AuroraMiddleware/glibc
1
0
Fork 0
mirror of https://sourceware.org/git/glibc.git synced 2024-11-22 21:10:07 +00:00
Code Issues Packages Projects Releases Wiki Activity
4b0fa403ab
glibc/sysdeps/x86/cpu-features.c
H.J. Lu 45ff34638f x86: Add x86_64 to x86-64 HWCAP [BZ #22093] 
Before glibc 2.26, ld.so set dl_platform to "x86_64" and searched the
"x86_64" subdirectory when loading a shared library.  ld.so in glibc
2.26 was changed to set dl_platform to "haswell" or "xeon_phi", based
on supported ISAs.  This led to shared library loading failure for
shared libraries placed under the "x86_64" subdirectory.

This patch adds "x86_64" to x86-64 dl_hwcap so that ld.so will always
search the "x86_64" subdirectory when loading a shared library.

NB: We can't set x86-64 dl_platform to "x86-64" since ld.so will skip
the "haswell" and "xeon_phi" subdirectories on "haswell" and "xeon_phi"
machines.

Tested on i686 and x86-64.

	[BZ #22093]
	* sysdeps/x86/cpu-features.c (init_cpu_features): Initialize
	GLRO(dl_hwcap) to HWCAP_X86_64 for x86-64.
	* sysdeps/x86/dl-hwcap.h (HWCAP_COUNT): Updated.
	(HWCAP_IMPORTANT): Likewise.
	(HWCAP_X86_64): New enum.
	(HWCAP_X86_AVX512_1): Updated.
	* sysdeps/x86/dl-procinfo.c (_dl_x86_hwcap_flags): Add "x86_64".
	* sysdeps/x86_64/Makefile (tests): Add tst-x86_64-1.
	(modules-names): Add x86_64/tst-x86_64mod-1.
	(LDFLAGS-tst-x86_64mod-1.so): New.
	($(objpfx)tst-x86_64-1): Likewise.
	($(objpfx)x86_64/tst-x86_64mod-1.os): Likewise.
	(tst-x86_64-1-clean): Likewise.
	* sysdeps/x86_64/tst-x86_64-1.c: New file.
	* sysdeps/x86_64/tst-x86_64mod-1.c: Likewise.
2017-09-11 08:18:32 -07:00

387 lines
12 KiB
C
Raw Blame History

/* Initialize CPU feature data.
This file is part of the GNU C Library.
Copyright (C) 2008-2017 Free Software Foundation, Inc.
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, see
<http://www.gnu.org/licenses/>. */
#include <cpuid.h>
#include <cpu-features.h>
#include <dl-hwcap.h>
#if HAVE_TUNABLES
# define TUNABLE_NAMESPACE tune
# include <unistd.h> /* Get STDOUT_FILENO for _dl_printf. */
# include <elf/dl-tunables.h>
extern void TUNABLE_CALLBACK (set_hwcaps) (tunable_val_t *)
attribute_hidden;
#endif
static void
get_common_indeces (struct cpu_features *cpu_features,
unsigned int *family, unsigned int *model,
unsigned int *extended_model, unsigned int *stepping)
{
if (family)
{
unsigned int eax;
__cpuid (1, eax, cpu_features->cpuid[COMMON_CPUID_INDEX_1].ebx,
cpu_features->cpuid[COMMON_CPUID_INDEX_1].ecx,
cpu_features->cpuid[COMMON_CPUID_INDEX_1].edx);
cpu_features->cpuid[COMMON_CPUID_INDEX_1].eax = eax;
*family = (eax >> 8) & 0x0f;
*model = (eax >> 4) & 0x0f;
*extended_model = (eax >> 12) & 0xf0;
*stepping = eax & 0x0f;
if (*family == 0x0f)
{
*family += (eax >> 20) & 0xff;
*model += *extended_model;
}
}
if (cpu_features->max_cpuid >= 7)
__cpuid_count (7, 0,
cpu_features->cpuid[COMMON_CPUID_INDEX_7].eax,
cpu_features->cpuid[COMMON_CPUID_INDEX_7].ebx,
cpu_features->cpuid[COMMON_CPUID_INDEX_7].ecx,
cpu_features->cpuid[COMMON_CPUID_INDEX_7].edx);
/* Can we call xgetbv? */
if (CPU_FEATURES_CPU_P (cpu_features, OSXSAVE))
