linux-stable/arch/x86/kernel/cpu/cpuid-deps.c
Dave Hansen 30d02551ba x86/fpu: Optimize out sigframe xfeatures when in init state
tl;dr: AMX state is ~8k.  Signal frames can have space for this
~8k and each signal entry writes out all 8k even if it is zeros.
Skip writing zeros for AMX to speed up signal delivery by about
4% overall when AMX is in its init state.

This is a user-visible change to the sigframe ABI.

== Hardware XSAVE Background ==

XSAVE state components may be tracked by the processor as being
in their initial configuration.  Software can detect which
features are in this configuration by looking at the XSTATE_BV
field in an XSAVE buffer or with the XGETBV(1) instruction.

Both the XSAVE and XSAVEOPT instructions enumerate features s
being in the initial configuration via the XSTATE_BV field in the
XSAVE header,  However, XSAVEOPT declines to actually write
features in their initial configuration to the buffer.  XSAVE
writes the feature unconditionally, regardless of whether it is
in the initial configuration or not.

Basically, XSAVE users never need to inspect XSTATE_BV to
determine if the feature has been written to the buffer.
XSAVEOPT users *do* need to inspect XSTATE_BV.  They might also
need to clear out the buffer if they want to make an isolated
change to the state, like modifying one register.

== Software Signal / XSAVE Background ==

Signal frames have historically been written with XSAVE itself.
Each state is written in its entirety, regardless of being in its
initial configuration.

In other words, the signal frame ABI uses the XSAVE behavior, not
the XSAVEOPT behavior.

== Problem ==

This means that any application which has acquired permission to
use AMX via ARCH_REQ_XCOMP_PERM will write 8k of state to the
signal frame.  This 8k write will occur even when AMX was in its
initial configuration and software *knows* this because of
XSTATE_BV.

This problem also exists to a lesser degree with AVX-512 and its
2k of state.  However, AVX-512 use does not require
ARCH_REQ_XCOMP_PERM and is more likely to have existing users
which would be impacted by any change in behavior.

== Solution ==

Stop writing out AMX xfeatures which are in their initial state
to the signal frame.  This effectively makes the signal frame
XSAVE buffer look as if it were written with a combination of
XSAVEOPT and XSAVE behavior.  Userspace which handles XSAVEOPT-
style buffers should be able to handle this naturally.

For now, include only the AMX xfeatures: XTILE and XTILEDATA in
this new behavior.  These require new ABI to use anyway, which
makes their users very unlikely to be broken.  This XSAVEOPT-like
behavior should be expected for all future dynamic xfeatures.  It
may also be extended to legacy features like AVX-512 in the
future.

Only attempt this optimization on systems with dynamic features.
Disable dynamic feature support (XFD) if XGETBV1 is unavailable
by adding a CPUID dependency.

This has been measured to reduce the *overall* cycle cost of
signal delivery by about 4%.

Fixes: 2308ee57d9 ("x86/fpu/amx: Enable the AMX feature in 64-bit mode")
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: "Chang S. Bae" <chang.seok.bae@intel.com>
Link: https://lore.kernel.org/r/20211102224750.FA412E26@davehans-spike.ostc.intel.com
2021-11-03 22:42:35 +01:00

