From: Suresh Siddha Date: Thu, 6 Sep 2012 21:58:52 +0000 (-0700) Subject: x86, fpu: decouple non-lazy/eager fpu restore from xsave X-Git-Url: http://git.cdn.openwrt.org/?a=commitdiff_plain;h=5d2bd7009f306c82afddd1ca4d9763ad8473c216;p=openwrt%2Fstaging%2Fblogic.git x86, fpu: decouple non-lazy/eager fpu restore from xsave Decouple non-lazy/eager fpu restore policy from the existence of the xsave feature. Introduce a synthetic CPUID flag to represent the eagerfpu policy. "eagerfpu=on" boot paramter will enable the policy. Requested-by: H. Peter Anvin Requested-by: Linus Torvalds Signed-off-by: Suresh Siddha Link: http://lkml.kernel.org/r/1347300665-6209-2-git-send-email-suresh.b.siddha@intel.com Signed-off-by: H. Peter Anvin --- diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index ad7e2e5088c1..741d064fdc6a 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1833,6 +1833,10 @@ bytes respectively. Such letter suffixes can also be entirely omitted. and restore using xsave. The kernel will fallback to enabling legacy floating-point and sse state. + eagerfpu= [X86] + on enable eager fpu restore + off disable eager fpu restore + nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or wfi(ARM) instruction doesn't work correctly and not to use it. This is also useful when using JTAG debugger. diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 6b7ee5ff6820..5dd2b473ccff 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -97,6 +97,7 @@ #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 */ /* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */ #define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */ @@ -305,6 +306,7 @@ extern const char * const x86_power_flags[32]; #define cpu_has_perfctr_core boot_cpu_has(X86_FEATURE_PERFCTR_CORE) #define cpu_has_cx8 boot_cpu_has(X86_FEATURE_CX8) #define cpu_has_cx16 boot_cpu_has(X86_FEATURE_CX16) +#define cpu_has_eager_fpu boot_cpu_has(X86_FEATURE_EAGER_FPU) #if defined(CONFIG_X86_INVLPG) || defined(CONFIG_X86_64) # define cpu_has_invlpg 1 diff --git a/arch/x86/include/asm/fpu-internal.h b/arch/x86/include/asm/fpu-internal.h index 8ca0f9f45ac4..0ca72f0d4b41 100644 --- a/arch/x86/include/asm/fpu-internal.h +++ b/arch/x86/include/asm/fpu-internal.h @@ -38,6 +38,7 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka, extern unsigned int mxcsr_feature_mask; extern void fpu_init(void); +extern void eager_fpu_init(void); DECLARE_PER_CPU(struct task_struct *, fpu_owner_task); @@ -84,6 +85,11 @@ static inline int is_x32_frame(void) #define X87_FSW_ES (1 << 7) /* Exception Summary */ +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); @@ -99,6 +105,14 @@ static __always_inline __pure bool use_fxsr(void) return static_cpu_has(X86_FEATURE_FXSR); } +static inline void fx_finit(struct i387_fxsave_struct *fx) +{ + memset(fx, 0, xstate_size); + fx->cwd = 0x37f; + if (cpu_has_xmm) + fx->mxcsr = MXCSR_DEFAULT; +} + extern void __sanitize_i387_state(struct task_struct *); static inline void sanitize_i387_state(struct task_struct *tsk) @@ -291,13 +305,13 @@ static inline void __thread_set_has_fpu(struct task_struct *tsk) static inline void __thread_fpu_end(struct task_struct *tsk) { __thread_clear_has_fpu(tsk); - if (!use_xsave()) + if (!use_eager_fpu()) stts(); } static inline void __thread_fpu_begin(struct task_struct *tsk) { - if (!use_xsave()) + if (!use_eager_fpu()) clts(); __thread_set_has_fpu(tsk); } @@ -327,10 +341,14 @@ static inline void drop_fpu(struct task_struct *tsk) static inline void drop_init_fpu(struct task_struct *tsk) { - if (!use_xsave()) + if (!use_eager_fpu()) drop_fpu(tsk); - else - xrstor_state(init_xstate_buf, -1); + else { + if (use_xsave()) + xrstor_state(init_xstate_buf, -1); + else + fxrstor_checking(&init_xstate_buf->i387); + } } /* @@ -370,7 +388,7 @@ static inline fpu_switch_t switch_fpu_prepare(struct task_struct *old, struct ta * 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 = tsk_used_math(new) && (use_xsave() || + fpu.preload = tsk_used_math(new) && (use_eager_fpu() || new->fpu_counter > 5); if (__thread_has_fpu(old)) { if (!