xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge branch 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Browse Source 
Pull x86 FPU updates from Ingo Molnar:
 "The main changes in this cycle were:

   - do a large round of simplifications after all CPUs do 'eager' FPU
     context switching in v4.9: remove CR0 twiddling, remove leftover
     eager/lazy bts, etc (Andy Lutomirski)

   - more FPU code simplifications: remove struct fpu::counter, clarify
     nomenclature, remove unnecessary arguments/functions and better
     structure the code (Rik van Riel)"

* 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/fpu: Remove clts()
  x86/fpu: Remove stts()
  x86/fpu: Handle #NM without FPU emulation as an error
  x86/fpu, lguest: Remove CR0.TS support
  x86/fpu, kvm: Remove host CR0.TS manipulation
  x86/fpu: Remove irq_ts_save() and irq_ts_restore()
  x86/fpu: Stop saving and restoring CR0.TS in fpu__init_check_bugs()
  x86/fpu: Get rid of two redundant clts() calls
  x86/fpu: Finish excising 'eagerfpu'
  x86/fpu: Split old_fpu & new_fpu handling into separate functions
  x86/fpu: Remove 'cpu' argument from __cpu_invalidate_fpregs_state()
  x86/fpu: Split old & new FPU code paths
  x86/fpu: Remove __fpregs_(de)activate()
  x86/fpu: Rename lazy restore functions to "register state valid"
  x86/fpu, kvm: Remove KVM vcpu->fpu_counter
  x86/fpu: Remove struct fpu::counter
  x86/fpu: Remove use_eager_fpu()
  x86/fpu: Remove the XFEATURE_MASK_EAGER/LAZY distinction
  x86/fpu: Hard-disable lazy FPU mode
  x86/crypto, x86/fpu: Remove X86_FEATURE_EAGER_FPU #ifdef from the crc32c code
This commit is contained in:
Linus Torvalds
2016-12-12 14:27:49 -08:00
parent 535b2f73f6 064e6a8ba6
commit 518bacf5a5
38 changed files with 105 additions and 547 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
arch/x86/include/asm/cpufeatures.h
View File

@@ -104,7 +104,6 @@
#define X86_FEATURE_EXTD_APICID ( 3*32+26) /* has extended APICID (8 bits) */
#define X86_FEATURE_AMD_DCM ( 3*32+27) /* multi-node processor */
#define X86_FEATURE_APERFMPERF ( 3*32+28) /* APERFMPERF */
#define X86_FEATURE_EAGER_FPU ( 3*32+29) /* "eagerfpu" Non lazy FPU restore */
#define X86_FEATURE_NONSTOP_TSC_S3 ( 3*32+30) /* TSC doesn't stop in S3 state */
/* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */

10
arch/x86/include/asm/fpu/api.h
View File

@@ -26,16 +26,6 @@ extern void kernel_fpu_begin(void);
extern void kernel_fpu_end(void);
extern bool irq_fpu_usable(void);
/*
* Some instructions like VIA's padlock instructions generate a spurious
* DNA fault but don't modify SSE registers. And these instructions
* get used from interrupt context as well. To prevent these kernel instructions
* in interrupt context interacting wrongly with other user/kernel fpu usage, we
* should use them only in the context of irq_ts_save/restore()
*/
extern int irq_ts_save(void);
extern void irq_ts_restore(int TS_state);
/*
* Query the presence of one or more xfeatures. Works on any legacy CPU as well.
*