{
unsigned int xcrlow;
unsigned int xcrhigh;
asm ("xgetbv" : "=a" (xcrlow), "=d" (xcrhigh) : "c" (0));
/* Is YMM and XMM state usable? */
if ((xcrlow & (bit_YMM_state | bit_XMM_state)) ==
(bit_YMM_state | bit_XMM_state))
{
/* Determine if AVX is usable. */
if (CPU_FEATURES_CPU_P (cpu_features, AVX))
{
cpu_features->feature[index_arch_AVX_Usable]
|= bit_arch_AVX_Usable;
/* The following features depend on AVX being usable. */
/* Determine if AVX2 is usable. */
if (CPU_FEATURES_CPU_P (cpu_features, AVX2))
cpu_features->feature[index_arch_AVX2_Usable]
|= bit_arch_AVX2_Usable;
/* Determine if FMA is usable. */
if (CPU_FEATURES_CPU_P (cpu_features, FMA))
cpu_features->feature[index_arch_FMA_Usable]
|= bit_arch_FMA_Usable;
}
/* Check if OPMASK state, upper 256-bit of ZMM0-ZMM15 and
ZMM16-ZMM31 state are enabled. */
if ((xcrlow & (bit_Opmask_state | bit_ZMM0_15_state
| bit_ZMM16_31_state)) ==
(bit_Opmask_state | bit_ZMM0_15_state | bit_ZMM16_31_state))
{
/* Determine if AVX512F is usable. */
if (CPU_FEATURES_CPU_P (cpu_features, AVX512F))
{
cpu_features->feature[index_arch_AVX512F_Usable]
|= bit_arch_AVX512F_Usable;
/* Determine if AVX512DQ is usable. */
if (CPU_FEATURES_CPU_P (cpu_features, AVX512DQ))
cpu_features->feature[index_arch_AVX512DQ_Usable]
|= bit_arch_AVX512DQ_Usable;
}
}
}
}
}
static inline void
init_cpu_features (struct cpu_features *cpu_features)
{
unsigned int ebx, ecx, edx;
unsigned int family = 0;
unsigned int model = 0;
enum cpu_features_kind kind;
#if !HAS_CPUID
if (__get_cpuid_max (0, 0) == 0)
{
kind = arch_kind_other;
goto no_cpuid;
}
#endif
__cpuid (0, cpu_features->max_cpuid, ebx, ecx, edx);
/* This spells out "GenuineIntel". */
if (ebx == 0x756e6547 && ecx == 0x6c65746e && edx == 0x49656e69)
{
unsigned int extended_model, stepping;
kind = arch_kind_intel;
get_common_indeces (cpu_features, &family, &model, &extended_model,
&stepping);
if (family == 0x06)
{
model += extended_model;
switch (model)
{
case 0x1c:
case 0x26:
/* BSF is slow on Atom. */
cpu_features->feature[index_arch_Slow_BSF]
|= bit_arch_Slow_BSF;
break;
case 0x57:
/* Knights Landing. Enable Silvermont optimizations. */
case 0x5c:
case 0x5f:
/* Unaligned load versions are faster than SSSE3
on Goldmont. */
case 0x4c:
/* Airmont is a die shrink of Silvermont. */
case 0x37:
case 0x4a:
case 0x4d:
case 0x5a:
case 0x5d:
/* Unaligned load versions are faster than SSSE3
on Silvermont. */
#if index_arch_Fast_Unaligned_Load != index_arch_Prefer_PMINUB_for_stringop
# error index_arch_Fast_Unaligned_Load != index_arch_Prefer_PMINUB_for_stringop
#endif
#if index_arch_Fast_Unaligned_Load != index_arch_Slow_SSE4_2
# error index_arch_Fast_Unaligned_Load != index_arch_Slow_SSE4_2
#endif
#if index_arch_Fast_Unaligned_Load != index_arch_Fast_Unaligned_Copy
# error index_arch_Fast_Unaligned_Load != index_arch_Fast_Unaligned_Copy
#endif
cpu_features->feature[index_arch_Fast_Unaligned_Load]
|= (bit_arch_Fast_Unaligned_Load
| bit_arch_Fast_Unaligned_Copy
| bit_arch_Prefer_PMINUB_for_stringop
| bit_arch_Slow_SSE4_2);
break;
default:
/* Unknown family 0x06 processors. Assuming this is one
