diff options
48 files changed, 1610 insertions, 680 deletions
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 3db651fc8ec..d679cb2c79b 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -10,7 +10,7 @@ ifdef CONFIG_FTRACE # Do not profile debug and lowlevel utilities CFLAGS_REMOVE_tsc.o = -pg CFLAGS_REMOVE_rtc.o = -pg -CFLAGS_REMOVE_paravirt.o = -pg +CFLAGS_REMOVE_paravirt-spinlocks.o = -pg endif # @@ -89,7 +89,7 @@ obj-$(CONFIG_DEBUG_NX_TEST) += test_nx.o obj-$(CONFIG_VMI) += vmi_32.o vmiclock_32.o obj-$(CONFIG_KVM_GUEST) += kvm.o obj-$(CONFIG_KVM_CLOCK) += kvmclock.o -obj-$(CONFIG_PARAVIRT) += paravirt.o paravirt_patch_$(BITS).o +obj-$(CONFIG_PARAVIRT) += paravirt.o paravirt_patch_$(BITS).o paravirt-spinlocks.o obj-$(CONFIG_PARAVIRT_CLOCK) += pvclock.o obj-$(CONFIG_PCSPKR_PLATFORM) += pcspeaker.o diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 4e456bd955b..9983bc3f5d1 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -728,14 +728,3 @@ void __cpuinit cpu_init(void) mxcsr_feature_mask_init(); } -#ifdef CONFIG_HOTPLUG_CPU -void __cpuinit cpu_uninit(void) -{ - int cpu = raw_smp_processor_id(); - cpu_clear(cpu, cpu_initialized); - - /* lazy TLB state */ - per_cpu(cpu_tlbstate, cpu).state = 0; - per_cpu(cpu_tlbstate, cpu).active_mm = &init_mm; -} -#endif diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c index b68e21f06f4..6e388412a85 100644 --- a/arch/x86/kernel/ldt.c +++ b/arch/x86/kernel/ldt.c @@ -51,6 +51,8 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload) memset(newldt + oldsize * LDT_ENTRY_SIZE, 0, (mincount - oldsize) * LDT_ENTRY_SIZE); + paravirt_alloc_ldt(newldt, mincount); + #ifdef CONFIG_X86_64 /* CHECKME: Do we really need this ? */ wmb(); @@ -73,6 +75,7 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload) #endif } if (oldsize) { + paravirt_free_ldt(oldldt, oldsize); if (oldsize * LDT_ENTRY_SIZE > PAGE_SIZE) vfree(oldldt); else @@ -84,10 +87,13 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload) static inline int copy_ldt(mm_context_t *new, mm_context_t *old) { int err = alloc_ldt(new, old->size, 0); + int i; if (err < 0) return err; - memcpy(new->ldt, old->ldt, old->size * LDT_ENTRY_SIZE); + + for(i = 0; i < old->size; i++) + write_ldt_entry(new->ldt, i, old->ldt + i * LDT_ENTRY_SIZE); return 0; } @@ -124,6 +130,7 @@ void destroy_context(struct mm_struct *mm) if (mm == current->active_mm) clear_LDT(); #endif + paravirt_free_ldt(mm->context.ldt, mm->context.size); if (mm->context.size * LDT_ENTRY_SIZE > PAGE_SIZE) vfree(mm->context.ldt); else diff --git a/arch/x86/kernel/paravirt-spinlocks.c b/arch/x86/kernel/paravirt-spinlocks.c new file mode 100644 index 00000000000..0e9f1982b1d --- /dev/null +++ b/arch/x86/kernel/paravirt-spinlocks.c @@ -0,0 +1,37 @@ +/* + * Split spinlock implementation out into its own file, so it can be + * compiled in a FTRACE-compatible way. + */ +#include <linux/spinlock.h> +#include <linux/module.h> + +#include <asm/paravirt.h> + +static void default_spin_lock_flags(struct raw_spinlock *lock, unsigned long flags) +{ + __raw_spin_lock(lock); +} + +struct pv_lock_ops pv_lock_ops = { +#ifdef CONFIG_SMP + .spin_is_locked = __ticket_spin_is_locked, + .spin_is_contended = __ticket_spin_is_contended, + + .spin_lock = __ticket_spin_lock, + .spin_lock_flags = default_spin_lock_flags, + .spin_trylock = __ticket_spin_trylock, + .spin_unlock = __ticket_spin_unlock, +#endif +}; +EXPORT_SYMBOL(pv_lock_ops); + +void __init paravirt_use_bytelocks(void) +{ +#ifdef CONFIG_SMP + pv_lock_ops.spin_is_locked = __byte_spin_is_locked; + pv_lock_ops.spin_is_contended = __byte_spin_is_contended; + pv_lock_ops.spin_lock = __byte_spin_lock; + pv_lock_ops.spin_trylock = __byte_spin_trylock; + pv_lock_ops.spin_unlock = __byte_spin_unlock; +#endif +} diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c index 300da17e61c..7faea1817d0 100644 --- a/arch/x86/kernel/paravirt.c +++ b/arch/x86/kernel/paravirt.c @@ -268,17 +268,6 @@ enum paravirt_lazy_mode paravirt_get_lazy_mode(void) return __get_cpu_var(paravirt_lazy_mode); } -void __init paravirt_use_bytelocks(void) -{ -#ifdef CONFIG_SMP - pv_lock_ops.spin_is_locked = __byte_spin_is_locked; - pv_lock_ops.spin_is_contended = __byte_spin_is_contended; - pv_lock_ops.spin_lock = __byte_spin_lock; - pv_lock_ops.spin_trylock = __byte_spin_trylock; - pv_lock_ops.spin_unlock = __byte_spin_unlock; -#endif -} - struct pv_info pv_info = { .name = "bare hardware", .paravirt_enabled = 0, @@ -348,6 +337,10 @@ struct pv_cpu_ops pv_cpu_ops = { .write_ldt_entry = native_write_ldt_entry, .write_gdt_entry = native_write_gdt_entry, .write_idt_entry = native_write_idt_entry, + + .alloc_ldt = paravirt_nop, + .free_ldt = paravirt_nop, + .load_sp0 = native_load_sp0, #if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION) @@ -461,18 +454,6 @@ struct pv_mmu_ops pv_mmu_ops = { .set_fixmap = native_set_fixmap, }; -struct pv_lock_ops pv_lock_ops = { -#ifdef CONFIG_SMP - .spin_is_locked = __ticket_spin_is_locked, - .spin_is_contended = __ticket_spin_is_contended, - - .spin_lock = __ticket_spin_lock, - .spin_trylock = __ticket_spin_trylock, - .spin_unlock = __ticket_spin_unlock, -#endif -}; -EXPORT_SYMBOL(pv_lock_ops); - EXPORT_SYMBOL_GPL(pv_time_ops); EXPORT_SYMBOL (pv_cpu_ops); EXPORT_SYMBOL (pv_mmu_ops); diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 4b3cfdf5421..b76b38ff962 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -72,47 +72,12 @@ unsigned long thread_saved_pc(struct task_struct *tsk) return ((unsigned long *)tsk->thread.sp)[3]; } -#ifdef CONFIG_HOTPLUG_CPU -#include <asm/nmi.h> - -static void cpu_exit_clear(void) -{ - int cpu = raw_smp_processor_id(); - - idle_task_exit(); - - cpu_uninit(); - irq_ctx_exit(cpu); - - cpu_clear(cpu, cpu_callout_map); - cpu_clear(cpu, cpu_callin_map); - - numa_remove_cpu(cpu); - c1e_remove_cpu(cpu); -} - -/* We don't actually take CPU down, just spin without interrupts. */ -static inline void play_dead(void) -{ - /* This must be done before dead CPU ack */ - cpu_exit_clear(); - mb(); - /* Ack it */ - __get_cpu_var(cpu_state) = CPU_DEAD; - - /* - * With physical CPU hotplug, we should halt the cpu - */ - local_irq_disable(); - /* mask all interrupts, flush any and all caches, and halt */ - wbinvd_halt(); -} -#else +#ifndef CONFIG_SMP static inline void play_dead(void) { BUG(); } -#endif /* CONFIG_HOTPLUG_CPU */ +#endif /* * The idle thread. There's no useful work to be diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index e12e0e4dd25..ec27afa43d7 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -85,30 +85,12 @@ void exit_idle(void) __exit_idle(); } -#ifdef CONFIG_HOTPLUG_CPU -DECLARE_PER_CPU(int, cpu_state); - -#include <asm/nmi.h> -/* We halt the CPU with physical CPU hotplug */ -static inline void play_dead(void) -{ - idle_task_exit(); - c1e_remove_cpu(raw_smp_processor_id()); - - mb(); - /* Ack it */ - __get_cpu_var(cpu_state) = CPU_DEAD; - - local_irq_disable(); - /* mask all interrupts, flush any and all caches, and halt */ - wbinvd_halt(); -} -#else +#ifndef CONFIG_SMP static inline void play_dead(void) { BUG(); } -#endif /* CONFIG_HOTPLUG_CPU */ +#endif /* * The idle thread. There's no useful work to be diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 361b7a4c640..18f9b19f5f8 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -214,12 +214,16 @@ void smp_call_function_single_interrupt(struct pt_regs *regs) struct smp_ops smp_ops = { .smp_prepare_boot_cpu = native_smp_prepare_boot_cpu, .smp_prepare_cpus = native_smp_prepare_cpus, - .cpu_up = native_cpu_up, .smp_cpus_done = native_smp_cpus_done, .smp_send_stop = native_smp_send_stop, .smp_send_reschedule = native_smp_send_reschedule, + .cpu_up = native_cpu_up, + .cpu_die = native_cpu_die, + .cpu_disable = native_cpu_disable, + .play_dead = native_play_dead, + .send_call_func_ipi = native_send_call_func_ipi, .send_call_func_single_ipi = native_send_call_func_single_ipi, }; diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7985c5b3f91..06f1407d554 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -52,6 +52,7 @@ #include <asm/desc.h> #include <asm/nmi.h> #include <asm/irq.h> +#include <asm/idle.h> #include <asm/smp.h> #include <asm/trampoline.h> #include <asm/cpu.h> @@ -1346,25 +1347,9 @@ static void __ref remove_cpu_from_maps(int cpu) numa_remove_cpu(cpu); } -int __cpu_disable(void) +void cpu_disable_common(void) { int cpu = smp_processor_id(); - - /* - * Perhaps use cpufreq to drop frequency, but that could go - * into generic code. - * - * We won't take down the boot processor on i386 due to some - * interrupts only being able to be serviced by the BSP. - * Especially so if we're not using an IOAPIC -zwane - */ - if (cpu == 0) - return -EBUSY; - - if (nmi_watchdog == NMI_LOCAL_APIC) - stop_apic_nmi_watchdog(NULL); - clear_local_APIC(); - /* * HACK: * Allow any queued timer interrupts to get serviced @@ -1382,10 +1367,32 @@ int __cpu_disable(void) remove_cpu_from_maps(cpu); unlock_vector_lock(); fixup_irqs(cpu_online_map); +} + +int native_cpu_disable(void) +{ + int cpu = smp_processor_id(); + + /* + * Perhaps use cpufreq to drop frequency, but that could go + * into generic code. + * + * We won't take down the boot processor on i386 due to some + * interrupts only being able to be serviced by the BSP. + * Especially so if we're not using an IOAPIC -zwane + */ + if (cpu == 0) + return -EBUSY; + + if (nmi_watchdog == NMI_LOCAL_APIC) + stop_apic_nmi_watchdog(NULL); + clear_local_APIC(); + + cpu_disable_common(); return 0; } -void __cpu_die(unsigned int cpu) +void native_cpu_die(unsigned int cpu) { /* We don't do anything here: idle task is faking death itself. */ unsigned int i; @@ -1402,15 +1409,45 @@ void __cpu_die(unsigned int cpu) } printk(KERN_ERR "CPU %u didn't die...\n", cpu); } + +void play_dead_common(void) +{ + idle_task_exit(); + reset_lazy_tlbstate(); + irq_ctx_exit(raw_smp_processor_id()); + c1e_remove_cpu(raw_smp_processor_id()); + + mb(); + /* Ack it */ + __get_cpu_var(cpu_state) = CPU_DEAD; + + /* + * With physical CPU hotplug, we should halt the cpu + */ + local_irq_disable(); +} + +void native_play_dead(void) +{ + play_dead_common(); + wbinvd_halt(); +} + #else /* ... !CONFIG_HOTPLUG_CPU */ -int __cpu_disable(void) +int native_cpu_disable(void) { return -ENOSYS; } -void __cpu_die(unsigned int cpu) +void native_cpu_die(unsigned int cpu) { /* We said "no" in __cpu_disable */ BUG(); } + +void native_play_dead(void) +{ + BUG(); +} + #endif diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c index fec1ecedc9b..e00534b3353 100644 --- a/arch/x86/kernel/tlb_32.c +++ b/arch/x86/kernel/tlb_32.c @@ -241,3 +241,11 @@ void flush_tlb_all(void) on_each_cpu(do_flush_tlb_all, NULL, 1); } +void reset_lazy_tlbstate(void) +{ + int cpu = raw_smp_processor_id(); + + per_cpu(cpu_tlbstate, cpu).state = 0; + per_cpu(cpu_tlbstate, cpu).active_mm = &init_mm; +} + diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 455f3fe67b4..356ed2dec3a 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -915,15 +915,15 @@ LIST_HEAD(pgd_list); void vmalloc_sync_all(void) { -#ifdef CONFIG_X86_32 - unsigned long start = VMALLOC_START & PGDIR_MASK; unsigned long address; +#ifdef CONFIG_X86_32 if (SHARED_KERNEL_PMD) return; - BUILD_BUG_ON(TASK_SIZE & ~PGDIR_MASK); - for (address = start; address >= TASK_SIZE; address += PGDIR_SIZE) { + for (address = VMALLOC_START & PMD_MASK; + address >= TASK_SIZE && address < FIXADDR_TOP; + address += PMD_SIZE) { unsigned long flags; struct page *page; @@ -936,10 +936,8 @@ void vmalloc_sync_all(void) spin_unlock_irqrestore(&pgd_lock, flags); } #else /* CONFIG_X86_64 */ - unsigned long start = VMALLOC_START & PGDIR_MASK; - unsigned long address; - - for (address = start; address <= VMALLOC_END; address += PGDIR_SIZE) { + for (address = VMALLOC_START & PGDIR_MASK; address <= VMALLOC_END; + address += PGDIR_SIZE) { const pgd_t *pgd_ref = pgd_offset_k(address); unsigned long flags; struct page *page; diff --git a/arch/x86/xen/Kconfig b/arch/x86/xen/Kconfig index 3815e425f47..d3e68465ace 100644 --- a/arch/x86/xen/Kconfig +++ b/arch/x86/xen/Kconfig @@ -27,4 +27,12 @@ config XEN_MAX_DOMAIN_MEMORY config XEN_SAVE_RESTORE bool depends on PM - default y
