2005-01-06 22:40:27 +00:00
|
|
|
/* PLT trampolines. x86-64 version.
|
2015-01-02 16:28:19 +00:00
|
|
|
Copyright (C) 2004-2015 Free Software Foundation, Inc.
|
2005-01-06 22:40:27 +00:00
|
|
|
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
|
2012-02-09 23:18:22 +00:00
|
|
|
License along with the GNU C Library; if not, see
|
|
|
|
|
<http://www.gnu.org/licenses/>. */
|
2005-01-06 22:40:27 +00:00
|
|
|
|
2009-07-10 19:04:14 +00:00
|
|
|
#include <config.h>
|
2005-01-06 22:40:27 +00:00
|
|
|
#include <sysdep.h>
|
2009-07-10 19:04:14 +00:00
|
|
|
#include <link-defines.h>
|
2005-01-06 22:40:27 +00:00
|
|
|
|
2012-05-11 18:50:11 +00:00
|
|
|
#if (RTLD_SAVESPACE_SSE % 32) != 0
|
|
|
|
|
# error RTLD_SAVESPACE_SSE must be aligned to 32 bytes
|
|
|
|
|
#endif
|
|
|
|
|
|
2014-04-01 17:16:04 +00:00
|
|
|
/* Area on stack to save and restore registers used for parameter
|
|
|
|
|
passing when calling _dl_fixup. */
|
|
|
|
|
#ifdef __ILP32__
|
|
|
|
|
/* X32 saves RCX, RDX, RSI, RDI, R8 and R9 plus RAX. */
|
|
|
|
|
# define REGISTER_SAVE_AREA (8 * 7)
|
|
|
|
|
# define REGISTER_SAVE_RAX 0
|
2015-03-16 21:58:43 +00:00
|
|
|
# define PRESERVE_BND_REGS_PREFIX
|
2014-04-01 17:16:04 +00:00
|
|
|
#else
|
|
|
|
|
/* X86-64 saves RCX, RDX, RSI, RDI, R8 and R9 plus RAX as well as BND0,
|
|
|
|
|
BND1, BND2, BND3. */
|
|
|
|
|
# define REGISTER_SAVE_AREA (8 * 7 + 16 * 4)
|
|
|
|
|
/* Align bound register save area to 16 bytes. */
|
|
|
|
|
# define REGISTER_SAVE_BND0 0
|
|
|
|
|
# define REGISTER_SAVE_BND1 (REGISTER_SAVE_BND0 + 16)
|
|
|
|
|
# define REGISTER_SAVE_BND2 (REGISTER_SAVE_BND1 + 16)
|
|
|
|
|
# define REGISTER_SAVE_BND3 (REGISTER_SAVE_BND2 + 16)
|
|
|
|
|
# define REGISTER_SAVE_RAX (REGISTER_SAVE_BND3 + 16)
|
2015-03-16 21:58:43 +00:00
|
|
|
# ifdef HAVE_MPX_SUPPORT
|
|
|
|
|
# define PRESERVE_BND_REGS_PREFIX bnd
|
|
|
|
|
# else
|
|
|
|
|
# define PRESERVE_BND_REGS_PREFIX .byte 0xf2
|
|
|
|
|
# endif
|
2014-04-01 17:16:04 +00:00
|
|
|
#endif
|
|
|
|
|
#define REGISTER_SAVE_RCX (REGISTER_SAVE_RAX + 8)
|
|
|
|
|
#define REGISTER_SAVE_RDX (REGISTER_SAVE_RCX + 8)
|
|
|
|
|
#define REGISTER_SAVE_RSI (REGISTER_SAVE_RDX + 8)
|
|
|
|
|
#define REGISTER_SAVE_RDI (REGISTER_SAVE_RSI + 8)
|
|
|
|
|
#define REGISTER_SAVE_R8 (REGISTER_SAVE_RDI + 8)
|
|
|
|
|
#define REGISTER_SAVE_R9 (REGISTER_SAVE_R8 + 8)
|
|
|
|
|
|
2005-01-06 22:40:27 +00:00
|
|
|
.text
|
|
|
|
|
.globl _dl_runtime_resolve
|
|
|
|
|
.type _dl_runtime_resolve, @function
|
|
|
|
|
.align 16
|
|
|
|
|
cfi_startproc
|
|
|
|
|
_dl_runtime_resolve:
|
2011-08-20 12:58:44 +00:00
|
|
|
cfi_adjust_cfa_offset(16) # Incorporate PLT
|
2014-04-01 17:16:04 +00:00
|
|
|
subq $REGISTER_SAVE_AREA,%rsp
|
|
|
|
|
cfi_adjust_cfa_offset(REGISTER_SAVE_AREA)