of Core i3/i5/i7 processors if AVX is available. */
if (!CPU_FEATURES_CPU_P (cpu_features, AVX))
break;
case 0x1a:
case 0x1e:
case 0x1f:
case 0x25:
case 0x2c:
case 0x2e:
case 0x2f:
/* Rep string instructions, unaligned load, unaligned copy,
and pminub are fast on Intel Core i3, i5 and i7. */
#if index_arch_Fast_Rep_String != index_arch_Fast_Unaligned_Load
# error index_arch_Fast_Rep_String != index_arch_Fast_Unaligned_Load
#endif
#if index_arch_Fast_Rep_String != index_arch_Prefer_PMINUB_for_stringop
# error index_arch_Fast_Rep_String != index_arch_Prefer_PMINUB_for_stringop
#endif
#if index_arch_Fast_Rep_String != index_arch_Fast_Unaligned_Copy
# error index_arch_Fast_Rep_String != index_arch_Fast_Unaligned_Copy
#endif
cpu_features->feature[index_arch_Fast_Rep_String]
|= (bit_arch_Fast_Rep_String
| bit_arch_Fast_Unaligned_Load
| bit_arch_Fast_Unaligned_Copy
| bit_arch_Prefer_PMINUB_for_stringop);
break;
case 0x3f:
/* Xeon E7 v3 with stepping >= 4 has working TSX. */
if (stepping >= 4)
break;
case 0x3c:
case 0x45:
case 0x46:
/* Disable Intel TSX on Haswell processors (except Xeon E7 v3
with stepping >= 4) to avoid TSX on kernels that weren't
updated with the latest microcode package (which disables
broken feature by default). */
cpu_features->cpuid[index_cpu_RTM].reg_RTM &= ~bit_cpu_RTM;
break;
}
}
/* Unaligned load with 256-bit AVX registers are faster on
Intel processors with AVX2. */
if (CPU_FEATURES_ARCH_P (cpu_features, AVX2_Usable))
cpu_features->feature[index_arch_AVX_Fast_Unaligned_Load]
|= bit_arch_AVX_Fast_Unaligned_Load;
/* Since AVX512ER is unique to Xeon Phi, set Prefer_No_VZEROUPPER
if AVX512ER is available. Don't use AVX512 to avoid lower CPU
frequency if AVX512ER isn't available. */
if (CPU_FEATURES_CPU_P (cpu_features, AVX512ER))
cpu_features->feature[index_arch_Prefer_No_VZEROUPPER]
|= bit_arch_Prefer_No_VZEROUPPER;
else
cpu_features->feature[index_arch_Prefer_No_AVX512]
|= bit_arch_Prefer_No_AVX512;
/* To avoid SSE transition penalty, use _dl_runtime_resolve_slow.
If XGETBV suports ECX == 1, use _dl_runtime_resolve_opt.
Use _dl_runtime_resolve_opt only with AVX512F since it is
slower than _dl_runtime_resolve_slow with AVX. */
cpu_features->feature[index_arch_Use_dl_runtime_resolve_slow]
|= bit_arch_Use_dl_runtime_resolve_slow;
if (CPU_FEATURES_ARCH_P (cpu_features, AVX512F_Usable)
&& cpu_features->max_cpuid >= 0xd)
{
unsigned int eax;
__cpuid_count (0xd, 1, eax, ebx, ecx, edx);
if ((eax & (1 << 2)) != 0)
cpu_features->feature[index_arch_Use_dl_runtime_resolve_opt]
|= bit_arch_Use_dl_runtime_resolve_opt;
}
}
/* This spells out "AuthenticAMD". */
else if (ebx == 0x68747541 && ecx == 0x444d4163 && edx == 0x69746e65)
{
unsigned int extended_model, stepping;
kind = arch_kind_amd;
get_common_indeces (cpu_features, &family, &model, &extended_model,
&stepping);
ecx = cpu_features->cpuid[COMMON_CPUID_INDEX_1].ecx;
unsigned int eax;
__cpuid (0x80000000, eax, ebx, ecx, edx);
if (eax >= 0x80000001)
__cpuid (0x80000001,
cpu_features->cpuid[COMMON_CPUID_INDEX_80000001].eax,
cpu_features->cpuid[COMMON_CPUID_INDEX_80000001].ebx,