\ No newline at end of file + default y + +config XEN_DEBUG_FS + bool "Enable Xen debug and tuning parameters in debugfs" + depends on XEN && DEBUG_FS + default n + help + Enable statistics output and various tuning options in debugfs. + Enabling this option may incur a significant performance overhead.
\ No newline at end of file diff --git a/arch/x86/xen/Makefile b/arch/x86/xen/Makefile index 59c1e539aed..313947940a1 100644 --- a/arch/x86/xen/Makefile +++ b/arch/x86/xen/Makefile @@ -1,4 +1,12 @@ -obj-y := enlighten.o setup.o multicalls.o mmu.o \ +ifdef CONFIG_FTRACE +# Do not profile debug and lowlevel utilities +CFLAGS_REMOVE_spinlock.o = -pg +CFLAGS_REMOVE_time.o = -pg +CFLAGS_REMOVE_irq.o = -pg +endif + +obj-y := enlighten.o setup.o multicalls.o mmu.o irq.o \ time.o xen-asm_$(BITS).o grant-table.o suspend.o -obj-$(CONFIG_SMP) += smp.o +obj-$(CONFIG_SMP) += smp.o spinlock.o +obj-$(CONFIG_XEN_DEBUG_FS) += debugfs.o
\ No newline at end of file diff --git a/arch/x86/xen/debugfs.c b/arch/x86/xen/debugfs.c new file mode 100644 index 00000000000..b53225d2cac --- /dev/null +++ b/arch/x86/xen/debugfs.c @@ -0,0 +1,123 @@ +#include <linux/init.h> +#include <linux/debugfs.h> +#include <linux/module.h> + +#include "debugfs.h" + +static struct dentry *d_xen_debug; + +struct dentry * __init xen_init_debugfs(void) +{ + if (!d_xen_debug) { + d_xen_debug = debugfs_create_dir("xen", NULL); + + if (!d_xen_debug) + pr_warning("Could not create 'xen' debugfs directory\n"); + } + + return d_xen_debug; +} + +struct array_data +{ + void *array; + unsigned elements; +}; + +static int u32_array_open(struct inode *inode, struct file *file) +{ + file->private_data = NULL; + return nonseekable_open(inode, file); +} + +static size_t format_array(char *buf, size_t bufsize, const char *fmt, + u32 *array, unsigned array_size) +{ + size_t ret = 0; + unsigned i; + + for(i = 0; i < array_size; i++) { + size_t len; + + len = snprintf(buf, bufsize, fmt, array[i]); + len++; /* ' ' or '\n' */ + ret += len; + + if (buf) { + buf += len; + bufsize -= len; + buf[-1] = (i == array_size-1) ? '\n' : ' '; + } + } + + ret++; /* \0 */ + if (buf) + *buf = '\0'; + + return ret; +} + +static char *format_array_alloc(const char *fmt, u32 *array, unsigned array_size) +{ + size_t len = format_array(NULL, 0, fmt, array, array_size); + char *ret; + + ret = kmalloc(len, GFP_KERNEL); + if (ret == NULL) + return NULL; + + format_array(ret, len, fmt, array, array_size); + return ret; +} + +static ssize_t u32_array_read(struct file *file, char __user *buf, size_t len, + loff_t *ppos) +{ + struct inode *inode = file->f_path.dentry->d_inode; + struct array_data *data = inode->i_private; + size_t size; + + if (*ppos == 0) { + if (file->private_data) { + kfree(file->private_data); + file->private_data = NULL; + } + + file->private_data = format_array_alloc("%u", data->array, data->elements); + } + + size = 0; + if (file->private_data) + size = strlen(file->private_data); + + return simple_read_from_buffer(buf, len, ppos, file->private_data, size); +} + +static int xen_array_release(struct inode *inode, struct file *file) +{ + kfree(file->private_data); + + return 0; +} + +static struct file_operations u32_array_fops = { + .owner = THIS_MODULE, + .open = u32_array_open, + .release= xen_array_release, + .read = u32_array_read, +}; + +struct dentry *xen_debugfs_create_u32_array(const char *name, mode_t mode, + struct dentry *parent, + u32 *array, unsigned elements) +{ + struct array_data *data = kmalloc(sizeof(*data), GFP_KERNEL); + + if (data == NULL) + return NULL; + + data->array = array; + data->elements = elements; + + return debugfs_create_file(name, mode, parent, data, &u32_array_fops); +} diff --git a/arch/x86/xen/debugfs.h b/arch/x86/xen/debugfs.h new file mode 100644 index 00000000000..e2813208483 --- /dev/null +++ b/arch/x86/xen/debugfs.h @@ -0,0 +1,10 @@ +#ifndef _XEN_DEBUGFS_H +#define _XEN_DEBUGFS_H + +struct dentry * __init xen_init_debugfs(void); + +struct dentry *xen_debugfs_create_u32_array(const char *name, mode_t mode, + struct dentry *parent, + u32 *array, unsigned elements); + +#endif /* _XEN_DEBUGFS_H */ diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index a4e201b47f6..8ca2f88bde1 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -30,7 +30,6 @@ #include <xen/interface/xen.h> #include <xen/interface/physdev.h> #include <xen/interface/vcpu.h> -#include <xen/interface/sched.h> #include <xen/features.h> #include <xen/page.h> #include <xen/hvc-console.h> @@ -57,6 +56,9 @@ EXPORT_SYMBOL_GPL(hypercall_page); DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu); DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info); +enum xen_domain_type xen_domain_type = XEN_NATIVE; +EXPORT_SYMBOL_GPL(xen_domain_type); + /* * Identity map, in addition to plain kernel map. This needs to be * large enough to allocate page table pages to allocate the rest. @@ -226,103 +228,68 @@ static unsigned long xen_get_debugreg(int reg) return HYPERVISOR_get_debugreg(reg); } -static unsigned long xen_save_fl(void) +static void xen_leave_lazy(void) { - struct vcpu_info *vcpu; - unsigned long flags; - - vcpu = x86_read_percpu(xen_vcpu); - - /* flag has opposite sense of mask */ - flags = !vcpu->evtchn_upcall_mask; - - /* convert to IF type flag - -0 -> 0x00000000 - -1 -> 0xffffffff - */ - return (-flags) & X86_EFLAGS_IF; + paravirt_leave_lazy(paravirt_get_lazy_mode()); + xen_mc_flush(); } -static void xen_restore_fl(unsigned long flags) +static unsigned long xen_store_tr(void) { - struct vcpu_info *vcpu; - - /* convert from IF type flag */ - flags = !(flags & X86_EFLAGS_IF); - - /* There's a one instruction preempt window here. We need to - make sure we're don't switch CPUs between getting the vcpu - pointer and updating the mask. */ - preempt_disable(); - vcpu = x86_read_percpu(xen_vcpu); - vcpu->evtchn_upcall_mask = flags; - preempt_enable_no_resched(); - - /* Doesn't matter if we get preempted here, because any - pending event will get dealt with anyway. */ - - if (flags == 0) { - preempt_check_resched(); - barrier(); /* unmask then check (avoid races) */ - if (unlikely(vcpu->evtchn_upcall_pending)) - force_evtchn_callback(); - } + return 0; } -static void xen_irq_disable(void) +/* + * Set the page permissions for a particular virtual address. If the + * address is a vmalloc mapping (or other non-linear mapping), then + * find the linear mapping of the page and also set its protections to + * match. + */ +static void set_aliased_prot(void *v, pgprot_t prot) { - /* There's a one instruction preempt window here. We need to - make sure we're don't switch CPUs between getting the vcpu - pointer and updating the mask. */ - preempt_disable(); - x86_read_percpu(xen_vcpu)->evtchn_upcall_mask = 1; - preempt_enable_no_resched(); -} + int level; + pte_t *ptep; + pte_t pte; + unsigned long pfn; + struct page *page; -static void xen_irq_enable(void) -{ - struct vcpu_info *vcpu; + ptep = lookup_address((unsigned long)v, &level); + BUG_ON(ptep == NULL); - /* We don't need to worry about being preempted here, since - either a) interrupts are disabled, so no preemption, or b) - the caller is confused and is trying to re-enable interrupts - on an indeterminate processor. */ + pfn = pte_pfn(*ptep); + page = pfn_to_page(pfn); - vcpu = x86_read_percpu(xen_vcpu); - vcpu->evtchn_upcall_mask = 0; + pte = pfn_pte(pfn, prot); - /* Doesn't matter if we get preempted here, because any - pending event will get dealt with anyway. */ + if (HYPERVISOR_update_va_mapping((unsigned long)v, pte, 0)) + BUG(); - barrier(); /* unmask then check (avoid races) */ - if (unlikely(vcpu->evtchn_upcall_pending)) - force_evtchn_callback(); -} + if (!PageHighMem(page)) { + void *av = __va(PFN_PHYS(pfn)); -static void xen_safe_halt(void) -{ - /* Blocking includes an implicit local_irq_enable(). */ - if (HYPERVISOR_sched_op(SCHEDOP_block, NULL) != 0) - BUG(); + if (av != v) + if (HYPERVISOR_update_va_mapping((unsigned long)av, pte, 0)) + BUG(); + } else + kmap_flush_unused(); } -static void xen_halt(void) +static void xen_alloc_ldt(struct desc_struct *ldt, unsigned entries) { - if (irqs_disabled()) - HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL); - else - xen_safe_halt(); -} + const unsigned entries_per_page = PAGE_SIZE / LDT_ENTRY_SIZE; + int i; -static void xen_leave_lazy(void) -{ - paravirt_leave_lazy(paravirt_get_lazy_mode()); - xen_mc_flush(); + for(i = 0; i < entries; i += entries_per_page) + set_aliased_prot(ldt + i, PAGE_KERNEL_RO); } -static unsigned long xen_store_tr(void) +static void xen_free_ldt(struct desc_struct *ldt, unsigned entries) { - return 0; + const unsigned entries_per_page = PAGE_SIZE / LDT_ENTRY_SIZE; + int i; + + for(i = 0; i < entries; i += entries_per_page) + set_aliased_prot(ldt + i, PAGE_KERNEL); } static void xen_set_ldt(const void *addr, unsigned entries) @@ -425,8 +392,7 @@ static void xen_load_gs_index(unsigned int idx) static void xen_write_ldt_entry(struct desc_struct *dt, int entrynum, const void *ptr) { - unsigned long lp = (unsigned long)&dt[entrynum]; - xmaddr_t mach_lp = virt_to_machine(lp); + xmaddr_t mach_lp = arbitrary_virt_to_machine(&dt[entrynum]); u64 entry = *(u64 *)ptr; preempt_disable(); @@ -559,7 +525,7 @@ static void xen_write_gdt_entry(struct desc_struct *dt, int entry, } static void xen_load_sp0(struct tss_struct *tss, - struct thread_struct *thread) + struct thread_struct *thread) { struct multicall_space mcs = xen_mc_entry(0); MULTI_stack_switch(mcs.mc, __KERNEL_DS, thread->sp0); @@ -803,6 +769,19 @@ static int xen_write_msr_safe(unsigned int msr, unsigned low, unsigned high) ret = -EFAULT; break; #endif + + case MSR_STAR: + case MSR_CSTAR: + case MSR_LSTAR: + case MSR_SYSCALL_MASK: + case MSR_IA32_SYSENTER_CS: + case MSR_IA32_SYSENTER_ESP: + case MSR_IA32_SYSENTER_EIP: + /* Fast syscall setup is all done in hypercalls, so + these are all ignored. Stub them out here to stop + Xen console noise. */ + break; + default: ret = native_write_msr_safe(msr, low, high); } @@ -846,8 +825,8 @@ static void xen_alloc_ptpage(struct mm_struct *mm, u32 pfn, unsigned level) SetPagePinned(page); if (!PageHighMem(page)) { - make_lowmem_page_readonly(__va(PFN_PHYS(pfn))); - if (level == PT_PTE) + make_lowmem_page_readonly(__va(PFN_PHYS((unsigned long)pfn))); + if (level == PT_PTE && USE_SPLIT_PTLOCKS) pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn); } else /* make sure there are no stray mappings of @@ -915,7 +894,7 @@ static void xen_release_ptpage(u32 pfn, unsigned level) if (PagePinned(page)) { if (!PageHighMem(page)) { - if (level == PT_PTE) + if (level == PT_PTE && USE_SPLIT_PTLOCKS) pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn); make_lowmem_page_readwrite(__va(PFN_PHYS(pfn))); } @@ -1220,6 +1199,9 @@ static const struct pv_cpu_ops xen_cpu_ops __initdata = { .load_gs_index = xen_load_gs_index, #endif + .alloc_ldt = xen_alloc_ldt, + .free_ldt = xen_free_ldt, + .store_gdt = native_store_gdt, .store_idt = native_store_idt, .store_tr = xen_store_tr, @@ -1241,36 +1223,6 @@ static const struct pv_cpu_ops xen_cpu_ops __initdata = { }, }; -static void __init __xen_init_IRQ(void) -{ -#ifdef CONFIG_X86_64 - int i; - - /* Create identity vector->irq map */ - for(i = 0; i < NR_VECTORS; i++) { - int cpu; - - for_each_possible_cpu(cpu) - per_cpu(vector_irq, cpu)[i] = i; - } -#endif /* CONFIG_X86_64 */ - - xen_init_IRQ(); -} - -static const struct pv_irq_ops xen_irq_ops __initdata = { - .init_IRQ = __xen_init_IRQ, - .save_fl = xen_save_fl, - .restore_fl = xen_restore_fl, - .irq_disable = xen_irq_disable, - .irq_enable = xen_irq_enable, - .safe_halt = xen_safe_halt, - .halt = xen_halt, -#ifdef CONFIG_X86_64 - .adjust_exception_frame = xen_adjust_exception_frame, -#endif -}; - static const struct pv_apic_ops xen_apic_ops __initdata = { #ifdef CONFIG_X86_LOCAL_APIC .apic_write = xen_apic_write, @@ -1664,6 +1616,8 @@ asmlinkage void __init xen_start_kernel(void) if (!xen_start_info) return; + xen_domain_type = XEN_PV_DOMAIN; + BUG_ON(memcmp(xen_start_info->magic, "xen-3", 5) != 0); xen_setup_features(); @@ -1673,10 +1627,11 @@ asmlinkage void __init xen_start_kernel(void) pv_init_ops = xen_init_ops; pv_time_ops = xen_time_ops; pv_cpu_ops = xen_cpu_ops; - pv_irq_ops = xen_irq_ops; pv_apic_ops = xen_apic_ops; pv_mmu_ops = xen_mmu_ops; + xen_init_irq_ops(); + if (xen_feature(XENFEAT_mmu_pt_update_preserve_ad)) { pv_mmu_ops.ptep_modify_prot_start = xen_ptep_modify_prot_start; pv_mmu_ops.ptep_modify_prot_commit = xen_ptep_modify_prot_commit; @@ -1700,7 +1655,7 @@ asmlinkage void __init xen_start_kernel(void) /* Prevent unwanted bits from being set in PTEs. */ __supported_pte_mask &= ~_PAGE_GLOBAL; - if (!is_initial_xendomain()) + if (!xen_initial_domain()) __supported_pte_mask &= ~(_PAGE_PWT | _PAGE_PCD); /* Don't do the full vcpu_info placement stuff until we have a @@ -1735,7 +1690,7 @@ asmlinkage void __init xen_start_kernel(void) boot_params.hdr.ramdisk_size = xen_start_info->mod_len; boot_params.hdr.cmd_line_ptr = __pa(xen_start_info->cmd_line); - if (!is_initial_xendomain()) { + if (!xen_initial_domain()) { add_preferred_console("xenboot", 0, NULL); add_preferred_console("tty", 0, NULL); add_preferred_console("hvc", 0, NULL); diff --git a/arch/x86/xen/irq.c b/arch/x86/xen/irq.c new file mode 100644 index 00000000000..28b85ab8422 --- /dev/null +++ b/arch/x86/xen/irq.c @@ -0,0 +1,143 @@ +#include <linux/hardirq.h> + +#include <xen/interface/xen.h> +#include <xen/interface/sched.h> +#include <xen/interface/vcpu.h> + +#include <asm/xen/hypercall.h> +#include <asm/xen/hypervisor.h> + +#include "xen-ops.h" + +/* + * Force a proper event-channel callback from Xen after clearing the + * callback mask. We do this in a very simple manner, by making a call + * down into Xen. The pending flag will be checked by Xen on return. + */ +void xen_force_evtchn_callback(void) +{ + (void)HYPERVISOR_xen_version(0, NULL); +} + +static void __init __xen_init_IRQ(void) +{ +#ifdef CONFIG_X86_64 + int i; + + /* Create identity vector->irq map */ + for(i = 0; i < NR_VECTORS; i++) { + int cpu; + + for_each_possible_cpu(cpu) + per_cpu(vector_irq, cpu)[i] = i; + } +#endif /* CONFIG_X86_64 */ + + xen_init_IRQ(); +} + +static unsigned long xen_save_fl(void) +{ + struct vcpu_info *vcpu; + unsigned long flags; + + vcpu = x86_read_percpu(xen_vcpu); + + /* flag has opposite sense of mask */ + flags = !vcpu->evtchn_upcall_mask; + + /* convert to IF type flag + -0 -> 0x00000000 + -1 -> 0xffffffff + */ + return (-flags) & X86_EFLAGS_IF; +} + +static void xen_restore_fl(unsigned long flags) +{ + struct vcpu_info *vcpu; + + /* convert from IF type flag */ + flags = !(flags & X86_EFLAGS_IF); + + /* There's a one instruction preempt window here. We need to + make sure we're don't switch CPUs between getting the vcpu + pointer and updating the mask. */ + preempt_disable(); + vcpu = x86_read_percpu(xen_vcpu); + vcpu->evtchn_upcall_mask = flags; + preempt_enable_no_resched(); + + /* Doesn't matter if we get preempted here, because any + pending event will get dealt with anyway. */ + + if (flags == 0) { + preempt_check_resched(); + barrier(); /* unmask then check (avoid races) */ + if (unlikely(vcpu->evtchn_upcall_pending)) + xen_force_evtchn_callback(); + } +} + +static void xen_irq_disable(void) +{ + /* There's a one instruction preempt window here. We need to + make sure we're don't switch CPUs between getting the vcpu + pointer and updating the mask. */ + preempt_disable(); + x86_read_percpu(xen_vcpu)->evtchn_upcall_mask = 1; + preempt_enable_no_resched(); +} + +static void xen_irq_enable(void) +{ + struct vcpu_info *vcpu; + + /* We don't need to worry about being preempted here, since + either a) interrupts are disabled, so no preemption, or b) + the caller is confused and is trying to re-enable interrupts + on an indeterminate processor. */ + + vcpu = x86_read_percpu(xen_vcpu); + vcpu->evtchn_upcall_mask = 0; + + /* Doesn't matter if we get preempted here, because any + pending event will get dealt with anyway. */ + + barrier(); /* unmask then check (avoid races) */ + if (unlikely(vcpu->evtchn_upcall_pending)) + xen_force_evtchn_callback(); +} + +static void xen_safe_halt(void) +{ + /* Blocking includes an implicit local_irq_enable(). */ + if (HYPERVISOR_sched_op(SCHEDOP_block, NULL) != 0) + BUG(); +} + +static void xen_halt(void) +{ + if (irqs_disabled()) + HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL); + else + xen_safe_halt(); +} + +static const struct pv_irq_ops xen_irq_ops __initdata = { + .init_IRQ = __xen_init_IRQ, + .save_fl = xen_save_fl, + .restore_fl = xen_restore_fl, + .irq_disable = xen_irq_disable, + .irq_enable = xen_irq_enable, + .safe_halt = xen_safe_halt, + .halt = xen_halt, +#ifdef CONFIG_X86_64 + .adjust_exception_frame = xen_adjust_exception_frame, +#endif +}; + +void __init xen_init_irq_ops() +{ + pv_irq_ops = xen_irq_ops; +} diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index aa37469da69..64e58681767 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -40,6 +40,7 @@ */ #include <linux/sched.h> #include <linux/highmem.h> +#include <linux/debugfs.h> #include <linux/bug.h> #include <asm/pgtable.h> @@ -57,6 +58,61 @@ #include "multicalls.h" #include "mmu.h" +#include "debugfs.h" + +#define MMU_UPDATE_HISTO 30 + +#ifdef CONFIG_XEN_DEBUG_FS + +static struct { + u32 pgd_update; + u32 pgd_update_pinned; + u32 pgd_update_batched; + + u32 pud_update; + u32 pud_update_pinned; + u32 pud_update_batched; + + u32 pmd_update; + u32 pmd_update_pinned; + u32 pmd_update_batched; + + u32 pte_update; + u32 pte_update_pinned; + u32 pte_update_batched; + + u32 mmu_update; + u32 mmu_update_extended; + u32 mmu_update_histo[MMU_UPDATE_HISTO]; + + u32 prot_commit; + u32 prot_commit_batched; + + u32 set_pte_at; + u32 set_pte_at_batched; + u32 set_pte_at_pinned; + u32 set_pte_at_current; + u32 set_pte_at_kernel; +} mmu_stats; + +static u8 zero_stats; + +static inline void check_zero(void) +{ + if (unlikely(zero_stats)) { + memset(&mmu_stats, 0, sizeof(mmu_stats)); + zero_stats = 0; + } +} + +#define ADD_STATS(elem, val) \ + do { check_zero(); mmu_stats.elem += (val); } while(0) + +#else /* !CONFIG_XEN_DEBUG_FS */ + +#define ADD_STATS(elem, val) do { (void)(val); } while(0) + +#endif /* CONFIG_XEN_DEBUG_FS */ /* * Just beyond the highest usermode address. STACK_TOP_MAX has a @@ -229,25 +285,35 @@ void make_lowmem_page_readwrite(void *vaddr) } -static bool page_pinned(void *ptr) +static bool xen_page_pinned(void *ptr) { struct page *page = virt_to_page(ptr); return PagePinned(page); } -static void extend_mmu_update(const struct mmu_update *update) +static void xen_extend_mmu_update(const struct mmu_update *update) { struct multicall_space mcs; struct mmu_update *u; mcs = xen_mc_extend_args(__HYPERVISOR_mmu_update, sizeof(*u)); - if (mcs.mc != NULL) + if (mcs.mc != NULL) { + ADD_STATS(mmu_update_extended, 1); + ADD_STATS(mmu_update_histo[mcs.mc->args[1]], -1); + mcs.mc->args[1]++; - else { + + if (mcs.mc->args[1] < MMU_UPDATE_HISTO) + ADD_STATS(mmu_update_histo[mcs.mc->args[1]], 1); + else + ADD_STATS(mmu_update_histo[0], 1); + } else { + ADD_STATS(mmu_update, 1); mcs = __xen_mc_entry(sizeof(*u)); MULTI_mmu_update(mcs.mc, mcs.args, 1, NULL, DOMID_SELF); + ADD_STATS(mmu_update_histo[1], 1); } u = mcs.args; @@ -265,7 +331,9 @@ void xen_set_pmd_hyper(pmd_t *ptr, pmd_t val) /* ptr may be ioremapped for 64-bit pagetable setup */ u.ptr = arbitrary_virt_to_machine(ptr).maddr; u.val = pmd_val_ma(val); - extend_mmu_update(&u); + xen_extend_mmu_update(&u); + + ADD_STATS(pmd_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU); xen_mc_issue(PARAVIRT_LAZY_MMU); @@ -274,13 +342,17 @@ void xen_set_pmd_hyper(pmd_t *ptr, pmd_t val) void xen_set_pmd(pmd_t *ptr, pmd_t val) { + ADD_STATS(pmd_update, 1); + /* If page is not pinned, we can just update the entry directly */ - if (!page_pinned(ptr)) { + if (!xen_page_pinned(ptr)) { *ptr = val; return; } + ADD_STATS(pmd_update_pinned, 1); + xen_set_pmd_hyper(ptr, val); } @@ -300,12 +372,18 @@ void xen_set_pte_at(struct mm_struct *mm, unsigned long addr, if (mm == &init_mm) preempt_disable(); + ADD_STATS(set_pte_at, 1); +// ADD_STATS(set_pte_at_pinned, xen_page_pinned(ptep)); + ADD_STATS(set_pte_at_current, mm == current->mm); + ADD_STATS(set_pte_at_kernel, mm == &init_mm); + if (mm == current->mm || mm == &init_mm) { if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU) { struct multicall_space mcs; mcs = xen_mc_entry(0); MULTI_update_va_mapping(mcs.mc, addr, pteval, 0); + ADD_STATS(set_pte_at_batched, 1); xen_mc_issue(PARAVIRT_LAZY_MMU); goto out; } else @@ -334,7 +412,10 @@ void xen_ptep_modify_prot_commit(struct mm_struct *mm, unsigned long addr, u.ptr = virt_to_machine(ptep).maddr | MMU_PT_UPDATE_PRESERVE_AD; u.val = pte_val_ma(pte); - extend_mmu_update(&u); + xen_extend_mmu_update(&u); + + ADD_STATS(prot_commit, 1); + ADD_STATS(prot_commit_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU); xen_mc_issue(PARAVIRT_LAZY_MMU); } @@ -400,7 +481,9 @@ void xen_set_pud_hyper(pud_t *ptr, pud_t val) /* ptr may be ioremapped for 64-bit pagetable setup */ u.ptr = arbitrary_virt_to_machine(ptr).maddr; u.val = pud_val_ma(val); - extend_mmu_update(&u); + xen_extend_mmu_update(&u); + + ADD_STATS(pud_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU); xen_mc_issue(PARAVIRT_LAZY_MMU); @@ -409,18 +492,26 @@ void xen_set_pud_hyper(pud_t *ptr, pud_t val) void xen_set_pud(pud_t *ptr, pud_t val) { + ADD_STATS(pud_update, 1); + /* If page is not pinned, we can just update the entry directly */ - if (!page_pinned(ptr)) { + if (!xen_page_pinned(ptr)) { *ptr = val; return; } + ADD_STATS(pud_update_pinned, 1); + xen_set_pud_hyper(ptr, val); } void xen_set_pte(pte_t *ptep, pte_t pte) { + ADD_STATS(pte_update, 1); +// ADD_STATS(pte_update_pinned, xen_page_pinned(ptep)); + ADD_STATS(pte_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU); + #ifdef CONFIG_X86_PAE ptep->pte_high = pte.pte_high; smp_wmb(); @@ -490,7 +581,7 @@ static void __xen_set_pgd_hyper(pgd_t *ptr, pgd_t val) u.ptr = virt_to_machine(ptr).maddr; u.val = pgd_val_ma(val); - extend_mmu_update(&u); + xen_extend_mmu_update(&u); } /* @@ -517,17 +608,22 @@ void xen_set_pgd(pgd_t *ptr, pgd_t val) { pgd_t *user_ptr = xen_get_user_pgd(ptr); + ADD_STATS(pgd_update, 1); + /* If page is not pinned, we can just update the entry directly */ - if (!page_pinned(ptr)) { + if (!xen_page_pinned(ptr)) { *ptr = val; if (user_ptr) { - WARN_ON(page_pinned(user_ptr)); + WARN_ON(xen_page_pinned(user_ptr)); *user_ptr = val; } return; } + ADD_STATS(pgd_update_pinned, 1); + ADD_STATS(pgd_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU); + /* If it's pinned, then we can at least batch the kernel and user updates together. */ xen_mc_batch(); @@ -555,8 +651,8 @@ void xen_set_pgd(pgd_t *ptr, pgd_t val) * For 64-bit, we must skip the Xen hole in the middle of the address * space, just after the big x86-64 virtual hole. */ -static int pgd_walk(pgd_t *pgd, int (*func)(struct page *, enum pt_level), - unsigned long limit) +static int xen_pgd_walk(pgd_t *pgd, int (*func)(struct page *, enum pt_level), + unsigned long limit) { int flush = 0; unsigned hole_low, hole_high; @@ -590,8 +686,6 @@ static int pgd_walk(pgd_t *pgd, int (*func)(struct page *, enum pt_level), pmdidx_limit = 0; #endif - flush |= (*func)(virt_to_page(pgd), PT_PGD); - for (pgdidx = 0; pgdidx <= pgdidx_limit; pgdidx++) { pud_t *pud; @@ -637,16 +731,22 @@ static int pgd_walk(pgd_t *pgd, int (*func)(struct page *, enum pt_level), } } } + out: + /* Do the top level last, so that the callbacks can use it as + a cue to do final things like tlb flushes. */ + flush |= (*func)(virt_to_page(pgd), PT_PGD); return flush; } -static spinlock_t *lock_pte(struct page *page) +/* If we're using split pte locks, then take the page's lock and + return a pointer to it. Otherwise return NULL. */ +static spinlock_t *xen_pte_lock(struct page *page) { spinlock_t *ptl = NULL; -#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#if USE_SPLIT_PTLOCKS ptl = __pte_lockptr(page); spin_lock(ptl); #endif @@ -654,7 +754,7 @@ static spinlock_t *lock_pte(struct page *page) return ptl; } -static void do_unlock(void *v) +static void xen_pte_unlock(void *v) { spinlock_t *ptl = v; spin_unlock(ptl); @@ -672,7 +772,7 @@ static void xen_do_pin(unsigned level, unsigned long pfn) MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); } -static int pin_page(struct page *page, enum pt_level level) +static int xen_pin_page(struct page *page, enum pt_level level) { unsigned pgfl = TestSetPagePinned(page); int flush; @@ -691,21 +791,40 @@ static int pin_page(struct page *page, enum pt_level level) flush = 0; + /* + * We need to hold the pagetable lock between the time + * we make the pagetable RO and when we actually pin + * it. If we don't, then other users may come in and + * attempt to update the pagetable by writing it, + * which will fail because the memory is RO but not + * pinned, so Xen won't do the trap'n'emulate. + * + * If we're using split pte locks, we can't hold the + * entire pagetable's worth of locks during the + * traverse, because we may wrap the preempt count (8 + * bits). The solution is to mark RO and pin each PTE + * page while holding the lock. This means the number + * of locks we end up holding is never more than a + * batch size (~32 entries, at present). + * + * If we're not using split pte locks, we needn't pin + * the PTE pages independently, because we're + * protected by the overall pagetable lock. + */ ptl = NULL; if (level == PT_PTE) - ptl = lock_pte(page); + ptl = xen_pte_lock(page); MULTI_update_va_mapping(mcs.mc, (unsigned long)pt, pfn_pte(pfn, PAGE_KERNEL_RO), level == PT_PGD ? UVMF_TLB_FLUSH : 0); - if (level == PT_PTE) + if (ptl) { xen_do_pin(MMUEXT_PIN_L1_TABLE, pfn); - if (ptl) { /* Queue a deferred unlock for when this batch is completed. */ - xen_mc_callback(do_unlock, ptl); + xen_mc_callback(xen_pte_unlock, ptl); } } @@ -719,7 +838,7 @@ void xen_pgd_pin(pgd_t *pgd) { xen_mc_batch(); - if (pgd_walk(pgd, pin_page, USER_LIMIT)) { + if (xen_pgd_walk(pgd, xen_pin_page, USER_LIMIT)) { /* re-enable interrupts for kmap_flush_unused */ xen_mc_issue(0); kmap_flush_unused(); @@ -733,14 +852,14 @@ void xen_pgd_pin(pgd_t *pgd) xen_do_pin(MMUEXT_PIN_L4_TABLE, PFN_DOWN(__pa(pgd))); if (user_pgd) { - pin_page(virt_to_page(user_pgd), PT_PGD); + xen_pin_page(virt_to_page(user_pgd), PT_PGD); xen_do_pin(MMUEXT_PIN_L4_TABLE, PFN_DOWN(__pa(user_pgd))); } } #else /* CONFIG_X86_32 */ #ifdef CONFIG_X86_PAE /* Need to make sure unshared kernel PMD is pinnable */ - pin_page(virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), PT_PMD); + xen_pin_page(virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), PT_PMD); #endif xen_do_pin(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(pgd))); #endif /* CONFIG_X86_64 */ @@ -775,7 +894,7 @@ void xen_mm_pin_all(void) * that's before we have page structures to store the bits. So do all * the book-keeping now. */ -static __init int mark_pinned(struct page *page, enum pt_level level) +static __init int xen_mark_pinned(struct page *page, enum pt_level level) { SetPagePinned(page); return 0; @@ -783,10 +902,10 @@ static __init int mark_pinned(struct page *page, enum pt_level level) void __init xen_mark_init_mm_pinned(void) { - pgd_walk(init_mm.pgd, mark_pinned, FIXADDR_TOP); + xen_pgd_walk(init_mm.pgd, xen_mark_pinned, FIXADDR_TOP); } -static int unpin_page(struct page *page, enum pt_level level) +static int xen_unpin_page(struct page *page, enum pt_level level) { unsigned pgfl = TestClearPagePinned(page); @@ -796,10 +915,18 @@ static int unpin_page(struct page *page, enum pt_level level) spinlock_t *ptl = NULL; struct multicall_space mcs; + /* + * Do the converse to pin_page. If we're using split + * pte locks, we must be holding the lock for while + * the pte page is unpinned but still RO to prevent + * concurrent updates from seeing it in this + * partially-pinned state. + */ if (level == PT_PTE) { - ptl = lock_pte(page); + ptl = xen_pte_lock(page); - xen_do_pin(MMUEXT_UNPIN_TABLE, pfn); + if (ptl) + xen_do_pin(MMUEXT_UNPIN_TABLE, pfn); } mcs = __xen_mc_entry(0); @@ -810,7 +937,7 @@ static int unpin_page(struct page *page, enum pt_level level) if (ptl) { /* unlock when batch completed */ - xen_mc_callback(do_unlock, ptl); + xen_mc_callback(xen_pte_unlock, ptl); } } @@ -830,17 +957,17 @@ static void xen_pgd_unpin(pgd_t *pgd) if (user_pgd) { xen_do_pin(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(user_pgd))); - unpin_page(virt_to_page(user_pgd), PT_PGD); + xen_unpin_page(virt_to_page(user_pgd), PT_PGD); } } #endif #ifdef CONFIG_X86_PAE /* Need to make sure unshared kernel PMD is unpinned */ - pin_page(virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), PT_PMD); + xen_unpin_page(virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), PT_PMD); #endif - pgd_walk(pgd, unpin_page, USER_LIMIT); + xen_pgd_walk(pgd, xen_unpin_page, USER_LIMIT); xen_mc_issue(0); } @@ -907,7 +1034,7 @@ static void drop_other_mm_ref(void *info) } } -static void drop_mm_ref(struct mm_struct *mm) +static void xen_drop_mm_ref(struct mm_struct *mm) { cpumask_t mask; unsigned cpu; @@ -937,7 +1064,7 @@ static void drop_mm_ref(struct mm_struct *mm) smp_call_function_mask(mask, drop_other_mm_ref, mm, 1); } #else -static void drop_mm_ref(struct mm_struct *mm) +static void xen_drop_mm_ref(struct mm_struct *mm) { if (current->active_mm == mm) load_cr3(swapper_pg_dir); @@ -961,14 +1088,77 @@ static void drop_mm_ref(struct mm_struct *mm) void xen_exit_mmap(struct mm_struct *mm) { get_cpu(); /* make sure we don't move around */ - drop_mm_ref(mm); + xen_drop_mm_ref(mm); put_cpu(); spin_lock(&mm->page_table_lock); /* pgd may not be pinned in the error exit path of execve */ - if (page_pinned(mm->pgd)) + if (xen_page_pinned(mm->pgd)) xen_pgd_unpin(mm->pgd); spin_unlock(&mm->page_table_lock); } + +#ifdef CONFIG_XEN_DEBUG_FS + +static struct dentry *d_mmu_debug; + +static int __init xen_mmu_debugfs(void) +{ + struct dentry *d_xen = xen_init_debugfs(); + + if (d_xen == NULL) + return -ENOMEM; + + d_mmu_debug = debugfs_create_dir("mmu", d_xen); + + debugfs_create_u8("zero_stats", 0644, d_mmu_debug, &zero_stats); + + debugfs_create_u32("pgd_update", 0444, d_mmu_debug, &mmu_stats.pgd_update); + debugfs_create_u32("pgd_update_pinned", 0444, d_mmu_debug, + &mmu_stats.pgd_update_pinned); + debugfs_create_u32("pgd_update_batched", 0444, d_mmu_debug, + &mmu_stats.pgd_update_pinned); + + debugfs_create_u32("pud_update", 0444, d_mmu_debug, &mmu_stats.pud_update); + debugfs_create_u32("pud_update_pinned", 0444, d_mmu_debug, + &mmu_stats.pud_update_pinned); + debugfs_create_u32("pud_update_batched", 0444, d_mmu_debug, + &mmu_stats.pud_update_pinned); + + debugfs_create_u32("pmd_update", 0444, d_mmu_debug, &mmu_stats.pmd_update); + debugfs_create_u32("pmd_update_pinned", 0444, d_mmu_debug, + &mmu_stats.pmd_update_pinned); + debugfs_create_u32("pmd_update_batched", 0444, d_mmu_debug, + &mmu_stats.pmd_update_pinned); + + debugfs_create_u32("pte_update", 0444, d_mmu_debug, &mmu_stats.pte_update); +// debugfs_create_u32("pte_update_pinned", 0444, d_mmu_debug, +// &mmu_stats.pte_update_pinned); + debugfs_create_u32("pte_update_batched", 0444, d_mmu_debug, + &mmu_stats.pte_update_pinned); + + debugfs_create_u32("mmu_update", 0444, d_mmu_debug, &mmu_stats.mmu_update); + debugfs_create_u32("mmu_update_extended", 0444, d_mmu_debug, + &mmu_stats.mmu_update_extended); + xen_debugfs_create_u32_array("mmu_update_histo", 0444, d_mmu_debug, + mmu_stats.mmu_update_histo, 20); + + debugfs_create_u32("set_pte_at", 0444, d_mmu_debug, &mmu_stats.set_pte_at); + debugfs_create_u32("set_pte_at_batched", 0444, d_mmu_debug, + &mmu_stats.set_pte_at_batched); + debugfs_create_u32("set_pte_at_current", 0444, d_mmu_debug, + &mmu_stats.set_pte_at_current); + debugfs_create_u32("set_pte_at_kernel", 0444, d_mmu_debug, + &mmu_stats.set_pte_at_kernel); + + debugfs_create_u32("prot_commit", 0444, d_mmu_debug, &mmu_stats.prot_commit); + debugfs_create_u32("prot_commit_batched", 0444, d_mmu_debug, + &mmu_stats.prot_commit_batched); + + return 0; +} +fs_initcall(xen_mmu_debugfs); + +#endif /* CONFIG_XEN_DEBUG_FS */ diff --git a/arch/x86/xen/multicalls.c b/arch/x86/xen/multicalls.c index 9efd1c6c977..8ea8a0d0b0d 100644 --- a/arch/x86/xen/multicalls.c +++ b/arch/x86/xen/multicalls.c @@ -21,16 +21,20 @@ */ #include <linux/percpu.h> #include <linux/hardirq.h> +#include <linux/debugfs.h> #include <asm/xen/hypercall.h> #include "multicalls.h" +#include "debugfs.h" + +#define MC_BATCH 32 #define MC_DEBUG 1 -#define MC_BATCH 32 #define MC_ARGS (MC_BATCH * 16) + struct mc_buffer { struct multicall_entry entries[MC_BATCH]; #if MC_DEBUG @@ -47,6 +51,76 @@ struct mc_buffer { static DEFINE_PER_CPU(struct mc_buffer, mc_buffer); DEFINE_PER_CPU(unsigned long, xen_mc_irq_flags); +/* flush reasons 0- slots, 1- args, 2- callbacks */ +enum flush_reasons +{ + FL_SLOTS, + FL_ARGS, + FL_CALLBACKS, + + FL_N_REASONS +}; + +#ifdef CONFIG_XEN_DEBUG_FS +#define NHYPERCALLS 40 /* not really */ + +static struct { + unsigned histo[MC_BATCH+1]; + + unsigned issued; + unsigned arg_total; + unsigned hypercalls; + unsigned histo_hypercalls[NHYPERCALLS]; + + unsigned flush[FL_N_REASONS]; +} mc_stats; + +static u8 zero_stats; + +static inline void check_zero(void) +{ + if (unlikely(zero_stats)) { + memset(&mc_stats, 0, sizeof(mc_stats)); + zero_stats = 0; + } +} + +static void mc_add_stats(const struct mc_buffer *mc) +{ + int i; + + check_zero(); + + mc_stats.issued++; + mc_stats.hypercalls += mc->mcidx; + mc_stats.arg_total += mc->argidx; + + mc_stats.histo[mc->mcidx]++; + for(i = 0; i < mc->mcidx; i++) { + unsigned op = mc->entries[i].op; + if (op < NHYPERCALLS) + mc_stats.histo_hypercalls[op]++; + } +} + +static void mc_stats_flush(enum flush_reasons idx) +{ + check_zero(); + + mc_stats.flush[idx]++; +} + +#else /* !CONFIG_XEN_DEBUG_FS */ + +static inline void mc_add_stats(const struct mc_buffer *mc) +{ +} + +static inline void mc_stats_flush(enum flush_reasons idx) +{ +} +#endif /* CONFIG_XEN_DEBUG_FS */ + void xen_mc_flush(void) { struct mc_buffer *b = &__get_cpu_var(mc_buffer); @@ -60,6 +134,8 @@ void xen_mc_flush(void) something in the middle */ local_irq_save(flags); + mc_add_stats(b); + if (b->mcidx) { #if MC_DEBUG memcpy(b->debug, b->entries, @@ -115,6 +191,7 @@ struct multicall_space __xen_mc_entry(size_t args) if (b->mcidx == MC_BATCH || (argidx + args) > MC_ARGS) { + mc_stats_flush(b->mcidx == MC_BATCH ? FL_SLOTS : FL_ARGS); xen_mc_flush(); argidx = roundup(b->argidx, sizeof(u64)); } @@ -158,10 +235,44 @@ void xen_mc_callback(void (*fn)(void *), void *data) struct mc_buffer *b = &__get_cpu_var(mc_buffer); struct callback *cb; - if (b->cbidx == MC_BATCH) + if (b->cbidx == MC_BATCH) { + mc_stats_flush(FL_CALLBACKS); xen_mc_flush(); + } cb = &b->callbacks[b->cbidx++]; cb->fn = fn; cb->data = data; } + +#ifdef CONFIG_XEN_DEBUG_FS + +static struct dentry *d_mc_debug; + +static int __init xen_mc_debugfs(void) +{ + struct dentry *d_xen = xen_init_debugfs(); + + if (d_xen == NULL) + return -ENOMEM; + + d_mc_debug = debugfs_create_dir("multicalls", d_xen); + + debugfs_create_u8("zero_stats", 