diff options
Diffstat (limited to 'sysdeps/x86_64/dl-trampoline.S')
| -rw-r--r-- | sysdeps/x86_64/dl-trampoline.S | 189 |
1 files changed, 189 insertions, 0 deletions
diff --git a/sysdeps/x86_64/dl-trampoline.S b/sysdeps/x86_64/dl-trampoline.S new file mode 100644 index 0000000000..c1686dae10 --- /dev/null +++ b/sysdeps/x86_64/dl-trampoline.S @@ -0,0 +1,189 @@ +/* PLT trampolines. x86-64 version. + Copyright (C) 2004, 2005 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#include <sysdep.h> + + .text + .globl _dl_runtime_resolve + .type _dl_runtime_resolve, @function + .align 16 + cfi_startproc +_dl_runtime_resolve: + subq $56,%rsp + cfi_adjust_cfa_offset(72) # Incorporate PLT + movq %rax,(%rsp) # Preserve registers otherwise clobbered. + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %rsi, 24(%rsp) + movq %rdi, 32(%rsp) + movq %r8, 40(%rsp) + movq %r9, 48(%rsp) + movq 64(%rsp), %rsi # Copy args pushed by PLT in register. + movq %rsi, %r11 # Multiply by 24 + addq %r11, %rsi + addq %r11, %rsi + shlq $3, %rsi + movq 56(%rsp), %rdi # %rdi: link_map, %rsi: reloc_offset + call _dl_fixup # Call resolver. + movq %rax, %r11 # Save return value + movq 48(%rsp), %r9 # Get register content back. + movq 40(%rsp), %r8 + movq 32(%rsp), %rdi + movq 24(%rsp), %rsi + movq 16(%rsp), %rdx + movq 8(%rsp), %rcx + movq (%rsp), %rax + addq $72, %rsp # Adjust stack(PLT did 2 pushes) + cfi_adjust_cfa_offset(-72) + jmp *%r11 # Jump to function address. + cfi_endproc + .size _dl_runtime_resolve, .-_dl_runtime_resolve + + +#ifndef PROF + .globl _dl_runtime_profile + .type _dl_runtime_profile, @function + .align 16 + cfi_startproc +_dl_runtime_profile: + subq $80, %rsp + cfi_adjust_cfa_offset(96) # Incorporate PLT + movq %rax, (%rsp) # Preserve registers otherwise clobbered. + movq %rdx, 8(%rsp) + movq %r8, 16(%rsp) + movq %r9, 24(%rsp) + movq %rcx, 32(%rsp) + movq %rsi, 40(%rsp) + movq %rdi, 48(%rsp) + movq %rbp, 56(%rsp) # Information for auditors. + leaq 96(%rsp), %rax + movq %rax, 64(%rsp) + leaq 8(%rsp), %rcx + movq 96(%rsp), %rdx # Load return address if needed + movq 88(%rsp), %rsi # Copy args pushed by PLT in register. + movq %rsi,%r11 # Multiply by 24 + addq %r11,%rsi + addq %r11,%rsi + shlq $3, %rsi + movq 80(%rsp), %rdi # %rdi: link_map, %rsi: reloc_offset + leaq 72(%rsp), %r8 + call _dl_profile_fixup # Call resolver. + movq %rax, %r11 # Save return value + movq 8(%rsp), %rdx # Get back register content. + movq 16(%rsp), %r8 + movq 24(%rsp), %r9 + movq (%rsp),%rax + movq 72(%rsp), %r10 + testq %r10, %r10 + jns 1f + movq 32(%rsp), %rcx + movq 40(%rsp), %rsi + movq 48(%rsp), %rdi + addq $96,%rsp # Adjust stack + cfi_adjust_cfa_offset (-96) + jmp *%r11 # Jump to function address. + + /* + +96 return address + +88 PLT2 + +80 PLT1 + +72 free + +64 %rsp + +56 %rbp + +48 %rdi + +40 %rsi + +32 %rcx + +24 %r9 + +16 %r8 + +8 %rdx + %esp %rax + */ + cfi_adjust_cfa_offset (96) +1: movq %rbx, 72(%rsp) + cfi_rel_offset (1, 72) + leaq 104(%rsp), %rsi + movq %rsp, %rbx + cfi_def_cfa_register (1) + subq %r10, %rsp + movq %rsp, %rdi + movq %r10, %rcx + shrq $3, %rcx + rep + movsq + andq $0xfffffffffffffff0, %rsp + movq 32(%rbx), %rcx + movq 40(%rbx), %rsi + movq 48(%rbx), %rdi + call *%r11 + movq %rbx, %rsp + cfi_def_cfa_register (7) + subq $72, %rsp + cfi_adjust_cfa_offset (72) + movq %rsp, %rcx + movq %rax, (%rcx) + movq %rdx, 8(%rcx) + /* Even though the stack is correctly aligned to allow using movaps + we use movups. Some callers might provide an incorrectly aligned + stack and we do not want to have it blow up here. */ + movups %xmm0, 16(%rcx) + movups %xmm1, 32(%rcx) + fstpt 48(%rcx) + fstpt 64(%rcx) + /* + +168 return address + +160 PLT2 + +152 PLT1 + +144 free + +136 %rsp + +128 %rbp + +120 %rdi + +112 %rsi + +104 %rcx + +96 %r9 + +88 %r8 + +80 %rdx + +64 %st1 result + +48 %st result + +32 %xmm1 result + +16 %xmm0 result + +8 %rdx result + %esp %rax result + */ + leaq 80(%rsp), %rdx + movq 144(%rsp), %rbx + cfi_restore (1) + movq 160(%rsp), %rsi # Copy args pushed by PLT in register. + movq %rsi,%r11 # Multiply by 24 + addq %r11,%rsi + addq %r11,%rsi + shlq $3, %rsi + movq 152(%rsp), %rdi # %rdi: link_map, %rsi: reloc_offset + call _dl_call_pltexit + movq (%rsp), %rax + movq 8(%rsp), %rdx + movups 16(%rsp), %xmm0 + movups 32(%rsp), %xmm1 + fldt 64(%rsp) + fldt 48(%rsp) + addq $168, %rsp + cfi_adjust_cfa_offset (-168) + retq + cfi_endproc + .size _dl_runtime_profile, .-_dl_runtime_profile +#endif |