From c935cd62d3fe985d7f0ebea185d2759e8992e96f Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Wed, 17 Jun 2020 17:17:19 +1000 Subject: [PATCH 1/4] lockdep: Split header file into lockdep and lockdep_types There is a header file inclusion loop between asm-generic/bug.h and linux/kernel.h. This causes potential compile failurs depending on the which file is included first. One way of breaking this loop is to stop spinlock_types.h from including lockdep.h. This patch splits lockdep.h into two files for this purpose. Signed-off-by: Herbert Xu Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Sergey Senozhatsky Reviewed-by: Andy Shevchenko Acked-by: Petr Mladek Acked-by: Steven Rostedt (VMware) Link: https://lkml.kernel.org/r/E1jlSJz-0003hE-8g@fornost.hmeau.com --- include/linux/lockdep.h | 178 +----------------------------- include/linux/lockdep_types.h | 196 +++++++++++++++++++++++++++++++++ include/linux/spinlock.h | 1 + include/linux/spinlock_types.h | 2 +- 4 files changed, 200 insertions(+), 177 deletions(-) create mode 100644 include/linux/lockdep_types.h diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h index 8fce5c98a4b0..3b73cf84f77d 100644 --- a/include/linux/lockdep.h +++ b/include/linux/lockdep.h @@ -10,181 +10,20 @@ #ifndef __LINUX_LOCKDEP_H #define __LINUX_LOCKDEP_H +#include + struct task_struct; -struct lockdep_map; /* for sysctl */ extern int prove_locking; extern int lock_stat; -#define MAX_LOCKDEP_SUBCLASSES 8UL - -#include - -enum lockdep_wait_type { - LD_WAIT_INV = 0, /* not checked, catch all */ - - LD_WAIT_FREE, /* wait free, rcu etc.. */ - LD_WAIT_SPIN, /* spin loops, raw_spinlock_t etc.. */ - -#ifdef CONFIG_PROVE_RAW_LOCK_NESTING - LD_WAIT_CONFIG, /* CONFIG_PREEMPT_LOCK, spinlock_t etc.. */ -#else - LD_WAIT_CONFIG = LD_WAIT_SPIN, -#endif - LD_WAIT_SLEEP, /* sleeping locks, mutex_t etc.. */ - - LD_WAIT_MAX, /* must be last */ -}; - #ifdef CONFIG_LOCKDEP #include -#include #include #include -/* - * We'd rather not expose kernel/lockdep_states.h this wide, but we do need - * the total number of states... :-( - */ -#define XXX_LOCK_USAGE_STATES (1+2*4) - -/* - * NR_LOCKDEP_CACHING_CLASSES ... Number of classes - * cached in the instance of lockdep_map - * - * Currently main class (subclass == 0) and signle depth subclass - * are cached in lockdep_map. This optimization is mainly targeting - * on rq->lock. double_rq_lock() acquires this highly competitive with - * single depth. - */ -#define NR_LOCKDEP_CACHING_CLASSES 2 - -/* - * A lockdep key is associated with each lock object. For static locks we use - * the lock address itself as the key. Dynamically allocated lock objects can - * have a statically or dynamically allocated key. Dynamically allocated lock - * keys must be registered before being used and must be unregistered before - * the key memory is freed. - */ -struct lockdep_subclass_key { - char __one_byte; -} __attribute__ ((__packed__)); - -/* hash_entry is used to keep track of dynamically allocated keys. */ -struct lock_class_key { - union { - struct hlist_node hash_entry; - struct lockdep_subclass_key subkeys[MAX_LOCKDEP_SUBCLASSES]; - }; -}; - -extern struct lock_class_key __lockdep_no_validate__; - -struct lock_trace; - -#define LOCKSTAT_POINTS 4 - -/* - * The lock-class itself. The order of the structure members matters. - * reinit_class() zeroes the key member and all subsequent members. - */ -struct lock_class { - /* - * class-hash: - */ - struct hlist_node hash_entry; - - /* - * Entry in all_lock_classes when in use. Entry in free_lock_classes - * when not in use. Instances that are being freed are on one of the - * zapped_classes lists. - */ - struct list_head lock_entry; - - /* - * These fields represent a directed graph of lock dependencies, - * to every node we attach a list of "forward" and a list of - * "backward" graph nodes. - */ - struct list_head locks_after, locks_before; - - const struct lockdep_subclass_key *key; - unsigned int subclass; - unsigned int dep_gen_id; - - /* - * IRQ/softirq usage tracking bits: - */ - unsigned long usage_mask; - const struct lock_trace *usage_traces[XXX_LOCK_USAGE_STATES]; - - /* - * Generation counter, when doing certain classes of graph walking, - * to ensure that we check one node only once: - */ - int