0644, d_mc_debug, &zero_stats); + + debugfs_create_u32("batches", 0444, d_mc_debug, &mc_stats.issued); + debugfs_create_u32("hypercalls", 0444, d_mc_debug, &mc_stats.hypercalls); + debugfs_create_u32("arg_total", 0444, d_mc_debug, &mc_stats.arg_total); + + xen_debugfs_create_u32_array("batch_histo", 0444, d_mc_debug, + mc_stats.histo, MC_BATCH); + xen_debugfs_create_u32_array("hypercall_histo", 0444, d_mc_debug, + mc_stats.histo_hypercalls, NHYPERCALLS); + xen_debugfs_create_u32_array("flush_reasons", 0444, d_mc_debug, + mc_stats.flush, FL_N_REASONS); + + return 0; +} +fs_initcall(xen_mc_debugfs); + +#endif /* CONFIG_XEN_DEBUG_FS */ diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index d8faf79a0a1..d77da613b1d 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -11,11 +11,8 @@ * useful topology information for the kernel to make use of. As a * result, all CPUs are treated as if they're single-core and * single-threaded. - * - * This does not handle HOTPLUG_CPU yet. */ #include <linux/sched.h> -#include <linux/kernel_stat.h> #include <linux/err.h> #include <linux/smp.h> @@ -36,8 +33,6 @@ #include "xen-ops.h" #include "mmu.h" -static void __cpuinit xen_init_lock_cpu(int cpu); - cpumask_t xen_cpu_initialized_map; static DEFINE_PER_CPU(int, resched_irq); @@ -64,11 +59,12 @@ static irqreturn_t xen_reschedule_interrupt(int irq, void *dev_id) return IRQ_HANDLED; } -static __cpuinit void cpu_bringup_and_idle(void) +static __cpuinit void cpu_bringup(void) { int cpu = smp_processor_id(); cpu_init(); + touch_softlockup_watchdog(); preempt_disable(); xen_enable_sysenter(); @@ -89,6 +85,11 @@ static __cpuinit void cpu_bringup_and_idle(void) local_irq_enable(); wmb(); /* make sure everything is out */ +} + +static __cpuinit void cpu_bringup_and_idle(void) +{ + cpu_bringup(); cpu_idle(); } @@ -212,8 +213,6 @@ static void __init xen_smp_prepare_cpus(unsigned int max_cpus) cpu_set(cpu, cpu_present_map); } - - //init_xenbus_allowed_cpumask(); } static __cpuinit int @@ -281,12 +280,6 @@ static int __cpuinit xen_cpu_up(unsigned int cpu) struct task_struct *idle = idle_task(cpu); int rc; -#if 0 - rc = cpu_up_check(cpu); - if (rc) - return rc; -#endif - #ifdef CONFIG_X86_64 /* Allocate node local memory for AP pdas */ WARN_ON(cpu == 0); @@ -339,6 +332,60 @@ static void xen_smp_cpus_done(unsigned int max_cpus) { } +#ifdef CONFIG_HOTPLUG_CPU +static int xen_cpu_disable(void) +{ + unsigned int cpu = smp_processor_id(); + if (cpu == 0) + return -EBUSY; + + cpu_disable_common(); + + load_cr3(swapper_pg_dir); + return 0; +} + +static void xen_cpu_die(unsigned int cpu) +{ + while (HYPERVISOR_vcpu_op(VCPUOP_is_up, cpu, NULL)) { + current->state = TASK_UNINTERRUPTIBLE; + schedule_timeout(HZ/10); + } + unbind_from_irqhandler(per_cpu(resched_irq, cpu), NULL); + unbind_from_irqhandler(per_cpu(callfunc_irq, cpu), NULL); + unbind_from_irqhandler(per_cpu(debug_irq, cpu), NULL); + unbind_from_irqhandler(per_cpu(callfuncsingle_irq, cpu), NULL); + xen_uninit_lock_cpu(cpu); + xen_teardown_timer(cpu); + + if (num_online_cpus() == 1) + alternatives_smp_switch(0); +} + +static void xen_play_dead(void) +{ + play_dead_common(); + HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL); + cpu_bringup(); +} + +#else /* !CONFIG_HOTPLUG_CPU */ +static int xen_cpu_disable(void) +{ + return -ENOSYS; +} + +static void xen_cpu_die(unsigned int cpu) +{ + BUG(); +} + +static void xen_play_dead(void) +{ + BUG(); +} + +#endif static void stop_self(void *v) { int cpu = smp_processor_id(); @@ -419,176 +466,16 @@ static irqreturn_t xen_call_function_single_interrupt(int irq, void *dev_id) return IRQ_HANDLED; } -struct xen_spinlock { - unsigned char lock; /* 0 -> free; 1 -> locked */ - unsigned short spinners; /* count of waiting cpus */ -}; - -static int xen_spin_is_locked(struct raw_spinlock *lock) -{ - struct xen_spinlock *xl = (struct xen_spinlock *)lock; - - return xl->lock != 0; -} - -static int xen_spin_is_contended(struct raw_spinlock *lock) -{ - struct xen_spinlock *xl = (struct xen_spinlock *)lock; - - /* Not strictly true; this is only the count of contended - lock-takers entering the slow path. */ - return xl->spinners != 0; -} - -static int xen_spin_trylock(struct raw_spinlock *lock) -{ - struct xen_spinlock *xl = (struct xen_spinlock *)lock; - u8 old = 1; - - asm("xchgb %b0,%1" - : "+q" (old), "+m" (xl->lock) : : "memory"); - - return old == 0; -} - -static DEFINE_PER_CPU(int, lock_kicker_irq) = -1; -static DEFINE_PER_CPU(struct xen_spinlock *, lock_spinners); - -static inline void spinning_lock(struct xen_spinlock *xl) -{ - __get_cpu_var(lock_spinners) = xl; - wmb(); /* set lock of interest before count */ - asm(LOCK_PREFIX " incw %0" - : "+m" (xl->spinners) : : "memory"); -} - -static inline void unspinning_lock(struct xen_spinlock *xl) -{ - asm(LOCK_PREFIX " decw %0" - : "+m" (xl->spinners) : : "memory"); - wmb(); /* decrement count before clearing lock */ - __get_cpu_var(lock_spinners) = NULL; -} - -static noinline int xen_spin_lock_slow(struct raw_spinlock *lock) -{ - struct xen_spinlock *xl = (struct xen_spinlock *)lock; - int irq = __get_cpu_var(lock_kicker_irq); - int ret; - - /* If kicker interrupts not initialized yet, just spin */ - if (irq == -1) - return 0; - - /* announce we're spinning */ - spinning_lock(xl); - - /* clear pending */ - xen_clear_irq_pending(irq); - - /* check again make sure it didn't become free while - we weren't looking */ - ret = xen_spin_trylock(lock); - if (ret) - goto out; - - /* block until irq becomes pending */ - xen_poll_irq(irq); - kstat_this_cpu.irqs[irq]++; - -out: - unspinning_lock(xl); - return ret; -} - -static void xen_spin_lock(struct raw_spinlock *lock) -{ - struct xen_spinlock *xl = (struct xen_spinlock *)lock; - int timeout; - u8 oldval; - - do { - timeout = 1 << 10; - - asm("1: xchgb %1,%0\n" - " testb %1,%1\n" - " jz 3f\n" - "2: rep;nop\n" - " cmpb $0,%0\n" - " je 1b\n" - " dec %2\n" - " jnz 2b\n" - "3:\n" - : "+m" (xl->lock), "=q" (oldval), "+r" (timeout) - : "1" (1) - : "memory"); - - } while (unlikely(oldval != 0 && !xen_spin_lock_slow(lock))); -} - -static noinline void xen_spin_unlock_slow(struct xen_spinlock *xl) -{ - int cpu; - - for_each_online_cpu(cpu) { - /* XXX should mix up next cpu selection */ - if (per_cpu(lock_spinners, cpu) == xl) { - xen_send_IPI_one(cpu, XEN_SPIN_UNLOCK_VECTOR); - break; - } - } -} - -static void xen_spin_unlock(struct raw_spinlock *lock) -{ - struct xen_spinlock *xl = (struct xen_spinlock *)lock; - - smp_wmb(); /* make sure no writes get moved after unlock */ - xl->lock = 0; /* release lock */ - - /* make sure unlock happens before kick */ - barrier(); - - if (unlikely(xl->spinners)) - xen_spin_unlock_slow(xl); -} - -static __cpuinit void xen_init_lock_cpu(int cpu) -{ - int irq; - const char *name; - - name = kasprintf(GFP_KERNEL, "spinlock%d", cpu); - irq = bind_ipi_to_irqhandler(XEN_SPIN_UNLOCK_VECTOR, - cpu, - xen_reschedule_interrupt, - IRQF_DISABLED|IRQF_PERCPU|IRQF_NOBALANCING, - name, - NULL); - - if (irq >= 0) { - disable_irq(irq); /* make sure it's never delivered */ - per_cpu(lock_kicker_irq, cpu) = irq; - } - - printk("cpu %d spinlock event irq %d\n", cpu, irq); -} - -static void __init xen_init_spinlocks(void) -{ - pv_lock_ops.spin_is_locked = xen_spin_is_locked; - pv_lock_ops.spin_is_contended = xen_spin_is_contended; - pv_lock_ops.spin_lock = xen_spin_lock; - pv_lock_ops.spin_trylock = xen_spin_trylock; - pv_lock_ops.spin_unlock = xen_spin_unlock; -} - 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, + .cpu_up = xen_cpu_up, + .cpu_die = xen_cpu_die, + .cpu_disable = xen_cpu_disable, + .play_dead = xen_play_dead, + .smp_send_stop = xen_smp_send_stop, .smp_send_reschedule = xen_smp_send_reschedule, diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c new file mode 100644 index 00000000000..dd71e3a021c --- /dev/null +++ b/arch/x86/xen/spinlock.c @@ -0,0 +1,428 @@ +/* + * Split spinlock implementation out into its own file, so it can be + * compiled in a FTRACE-compatible way. + */ +#include <linux/kernel_stat.h> +#include <linux/spinlock.h> +#include <linux/debugfs.h> +#include <linux/log2.h> + +#include <asm/paravirt.h> + +#include <xen/interface/xen.h> +#include <xen/events.h> + +#include "xen-ops.h" +#include "debugfs.h" + +#ifdef CONFIG_XEN_DEBUG_FS +static struct xen_spinlock_stats +{ + u64 taken; + u32 taken_slow; + u32 taken_slow_nested; + u32 taken_slow_pickup; + u32 taken_slow_spurious; + u32 taken_slow_irqenable; + + u64 released; + u32 released_slow; + u32 released_slow_kicked; + +#define HISTO_BUCKETS 30 + u32 histo_spin_total[HISTO_BUCKETS+1]; + u32 histo_spin_spinning[HISTO_BUCKETS+1]; + u32 histo_spin_blocked[HISTO_BUCKETS+1]; + + u64 time_total; + u64 time_spinning; + u64 time_blocked; +} spinlock_stats; + +static u8 zero_stats; + +static unsigned lock_timeout = 1 << 10; +#define TIMEOUT lock_timeout + +static inline void check_zero(void) +{ + if (unlikely(zero_stats)) { + memset(&spinlock_stats, 0, sizeof(spinlock_stats)); + zero_stats = 0; + } +} + +#define ADD_STATS(elem, val) \ + do { check_zero(); spinlock_stats.elem += (val); } while(0) + +static inline u64 spin_time_start(void) +{ + return xen_clocksource_read(); +} + +static void __spin_time_accum(u64 delta, u32 *array) +{ + unsigned index = ilog2(delta); + + check_zero(); + + if (index < HISTO_BUCKETS) + array[index]++; + else + array[HISTO_BUCKETS]++; +} + +static inline void spin_time_accum_spinning(u64 start) +{ + u32 delta = xen_clocksource_read() - start; + + __spin_time_accum(delta, spinlock_stats.histo_spin_spinning); + spinlock_stats.time_spinning += delta; +} + +static inline void spin_time_accum_total(u64 start) +{ + u32 delta = xen_clocksource_read() - start; + + __spin_time_accum(delta, spinlock_stats.histo_spin_total); + spinlock_stats.time_total += delta; +} + +static inline void spin_time_accum_blocked(u64 start) +{ + u32 delta = xen_clocksource_read() - start; + + __spin_time_accum(delta, spinlock_stats.histo_spin_blocked); + spinlock_stats.time_blocked += delta; +} +#else /* !CONFIG_XEN_DEBUG_FS */ +#define TIMEOUT (1 << 10) +#define ADD_STATS(elem, val) do { (void)(val); } while(0) + +static inline u64 spin_time_start(void) +{ + return 0; +} + +static inline void spin_time_accum_total(u64 start) +{ +} +static inline void spin_time_accum_spinning(u64 start) +{ +} +static inline void spin_time_accum_blocked(u64 start) +{ +} +#endif /* CONFIG_XEN_DEBUG_FS */ + +struct xen_spinlock { + unsigned char lock; /* 0 -> free; 1 -> locked */ + unsigned short spinners; /* count of waiting cpus */ +}; + +static int xen_spin_is_locked(struct raw_spinlock *lock) +{ + struct xen_spinlock *xl = (struct xen_spinlock *)lock; + + return xl->lock != 0; +} + +static int xen_spin_is_contended(struct raw_spinlock *lock) +{ + struct xen_spinlock *xl = (struct xen_spinlock *)lock; + + /* Not strictly true; this is only the count of contended + lock-takers entering the slow path. */ + return xl->spinners != 0; +} + +static int xen_spin_trylock(struct raw_spinlock *lock) +{ + struct xen_spinlock *xl = (struct xen_spinlock *)lock; + u8 old = 1; + + asm("xchgb %b0,%1" + : "+q" (old), "+m" (xl->lock) : : "memory"); + + return old == 0; +} + +static DEFINE_PER_CPU(int, lock_kicker_irq) = -1; +static DEFINE_PER_CPU(struct xen_spinlock *, lock_spinners); + +/* + * Mark a cpu as interested in a lock. Returns the CPU's previous + * lock of interest, in case we got preempted by an interrupt. + */ +static inline struct xen_spinlock *spinning_lock(struct xen_spinlock *xl) +{ + struct xen_spinlock *prev; + + prev = __get_cpu_var(lock_spinners); + __get_cpu_var(lock_spinners) = xl; + + wmb(); /* set lock of interest before count */ + + asm(LOCK_PREFIX " incw %0" + : "+m" (xl->spinners) : : "memory"); + + return prev; +} + +/* + * Mark a cpu as no longer interested in a lock. Restores previous + * lock of