cpu_features->cpuid[COMMON_CPUID_INDEX_80000001].ecx,
cpu_features->cpuid[COMMON_CPUID_INDEX_80000001].edx);
if (HAS_ARCH_FEATURE (AVX_Usable))
{
/* Since the FMA4 bit is in COMMON_CPUID_INDEX_80000001 and
FMA4 requires AVX, determine if FMA4 is usable here. */
if (CPU_FEATURES_CPU_P (cpu_features, FMA4))
cpu_features->feature[index_arch_FMA4_Usable]
|= bit_arch_FMA4_Usable;
}
if (family == 0x15)
{
#if index_arch_Fast_Unaligned_Load != index_arch_Fast_Copy_Backward
# error index_arch_Fast_Unaligned_Load != index_arch_Fast_Copy_Backward
#endif
/* "Excavator" */
if (model >= 0x60 && model <= 0x7f)
cpu_features->feature[index_arch_Fast_Unaligned_Load]
|= (bit_arch_Fast_Unaligned_Load
| bit_arch_Fast_Copy_Backward);
}
}
else
{
kind = arch_kind_other;
get_common_indeces (cpu_features, NULL, NULL, NULL, NULL);
}
/* Support i586 if CX8 is available. */
if (CPU_FEATURES_CPU_P (cpu_features, CX8))
cpu_features->feature[index_arch_I586] |= bit_arch_I586;
/* Support i686 if CMOV is available. */
if (CPU_FEATURES_CPU_P (cpu_features, CMOV))
cpu_features->feature[index_arch_I686] |= bit_arch_I686;
#if !HAS_CPUID
no_cpuid:
#endif
cpu_features->family = family;
cpu_features->model = model;
cpu_features->kind = kind;
#if HAVE_TUNABLES
TUNABLE_GET (hwcaps, tunable_val_t *, TUNABLE_CALLBACK (set_hwcaps));
cpu_features->non_temporal_threshold
= TUNABLE_GET (x86_non_temporal_threshold, long int, NULL);
cpu_features->data_cache_size
= TUNABLE_GET (x86_data_cache_size, long int, NULL);
cpu_features->shared_cache_size
= TUNABLE_GET (x86_shared_cache_size, long int, NULL);
#endif
/* Reuse dl_platform, dl_hwcap and dl_hwcap_mask for x86. */
GLRO(dl_platform) = NULL;
#if !HAVE_TUNABLES && defined SHARED
/* The glibc.tune.hwcap_mask tunable is initialized already, so no need to do
this. */
GLRO(dl_hwcap_mask) = HWCAP_IMPORTANT;
#endif
#ifdef __x86_64__
GLRO(dl_hwcap) = HWCAP_X86_64;
if (cpu_features->kind == arch_kind_intel)
{
if (CPU_FEATURES_ARCH_P (cpu_features, AVX512F_Usable)
&& CPU_FEATURES_CPU_P (cpu_features, AVX512CD))
{
if (CPU_FEATURES_CPU_P (cpu_features, AVX512ER))
{
if (CPU_FEATURES_CPU_P (cpu_features, AVX512PF))
GLRO(dl_platform) = "xeon_phi";
}
else
{
if (CPU_FEATURES_CPU_P (cpu_features, AVX512BW)
&& CPU_FEATURES_CPU_P (cpu_features, AVX512DQ)
&& CPU_FEATURES_CPU_P (cpu_features, AVX512VL))
GLRO(dl_hwcap) |= HWCAP_X86_AVX512_1;
}
}
if (GLRO(dl_platform) == NULL
&& CPU_FEATURES_ARCH_P (cpu_features, AVX2_Usable)
&& CPU_FEATURES_ARCH_P (cpu_features, FMA_Usable)
&& CPU_FEATURES_CPU_P (cpu_features, BMI1)
&& CPU_FEATURES_CPU_P (cpu_features, BMI2)
&& CPU_FEATURES_CPU_P (cpu_features, LZCNT)
&& CPU_FEATURES_CPU_P (cpu_features, MOVBE)
&& CPU_FEATURES_CPU_P (cpu_features, POPCNT))
GLRO(dl_platform) = "haswell";
}
#else
GLRO(dl_hwcap) = 0;
if (CPU_FEATURES_CPU_P (cpu_features, SSE2))
GLRO(dl_hwcap) |= HWCAP_X86_SSE2;
if (CPU_FEATURES_ARCH_P (cpu_features, I686))
GLRO(dl_platform) = "i686";
else if (CPU_FEATURES_ARCH_P (cpu_features, I586))
GLRO(dl_platform) = "i586";
#endif
}
Reference in New Issue View Git Blame Copy Permalink