139
arch/x86/include/asm/fpu/internal.h
View File

@@ -60,11 +60,6 @@ extern u64 fpu__get_supported_xfeatures_mask(void);
/*
* FPU related CPU feature flag helper routines:
*/
static __always_inline __pure bool use_eager_fpu(void)
{
return static_cpu_has(X86_FEATURE_EAGER_FPU);
}
static __always_inline __pure bool use_xsaveopt(void)
{
return static_cpu_has(X86_FEATURE_XSAVEOPT);
@@ -484,42 +479,42 @@ extern int copy_fpstate_to_sigframe(void __user *buf, void __user *fp, int size)
DECLARE_PER_CPU(struct fpu *, fpu_fpregs_owner_ctx);
/*
* Must be run with preemption disabled: this clears the fpu_fpregs_owner_ctx,
* on this CPU.
* The in-register FPU state for an FPU context on a CPU is assumed to be
* valid if the fpu->last_cpu matches the CPU, and the fpu_fpregs_owner_ctx
* matches the FPU.
*
* This will disable any lazy FPU state restore of the current FPU state,
* but if the current thread owns the FPU, it will still be saved by.
* If the FPU register state is valid, the kernel can skip restoring the
* FPU state from memory.
*
* Any code that clobbers the FPU registers or updates the in-memory
* FPU state for a task MUST let the rest of the kernel know that the
* FPU registers are no longer valid for this task.
*
* Either one of these invalidation functions is enough. Invalidate
* a resource you control: CPU if using the CPU for something else
* (with preemption disabled), FPU for the current task, or a task that
* is prevented from running by the current task.
*/
static inline void __cpu_disable_lazy_restore(unsigned int cpu)
static inline void __cpu_invalidate_fpregs_state(void)
{
per_cpu(fpu_fpregs_owner_ctx, cpu) = NULL;
__this_cpu_write(fpu_fpregs_owner_ctx, NULL);
}
static inline int fpu_want_lazy_restore(struct fpu *fpu, unsigned int cpu)
static inline void __fpu_invalidate_fpregs_state(struct fpu *fpu)
{
fpu->last_cpu = -1;
}
static inline int fpregs_state_valid(struct fpu *fpu, unsigned int cpu)
{
return fpu == this_cpu_read_stable(fpu_fpregs_owner_ctx) && cpu == fpu->last_cpu;
}
/*
* Wrap lazy FPU TS handling in a 'hw fpregs activation/deactivation'
* idiom, which is then paired with the sw-flag (fpregs_active) later on:
* These generally need preemption protection to work,
* do try to avoid using these on their own:
*/
static inline void __fpregs_activate_hw(void)
{
if (!use_eager_fpu())
clts();
}
static inline void __fpregs_deactivate_hw(void)
{
if (!use_eager_fpu())
stts();
}
/* Must be paired with an 'stts' (fpregs_deactivate_hw()) after! */
static inline void __fpregs_deactivate(struct fpu *fpu)
static inline void fpregs_deactivate(struct fpu *fpu)
{
WARN_ON_FPU(!fpu->fpregs_active);
@@ -528,8 +523,7 @@ static inline void __fpregs_deactivate(struct fpu *fpu)
trace_x86_fpu_regs_deactivated(fpu);
}
/* Must be paired with a 'clts' (fpregs_activate_hw()) before! */
static inline void __fpregs_activate(struct fpu *fpu)
static inline void fpregs_activate(struct fpu *fpu)
{
WARN_ON_FPU(fpu->fpregs_active);
@@ -553,52 +547,20 @@ static inline int fpregs_active(void)
return current->thread.fpu.fpregs_active;
}
/*
* Encapsulate the CR0.TS handling together with the
* software flag.
*
* These generally need preemption protection to work,
* do try to avoid using these on their own.
*/
static inline void fpregs_activate(struct fpu *fpu)
{
__fpregs_activate_hw();
__fpregs_activate(fpu);
}
static inline void fpregs_deactivate(struct fpu *fpu)
{
__fpregs_deactivate(fpu);
__fpregs_deactivate_hw();
}
/*
* FPU state switching for scheduling.
*
* This is a two-stage process:
*
* - switch_fpu_prepare() saves the old state and
* sets the new state of the CR0.TS bit. This is
* done within the context of the old process.
* - switch_fpu_prepare() saves the old state.
* This is done within the context of the old process.
*
* - switch_fpu_finish() restores the new state as
* necessary.
*/
typedef struct { int preload; } fpu_switch_t;
static inline fpu_switch_t
switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
static inline void
switch_fpu_prepare(struct fpu *old_fpu, int cpu)
{
fpu_switch_t fpu;
/*
* If the task has used the math, pre-load the FPU on xsave processors
* or if the past 5 consecutive context-switches used math.
*/
fpu.preload = static_cpu_has(X86_FEATURE_FPU) &&
new_fpu->fpstate_active &&
(use_eager_fpu() || new_fpu->counter > 5);
if (old_fpu->fpregs_active) {
if (!copy_fpregs_to_fpstate(old_fpu))
old_fpu->last_cpu = -1;
@@ -608,29 +570,8 @@ switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
/* But leave fpu_fpregs_owner_ctx! */
old_fpu->fpregs_active = 0;
trace_x86_fpu_regs_deactivated(old_fpu);
/* Don't change CR0.TS if we just switch! */
if (fpu.preload) {
new_fpu->counter++;
__fpregs_activate(new_fpu);
trace_x86_fpu_regs_activated(new_fpu);
prefetch(&new_fpu->state);
} else {
__fpregs_deactivate_hw();
}
} else {
old_fpu->counter = 0;
} else
old_fpu->last_cpu = -1;
if (fpu.preload) {
new_fpu->counter++;
if (fpu_want_lazy_restore(new_fpu, cpu))
fpu.preload = 0;
else
prefetch(&new_fpu->state);
fpregs_activate(new_fpu);
}
}
return fpu;
}
/*
@@ -638,15 +579,19 @@ switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
*/
/*
* By the time this gets called, we've already cleared CR0.TS and
* given the process the FPU if we are going to preload the FPU
* state - all we need to do is to conditionally restore the register
* state itself.
* Set up the userspace FPU context for the new task, if the task
* has used the FPU.
*/
static inline void switch_fpu_finish(struct fpu *new_fpu, fpu_switch_t fpu_switch)
static inline void switch_fpu_finish(struct fpu *new_fpu, int cpu)
{
if (fpu_switch.preload)
copy_kernel_to_fpregs(&new_fpu->state);
bool preload = static_cpu_has(X86_FEATURE_FPU) &&
new_fpu->fpstate_active;
if (preload) {
if (!fpregs_state_valid(new_fpu, cpu))
copy_kernel_to_fpregs(&new_fpu->state);
fpregs_activate(new_fpu);
}
}
/*