|
|
|
|
|
# Preserve registers otherwise clobbered.
|
|
|
|
|
movq %rax, REGISTER_SAVE_RAX(%rsp)
|
|
|
|
|
movq %rcx, REGISTER_SAVE_RCX(%rsp)
|
|
|
|
|
movq %rdx, REGISTER_SAVE_RDX(%rsp)
|
|
|
|
|
movq %rsi, REGISTER_SAVE_RSI(%rsp)
|
|
|
|
|
movq %rdi, REGISTER_SAVE_RDI(%rsp)
|
|
|
|
|
movq %r8, REGISTER_SAVE_R8(%rsp)
|
|
|
|
|
movq %r9, REGISTER_SAVE_R9(%rsp)
|
|
|
|
|
#ifndef __ILP32__
|
|
|
|
|
# We also have to preserve bound registers. These are nops if
|
|
|
|
|
# Intel MPX isn't available or disabled.
|
|
|
|
|
# ifdef HAVE_MPX_SUPPORT
|
|
|
|
|
bndmov %bnd0, REGISTER_SAVE_BND0(%rsp)
|
|
|
|
|
bndmov %bnd1, REGISTER_SAVE_BND1(%rsp)
|
|
|
|
|
bndmov %bnd2, REGISTER_SAVE_BND2(%rsp)
|
|
|
|
|
bndmov %bnd3, REGISTER_SAVE_BND3(%rsp)
|
|
|
|
|
# else
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x44,0x24,REGISTER_SAVE_BND0
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x4c,0x24,REGISTER_SAVE_BND1
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x54,0x24,REGISTER_SAVE_BND2
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x5c,0x24,REGISTER_SAVE_BND3
|
|
|
|
|
# endif
|
|
|
|
|
#endif
|
|
|
|
|
# Copy args pushed by PLT in register.
|
|
|
|
|
# %rdi: link_map, %rsi: reloc_index
|
|
|
|
|
movq (REGISTER_SAVE_AREA + 8)(%rsp), %rsi
|
|
|
|
|
movq REGISTER_SAVE_AREA(%rsp), %rdi
|
2005-01-06 22:40:27 +00:00
|
|
|
call _dl_fixup # Call resolver.