140 lines
4.9 KiB
C

/* Declare dependencies between CPUIDs */
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/module.h>
#include <asm/cpufeature.h>
struct cpuid_dep {
unsigned int feature;
unsigned int depends;
};
/*
* Table of CPUID features that depend on others.
*
* This only includes dependencies that can be usefully disabled, not
* features part of the base set (like FPU).
*
* Note this all is not __init / __initdata because it can be
* called from cpu hotplug. It shouldn't do anything in this case,
* but it's difficult to tell that to the init reference checker.
*/
static const struct cpuid_dep cpuid_deps[] = {
{ X86_FEATURE_FXSR, X86_FEATURE_FPU },
{ X86_FEATURE_XSAVEOPT, X86_FEATURE_XSAVE },
{ X86_FEATURE_XSAVEC, X86_FEATURE_XSAVE },
{ X86_FEATURE_XSAVES, X86_FEATURE_XSAVE },
{ X86_FEATURE_AVX, X86_FEATURE_XSAVE },
{ X86_FEATURE_PKU, X86_FEATURE_XSAVE },
{ X86_FEATURE_MPX, X86_FEATURE_XSAVE },
{ X86_FEATURE_XGETBV1, X86_FEATURE_XSAVE },
{ X86_FEATURE_CMOV, X86_FEATURE_FXSR },
{ X86_FEATURE_MMX, X86_FEATURE_FXSR },
{ X86_FEATURE_MMXEXT, X86_FEATURE_MMX },
{ X86_FEATURE_FXSR_OPT, X86_FEATURE_FXSR },
{ X86_FEATURE_XSAVE, X86_FEATURE_FXSR },
{ X86_FEATURE_XMM, X86_FEATURE_FXSR },
{ X86_FEATURE_XMM2, X86_FEATURE_XMM },
{ X86_FEATURE_XMM3, X86_FEATURE_XMM2 },
{ X86_FEATURE_XMM4_1, X86_FEATURE_XMM2 },
{ X86_FEATURE_XMM4_2, X86_FEATURE_XMM2 },
{ X86_FEATURE_XMM3, X86_FEATURE_XMM2 },
{ X86_FEATURE_PCLMULQDQ, X86_FEATURE_XMM2 },
{ X86_FEATURE_SSSE3, X86_FEATURE_XMM2, },
{ X86_FEATURE_F16C, X86_FEATURE_XMM2, },
{ X86_FEATURE_AES, X86_FEATURE_XMM2 },
{ X86_FEATURE_SHA_NI, X86_FEATURE_XMM2 },
{ X86_FEATURE_FMA, X86_FEATURE_AVX },
{ X86_FEATURE_AVX2, X86_FEATURE_AVX, },
{ X86_FEATURE_AVX512F, X86_FEATURE_AVX, },
{ X86_FEATURE_AVX512IFMA, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512PF, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512ER, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512CD, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512DQ, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512BW, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512VL, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512VBMI, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512_VBMI2, X86_FEATURE_AVX512VL },
{ X86_FEATURE_GFNI, X86_FEATURE_AVX512VL },
{ X86_FEATURE_VAES, X86_FEATURE_AVX512VL },
{ X86_FEATURE_VPCLMULQDQ, X86_FEATURE_AVX512VL },
{ X86_FEATURE_AVX512_VNNI, X86_FEATURE_AVX512VL },
{ X86_FEATURE_AVX512_BITALG, X86_FEATURE_AVX512VL },
{ X86_FEATURE_AVX512_4VNNIW, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512_4FMAPS, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512_VPOPCNTDQ, X86_FEATURE_AVX512F },
{ X86_FEATURE_AVX512_VP2INTERSECT, X86_FEATURE_AVX512VL },
{ X86_FEATURE_CQM_OCCUP_LLC, X86_FEATURE_CQM_LLC },
{ X86_FEATURE_CQM_MBM_TOTAL, X86_FEATURE_CQM_LLC },
{ X86_FEATURE_CQM_MBM_LOCAL, X86_FEATURE_CQM_LLC },
{ X86_FEATURE_AVX512_BF16, X86_FEATURE_AVX512VL },
{ X86_FEATURE_AVX512_FP16, X86_FEATURE_AVX512BW },
{ X86_FEATURE_ENQCMD, X86_FEATURE_XSAVES },
{ X86_FEATURE_PER_THREAD_MBA, X86_FEATURE_MBA },
{ X86_FEATURE_SGX_LC, X86_FEATURE_SGX },
{ X86_FEATURE_SGX1, X86_FEATURE_SGX },
{ X86_FEATURE_SGX2, X86_FEATURE_SGX1 },
{ X86_FEATURE_XFD, X86_FEATURE_XSAVES },
{ X86_FEATURE_XFD, X86_FEATURE_XGETBV1 },
{ X86_FEATURE_AMX_TILE, X86_FEATURE_XFD },
{}
};
static inline void clear_feature(struct cpuinfo_x86 *c, unsigned int feature)
{
/*
* Note: This could use the non atomic __*_bit() variants, but the
* rest of the cpufeature code uses atomics as well, so keep it for
* consistency. Cleanup all of it separately.
*/
if (!c) {
clear_cpu_cap(&boot_cpu_data, feature);
set_bit(feature, (unsigned long *)cpu_caps_cleared);
} else {
clear_bit(feature, (unsigned long *)c->x86_capability);
}
}
/* Take the capabilities and the BUG bits into account */
#define MAX_FEATURE_BITS ((NCAPINTS + NBUGINTS) * sizeof(u32) * 8)
static void do_clear_cpu_cap(struct cpuinfo_x86 *c, unsigned int feature)
{
DECLARE_BITMAP(disable, MAX_FEATURE_BITS);
const struct cpuid_dep *d;
bool changed;
if (WARN_ON(feature >= MAX_FEATURE_BITS))
return;
clear_feature(c, feature);
/* Collect all features to disable, handling dependencies */
memset(disable, 0, sizeof(disable));
__set_bit(feature, disable);
/* Loop until we get a stable state. */
do {
changed = false;
for (d = cpuid_deps; d->feature; d++) {
if (!test_bit(d->depends, disable))
continue;
if (__test_and_set_bit(d->feature, disable))
continue;
changed = true;
clear_feature(c, d->feature);
}
} while (changed);
}
void clear_cpu_cap(struct cpuinfo_x86 *c, unsigned int feature)
{
do_clear_cpu_cap(c, feature);
}
void setup_clear_cpu_cap(unsigned int feature)
{
do_clear_cpu_cap(NULL, feature);
}