interest (NULL for none). + */ +static inline void unspinning_lock(struct xen_spinlock *xl, struct xen_spinlock *prev) +{ + asm(LOCK_PREFIX " decw %0" + : "+m" (xl->spinners) : : "memory"); + wmb(); /* decrement count before restoring lock */ + __get_cpu_var(lock_spinners) = prev; +} + +static noinline int xen_spin_lock_slow(struct raw_spinlock *lock, bool irq_enable) +{ + struct xen_spinlock *xl = (struct xen_spinlock *)lock; + struct xen_spinlock *prev; + int irq = __get_cpu_var(lock_kicker_irq); + int ret; + unsigned long flags; + u64 start; + + /* If kicker interrupts not initialized yet, just spin */ + if (irq == -1) + return 0; + + start = spin_time_start(); + + /* announce we're spinning */ + prev = spinning_lock(xl); + + flags = __raw_local_save_flags(); + if (irq_enable) { + ADD_STATS(taken_slow_irqenable, 1); + raw_local_irq_enable(); + } + + ADD_STATS(taken_slow, 1); + ADD_STATS(taken_slow_nested, prev != NULL); + + do { + /* clear pending */ + xen_clear_irq_pending(irq); + + /* check again make sure it didn't become free while + we weren't looking */ + ret = xen_spin_trylock(lock); + if (ret) { + ADD_STATS(taken_slow_pickup, 1); + + /* + * If we interrupted another spinlock while it + * was blocking, make sure it doesn't block + * without rechecking the lock. + */ + if (prev != NULL) + xen_set_irq_pending(irq); + goto out; + } + + /* + * Block until irq becomes pending. If we're + * interrupted at this point (after the trylock but + * before entering the block), then the nested lock + * handler guarantees that the irq will be left + * pending if there's any chance the lock became free; + * xen_poll_irq() returns immediately if the irq is + * pending. + */ + xen_poll_irq(irq); + ADD_STATS(taken_slow_spurious, !xen_test_irq_pending(irq)); + } while (!xen_test_irq_pending(irq)); /* check for spurious wakeups */ + + kstat_this_cpu.irqs[irq]++; + +out: + raw_local_irq_restore(flags); + unspinning_lock(xl, prev); + spin_time_accum_blocked(start); + + return ret; +} + +static inline void __xen_spin_lock(struct raw_spinlock *lock, bool irq_enable) +{ + struct xen_spinlock *xl = (struct xen_spinlock *)lock; + unsigned timeout; + u8 oldval; + u64 start_spin; + + ADD_STATS(taken, 1); + + start_spin = spin_time_start(); + + do { + u64 start_spin_fast = spin_time_start(); + + timeout = TIMEOUT; + + asm("1: xchgb %1,%0\n" + " testb %1,%1\n" + " jz 3f\n" + "2: rep;nop\n" + " cmpb $0,%0\n" + " je 1b\n" + " dec %2\n" + " jnz 2b\n" + "3:\n" + : "+m" (xl->lock), "=q" (oldval), "+r" (timeout) + : "1" (1) + : "memory"); + + spin_time_accum_spinning(start_spin_fast); + + } while (unlikely(oldval != 0 && + (TIMEOUT == ~0 || !xen_spin_lock_slow(lock, irq_enable)))); + + spin_time_accum_total(start_spin); +} + +static void xen_spin_lock(struct raw_spinlock *lock) +{ + __xen_spin_lock(lock, false); +} + +static void xen_spin_lock_flags(struct raw_spinlock *lock, unsigned long flags) +{ + __xen_spin_lock(lock, !raw_irqs_disabled_flags(flags)); +} + +static noinline void xen_spin_unlock_slow(struct xen_spinlock *xl) +{ + int cpu; + + ADD_STATS(released_slow, 1); + + for_each_online_cpu(cpu) { + /* XXX should mix up next cpu selection */ + if (per_cpu(lock_spinners, cpu) == xl) { + ADD_STATS(released_slow_kicked, 1); + xen_send_IPI_one(cpu, XEN_SPIN_UNLOCK_VECTOR); + break; + } + } +} + +static void xen_spin_unlock(struct raw_spinlock *lock) +{ + struct xen_spinlock *xl = (struct xen_spinlock *)lock; + + ADD_STATS(released, 1); + + smp_wmb(); /* make sure no writes get moved after unlock */ + xl->lock = 0; /* release lock */ + + /* make sure unlock happens before kick */ + barrier(); + + if (unlikely(xl->spinners)) + xen_spin_unlock_slow(xl); +} + +static irqreturn_t dummy_handler(int irq, void *dev_id) +{ + BUG(); + return IRQ_HANDLED; +} + +void __cpuinit xen_init_lock_cpu(int cpu) +{ + int irq; + const char *name; + + name = kasprintf(GFP_KERNEL, "spinlock%d", cpu); + irq = bind_ipi_to_irqhandler(XEN_SPIN_UNLOCK_VECTOR, + cpu, + dummy_handler, + IRQF_DISABLED|IRQF_PERCPU|IRQF_NOBALANCING, + name, + NULL); + + if (irq >= 0) { + disable_irq(irq); /* make sure it's never delivered */ + per_cpu(lock_kicker_irq, cpu) = irq; + } + + printk("cpu %d spinlock event irq %d\n", cpu, irq); +} + +void xen_uninit_lock_cpu(int cpu) +{ + unbind_from_irqhandler(per_cpu(lock_kicker_irq, cpu), NULL); +} + +void __init xen_init_spinlocks(void) +{ + pv_lock_ops.spin_is_locked = xen_spin_is_locked; + pv_lock_ops.spin_is_contended = xen_spin_is_contended; + pv_lock_ops.spin_lock = xen_spin_lock; + pv_lock_ops.spin_lock_flags = xen_spin_lock_flags; + pv_lock_ops.spin_trylock = xen_spin_trylock; + pv_lock_ops.spin_unlock = xen_spin_unlock; +} + +#ifdef CONFIG_XEN_DEBUG_FS + +static struct dentry *d_spin_debug; + +static int __init xen_spinlock_debugfs(void) +{ + struct dentry *d_xen = xen_init_debugfs(); + + if (d_xen == NULL) + return -ENOMEM; + + d_spin_debug = debugfs_create_dir("spinlocks", d_xen); + + debugfs_create_u8("zero_stats", 0644, d_spin_debug, &zero_stats); + + debugfs_create_u32("timeout", 0644, d_spin_debug, &lock_timeout); + + debugfs_create_u64("taken", 0444, d_spin_debug, &spinlock_stats.taken); + debugfs_create_u32("taken_slow", 0444, d_spin_debug, + &spinlock_stats.taken_slow); + debugfs_create_u32("taken_slow_nested", 0444, d_spin_debug, + &spinlock_stats.taken_slow_nested); + debugfs_create_u32("taken_slow_pickup", 0444, d_spin_debug, + &spinlock_stats.taken_slow_pickup); + debugfs_create_u32("taken_slow_spurious", 0444, d_spin_debug, + &spinlock_stats.taken_slow_spurious); + debugfs_create_u32("taken_slow_irqenable", 0444, d_spin_debug, + &spinlock_stats.taken_slow_irqenable); + + debugfs_create_u64("released", 0444, d_spin_debug, &spinlock_stats.released); + debugfs_create_u32("released_slow", 0444, d_spin_debug, + &spinlock_stats.released_slow); + debugfs_create_u32("released_slow_kicked", 0444, d_spin_debug, + &spinlock_stats.released_slow_kicked); + + debugfs_create_u64("time_spinning", 0444, d_spin_debug, + &spinlock_stats.time_spinning); + debugfs_create_u64("time_blocked", 0444, d_spin_debug, + &spinlock_stats.time_blocked); + debugfs_create_u64("time_total", 0444, d_spin_debug, + &spinlock_stats.time_total); + + xen_debugfs_create_u32_array("histo_total", 0444, d_spin_debug, + spinlock_stats.histo_spin_total, HISTO_BUCKETS + 1); + xen_debugfs_create_u32_array("histo_spinning", 0444, d_spin_debug, + spinlock_stats.histo_spin_spinning, HISTO_BUCKETS + 1); + xen_debugfs_create_u32_array("histo_blocked", 0444, d_spin_debug, + spinlock_stats.histo_spin_blocked, HISTO_BUCKETS + 1); + + return 0; +} +fs_initcall(xen_spinlock_debugfs); + +#endif /* CONFIG_XEN_DEBUG_FS */ diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index 685b77470fc..004ba86326a 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -30,8 +30,6 @@ #define TIMER_SLOP 100000 #define NS_PER_TICK (1000000000LL / HZ) -static cycle_t xen_clocksource_read(void); - /* runstate info updated by Xen */ static DEFINE_PER_CPU(struct vcpu_runstate_info, runstate); @@ -213,7 +211,7 @@ unsigned long xen_tsc_khz(void) return xen_khz; } -static cycle_t xen_clocksource_read(void) +cycle_t xen_clocksource_read(void) { struct pvclock_vcpu_time_info *src; cycle_t ret; @@ -452,6 +450,14 @@ void xen_setup_timer(int cpu) setup_runstate_info(cpu); } +void xen_teardown_timer(int cpu) +{ + struct clock_event_device *evt; + BUG_ON(cpu == 0); + evt = &per_cpu(xen_clock_events, cpu); + unbind_from_irqhandler(evt->irq, NULL); +} + void xen_setup_cpu_clockevents(void) { BUG_ON(preemptible()); diff --git a/arch/x86/xen/xen-asm_32.S b/arch/x86/xen/xen-asm_32.S index 2497a30f41d..42786f59d9c 100644 --- a/arch/x86/xen/xen-asm_32.S +++ b/arch/x86/xen/xen-asm_32.S @@ -298,7 +298,7 @@ check_events: push %eax push %ecx push %edx - call force_evtchn_callback + call xen_force_evtchn_callback pop %edx pop %ecx pop %eax diff --git a/arch/x86/xen/xen-asm_64.S b/arch/x86/xen/xen-asm_64.S index 7f58304fafb..3b9bda46487 100644 --- a/arch/x86/xen/xen-asm_64.S +++ b/arch/x86/xen/xen-asm_64.S @@ -122,7 +122,7 @@ check_events: push %r9 push %r10 push %r11 - call force_evtchn_callback + call xen_force_evtchn_callback pop %r11 pop %r10 pop %r9 diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h index dd3c23152a2..d7422dc2a55 100644 --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -2,6 +2,7 @@ #define XEN_OPS_H #include <linux/init.h> +#include <linux/clocksource.h> #include <linux/irqreturn.h> #include <xen/xen-ops.h> @@ -31,7 +32,10 @@ void xen_vcpu_restore(void); void __init xen_build_dynamic_phys_to_machine(void); +void xen_init_irq_ops(void); void xen_setup_timer(int cpu); +void xen_teardown_timer(int cpu); +cycle_t xen_clocksource_read(void); void xen_setup_cpu_clockevents(void); unsigned long xen_tsc_khz(void); void __init xen_time_init(void); @@ -50,6 +54,10 @@ void __init xen_setup_vcpu_info_placement(void); #ifdef CONFIG_SMP void xen_smp_init(void); +void __init xen_init_spinlocks(void); +__cpuinit void xen_init_lock_cpu(int cpu); +void xen_uninit_lock_cpu(int cpu); + extern cpumask_t xen_cpu_initialized_map; #else static inline void xen_smp_init(void) {} diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index 3ca643cafcc..d5e75325515 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -1032,7 +1032,7 @@ static struct xenbus_driver blkfront = { static int __init xlblk_init(void) { - if (!is_running_on_xen()) + if (!xen_domain()) return -ENODEV; if (register_blkdev(XENVBD_MAJOR, DEV_NAME)) { diff --git a/drivers/char/hvc_xen.c b/drivers/char/hvc_xen.c index 6b70aa66a58..538ceea5e7d 100644 --- a/drivers/char/hvc_xen.c +++ b/drivers/char/hvc_xen.c @@ -108,8 +108,8 @@ static int __init xen_init(void) { struct hvc_struct *hp; - if (!is_running_on_xen() || - is_initial_xendomain() || + if (!xen_pv_domain() || + xen_initial_domain() || !xen_start_info->console.domU.evtchn) return -ENODEV; @@ -142,7 +142,7 @@ static void __exit xen_fini(void) static int xen_cons_init(void) { - if (!is_running_on_xen()) + if (!xen_pv_domain()) return 0; hvc_instantiate(HVC_COOKIE, 0, &hvc_ops); diff --git a/drivers/input/xen-kbdfront.c b/drivers/input/xen-kbdfront.c index 9ce3b3baf3a..3ab6362f043 100644 --- a/drivers/input/xen-kbdfront.c +++ b/drivers/input/xen-kbdfront.c @@ -335,11 +335,11 @@ static struct xenbus_driver xenkbd = { static int __init xenkbd_init(void) { - if (!is_running_on_xen()) + if (!xen_domain()) return -ENODEV; /* Nothing to do if running in dom0. */ - if (is_initial_xendomain()) + if (xen_initial_domain()) return -ENODEV; return xenbus_register_frontend(&xenkbd); diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c index c749bdba214..3c3dd403f5d 100644 --- a/drivers/net/xen-netfront.c +++ b/drivers/net/xen-netfront.c @@ -1794,10 +1794,10 @@ static struct xenbus_driver netfront = { static int __init netif_init(void) { - if (!is_running_on_xen()) + if (!xen_domain()) return -ENODEV; - if (is_initial_xendomain()) + if (xen_initial_domain()) return 0; printk(KERN_INFO "Initialising Xen virtual ethernet driver.