diff options
Diffstat (limited to 'arch/x86/xen/smp.c')
-rw-r--r-- | arch/x86/xen/smp.c | 260 |
1 files changed, 128 insertions, 132 deletions
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index d2e3c20127d..f702199312a 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -36,27 +36,14 @@ #include "mmu.h" cpumask_t xen_cpu_initialized_map; -static DEFINE_PER_CPU(int, resched_irq) = -1; -static DEFINE_PER_CPU(int, callfunc_irq) = -1; -static DEFINE_PER_CPU(int, debug_irq) = -1; -/* - * Structure and data for smp_call_function(). This is designed to minimise - * static memory requirements. It also looks cleaner. - */ -static DEFINE_SPINLOCK(call_lock); - -struct call_data_struct { - void (*func) (void *info); - void *info; - atomic_t started; - atomic_t finished; - int wait; -}; +static DEFINE_PER_CPU(int, resched_irq); +static DEFINE_PER_CPU(int, callfunc_irq); +static DEFINE_PER_CPU(int, callfuncsingle_irq); +static DEFINE_PER_CPU(int, debug_irq) = -1; static irqreturn_t xen_call_function_interrupt(int irq, void *dev_id); - -static struct call_data_struct *call_data; +static irqreturn_t xen_call_function_single_interrupt(int irq, void *dev_id); /* * Reschedule call back. Nothing to do, @@ -79,13 +66,22 @@ static __cpuinit void cpu_bringup_and_idle(void) int cpu = smp_processor_id(); cpu_init(); + preempt_disable(); + xen_enable_sysenter(); + xen_enable_syscall(); - preempt_disable(); - per_cpu(cpu_state, cpu) = CPU_ONLINE; + cpu = smp_processor_id(); + smp_store_cpu_info(cpu); + cpu_data(cpu).x86_max_cores = 1; + set_cpu_sibling_map(cpu); xen_setup_cpu_clockevents(); + cpu_set(cpu, cpu_online_map); + x86_write_percpu(cpu_state, CPU_ONLINE); + wmb(); + /* We can take interrupts now: we're officially "up". */ local_irq_enable(); @@ -128,6 +124,17 @@ static int xen_smp_intr_init(unsigned int cpu) goto fail; per_cpu(debug_irq, cpu) = rc; + callfunc_name = kasprintf(GFP_KERNEL, "callfuncsingle%d", cpu); + rc = bind_ipi_to_irqhandler(XEN_CALL_FUNCTION_SINGLE_VECTOR, + cpu, + xen_call_function_single_interrupt, + IRQF_DISABLED|IRQF_PERCPU|IRQF_NOBALANCING, + callfunc_name, + NULL); + if (rc < 0) + goto fail; + per_cpu(callfuncsingle_irq, cpu) = rc; + return 0; fail: @@ -137,59 +144,43 @@ static int xen_smp_intr_init(unsigned int cpu) unbind_from_irqhandler(per_cpu(callfunc_irq, cpu), NULL); if (per_cpu(debug_irq, cpu) >= 0) unbind_from_irqhandler(per_cpu(debug_irq, cpu), NULL); + if (per_cpu(callfuncsingle_irq, cpu) >= 0) + unbind_from_irqhandler(per_cpu(callfuncsingle_irq, cpu), NULL); + return rc; } -void __init xen_fill_possible_map(void) +static void __init xen_fill_possible_map(void) { int i, rc; for (i = 0; i < NR_CPUS; i++) { rc = HYPERVISOR_vcpu_op(VCPUOP_is_up, i, NULL); - if (rc >= 0) + if (rc >= 0) { + num_processors++; cpu_set(i, cpu_possible_map); + } } } -void __init xen_smp_prepare_boot_cpu(void) +static void __init xen_smp_prepare_boot_cpu(void) { - int cpu; - BUG_ON(smp_processor_id() != 0); native_smp_prepare_boot_cpu(); /* We've switched to the "real" per-cpu gdt, so make sure the old memory can be recycled */ - make_lowmem_page_readwrite(&per_cpu__gdt_page); - - for_each_possible_cpu(cpu) { - cpus_clear(per_cpu(cpu_sibling_map, cpu)); - /* - * cpu_core_map lives in a per cpu area that is cleared - * when the per cpu array is allocated. - * - * cpus_clear(per_cpu(cpu_core_map, cpu)); - */ - } + make_lowmem_page_readwrite(&per_cpu_var(gdt_page)); xen_setup_vcpu_info_placement(); } -void __init xen_smp_prepare_cpus(unsigned int max_cpus) +static void __init xen_smp_prepare_cpus(unsigned int max_cpus) { unsigned