),%rbx - leaq 24(%rsp),%rsp + leaq 280+48(%rsp),%rsi +.cfi_def_cfa %rsi,8 + movq -8(%rsi),%rbx +.cfi_restore %rbx + leaq (%rsi),%rsp +.cfi_def_cfa_register %rsp .Lgmult_epilogue: .byte 0xf3,0xc3 +.cfi_endproc .size gcm_gmult_4bit,.-gcm_gmult_4bit .globl gcm_ghash_4bit .type gcm_ghash_4bit,@function .align 16 gcm_ghash_4bit: +.cfi_startproc pushq %rbx +.cfi_adjust_cfa_offset 8 +.cfi_offset %rbx,-16 pushq %rbp +.cfi_adjust_cfa_offset 8 +.cfi_offset %rbp,-24 pushq %r12 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r12,-32 pushq %r13 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r13,-40 pushq %r14 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r14,-48 pushq %r15 +.cfi_adjust_cfa_offset 8 +.cfi_offset %r15,-56 subq $280,%rsp +.cfi_adjust_cfa_offset 280 .Lghash_prologue: movq %rdx,%r14 movq %rcx,%r15 @@ -683,21 +720,31 @@ gcm_ghash_4bit: movq %r8,8(%rdi) movq %r9,(%rdi) - leaq 280(%rsp),%rsi - movq 0(%rsi),%r15 - movq 8(%rsi),%r14 - movq 16(%rsi),%r13 - movq 24(%rsi),%r12 - movq 32(%rsi),%rbp - movq 40(%rsi),%rbx - leaq 48(%rsi),%rsp + leaq 280+48(%rsp),%rsi +.cfi_def_cfa %rsi,8 + movq -48(%rsi),%r15 +.cfi_restore %r15 + movq -40(%rsi),%r14 +.cfi_restore %r14 + movq -32(%rsi),%r13 +.cfi_restore %r13 + movq -24(%rsi),%r12 +.cfi_restore %r12 + movq -16(%rsi),%rbp +.cfi_restore %rbp + movq -8(%rsi),%rbx +.cfi_restore %rbx + leaq 0(%rsi),%rsp +.cfi_def_cfa_register %rsp .Lghash_epilogue: .byte 0xf3,0xc3 +.cfi_endproc .size gcm_ghash_4bit,.-gcm_ghash_4bit .globl gcm_init_clmul .type gcm_init_clmul,@function .align 16 gcm_init_clmul: +.cfi_startproc .L_init_clmul: movdqu (%rsi),%xmm2 pshufd $78,%xmm2,%xmm2 @@ -849,11 +896,13 @@ gcm_init_clmul: .byte 102,15,58,15,227,8 movdqu %xmm4,80(%rdi) .byte 0xf3,0xc3 +.cfi_endproc .size gcm_init_clmul,.-gcm_init_clmul .globl gcm_gmult_clmul .type gcm_gmult_clmul,@function .align 16 gcm_gmult_clmul: +.cfi_startproc .L_gmult_clmul: movdqu (%rdi),%xmm0 movdqa .Lbswap_mask(%rip),%xmm5 @@ -900,11 +949,13 @@ gcm_gmult_clmul: .byte 102,15,56,0,197 movdqu %xmm0,(%rdi) .byte 0xf3,0xc3 +.cfi_endproc .size gcm_gmult_clmul,.-gcm_gmult_clmul .globl gcm_ghash_clmul .type gcm_ghash_clmul,@function .align 32 gcm_ghash_clmul: +.cfi_startproc .L_ghash_clmul: movdqa .Lbswap_mask(%rip),%xmm10 @@ -1283,11 +1334,13 @@ gcm_ghash_clmul: .byte 102,65,15,56,0,194 movdqu %xmm0,(%rdi) .byte 0xf3,0xc3 +.cfi_endproc .size gcm_ghash_clmul,.-gcm_ghash_clmul .globl gcm_init_avx .type gcm_init_avx,@function .align 32 gcm_init_avx: +.cfi_startproc vzeroupper vmovdqu (%rsi),%xmm2 @@ -1390,17 +1443,21 @@ gcm_init_avx: vzeroupper .byte 0xf3,0xc3 +.cfi_endproc .size gcm_init_avx,.-gcm_init_avx .globl gcm_gmult_avx .type gcm_gmult_avx,@function .align 32 gcm_gmult_avx: +.cfi_startproc jmp .L_gmult_clmul +.cfi_endproc .size gcm_gmult_avx,.-gcm_gmult_avx .globl gcm_ghash_avx .type gcm_ghash_avx,@function .align 32 gcm_ghash_avx: +.cfi_startproc vzeroupper vmovdqu (%rdi),%xmm10 @@ -1772,6 +1829,7 @@ gcm_ghash_avx: vmovdqu %xmm10,(%rdi) vzeroupper .byte 0xf3,0xc3 +.cfi_endproc .size gcm_ghash_avx,.-gcm_ghash_avx .align 64 .Lbswap_mask: |