|
|
|
|
|
movq %rax, %r11 # Save return value
|
2014-04-01 17:16:04 +00:00
|
|
|
#ifndef __ILP32__
|
|
|
|
|
# Restore bound registers. These are nops if Intel MPX isn't
|
|
|
|
|
# avaiable or disabled.
|
|
|
|
|
# ifdef HAVE_MPX_SUPPORT
|
|
|
|
|
bndmov REGISTER_SAVE_BND3(%rsp), %bnd3
|
|
|
|
|
bndmov REGISTER_SAVE_BND2(%rsp), %bnd2
|
|
|
|
|
bndmov REGISTER_SAVE_BND1(%rsp), %bnd1
|
|
|
|
|
bndmov REGISTER_SAVE_BND0(%rsp), %bnd0
|
|
|
|
|
# else
|
|
|
|
|
.byte 0x66,0x0f,0x1a,0x5c,0x24,REGISTER_SAVE_BND3
|
|
|
|
|
.byte 0x66,0x0f,0x1a,0x54,0x24,REGISTER_SAVE_BND2
|
|
|
|
|
.byte 0x66,0x0f,0x1a,0x4c,0x24,REGISTER_SAVE_BND1
|
|
|
|
|
.byte 0x66,0x0f,0x1a,0x44,0x24,REGISTER_SAVE_BND0
|
|
|
|
|
# endif
|
|
|
|
|
#endif
|
|
|
|
|
# Get register content back.
|
|
|
|
|
movq REGISTER_SAVE_R9(%rsp), %r9
|
|
|
|
|
movq REGISTER_SAVE_R8(%rsp), %r8
|
|
|
|
|
movq REGISTER_SAVE_RDI(%rsp), %rdi
|
|
|
|
|
movq REGISTER_SAVE_RSI(%rsp), %rsi
|
|
|
|
|
movq REGISTER_SAVE_RDX(%rsp), %rdx
|
|
|
|
|
movq REGISTER_SAVE_RCX(%rsp), %rcx
|
|
|
|
|
movq REGISTER_SAVE_RAX(%rsp), %rax
|
|
|
|
|
# Adjust stack(PLT did 2 pushes)
|
|
|
|
|
addq $(REGISTER_SAVE_AREA + 16), %rsp
|
|
|
|
|
cfi_adjust_cfa_offset(-(REGISTER_SAVE_AREA + 16))
|
2015-03-16 21:58:43 +00:00
|
|
|
# Preserve bound registers.
|
|
|
|
|
PRESERVE_BND_REGS_PREFIX
|
2005-01-06 22:40:27 +00:00
|
|
|
jmp *%r11 # Jump to function address.
|
|
|
|
|
cfi_endproc
|
|
|
|
|
.size _dl_runtime_resolve, .-_dl_runtime_resolve
|
|
|
|
|
|
|
|
|
|
|
2005-07-07 02:39:45 +00:00
|
|
|
#ifndef PROF
|
2005-01-06 22:40:27 +00:00
|
|
|
.globl _dl_runtime_profile
|
|
|
|
|
.type _dl_runtime_profile, @function
|
|
|
|
|
.align 16
|
|
|
|
|
cfi_startproc
|
2009-03-14 23:57:33 +00:00
|
|
|
|
2005-01-06 22:40:27 +00:00
|
|
|
_dl_runtime_profile:
|
2009-07-29 15:50:03 +00:00
|
|
|
cfi_adjust_cfa_offset(16) # Incorporate PLT
|
2009-03-14 23:57:33 +00:00
|
|
|
/* The La_x86_64_regs data structure pointed to by the
|
|
|
|
|
fourth paramater must be 16-byte aligned. This must
|
|
|
|
|
be explicitly enforced. We have the set up a dynamically
|
|
|
|
|
sized stack frame. %rbx points to the top half which
|
|
|
|
|
has a fixed size and preserves the original stack pointer. */
|
|
|
|
|
|
|
|
|
|
subq $32, %rsp # Allocate the local storage.
|
2009-07-29 15:50:03 +00:00
|
|
|
cfi_adjust_cfa_offset(32)
|
2009-03-14 23:57:33 +00:00
|
|
|
movq %rbx, (%rsp)
|
|
|
|
|
cfi_rel_offset(%rbx, 0)
|
|
|
|
|
|
|
|
|
|
/* On the stack:
|
|
|
|
|
56(%rbx) parameter #1
|
|
|
|
|
48(%rbx) return address
|
|
|
|
|
|
|
|
|
|
40(%rbx) reloc index
|
|
|
|
|
32(%rbx) link_map
|
|
|
|
|
|
|
|
|
|
24(%rbx) La_x86_64_regs pointer
|
|
|
|
|
16(%rbx) framesize
|
|
|
|
|
8(%rbx) rax
|
|
|
|
|
(%rbx) rbx
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
movq %rax, 8(%rsp)
|
|
|
|
|
movq %rsp, %rbx
|
|
|
|
|
cfi_def_cfa_register(%rbx)
|
|
|
|
|
|
|
|
|
|
/* Actively align the La_x86_64_regs structure. */
|
|
|
|
|
andq $0xfffffffffffffff0, %rsp
|
2014-03-13 18:10:22 +00:00
|
|
|
# if defined HAVE_AVX_SUPPORT || defined HAVE_AVX512_ASM_SUPPORT
|
2009-07-10 19:04:14 +00:00
|
|
|
/* sizeof(La_x86_64_regs). Need extra space for 8 SSE registers
|
|
|
|
|
to detect if any xmm0-xmm7 registers are changed by audit
|
|
|
|
|
module. */
|
|
|
|
|
subq $(LR_SIZE + XMM_SIZE*8), %rsp
|
2009-07-16 00:41:36 +00:00
|
|
|
# else
|
2009-07-10 19:04:14 +00:00
|
|
|
subq $LR_SIZE, %rsp # sizeof(La_x86_64_regs)
|
2009-07-16 00:41:36 +00:00
|
|
|
# endif
|
2009-03-14 23:57:33 +00:00
|
|
|
movq %rsp, 24(%rbx)
|
|
|
|
|
|
2009-07-10 19:04:14 +00:00
|
|
|
/* Fill the La_x86_64_regs structure. */
|
|
|
|
|
movq %rdx, LR_RDX_OFFSET(%rsp)
|
|
|
|
|
movq %r8, LR_R8_OFFSET(%rsp)
|
|
|
|
|
movq %r9, LR_R9_OFFSET(%rsp)
|
|
|
|
|
movq %rcx, LR_RCX_OFFSET(%rsp)
|
|
|
|
|
movq %rsi, LR_RSI_OFFSET(%rsp)
|
|
|
|
|
movq %rdi, LR_RDI_OFFSET(%rsp)
|
|
|
|
|
movq %rbp, LR_RBP_OFFSET(%rsp)
|
2005-01-06 22:40:27 +00:00
|
|
|
|
2009-07-16 00:41:36 +00:00
|
|
|
leaq 48(%rbx), %rax
|
|
|
|
|
movq %rax, LR_RSP_OFFSET(%rsp)
|
|
|
|
|
|
|
|
|
|
/* We always store the XMM registers even if AVX is available.