cpu; - for_each_possible_cpu(cpu) { - cpus_clear(per_cpu(cpu_sibling_map, cpu)); - /* - * cpu_core_ map will be zeroed when the per - * cpu area is allocated. - * - * cpus_clear(per_cpu(cpu_core_map, cpu)); - */ - } - smp_store_cpu_info(0); + cpu_data(0).x86_max_cores = 1; set_cpu_sibling_map(0); if (xen_smp_intr_init(0)) @@ -224,7 +215,7 @@ static __cpuinit int cpu_initialize_context(unsigned int cpu, struct task_struct *idle) { struct vcpu_guest_context *ctxt; - struct gdt_page *gdt = &per_cpu(gdt_page, cpu); + struct desc_struct *gdt; if (cpu_test_and_set(cpu, xen_cpu_initialized_map)) return 0; @@ -233,12 +224,15 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle) if (ctxt == NULL) return -ENOMEM; + gdt = get_cpu_gdt_table(cpu); + ctxt->flags = VGCF_IN_KERNEL; ctxt->user_regs.ds = __USER_DS; ctxt->user_regs.es = __USER_DS; - ctxt->user_regs.fs = __KERNEL_PERCPU; - ctxt->user_regs.gs = 0; ctxt->user_regs.ss = __KERNEL_DS; +#ifdef CONFIG_X86_32 + ctxt->user_regs.fs = __KERNEL_PERCPU; +#endif ctxt->user_regs.eip = (unsigned long)cpu_bringup_and_idle; ctxt->user_regs.eflags = 0x1000; /* IOPL_RING1 */ @@ -248,11 +242,11 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle) ctxt->ldt_ents = 0; - BUG_ON((unsigned long)gdt->gdt & ~PAGE_MASK); - make_lowmem_page_readonly(gdt->gdt); + BUG_ON((unsigned long)gdt & ~PAGE_MASK); + make_lowmem_page_readonly(gdt); - ctxt->gdt_frames[0] = virt_to_mfn(gdt->gdt); - ctxt->gdt_ents = ARRAY_SIZE(gdt->gdt); + ctxt->gdt_frames[0] = virt_to_mfn(gdt); + ctxt->gdt_ents = GDT_ENTRIES; ctxt->user_regs.cs = __KERNEL_CS; ctxt->user_regs.esp = idle->thread.sp0 - sizeof(struct pt_regs); @@ -260,9 +254,11 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle) ctxt->kernel_ss = __KERNEL_DS; ctxt->kernel_sp = idle->thread.sp0; +#ifdef CONFIG_X86_32 ctxt->event_callback_cs = __KERNEL_CS; - ctxt->event_callback_eip = (unsigned long)xen_hypervisor_callback; ctxt->failsafe_callback_cs = __KERNEL_CS; +#endif + ctxt->event_callback_eip = (unsigned long)xen_hypervisor_callback; ctxt->failsafe_callback_eip = (unsigned long)xen_failsafe_callback; per_cpu(xen_cr3, cpu) = __pa(swapper_pg_dir); @@ -275,7 +271,7 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle) return 0; } -int __cpuinit xen_cpu_up(unsigned int cpu) +static int __cpuinit xen_cpu_up(unsigned int cpu) { struct task_struct *idle = idle_task(cpu); int rc; @@ -286,11 +282,28 @@ int __cpuinit xen_cpu_up(unsigned int cpu) return rc; #endif +#ifdef CONFIG_X86_64 + /* Allocate node local memory for AP pdas */ + WARN_ON(cpu == 0); + if (cpu > 0) { + rc = get_local_pda(cpu); + if (rc) + return rc; + } +#endif + +#ifdef CONFIG_X86_32 init_gdt(cpu); per_cpu(current_task, cpu) = idle; irq_ctx_init(cpu); +#else + cpu_pda(cpu)->pcurrent = idle; + clear_tsk_thread_flag(idle, TIF_FORK); +#endif xen_setup_timer(cpu); + per_cpu(cpu_state, cpu) = CPU_UP_PREPARE; + /* make sure interrupts start blocked */ per_cpu(xen_vcpu, cpu)->evtchn_upcall_mask = 1; @@ -305,20 +318,18 @@ int __cpuinit xen_cpu_up(unsigned int cpu) if (rc) return rc; - smp_store_cpu_info(cpu); - set_cpu_sibling_map(cpu); - /* This must be done before setting cpu_online_map */ - wmb(); - - cpu_set(cpu, cpu_online_map); - rc = HYPERVISOR_vcpu_op(VCPUOP_up, cpu, NULL); BUG_ON(rc); + while(per_cpu(cpu_state, cpu) != CPU_ONLINE) { + HYPERVISOR_sched_op(SCHEDOP_yield, 0); + barrier(); + } + return 0; } -void xen_smp_cpus_done(unsigned int max_cpus) +static void xen_smp_cpus_done(unsigned int max_cpus) { } @@ -334,17 +345,16 @@ static void stop_self(void *v) BUG(); } -void xen_smp_send_stop(void) +static void xen_smp_send_stop(void) { - smp_call_function(stop_self, NULL, 