34
arch/x86/include/asm/fpu/types.h
View File

@@ -321,17 +321,6 @@ struct fpu {
*/
unsigned char fpregs_active;
/*
* @counter:
*
* This counter contains the number of consecutive context switches
* during which the FPU stays used. If this is over a threshold, the
* lazy FPU restore logic becomes eager, to save the trap overhead.
* This is an unsigned char so that after 256 iterations the counter
* wraps and the context switch behavior turns lazy again; this is to
* deal with bursty apps that only use the FPU for a short time:
*/
unsigned char counter;
/*
* @state:
*
@@ -340,29 +329,6 @@ struct fpu {
* the registers in the FPU are more recent than this state
* copy. If the task context-switches away then they get
* saved here and represent the FPU state.
*
* After context switches there may be a (short) time period
* during which the in-FPU hardware registers are unchanged
* and still perfectly match this state, if the tasks
* scheduled afterwards are not using the FPU.
*
* This is the 'lazy restore' window of optimization, which
* we track though 'fpu_fpregs_owner_ctx' and 'fpu->last_cpu'.
*
* We detect whether a subsequent task uses the FPU via setting
* CR0::TS to 1, which causes any FPU use to raise a #NM fault.
*
* During this window, if the task gets scheduled again, we
* might be able to skip having to do a restore from this
* memory buffer to the hardware registers - at the cost of
* incurring the overhead of #NM fault traps.
*
* Note that on modern CPUs that support the XSAVEOPT (or other
* optimized XSAVE instructions), we don't use #NM traps anymore,
* as the hardware can track whether FPU registers need saving
* or not. On such CPUs we activate the non-lazy ('eagerfpu')
* logic, which unconditionally saves/restores all FPU state
* across context switches. (if FPU state exists.)
*/
union fpregs_state state;
/*