|
2013-08-30 12:32:07 +00:00
|
|
|
This is to provide backward binary compatibility for existing
|
2009-07-16 00:41:36 +00:00
|
|
|
audit modules. */
|
|
|
|
|
movaps %xmm0, (LR_XMM_OFFSET)(%rsp)
|
|
|
|
|
movaps %xmm1, (LR_XMM_OFFSET + XMM_SIZE)(%rsp)
|
|
|
|
|
movaps %xmm2, (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp)
|
|
|
|
|
movaps %xmm3, (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp)
|
|
|
|
|
movaps %xmm4, (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp)
|
|
|
|
|
movaps %xmm5, (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp)
|
|
|
|
|
movaps %xmm6, (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp)
|
|
|
|
|
movaps %xmm7, (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp)
|
|
|
|
|
|
2014-04-16 21:43:16 +00:00
|
|
|
# ifndef __ILP32__
|
|
|
|
|
# ifdef HAVE_MPX_SUPPORT
|
|
|
|
|
bndmov %bnd0, (LR_BND_OFFSET)(%rsp) # Preserve bound
|
|
|
|
|
bndmov %bnd1, (LR_BND_OFFSET + BND_SIZE)(%rsp) # registers. Nops if
|
|
|
|
|
bndmov %bnd2, (LR_BND_OFFSET + BND_SIZE*2)(%rsp) # MPX not available
|
|
|
|
|
bndmov %bnd3, (LR_BND_OFFSET + BND_SIZE*3)(%rsp) # or disabled.
|
|
|
|
|
# else
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x84,0x24;.long (LR_BND_OFFSET)
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x8c,0x24;.long (LR_BND_OFFSET + BND_SIZE)
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x84,0x24;.long (LR_BND_OFFSET + BND_SIZE*2)
|
|
|
|
|
.byte 0x66,0x0f,0x1b,0x8c,0x24;.long (LR_BND_OFFSET + BND_SIZE*3)
|
|
|
|
|
# endif
|
|
|
|
|
# endif
|
|
|
|
|
|
2014-03-13 18:10:22 +00:00
|
|
|
# if defined HAVE_AVX_SUPPORT || defined HAVE_AVX512_ASM_SUPPORT
|
2009-07-16 00:41:36 +00:00
|
|
|
.data
|
|
|
|
|
L(have_avx):
|
|
|
|
|
.zero 4
|
|
|
|
|
.size L(have_avx), 4
|
|
|
|
|
.previous
|
2009-03-14 23:57:33 +00:00
|
|
|
|
2009-07-16 00:41:36 +00:00
|
|
|
cmpl $0, L(have_avx)(%rip)
|
2014-03-13 18:10:22 +00:00
|
|
|
jne L(defined)
|
2009-07-16 00:41:36 +00:00
|
|
|
movq %rbx, %r11 # Save rbx
|
|
|
|
|
movl $1, %eax
|
|
|
|
|
cpuid
|
|
|
|
|
movq %r11,%rbx # Restore rbx
|
2011-07-21 01:21:03 +00:00
|
|
|
xorl %eax, %eax
|
|
|
|
|
// AVX and XSAVE supported?