0, 0); + smp_call_function(stop_self, NULL, 0); } -void xen_smp_send_reschedule(int cpu) +static void xen_smp_send_reschedule(int cpu) { xen_send_IPI_one(cpu, XEN_RESCHEDULE_VECTOR); } - static void xen_send_IPI_mask(cpumask_t mask, enum ipi_vector vector) { unsigned cpu; @@ -355,83 +365,69 @@ static void xen_send_IPI_mask(cpumask_t mask, enum ipi_vector vector) xen_send_IPI_one(cpu, vector); } +static void xen_smp_send_call_function_ipi(cpumask_t mask) +{ + int cpu; + + xen_send_IPI_mask(mask, XEN_CALL_FUNCTION_VECTOR); + + /* Make sure other vcpus get a chance to run if they need to. */ + for_each_cpu_mask(cpu, mask) { + if (xen_vcpu_stolen(cpu)) { + HYPERVISOR_sched_op(SCHEDOP_yield, 0); + break; + } + } +} + +static void xen_smp_send_call_function_single_ipi(int cpu) +{ + xen_send_IPI_mask(cpumask_of_cpu(cpu), XEN_CALL_FUNCTION_SINGLE_VECTOR); +} + static irqreturn_t xen_call_function_interrupt(int irq, void *dev_id) { - void (*func) (void *info) = call_data->func; - void *info = call_data->info; - int wait = call_data->wait; - - /* - * Notify initiating CPU that I've grabbed the data and am - * about to execute the function - */ - mb(); - atomic_inc(&call_data->started); - /* - * At this point the info structure may be out of scope unless wait==1 - */ irq_enter(); - (*func)(info); + generic_smp_call_function_interrupt(); +#ifdef CONFIG_X86_32 __get_cpu_var(irq_stat).irq_call_count++; +#else + add_pda(irq_call_count, 1); +#endif irq_exit(); - if (wait) { - mb(); /* commit everything before setting finished */ - atomic_inc(&call_data->finished); - } - return IRQ_HANDLED; } -int xen_smp_call_function_mask(cpumask_t mask, void (*func)(void *), - void *info, int wait) +static irqreturn_t xen_call_function_single_interrupt(int irq, void *dev_id) { - struct call_data_struct data; - int cpus, cpu; - bool yield; - - /* Holding any lock stops cpus from going down. */ - spin_lock(&call_lock); - - cpu_clear(smp_processor_id(), mask); - - cpus = cpus_weight(mask); - if (!cpus) { - spin_unlock(&call_lock); - return 0; - } - - /* Can deadlock when called with interrupts disabled */ - WARN_ON(irqs_disabled()); - - data.func = func; - data.info = info; - atomic_set(&data.started, 0); - data.wait = wait; - if (wait) - atomic_set(&data.finished, 0); - - call_data = &data; - mb(); /* write everything before IPI */ - - /* Send a message to other CPUs and wait for them to respond */ - xen_send_IPI_mask(mask, XEN_CALL_FUNCTION_VECTOR); + irq_enter(); + generic_smp_call_function_single_interrupt(); +#ifdef CONFIG_X86_32 + __get_cpu_var(irq_stat).irq_call_count++; +#else + add_pda(irq_call_count, 1); +#endif + irq_exit(); - /* Make sure other vcpus get a chance to run if they need to. */ - yield = false; - for_each_cpu_mask(cpu, mask) - if (xen_vcpu_stolen(cpu)) - yield = true; + return IRQ_HANDLED; +} - if (yield) - HYPERVISOR_sched_op(SCHEDOP_yield, 0); +static const struct smp_ops xen_smp_ops __initdata = { + .smp_prepare_boot_cpu = xen_smp_prepare_boot_cpu, + .smp_prepare_cpus = xen_smp_prepare_cpus, + .cpu_up = xen_cpu_up, + .smp_cpus_done = xen_smp_cpus_done, - /* Wait for response */ - while (atomic_read(&data.started) != cpus || - (wait && atomic_read(&data.finished) != cpus)) - cpu_relax(); + .smp_send_stop = xen_smp_send_stop, + .smp_send_reschedule = xen_smp_send_reschedule, - spin_unlock(&call_lock); + .send_call_func_ipi = xen_smp_send_call_function_ipi, + .send_call_func_single_ipi = xen_smp_send_call_function_single_ipi, +}; - return 0; +void __init xen_smp_init(void) +{ + smp_ops = xen_smp_ops; + xen_fill_possible_map(); } |