name_version; - const char *name; - - short wait_type_inner; - short wait_type_outer; - -#ifdef CONFIG_LOCK_STAT - unsigned long contention_point[LOCKSTAT_POINTS]; - unsigned long contending_point[LOCKSTAT_POINTS]; -#endif -} __no_randomize_layout; - -#ifdef CONFIG_LOCK_STAT -struct lock_time { - s64 min; - s64 max; - s64 total; - unsigned long nr; -}; - -enum bounce_type { - bounce_acquired_write, - bounce_acquired_read, - bounce_contended_write, - bounce_contended_read, - nr_bounce_types, - - bounce_acquired = bounce_acquired_write, - bounce_contended = bounce_contended_write, -}; - -struct lock_class_stats { - unsigned long contention_point[LOCKSTAT_POINTS]; - unsigned long contending_point[LOCKSTAT_POINTS]; - struct lock_time read_waittime; - struct lock_time write_waittime; - struct lock_time read_holdtime; - struct lock_time write_holdtime; - unsigned long bounces[nr_bounce_types]; -}; - -struct lock_class_stats lock_stats(struct lock_class *class); -void clear_lock_stats(struct lock_class *class); -#endif - -/* - * Map the lock object (the lock instance) to the lock-class object. - * This is embedded into specific lock instances: - */ -struct lockdep_map { - struct lock_class_key *key; - struct lock_class *class_cache[NR_LOCKDEP_CACHING_CLASSES]; - const char *name; - short wait_type_outer; /* can be taken in this context */ - short wait_type_inner; /* presents this context */ -#ifdef CONFIG_LOCK_STAT - int cpu; - unsigned long ip; -#endif -}; - static inline void lockdep_copy_map(struct lockdep_map *to, struct lockdep_map *from) { @@ -440,8 +279,6 @@ static inline void lock_set_subclass(struct lockdep_map *lock, extern void lock_downgrade(struct lockdep_map *lock, unsigned long ip); -struct pin_cookie { unsigned int val; }; - #define NIL_COOKIE (struct pin_cookie){ .val = 0U, } extern struct pin_cookie lock_pin_lock(struct lockdep_map *lock); @@ -520,10 +357,6 @@ static inline void lockdep_set_selftest_task(struct task_struct *task) # define lockdep_reset() do { debug_locks = 1; } while (0) # define lockdep_free_key_range(start, size) do { } while (0) # define lockdep_sys_exit() do { } while (0) -/* - * The class key takes no space if lockdep is disabled: - */ -struct lock_class_key { }; static inline void lockdep_register_key(struct lock_class_key *key) { @@ -533,11 +366,6 @@ static inline void lockdep_unregister_key(struct lock_class_key *key) { } -/* - * The lockdep_map takes no space if lockdep is disabled: - */ -struct lockdep_map { }; - #define lockdep_depth(tsk) (0) #define lockdep_is_held_type(l, r) (1) @@ -549,8 +377,6 @@ struct lockdep_map { }; #define lockdep_recursing(tsk) (0) -struct pin_cookie { }; - #define NIL_COOKIE (struct pin_cookie){ } #define lockdep_pin_lock(l) ({ struct pin_cookie cookie = { }; cookie; }) diff --git a/include/linux/lockdep_types.h b/include/linux/lockdep_types.h new file mode 100644 index 000000000000..7b9350624577 --- /dev/null +++ b/include/linux/lockdep_types.h @@ -0,0 +1,196 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Runtime locking correctness validator + * + * Copyright (C) 2006,2007 Red Hat, Inc., Ingo Molnar + * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra + * + * see Documentation/locking/lockdep-design.rst for more details. + */ +#ifndef __LINUX_LOCKDEP_TYPES_H +#define __LINUX_LOCKDEP_TYPES_H + +#include + +#define MAX_LOCKDEP_SUBCLASSES 8UL + +enum lockdep_wait_type { + LD_WAIT_INV = 0, /* not checked, catch all */ + + LD_WAIT_FREE, /* wait free, rcu etc.. */ + LD_WAIT_SPIN, /* spin loops, raw_spinlock_t etc.. */ + +#ifdef CONFIG_PROVE_RAW_LOCK_NESTING + LD_WAIT_CONFIG, /* CONFIG_PREEMPT_LOCK, spinlock_t etc.. */ +#else + LD_WAIT_CONFIG = LD_WAIT_SPIN, +#endif + LD_WAIT_SLEEP, /* sleeping locks, mutex_t etc.. */ + + LD_WAIT_MAX, /* must be last */ +}; + +#ifdef CONFIG_LOCKDEP + +#include + +/* + * We'd rather not expose kernel/lockdep_states.h this wide, but we do need + * the total number of states... :-( + */ +#define XXX_LOCK_USAGE_STATES (1+2*4) + +/* + * NR_LOCKDEP_CACHING_CLASSES ... Number of classes + * cached in the instance of lockdep_map + * + * Currently main class (subclass == 0) and signle depth subclass + * are cached in lockdep_map. This optimization is mainly targeting + * on rq->lock. double_rq_lock() acquires