|
2011-07-23 03:33:22 +00:00
|
|
|
andl $((1 << 28) | (1 << 27)), %ecx
|
|
|
|
|
cmpl $((1 << 28) | (1 << 27)), %ecx
|
2014-03-13 18:10:22 +00:00
|
|
|
jne 10f
|
|
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
// AVX512 supported in processor?
|
|
|
|
|
movq %rbx, %r11 # Save rbx
|
|
|
|
|
xorl %ecx, %ecx
|
|
|
|
|
mov $0x7, %eax
|
|
|
|
|
cpuid
|
|
|
|
|
andl $(1 << 16), %ebx
|
|
|
|
|
# endif
|
2011-07-21 01:21:03 +00:00
|
|
|
xorl %ecx, %ecx
|
|
|
|
|
// Get XFEATURE_ENABLED_MASK
|
|
|
|
|
xgetbv
|
2014-03-13 18:10:22 +00:00
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
test %ebx, %ebx
|
|
|
|
|
movq %r11, %rbx # Restore rbx
|
|
|
|
|
je 20f
|
|
|
|
|
// Verify that XCR0[7:5] = '111b' and
|
|
|
|
|
// XCR0[2:1] = '11b' which means
|
|
|
|
|
// that zmm state is enabled
|
|
|
|
|
andl $0xe6, %eax
|
|
|
|
|
cmpl $0xe6, %eax
|
|
|
|
|
jne 20f
|
|
|
|
|
movl %eax, L(have_avx)(%rip)
|
|
|
|
|
L(avx512):
|
|
|
|
|
# define RESTORE_AVX
|
|
|
|
|
# define VMOV vmovdqu64
|
|
|
|
|
# define VEC(i) zmm##i
|
|
|
|
|
# define MORE_CODE
|
|
|
|
|
# include "dl-trampoline.h"
|
|
|
|
|
# undef VMOV
|
|
|
|
|
# undef VEC
|
|
|
|
|
# undef RESTORE_AVX
|
|
|
|
|
# endif
|
|
|
|
|
20: andl $0x6, %eax
|
|
|
|
|
10: subl $0x5, %eax
|
2011-07-22 18:33:47 +00:00
|
|
|
movl %eax, L(have_avx)(%rip)
|
2009-07-16 00:41:36 +00:00
|
|
|
cmpl $0, %eax
|
|
|
|
|
|
2014-03-13 18:10:22 +00:00
|
|
|
L(defined):
|
|
|
|
|
js L(no_avx)
|
|
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
cmpl $0xe6, L(have_avx)(%rip)
|
|
|
|
|
je L(avx512)
|
|
|
|
|
# endif
|
2009-07-16 00:41:36 +00:00
|
|
|
|
2009-08-08 17:54:42 +00:00
|
|
|
# define RESTORE_AVX
|
2014-03-13 18:10:22 +00:00
|
|
|
# define VMOV vmovdqu
|
|
|
|
|
# define VEC(i) ymm##i
|
2011-08-20 12:56:30 +00:00
|
|
|
# define MORE_CODE
|
2009-08-08 17:54:42 +00:00
|
|
|
# include "dl-trampoline.h"
|
2009-07-16 00:41:36 +00:00
|
|
|
|
2009-08-08 17:54:42 +00:00
|
|
|
.align 16
|
|
|
|
|
L(no_avx):
|
2009-07-16 00:59:14 +00:00
|
|
|
# endif
|
2009-07-16 00:41:36 +00:00
|
|
|
|
2011-08-20 12:56:30 +00:00
|
|
|
# undef RESTORE_AVX
|
|
|
|
|
# include "dl-trampoline.h"
|
2009-07-16 00:41:36 +00:00
|
|
|
|
|
|
|
|
cfi_endproc
|
|
|
|
|
.size _dl_runtime_profile, .-_dl_runtime_profile
|
2005-07-07 02:39:45 +00:00
|
|
|
#endif
|
2009-07-29 15:33:03 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
#ifdef SHARED
|
|
|
|
|
.globl _dl_x86_64_save_sse
|
|
|
|
|
.type _dl_x86_64_save_sse, @function
|
|
|
|
|
.align 16
|
|
|
|
|
cfi_startproc
|
|
|
|
|
_dl_x86_64_save_sse:
|
2014-03-13 18:10:22 +00:00
|
|
|
# if defined HAVE_AVX_SUPPORT || defined HAVE_AVX512_ASM_SUPPORT
|
2009-07-29 15:33:03 +00:00
|
|
|
cmpl $0, L(have_avx)(%rip)
|
2014-03-13 18:10:22 +00:00
|
|
|
jne L(defined_5)
|
2009-07-29 15:33:03 +00:00
|
|
|
movq %rbx, %r11 # Save rbx
|
|
|
|
|
movl $1, %eax
|
|
|
|
|
cpuid
|
|
|
|
|
movq %r11,%rbx # Restore rbx
|
2011-07-22 18:33:47 +00:00
|
|
|
xorl %eax, %eax
|
|
|
|
|
// AVX and XSAVE supported?