\n"); @@ -1809,7 +1809,7 @@ module_init(netif_init); static void __exit netif_exit(void) { - if (is_initial_xendomain()) + if (xen_initial_domain()) return; xenbus_unregister_driver(&netfront); diff --git a/drivers/video/xen-fbfront.c b/drivers/video/xen-fbfront.c index 47ed39b52f9..a463b3dd837 100644 --- a/drivers/video/xen-fbfront.c +++ b/drivers/video/xen-fbfront.c @@ -680,11 +680,11 @@ static struct xenbus_driver xenfb = { static int __init xenfb_init(void) { - if (!is_running_on_xen()) + if (!xen_domain()) return -ENODEV; /* Nothing to do if running in dom0. */ - if (is_initial_xendomain()) + if (xen_initial_domain()) return -ENODEV; return xenbus_register_frontend(&xenfb); diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile index 363286c5429..d2a8fdf0e19 100644 --- a/drivers/xen/Makefile +++ b/drivers/xen/Makefile @@ -1,4 +1,5 @@ obj-y += grant-table.o features.o events.o manage.o obj-y += xenbus/ +obj-$(CONFIG_HOTPLUG_CPU) += cpu_hotplug.o obj-$(CONFIG_XEN_XENCOMM) += xencomm.o obj-$(CONFIG_XEN_BALLOON) += balloon.o diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c index 2e15da5459c..a51f3e17a5f 100644 --- a/drivers/xen/balloon.c +++ b/drivers/xen/balloon.c @@ -226,9 +226,8 @@ static int increase_reservation(unsigned long nr_pages) } set_xen_guest_handle(reservation.extent_start, frame_list); - reservation.nr_extents = nr_pages; - rc = HYPERVISOR_memory_op( - XENMEM_populate_physmap, &reservation); + reservation.nr_extents = nr_pages; + rc = HYPERVISOR_memory_op(XENMEM_populate_physmap, &reservation); if (rc < nr_pages) { if (rc > 0) { int ret; @@ -236,7 +235,7 @@ static int increase_reservation(unsigned long nr_pages) /* We hit the Xen hard limit: reprobe. */ reservation.nr_extents = rc; ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation, - &reservation); + &reservation); BUG_ON(ret != rc); } if (rc >= 0) @@ -420,7 +419,7 @@ static int __init balloon_init(void) unsigned long pfn; struct page *page; - if (!is_running_on_xen()) + if (!xen_pv_domain()) return -ENODEV; pr_info("xen_balloon: Initialising balloon driver.\n"); @@ -464,136 +463,13 @@ static void balloon_exit(void) module_exit(balloon_exit); -static void balloon_update_driver_allowance(long delta) -{ - unsigned long flags; - - spin_lock_irqsave(&balloon_lock, flags); - balloon_stats.driver_pages += delta; - spin_unlock_irqrestore(&balloon_lock, flags); -} - -static int dealloc_pte_fn( - pte_t *pte, struct page *pmd_page, unsigned long addr, void *data) -{ - unsigned long mfn = pte_mfn(*pte); - int ret; - struct xen_memory_reservation reservation = { - .nr_extents = 1, - .extent_order = 0, - .domid = DOMID_SELF - }; - set_xen_guest_handle(reservation.extent_start, &mfn); - set_pte_at(&init_mm, addr, pte, __pte_ma(0ull)); - set_phys_to_machine(__pa(addr) >> PAGE_SHIFT, INVALID_P2M_ENTRY); - ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation, &reservation); - BUG_ON(ret != 1); - return 0; -} - -static struct page **alloc_empty_pages_and_pagevec(int nr_pages) -{ - unsigned long vaddr, flags; - struct page *page, **pagevec; - int i, ret; - - pagevec = kmalloc(sizeof(page) * nr_pages, GFP_KERNEL); - if (pagevec == NULL) - return NULL; - - for (i = 0; i < nr_pages; i++) { - page = pagevec[i] = alloc_page(GFP_KERNEL); - if (page == NULL) - goto err; - - vaddr = (unsigned long)page_address(page); - - scrub_page(page); - - spin_lock_irqsave(&balloon_lock, flags); - - if (xen_feature(XENFEAT_auto_translated_physmap)) { - unsigned long gmfn = page_to_pfn(page); - struct xen_memory_reservation reservation = { - .nr_extents = 1, - .extent_order = 0, - .domid = DOMID_SELF - }; - set_xen_guest_handle(reservation.extent_start, &gmfn); - ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation, - &reservation); - if (ret == 1) - ret = 0; /* success */ - } else { - ret = apply_to_page_range(&init_mm, vaddr, PAGE_SIZE, - dealloc_pte_fn, NULL); - } - - if (ret != 0) { - spin_unlock_irqrestore(&balloon_lock, flags); - __free_page(page); - goto err; - } - - totalram_pages = --balloon_stats.current_pages; - - spin_unlock_irqrestore(&balloon_lock, flags); - } - - out: - schedule_work(&balloon_worker); - flush_tlb_all(); - return pagevec; - - err: - spin_lock_irqsave(&balloon_lock, flags); - while (--i >= 0) - balloon_append(pagevec[i]); - spin_unlock_irqrestore(&balloon_lock, flags); - kfree(pagevec); - pagevec = NULL; - goto out; -} - -static void free_empty_pages_and_pagevec(struct page **pagevec, int nr_pages) -{ - unsigned long flags; - int i; - - if (pagevec == NULL) - return; - - spin_lock_irqsave(&balloon_lock, flags); - for (i = 0; i < nr_pages; i++) { - BUG_ON(page_count(pagevec[i]) != 1); - balloon_append(pagevec[i]); - } - spin_unlock_irqrestore(&balloon_lock, flags); - - kfree(pagevec); - - schedule_work(&balloon_worker); -} - -static void balloon_release_driver_page(struct page *page) -{ - unsigned long flags; - - spin_lock_irqsave(&balloon_lock, flags); - balloon_append(page); - balloon_stats.driver_pages--; - spin_unlock_irqrestore(&balloon_lock, flags); - - schedule_work(&balloon_worker); -} - - -#define BALLOON_SHOW(name, format, args...) \ - static ssize_t show_##name(struct sys_device *dev, \ - char *buf) \ - { \ - return sprintf(buf, format, ##args); \ - } \ +#define BALLOON_SHOW(name, format, args...) \ + static ssize_t show_##name(struct sys_device *dev, \ + struct sysdev_attribute *attr, \ + char *buf) \ + { \ + return sprintf(buf, format, ##args); \ + } \ static SYSDEV_ATTR(name, S_IRUGO, show_##name, NULL) BALLOON_SHOW(current_kb, "%lu\n", PAGES2KB(balloon_stats.current_pages)); @@ -604,7 +480,8 @@ BALLOON_SHOW(hard_limit_kb, (balloon_stats.hard_limit!=~0UL) ? PAGES2KB(balloon_stats.hard_limit) : 0); BALLOON_SHOW(driver_kb, "%lu\n", PAGES2KB(balloon_stats.driver_pages)); -static ssize_t show_target_kb(struct sys_device *dev, char *buf) +static ssize_t show_target_kb(struct sys_device *dev, struct sysdev_attribute *attr, + char *buf) { return sprintf(buf, "%lu\n", PAGES2KB(balloon_stats.target_pages)); } @@ -614,19 +491,14 @@ static ssize_t store_target_kb(struct sys_device *dev, const char *buf, size_t count) { - char memstring[64], *endchar; + char *endchar; unsigned long long target_bytes; if (!capable(CAP_SYS_ADMIN)) return -EPERM; - if (count <= 1) - return -EBADMSG; /* runt */ - if (count > sizeof(memstring)) - return -EFBIG; /* too long */ - strcpy(memstring, buf); + target_bytes = memparse(buf, &endchar); - target_bytes = memparse(memstring, &endchar); balloon_set_new_target(target_bytes >> PAGE_SHIFT); return count; @@ -694,20 +566,4 @@ static int register_balloon(struct sys_device *sysdev) return error; } -static void unregister_balloon(struct sys_device *sysdev) -{ - int i; - - sysfs_remove_group(&sysdev->kobj, &balloon_info_group); - for (i = 0; i < ARRAY_SIZE(balloon_attrs); i++) - sysdev_remove_file(sysdev, balloon_attrs[i]); - sysdev_unregister(sysdev); - sysdev_class_unregister(&balloon_sysdev_class); -} - -static void balloon_sysfs_exit(void) -{ - unregister_balloon(&balloon_sysdev); -} - MODULE_LICENSE("GPL"); diff --git a/drivers/xen/cpu_hotplug.c b/drivers/xen/cpu_hotplug.c new file mode 100644 index 00000000000..565280ec1c6 --- /dev/null +++ b/drivers/xen/cpu_hotplug.c @@ -0,0 +1,90 @@ +#include <linux/notifier.h> + +#include <xen/xenbus.h> + +#include <asm-x86/xen/hypervisor.h> +#include <asm/cpu.h> + +static void enable_hotplug_cpu(int cpu) +{ + if (!cpu_present(cpu)) + arch_register_cpu(cpu); + + cpu_set(cpu, cpu_present_map); +} + +static void disable_hotplug_cpu(int cpu) +{ + if (cpu_present(cpu)) + arch_unregister_cpu(cpu); + + cpu_clear(cpu, cpu_present_map); +} + +static void vcpu_hotplug(unsigned int cpu) +{ + int err; + char dir[32], state[32]; + + if (!cpu_possible(cpu)) + return; + + sprintf(dir, "cpu/%u", cpu); + err = xenbus_scanf(XBT_NIL, dir, "availability", "%s", state); + if (err != 1) { + printk(KERN_ERR "XENBUS: Unable to read cpu state\n"); + return; + } + + if (strcmp(state, "online") == 0) { + enable_hotplug_cpu(cpu); + } else if (strcmp(state, "offline") == 0) { + (void)cpu_down(cpu); + disable_hotplug_cpu(cpu); + } else { + printk(KERN_ERR "XENBUS: unknown state(%s) on CPU%d\n", + state, cpu); + } +} + +static void handle_vcpu_hotplug_event(struct xenbus_watch *watch, + const char **vec, unsigned int len) +{ + unsigned int cpu; + char *cpustr; + const char *node = vec[XS_WATCH_PATH]; + + cpustr = strstr(node, "cpu/"); + if (cpustr != NULL) { + sscanf(cpustr, "cpu/%u", &cpu); + vcpu_hotplug(cpu); + } +} + +static int setup_cpu_watcher(struct notifier_block *notifier, + unsigned long event, void *data) +{ + static struct xenbus_watch cpu_watch = { + .node = "cpu", + .callback = handle_vcpu_hotplug_event}; + + (void)register_xenbus_watch(&cpu_watch); + + return NOTIFY_DONE; +} + +static int __init setup_vcpu_hotplug_event(void) +{ + static struct notifier_block xsn_cpu = { + .notifier_call = setup_cpu_watcher }; + + if (!xen_pv_domain()) + return -ENODEV; + + register_xenstore_notifier(&xsn_cpu); + + return 0; +} + +arch_initcall(setup_vcpu_hotplug_event); + diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 0e0c28574af..c3290bc186a 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -84,17 +84,6 @@ static int irq_bindcount[NR_IRQS]; /* Xen will never allocate port zero for any purpose. */ #define VALID_EVTCHN(chn) ((chn) != 0) -/* - * Force a proper event-channel callback from Xen after clearing the - * callback mask. We do this in a very simple manner, by making a call - * down into Xen. The pending flag will be checked by Xen on return. - */ -void force_evtchn_callback(void) -{ - (void)HYPERVISOR_xen_version(0, NULL); -} -EXPORT_SYMBOL_GPL(force_evtchn_callback); - static struct irq_chip xen_dynamic_chip; /* Constructor for packed IRQ information. */ @@ -175,6 +164,12 @@ static inline void set_evtchn(int port) sync_set_bit(port, &s->evtchn_pending[0]); } +static inline int test_evtchn(int port) +{ + struct shared_info *s = HYPERVISOR_shared_info; + return sync_test_bit(port, &s->evtchn_pending[0]); +} + /** * notify_remote_via_irq - send event to remote end of event channel via irq @@ -365,6 +360,10 @@ static void unbind_from_irq(unsigned int irq) per_cpu(virq_to_irq, cpu_from_evtchn(evtchn)) [index_from_irq(irq)] = -1; break; + case IRQT_IPI: + per_cpu(ipi_to_irq, cpu_from_evtchn(evtchn)) + [index_from_irq(irq)] = -1; + break; default: break; } @@ -743,6 +742,25 @@ void xen_clear_irq_pending(int irq) clear_evtchn(evtchn); } +void xen_set_irq_pending(int irq) +{ + int evtchn = evtchn_from_irq(irq); + + if (VALID_EVTCHN(evtchn)) + set_evtchn(evtchn); +} + +bool xen_test_irq_pending(int irq) +{ + int evtchn = evtchn_from_irq(irq); + bool ret = false; + + if (VALID_EVTCHN(evtchn)) + ret = test_evtchn(evtchn); + + return ret; +} + /* Poll waiting for an irq to become pending. In the usual case, the irq will be disabled so it won't deliver an interrupt. */ void xen_poll_irq(int irq) diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c index e9e11168616..06592b9da83 100644 --- a/drivers/xen/grant-table.c +++ b/drivers/xen/grant-table.c @@ -508,7 +508,7 @@ static int __devinit gnttab_init(void) unsigned int max_nr_glist_frames, nr_glist_frames; unsigned int nr_init_grefs; - if (!is_running_on_xen()) + if (!xen_domain()) return -ENODEV; nr_grant_frames = 1; diff --git a/drivers/xen/xenbus/xenbus_probe.c b/drivers/xen/xenbus/xenbus_probe.c index 57ceb5346b7..7f24a98a446 100644 --- a/drivers/xen/xenbus/xenbus_probe.c +++ b/drivers/xen/xenbus/xenbus_probe.c @@ -814,7 +814,7 @@ static int __init xenbus_probe_init(void) DPRINTK(""); err = -ENODEV; - if (!is_running_on_xen()) + if (!xen_domain()) goto out_error; /* Register ourselves with the kernel bus subsystem */ @@ -829,7 +829,7 @@ static int __init xenbus_probe_init(void) /* * Domain0 doesn't have a store_evtchn or store_mfn yet. */ - if (is_initial_xendomain()) { + if (xen_initial_domain()) { /* dom0 not yet supported */ } else { xenstored_ready = 1; @@ -846,7 +846,7 @@ static int __init xenbus_probe_init(void) goto out_unreg_back; } - if (!is_initial_xendomain()) + if (!xen_initial_domain()) xenbus_probe(NULL); return 0; @@ -937,7 +937,7 @@ static void wait_for_devices(struct xenbus_driver *xendrv) unsigned long timeout = jiffies + 10*HZ; struct device_driver *drv = xendrv ? &xendrv->driver : NULL; - if (!ready_to_wait_for_devices || !is_running_on_xen()) + if (!ready_to_wait_for_devices || !xen_domain()) return; while (exists_disconnected_device(drv)) { diff --git a/include/asm-x86/desc.h b/include/asm-x86/desc.h index a44c4dc7059..06f786f4b4f 100644 --- a/include/asm-x86/desc.h +++ b/include/asm-x86/desc.h @@ -24,6 +24,11 @@ static inline void fill_ldt(struct desc_struct *desc, desc->d = info->seg_32bit; desc->g = info->limit_in_pages; desc->base2 = (info->base_addr & 0xff000000) >> 24; + /* + * Don't allow setting of the lm bit. It is useless anyway + * because 64bit system calls require __USER_CS: + */ + desc->l = 0; } extern struct desc_ptr idt_descr; @@ -97,7 +102,15 @@ static inline int desc_empty(const void *ptr) native_write_gdt_entry(dt, entry, desc, type) #define write_idt_entry(dt, entry, g) \ native_write_idt_entry(dt, entry, g) -#endif + +static inline void paravirt_alloc_ldt(struct desc_struct *ldt, unsigned entries) +{ +} + +static inline void paravirt_free_ldt(struct desc_struct *ldt, unsigned entries) +{ +} +#endif /* CONFIG_PARAVIRT */ static inline void native_write_idt_entry(gate_desc *idt, int entry, const gate_desc *gate) diff --git a/include/asm-x86/paravirt.h b/include/asm-x86/paravirt.h index fbbde93f12d..8e9b1266898 100644 --- a/include/asm-x86/paravirt.h +++ b/include/asm-x86/paravirt.h @@ -124,6 +124,9 @@ struct pv_cpu_ops { int entrynum, const void *desc, int size); void (*write_idt_entry)(gate_desc *, int entrynum, const gate_desc *gate); + void (*alloc_ldt)(struct desc_struct *ldt, unsigned entries); + void (*free_ldt)(struct desc_struct *ldt, unsigned entries); + void (*load_sp0)(struct tss_struct *tss, struct thread_struct *t); void (*set_iopl_mask)(unsigned mask); @@ -330,6 +333,7 @@ struct pv_lock_ops { int (*spin_is_locked)(struct raw_spinlock *lock); int (*spin_is_contended)(struct raw_spinlock *lock); void (*spin_lock)(struct raw_spinlock *lock); + void (*spin_lock_flags)(struct raw_spinlock *lock, unsigned long flags); int (*spin_trylock)(struct raw_spinlock *lock); void (*spin_unlock)(struct raw_spinlock *lock); }; @@ -824,6 +828,16 @@ do { \ (aux) = __aux; \ } while (0) +static inline void paravirt_alloc_ldt(struct desc_struct *ldt, unsigned entries) +{ + PVOP_VCALL2(pv_cpu_ops.alloc_ldt, ldt, entries); +} + +static inline void paravirt_free_ldt(struct desc_struct *ldt, unsigned entries) +{ + PVOP_VCALL2(pv_cpu_ops.free_ldt, ldt, entries); +} + static inline void load_TR_desc(void) { PVOP_VCALL0(pv_cpu_ops.load_tr_desc); @@ -1401,6 +1415,12 @@ static __always_inline void __raw_spin_lock(struct raw_spinlock *lock) PVOP_VCALL1(pv_lock_ops.spin_lock, lock); } +static __always_inline void __raw_spin_lock_flags(struct raw_spinlock *lock, + unsigned long flags) +{ + PVOP_VCALL2(pv_lock_ops.spin_lock_flags, lock, flags); +} + static __always_inline int __raw_spin_trylock(struct raw_spinlock *lock) { return PVOP_CALL1(int, pv_lock_ops.spin_trylock, lock); diff --git a/include/asm-x86/smp.h b/include/asm-x86/smp.h index 3c877f74f27..30b5146cc43 100644 --- a/include/asm-x86/smp.h +++ b/include/asm-x86/smp.h @@ -47,12 +47,16 @@ extern struct { struct smp_ops { void (*smp_prepare_boot_cpu)(void); void (*smp_prepare_cpus)(unsigned max_cpus); - int (*cpu_up)(unsigned cpu); void (*smp_cpus_done)(unsigned max_cpus); void (*smp_send_stop)(void); void (*smp_send_reschedule)(int cpu); + int (*cpu_up)(unsigned cpu); + int (*cpu_disable)(void); + void (*cpu_die)(unsigned int cpu); + void (*play_dead)(void); + void (*send_call_func_ipi)(cpumask_t mask); void (*send_call_func_single_ipi)(int cpu); }; @@ -91,6 +95,21 @@ static inline int __cpu_up(unsigned int cpu) return smp_ops.cpu_up(cpu); } +static inline int __cpu_disable(void) +{ + return smp_ops.cpu_disable(); +} + +static inline void __cpu_die(unsigned int cpu) +{ + smp_ops.cpu_die(cpu); +} + +static inline void play_dead(void) +{ + smp_ops.play_dead(); +} + static inline void smp_send_reschedule(int cpu) { smp_ops.smp_send_reschedule(cpu); @@ -106,16 +125,19 @@ static inline void arch_send_call_function_ipi(cpumask_t mask) smp_ops.send_call_func_ipi(mask); } +void cpu_disable_common(void); void native_smp_prepare_boot_cpu(void); void native_smp_prepare_cpus(unsigned int max_cpus); void native_smp_cpus_done(unsigned int max_cpus); int native_cpu_up(unsigned int cpunum); +int native_cpu_disable(void); +void native_cpu_die(unsigned int cpu); +void native_play_dead(void); +void play_dead_common(void); + void native_send_call_func_ipi(cpumask_t mask); void native_send_call_func_single_ipi(int cpu); -extern int __cpu_disable(void); -extern void __cpu_die(unsigned int cpu); - void smp_store_cpu_info(int id); #define cpu_physical_id(cpu) per_cpu(x86_cpu_to_apicid, cpu) @@ -200,9 +222,5 @@ static inline int hard_smp_processor_id(void) #endif /* CONFIG_X86_LOCAL_APIC */ -#ifdef CONFIG_HOTPLUG_CPU -extern void cpu_uninit(void); -#endif - #endif /* __ASSEMBLY__ */ #endif diff --git a/include/asm-x86/spinlock.h b/include/asm-x86/spinlock.h index e39c790dbfd..b755ea86367 100644 --- a/include/asm-x86/spinlock.h +++ b/include/asm-x86/spinlock.h @@ -182,8 +182,6 @@ static __always_inline void __ticket_spin_unlock(raw_spinlock_t *lock) } #endif -#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock) - #ifdef CONFIG_PARAVIRT /* * Define virtualization-friendly old-style lock byte lock, for use in @@ -272,6 +270,13 @@ static __always_inline void __raw_spin_unlock(raw_spinlock_t *lock) { __ticket_spin_unlock(lock); } + +static __always_inline void __raw_spin_lock_flags(raw_spinlock_t *lock, + unsigned long flags) +{ + __raw_spin_lock(lock); +} + #endif /* CONFIG_PARAVIRT */ static inline void __raw_spin_unlock_wait(raw_spinlock_t *lock) diff --git a/include/asm-x86/tlbflush.h b/include/asm-x86/tlbflush.h index 35c76ceb9f4..0e7bbb54911 100644 --- a/include/asm-x86/tlbflush.h +++ b/include/asm-x86/tlbflush.h @@ -119,6 +119,10 @@ static inline void native_flush_tlb_others(const cpumask_t *cpumask, { } +static inline void reset_lazy_tlbstate(void) +{ +} + #else /* SMP */ #include <asm/smp.h> @@ -151,6 +155,12 @@ struct tlb_state { char __cacheline_padding[L1_CACHE_BYTES-8]; }; DECLARE_PER_CPU(struct tlb_state, cpu_tlbstate); + +void reset_lazy_tlbstate(void); +#else +static inline void reset_lazy_tlbstate(void) +{ +} #endif #endif /* SMP */ diff --git a/include/asm-x86/xen/hypervisor.h b/include/asm-x86/xen/hypervisor.h index 04ee0610014..fca066febc3 100644 --- a/include/asm-x86/xen/hypervisor.h +++ b/include/asm-x86/xen/hypervisor.h @@ -54,7 +54,6 @@ /* arch/i386/kernel/setup.c */ extern struct shared_info *HYPERVISOR_shared_info; extern struct start_info *xen_start_info; -#define is_initial_xendomain() (xen_start_info->flags & SIF_INITDOMAIN) /* arch/i386/mach-xen/evtchn.c */ /* Force a proper event-channel callback from Xen. */ @@ -67,6 +66,17 @@ u64 jiffies_to_st(unsigned long jiffies); #define MULTI_UVMFLAGS_INDEX 3 #define MULTI_UVMDOMID_INDEX 4 -#define is_running_on_xen() (xen_start_info ? 1 : 0) +enum xen_domain_type { + XEN_NATIVE, + XEN_PV_DOMAIN, + XEN_HVM_DOMAIN, +}; + +extern enum xen_domain_type xen_domain_type; + +#define xen_domain() (xen_domain_type != XEN_NATIVE) +#define xen_pv_domain() (xen_domain_type == XEN_PV_DOMAIN) +#define xen_initial_domain() (xen_pv_domain() && xen_start_info->flags & SIF_INITDOMAIN) +#define xen_hvm_domain() (xen_domain_type == XEN_HVM_DOMAIN) #endif /* __HYPERVISOR_H__ */ diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 2651f805ba6..75d81f157d2 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -182,7 +182,7 @@ extern int vsscanf(const char *, const char *, va_list) extern int get_option(char **str, int *pint); extern char *get_options(const char *str, int nints, int *ints); -extern unsigned long long memparse(char *ptr, char **retptr); +extern unsigned long long memparse(const char *ptr, char **retptr); extern int core_kernel_text(unsigned long addr); extern int __kernel_text_address(unsigned long addr); diff --git a/include/linux/mm.h b/include/linux/mm.h index 72a15dc26bb..4194bf8e4f6 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -919,7 +919,7 @@ static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long a } #endif /* CONFIG_MMU && !__ARCH_HAS_4LEVEL_HACK */ -#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#if USE_SPLIT_PTLOCKS /* * We tuck a spinlock to guard each pagetable page into its struct page, * at page->private, with BUILD_BUG_ON to make sure that this will not @@ -932,14 +932,14 @@ static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long a } while (0) #define pte_lock_deinit(page) ((page)->mapping = NULL) #define pte_lockptr(mm, pmd) ({(void)(mm); __pte_lockptr(pmd_page(*(pmd)));}) -#else +#else /* !USE_SPLIT_PTLOCKS */ /* * We use mm->page_table_lock to guard all pagetable pages of the mm. */ #define pte_lock_init(page) do {} while (0) #define pte_lock_deinit(page) do {} while (0) #define pte_lockptr(mm, pmd) ({(void)(pmd); &(mm)->page_table_lock;}) -#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#endif /* USE_SPLIT_PTLOCKS */ static inline void pgtable_page_ctor(struct page *page) { diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index bf334138c7c..9d49fa36bbe 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -21,11 +21,13 @@ struct address_space; -#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#define USE_SPLIT_PTLOCKS (NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS) + +#if USE_SPLIT_PTLOCKS typedef atomic_long_t mm_counter_t; -#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#else /* !USE_SPLIT_PTLOCKS */ typedef unsigned long mm_counter_t; -#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#endif /* !USE_SPLIT_PTLOCKS */ /* * Each physical page in the system has a struct page associated with @@ -65,7 +67,7 @@ struct page { * see PAGE_MAPPING_ANON below. */ }; -#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#if USE_SPLIT_PTLOCKS spinlock_t ptl; #endif struct kmem_cache *slab; /* SLUB: Pointer to slab */ diff --git a/include/linux/sched.h b/include/linux/sched.h index 3d9120c5ad1..272c35309df 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -352,7 +352,7 @@ arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr, extern void arch_unmap_area(struct mm_struct *, unsigned long); extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long); -#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#if USE_SPLIT_PTLOCKS /* * The mm counters are not protected by its page_table_lock, * so must be incremented atomically. @@ -363,7 +363,7 @@ extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long); #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member) #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member) -#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#else /* !USE_SPLIT_PTLOCKS */ /* * The mm counters are protected by its page_table_lock, * so can be incremented directly. @@ -374,7 +374,7 @@ extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long); #define inc_mm_counter(mm, member) (mm)->_##member++ #define dec_mm_counter(mm, member) (mm)->_##member-- -#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#endif /* !USE_SPLIT_PTLOCKS */ #define get_mm_rss(mm) \ (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss)) diff --git a/include/xen/events.h b/include/xen/events.h index 4680ff3fbc9..0d5f1adc036 100644 --- a/include/xen/events.h +++ b/include/xen/events.h @@ -46,6 +46,8 @@ extern void xen_irq_resume(void); /* Clear an irq's pending state, in preparation for polling on it */ void xen_clear_irq_pending(int irq); +void xen_set_irq_pending(int irq); +bool xen_test_irq_pending(int irq); /* Poll waiting for an irq to become pending. In the usual case, the irq will be disabled so it won't deliver an interrupt. */ diff --git a/lib/cmdline.c b/lib/cmdline.c index 5ba8a942a47..f5f3ad8b62f 100644 --- a/lib/cmdline.c +++ b/lib/cmdline.c @@ -126,7 +126,7 @@ char *get_options(const char *str, int nints, int *ints) * megabyte, or one gigabyte, respectively. */ -unsigned long long memparse(char *ptr, char **retptr) +unsigned long long memparse(const char *ptr, char **retptr) { char *endptr; /* local pointer to end of parsed string */ |