17
arch/x86/include/asm/fpu/xstate.h
View File

@@ -21,21 +21,16 @@
/* Supervisor features */
#define XFEATURE_MASK_SUPERVISOR (XFEATURE_MASK_PT)
/* Supported features which support lazy state saving */
#define XFEATURE_MASK_LAZY (XFEATURE_MASK_FP | \
/* All currently supported features */
#define XCNTXT_MASK (XFEATURE_MASK_FP | \
XFEATURE_MASK_SSE | \
XFEATURE_MASK_YMM | \
XFEATURE_MASK_OPMASK | \
XFEATURE_MASK_ZMM_Hi256 | \
XFEATURE_MASK_Hi16_ZMM)
/* Supported features which require eager state saving */
#define XFEATURE_MASK_EAGER (XFEATURE_MASK_BNDREGS | \
XFEATURE_MASK_BNDCSR | \
XFEATURE_MASK_PKRU)
/* All currently supported features */
#define XCNTXT_MASK (XFEATURE_MASK_LAZY | XFEATURE_MASK_EAGER)
XFEATURE_MASK_Hi16_ZMM | \
XFEATURE_MASK_PKRU | \
XFEATURE_MASK_BNDREGS | \
XFEATURE_MASK_BNDCSR)
#ifdef CONFIG_X86_64
#define REX_PREFIX "0x48, "

1
arch/x86/include/asm/lguest_hcall.h
View File

@@ -9,7 +9,6 @@
#define LHCALL_FLUSH_TLB 5
#define LHCALL_LOAD_IDT_ENTRY 6
#define LHCALL_SET_STACK 7
#define LHCALL_TS 8
#define LHCALL_SET_CLOCKEVENT 9
#define LHCALL_HALT 10
#define LHCALL_SET_PMD 13

5
arch/x86/include/asm/paravirt.h
View File

@@ -41,11 +41,6 @@ static inline void set_debugreg(unsigned long val, int reg)
PVOP_VCALL2(pv_cpu_ops.set_debugreg, reg, val);
}
static inline void clts(void)
{
PVOP_VCALL0(pv_cpu_ops.clts);
}
static inline unsigned long read_cr0(void)
{
return PVOP_CALL0(unsigned long, pv_cpu_ops.read_cr0);

2
arch/x86/include/asm/paravirt_types.h
View File

@@ -103,8 +103,6 @@ struct pv_cpu_ops {
unsigned long (*get_debugreg)(int regno);
void (*set_debugreg)(int regno, unsigned long value);
void (*clts)(void);
unsigned long (*read_cr0)(void);
void (*write_cr0)(unsigned long);

13
arch/x86/include/asm/special_insns.h
View File

@@ -6,11 +6,6 @@
#include <asm/nops.h>
static inline void native_clts(void)
{
asm volatile("clts");
}
/*
* Volatile isn't enough to prevent the compiler from reordering the
* read/write functions for the control registers and messing everything up.
@@ -208,16 +203,8 @@ static inline void load_gs_index(unsigned selector)
#endif
/* Clear the 'TS' bit */
static inline void clts(void)
{
native_clts();
}
#endif/* CONFIG_PARAVIRT */
#define stts() write_cr0(read_cr0() | X86_CR0_TS)
static inline void clflush(volatile void *__p)
{
asm volatile("clflush %0" : "+m" (*(volatile char __force *)__p));

5
arch/x86/include/asm/trace/fpu.h
View File

@@ -14,7 +14,6 @@ DECLARE_EVENT_CLASS(x86_fpu,
__field(struct fpu *, fpu)
__field(bool, fpregs_active)
__field(bool, fpstate_active)
__field(int, counter)
__field(u64, xfeatures)
__field(u64, xcomp_bv)
),
@@ -23,17 +22,15 @@ DECLARE_EVENT_CLASS(x86_fpu,
__entry->fpu = fpu;
__entry->fpregs_active = fpu->fpregs_active;
__entry->fpstate_active = fpu->fpstate_active;
__entry->counter = fpu->counter;
if (boot_cpu_has(X86_FEATURE_OSXSAVE)) {
__entry->xfeatures = fpu->state.xsave.header.xfeatures;
__entry->xcomp_bv = fpu->state.xsave.header.xcomp_bv;
}
),
TP_printk("x86/fpu: %p fpregs_active: %d fpstate_active: %d counter: %d xfeatures: %llx xcomp_bv: %llx",
TP_printk("x86/fpu: %p fpregs_active: %d fpstate_active: %d xfeatures: %llx xcomp_bv: %llx",
__entry->fpu,
__entry->fpregs_active,
__entry->fpstate_active,
__entry->counter,
__entry->xfeatures,
__entry->xcomp_bv
)