|
2011-07-23 03:33:22 +00:00
|
|
|
andl $((1 << 28) | (1 << 27)), %ecx
|
|
|
|
|
cmpl $((1 << 28) | (1 << 27)), %ecx
|
2014-03-13 18:10:22 +00:00
|
|
|
jne 1f
|
|
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
// AVX512 supported in a processor?
|
|
|
|
|
movq %rbx, %r11 # Save rbx
|
|
|
|
|
xorl %ecx,%ecx
|
|
|
|
|
mov $0x7,%eax
|
|
|
|
|
cpuid
|
|
|
|
|
andl $(1 << 16), %ebx
|
|
|
|
|
# endif
|
2011-07-22 18:33:47 +00:00
|
|
|
xorl %ecx, %ecx
|
|
|
|
|
// Get XFEATURE_ENABLED_MASK
|
|
|
|
|
xgetbv
|
2014-03-13 18:10:22 +00:00
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
test %ebx, %ebx
|
|
|
|
|
movq %r11, %rbx # Restore rbx
|
|
|
|
|
je 2f
|
|
|
|
|
// Verify that XCR0[7:5] = '111b' and
|
|
|
|
|
// XCR0[2:1] = '11b' which means
|
|
|
|
|
// that zmm state is enabled
|
|
|
|
|
andl $0xe6, %eax
|
|
|
|
|
movl %eax, L(have_avx)(%rip)
|
|
|
|
|
cmpl $0xe6, %eax
|
|
|
|
|
je L(avx512_5)
|
|
|
|
|
# endif
|
|
|
|
|
|
|
|
|
|
2: andl $0x6, %eax
|
|
|
|
|
1: subl $0x5, %eax
|
2011-07-22 18:33:47 +00:00
|
|
|
movl %eax, L(have_avx)(%rip)
|
2009-07-29 15:33:03 +00:00
|
|
|
cmpl $0, %eax
|
|
|
|
|
|
2014-03-13 18:10:22 +00:00
|
|
|
L(defined_5):
|
|
|
|
|
js L(no_avx5)
|
|
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
cmpl $0xe6, L(have_avx)(%rip)
|
|
|
|
|
je L(avx512_5)
|
|
|
|
|
# endif
|
2009-07-29 15:33:03 +00:00
|
|
|
|
|
|
|
|
vmovdqa %ymm0, %fs:RTLD_SAVESPACE_SSE+0*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm1, %fs:RTLD_SAVESPACE_SSE+1*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm2, %fs:RTLD_SAVESPACE_SSE+2*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm3, %fs:RTLD_SAVESPACE_SSE+3*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm4, %fs:RTLD_SAVESPACE_SSE+4*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm5, %fs:RTLD_SAVESPACE_SSE+5*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm6, %fs:RTLD_SAVESPACE_SSE+6*YMM_SIZE
|
|
|
|
|
vmovdqa %ymm7, %fs:RTLD_SAVESPACE_SSE+7*YMM_SIZE
|
|
|
|
|
ret
|
2014-03-13 18:10:22 +00:00
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
L(avx512_5):
|
|
|
|
|
vmovdqu64 %zmm0, %fs:RTLD_SAVESPACE_SSE+0*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm1, %fs:RTLD_SAVESPACE_SSE+1*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm2, %fs:RTLD_SAVESPACE_SSE+2*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm3, %fs:RTLD_SAVESPACE_SSE+3*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm4, %fs:RTLD_SAVESPACE_SSE+4*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm5, %fs:RTLD_SAVESPACE_SSE+5*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm6, %fs:RTLD_SAVESPACE_SSE+6*ZMM_SIZE
|
|
|
|
|
vmovdqu64 %zmm7, %fs:RTLD_SAVESPACE_SSE+7*ZMM_SIZE
|
|
|
|
|
ret
|
|
|
|
|
# endif
|
2009-07-29 15:33:03 +00:00
|
|
|
L(no_avx5):
|
|
|
|
|
# endif
|
|
|
|
|
movdqa %xmm0, %fs:RTLD_SAVESPACE_SSE+0*XMM_SIZE
|
|
|
|
|
movdqa %xmm1, %fs:RTLD_SAVESPACE_SSE+1*XMM_SIZE
|
|
|
|
|
movdqa %xmm2, %fs:RTLD_SAVESPACE_SSE+2*XMM_SIZE
|
|
|
|