__save_init_fpu(old)) @@ -383,14 +401,14 @@ static inline fpu_switch_t switch_fpu_prepare(struct task_struct *old, struct ta new->fpu_counter++; __thread_set_has_fpu(new); prefetch(new->thread.fpu.state); - } else if (!use_xsave()) + } else if (!use_eager_fpu()) stts(); } else { old->fpu_counter = 0; old->thread.fpu.last_cpu = ~0; if (fpu.preload) { new->fpu_counter++; - if (!use_xsave() && fpu_lazy_restore(new, cpu)) + if (!use_eager_fpu() && fpu_lazy_restore(new, cpu)) fpu.preload = 0; else prefetch(new->thread.fpu.state); @@ -452,6 +470,14 @@ static inline void user_fpu_begin(void) preempt_enable(); } +static inline void __save_fpu(struct task_struct *tsk) +{ + if (use_xsave()) + xsave_state(&tsk->thread.fpu.state->xsave, -1); + else + fpu_fxsave(&tsk->thread.fpu); +} + /* * These disable preemption on their own and are safe */ @@ -459,8 +485,8 @@ static inline void save_init_fpu(struct task_struct *tsk) { WARN_ON_ONCE(!__thread_has_fpu(tsk)); - if (use_xsave()) { - xsave_state(&tsk->thread.fpu.state->xsave, -1); + if (use_eager_fpu()) { + __save_fpu(tsk); return; } @@ -526,11 +552,9 @@ static inline void fpu_free(struct fpu *fpu) static inline void fpu_copy(struct task_struct *dst, struct task_struct *src) { - if (use_xsave()) { - struct xsave_struct *xsave = &dst->thread.fpu.state->xsave; - - memset(&xsave->xsave_hdr, 0, sizeof(struct xsave_hdr_struct)); - xsave_state(xsave, -1); + if (use_eager_fpu()) { + memset(&dst->thread.fpu.state->xsave, 0, xstate_size); + __save_fpu(dst); } else { struct fpu *dfpu = &dst->thread.fpu; struct fpu *sfpu = &src->thread.fpu; diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index a5fbc3c5fccc..b0fe078614d8 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -1297,7 +1297,6 @@ void __cpuinit cpu_init(void) dbg_restore_debug_regs(); fpu_init(); - xsave_init(); raw_local_save_flags(kernel_eflags); @@ -1352,6 +1351,5 @@ void __cpuinit cpu_init(void) dbg_restore_debug_regs(); fpu_init(); - xsave_init(); } #endif diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index 528557470ddb..6782e3983865 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -22,9 +22,8 @@ /* * Were we in an interrupt that interrupted kernel mode? * - * For now, on xsave platforms we will return interrupted - * kernel FPU as not-idle. TBD: As we use non-lazy FPU restore - * for xsave platforms, ideally we can change the return value + * For now, with eagerfpu we will return interrupted kernel FPU + * state as not-idle. TBD: Ideally we can change the return value * to something like __thread_has_fpu(current). But we need to * be careful of doing __thread_clear_has_fpu() before saving * the FPU etc for supporting nested uses etc. For now, take @@ -38,7 +37,7 @@ */ static inline bool interrupted_kernel_fpu_idle(void) { - if (use_xsave()) + if (use_eager_fpu()) return 0; return !__thread_has_fpu(current) && @@ -84,7 +83,7 @@ void kernel_fpu_begin(void) __save_init_fpu(me); __thread_clear_has_fpu(me); /* We do 'stts()' in kernel_fpu_end() */ - } else if (!use_xsave()) { + } else if (!use_eager_fpu()) { this_cpu_write(fpu_owner_task, NULL); clts(); } @@ -93,7 +92,7 @@ EXPORT_SYMBOL(kernel_fpu_begin); void kernel_fpu_end(void) { - if (use_xsave()) + if (use_eager_fpu()) math_state_restore(); else stts(); @@ -122,7 +121,6 @@ static void __cpuinit mxcsr_feature_mask_init(void) { unsigned long mask = 0; - clts(); if (cpu_has_fxsr) { memset(&fx_scratch, 0, sizeof(struct i387_fxsave_struct)); asm volatile("fxsave %0" : : "m" (fx_scratch)); @@ -131,7 +129,6 @@ static void __cpuinit mxcsr_feature_mask_init(void) mask = 0x0000ffbf; } mxcsr_feature_mask &= mask; - stts(); } static void __cpuinit init_thread_xstate(void) @@ -185,9 +182,8 @@ void __cpuinit fpu_init(void) init_thread_xstate(); mxcsr_feature_mask_init(); - /* clean state in init */ - current_thread_info()->status = 0; - clear_used_math(); + xsave_init(); + eager_fpu_init(); } void fpu_finit(struct fpu *fpu) @@ -198,12 +194,7 @@ void fpu_finit(struct fpu *fpu) } if (cpu_has_fxsr) { - struct i387_fxsave_struct *fx = &fpu->state->fxsave; - - memset(fx, 0, xstate_size); - fx->cwd = 0x37f; - if (cpu_has_xmm) - fx->mxcsr = MXCSR_DEFAULT; + fx_finit(&fpu->state->fxsave); } else { struct i387_fsave_struct *fp = &fpu->state->fsave; memset(fp, 0, xstate_size); diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c21e30f8923b..dc3567e083f9 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -156,7 +156,7 @@ void flush_thread(void) * Free the FPU state for non xsave platforms. They get reallocated * lazily at the first use. */ - if (!use_xsave()) + if (!use_eager_fpu()) free_thread_xstate(tsk); } diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index ac7d5275f6e8..4f4aba0551b0 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -630,7 +630,7 @@ EXPORT_SYMBOL_GPL(math_state_restore); dotraplinkage void __kprobes do_device_not_available(struct pt_regs *regs, long error_code) { - BUG_ON(use_xsave()); + BUG_ON(use_eager_fpu()); #ifdef CONFIG_MATH_EMULATION if (read_cr0() & X86_CR0_EM) { diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c index e7752bd7cac8..c0afd2c43761 100644 --- a/arch/x86/kernel/xsave.c +++ b/arch/x86/kernel/xsave.c @@ -400,7 +400,7 @@ int __restore_xstate_sig(void __user *buf, void __user *buf_fx, int size) set_used_math(); } - if (use_xsave()) + if (use_eager_fpu()) math_state_restore(); return err; @@ -450,28 +450,10 @@ static void prepare_fx_sw_frame(void) */ static inline void xstate_enable(void) { - clts(); set_in_cr4(X86_CR4_OSXSAVE); xsetbv(XCR_XFEATURE_ENABLED_MASK, pcntxt_mask); } -/* - * This is same as math_state_restore(). But use_xsave() is not yet - * patched to use math_state_restore(). - */ -static inline void init_restore_xstate(void) -{ - init_fpu(current); - __thread_fpu_begin(current); - xrstor_state(init_xstate_buf, -1); -} - -static inline void xstate_enable_ap(void) -{ - xstate_enable(); - init_restore_xstate(); -} - /* * Record the offsets and sizes of different state managed by the xsave * memory layout. @@ -500,17 +482,20 @@ static void __init setup_xstate_features(void) /* * setup the xstate image representing the init state */ -static void __init setup_xstate_init(void) +static void __init setup_init_fpu_buf(void) { - setup_xstate_features(); - /* * Setup init_xstate_buf to represent the init state of * all the features managed by the xsave */ init_xstate_buf = alloc_bootmem_align(xstate_size, __alignof__(struct xsave_struct)); - init_xstate_buf->i387.mxcsr = MXCSR_DEFAULT; + fx_finit(&init_xstate_buf->i387); + + if (!cpu_has_xsave) + return; + + setup_xstate_features(); /* * Init all the features state with header_bv being 0x0 @@ -523,6 +508,17 @@ static void __init setup_xstate_init(void) xsave_state(init_xstate_buf, -1); } +static int disable_eagerfpu; +static int __init eager_fpu_setup(char *s) +{ + if (!strcmp(s, "on")) + setup_force_cpu_cap(X86_FEATURE_EAGER_FPU); + else if (!strcmp(s, "off")) + disable_eagerfpu = 1; + return 1; +} +__setup("eagerfpu=", eager_fpu_setup); + /* * Enable and initialize the xsave feature. */ @@ -559,15 +555,10 @@ static void __init xstate_enable_boot_cpu(void) update_regset_xstate_info(xstate_size, pcntxt_mask); prepare_fx_sw_frame(); - - setup_xstate_init(); + setup_init_fpu_buf(); pr_info("enabled xstate_bv 0x%llx, cntxt size 0x%x\n", pcntxt_mask, xstate_size); - - current->thread.fpu.state = - alloc_bootmem_align(xstate_size, __alignof__(struct xsave_struct)); - init_restore_xstate(); } /* @@ -586,6 +577,42 @@ void __cpuinit xsave_init(void) return; this_func = next_func; - next_func = xstate_enable_ap; + next_func = xstate_enable; this_func(); } + +static inline void __init eager_fpu_init_bp(void) +{ + current->thread.fpu.state = + alloc_bootmem_align(xstate_size, __alignof__(struct xsave_struct)); + if (!init_xstate_buf) + setup_init_fpu_buf(); +} + +void __cpuinit eager_fpu_init(void) +{ + static __refdata void (*boot_func)(void) = eager_fpu_init_bp; + + clear_used_math(); + current_thread_info()->status = 0; + if (!cpu_has_eager_fpu) { + stts(); + return; + } + + if (boot_func) { + boot_func(); + boot_func = NULL; + } + + /* + * This is same as math_state_restore(). But use_xsave() is + * not yet patched to use math_state_restore(). + */ + init_fpu(current); + __thread_fpu_begin(current); + if (cpu_has_xsave) + xrstor_state(init_xstate_buf, -1); + else + fxrstor_checking(&init_xstate_buf->i387); +}