|
movdqa %xmm3, %fs:RTLD_SAVESPACE_SSE+3*XMM_SIZE
|
|
|
|
|
movdqa %xmm4, %fs:RTLD_SAVESPACE_SSE+4*XMM_SIZE
|
|
|
|
|
movdqa %xmm5, %fs:RTLD_SAVESPACE_SSE+5*XMM_SIZE
|
|
|
|
|
movdqa %xmm6, %fs:RTLD_SAVESPACE_SSE+6*XMM_SIZE
|
|
|
|
|
movdqa %xmm7, %fs:RTLD_SAVESPACE_SSE+7*XMM_SIZE
|
|
|
|
|
ret
|
|
|
|
|
cfi_endproc
|
|
|
|
|
.size _dl_x86_64_save_sse, .-_dl_x86_64_save_sse
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
.globl _dl_x86_64_restore_sse
|
|
|
|
|
.type _dl_x86_64_restore_sse, @function
|
|
|
|
|
.align 16
|
|
|
|
|
cfi_startproc
|
|
|
|
|
_dl_x86_64_restore_sse:
|
2014-03-13 18:10:22 +00:00
|
|
|
# if defined HAVE_AVX_SUPPORT || defined HAVE_AVX512_ASM_SUPPORT
|
2009-07-29 15:33:03 +00:00
|
|
|
cmpl $0, L(have_avx)(%rip)
|
|
|
|
|
js L(no_avx6)
|
2014-03-13 18:10:22 +00:00
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
cmpl $0xe6, L(have_avx)(%rip)
|
|
|
|
|
je L(avx512_6)
|
|
|
|
|
# endif
|
2009-07-29 15:33:03 +00:00
|
|
|
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+0*YMM_SIZE, %ymm0
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+1*YMM_SIZE, %ymm1
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+2*YMM_SIZE, %ymm2
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+3*YMM_SIZE, %ymm3
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+4*YMM_SIZE, %ymm4
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+5*YMM_SIZE, %ymm5
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+6*YMM_SIZE, %ymm6
|
|
|
|
|
vmovdqa %fs:RTLD_SAVESPACE_SSE+7*YMM_SIZE, %ymm7
|
|
|
|
|
ret
|
2014-03-13 18:10:22 +00:00
|
|
|
# ifdef HAVE_AVX512_ASM_SUPPORT
|
|
|
|
|
L(avx512_6):
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+0*ZMM_SIZE, %zmm0
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+1*ZMM_SIZE, %zmm1
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+2*ZMM_SIZE, %zmm2
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+3*ZMM_SIZE, %zmm3
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+4*ZMM_SIZE, %zmm4
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+5*ZMM_SIZE, %zmm5
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+6*ZMM_SIZE, %zmm6
|
|
|
|
|
vmovdqu64 %fs:RTLD_SAVESPACE_SSE+7*ZMM_SIZE, %zmm7
|
|
|
|
|
ret
|
|
|
|
|
# endif
|
2009-07-29 15:33:03 +00:00
|
|
|
L(no_avx6):
|
|
|
|
|
# endif
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+0*XMM_SIZE, %xmm0
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+1*XMM_SIZE, %xmm1
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+2*XMM_SIZE, %xmm2
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+3*XMM_SIZE, %xmm3
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+4*XMM_SIZE, %xmm4
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+5*XMM_SIZE, %xmm5
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+6*XMM_SIZE, %xmm6
|
|
|
|
|
movdqa %fs:RTLD_SAVESPACE_SSE+7*XMM_SIZE, %xmm7
|
|
|
|
|
ret
|
|
|
|
|
cfi_endproc
|
|
|
|
|
.size _dl_x86_64_restore_sse, .-_dl_x86_64_restore_sse
|
|
|
|
|
#endif
|