this highly competitive with + * single depth. + */ +#define NR_LOCKDEP_CACHING_CLASSES 2 + +/* + * A lockdep key is associated with each lock object. For static locks we use + * the lock address itself as the key. Dynamically allocated lock objects can + * have a statically or dynamically allocated key. Dynamically allocated lock + * keys must be registered before being used and must be unregistered before + * the key memory is freed. + */ +struct lockdep_subclass_key { + char __one_byte; +} __attribute__ ((__packed__)); + +/* hash_entry is used to keep track of dynamically allocated keys. */ +struct lock_class_key { + union { + struct hlist_node hash_entry; + struct lockdep_subclass_key subkeys[MAX_LOCKDEP_SUBCLASSES]; + }; +}; + +extern struct lock_class_key __lockdep_no_validate__; + +struct lock_trace; + +#define LOCKSTAT_POINTS 4 + +/* + * The lock-class itself. The order of the structure members matters. + * reinit_class() zeroes the key member and all subsequent members. + */ +struct lock_class { + /* + * class-hash: + */ + struct hlist_node hash_entry; + + /* + * Entry in all_lock_classes when in use. Entry in free_lock_classes + * when not in use. Instances that are being freed are on one of the + * zapped_classes lists. + */ + struct list_head lock_entry; + + /* + * These fields represent a directed graph of lock dependencies, + * to every node we attach a list of "forward" and a list of + * "backward" graph nodes. + */ + struct list_head locks_after, locks_before; + + const struct lockdep_subclass_key *key; + unsigned int subclass; + unsigned int dep_gen_id; + + /* + * IRQ/softirq usage tracking bits: + */ + unsigned long usage_mask; + const struct lock_trace *usage_traces[XXX_LOCK_USAGE_STATES]; + + /* + * Generation counter, when doing certain classes of graph walking, + * to ensure that we check one node only once: + */ + int name_version; + const char *name; + + short wait_type_inner; + short wait_type_outer; + +#ifdef CONFIG_LOCK_STAT + unsigned long contention_point[LOCKSTAT_POINTS]; + unsigned long contending_point[LOCKSTAT_POINTS]; +#endif +} __no_randomize_layout; + +#ifdef CONFIG_LOCK_STAT +struct lock_time { + s64 min; + s64 max; + s64 total; + unsigned long nr; +}; + +enum bounce_type { + bounce_acquired_write, + bounce_acquired_read, + bounce_contended_write, + bounce_contended_read, + nr_bounce_types, + + bounce_acquired = bounce_acquired_write, + bounce_contended = bounce_contended_write, +}; + +struct lock_class_stats { + unsigned long contention_point[LOCKSTAT_POINTS]; + unsigned long contending_point[LOCKSTAT_POINTS]; + struct lock_time read_waittime; + struct lock_time write_waittime; + struct lock_time read_holdtime; + struct lock_time write_holdtime; + unsigned long bounces[nr_bounce_types]; +}; + +struct lock_class_stats lock_stats(struct lock_class *class); +void clear_lock_stats(struct lock_class *class); +#endif + +/* + * Map the lock object (the lock instance) to the lock-class object. + * This is embedded into specific lock instances: + */ +struct lockdep_map { + struct lock_class_key *key; + struct lock_class *class_cache[NR_LOCKDEP_CACHING_CLASSES]; + const char *name; + short wait_type_outer; /* can be taken in this context */ + short wait_type_inner; /* presents this context */ +#ifdef CONFIG_LOCK_STAT + int cpu; + unsigned long ip; +#endif +}; + +struct pin_cookie { unsigned int val; }; + +#else /* !CONFIG_LOCKDEP */ + +/* + * The class key takes no space if lockdep is disabled: + */ +struct lock_class_key { }; + +/* + * The lockdep_map takes no space if lockdep is disabled: + */ +struct lockdep_map { }; + +struct pin_cookie { }; + +#endif /* !LOCKDEP */ + +#endif /* __LINUX_LOCKDEP_TYPES_H */ diff --git a/include/linux/spinlock.h b/include/linux/spinlock.h index d3770b3f9d9a..f2f12d746dbd 100644 --- a/include/linux/spinlock.h +++ b/include/linux/spinlock.h @@ -56,6 +56,7 @@ #include #include #include +#include #include #include diff --git a/include/linux/spinlock_types.h b/include/linux/spinlock_types.h index 6102e6bff3ae..b981caafe8bf 100644 --- a/include/linux/spinlock_types.h +++ b/include/linux/spinlock_types.h @@ -15,7 +15,7 @@ # include #endif -#include +#include typedef struct raw_spinlock { arch_spinlock_t raw_lock; From 248591f5d257a19c1cba9ab9da3536bfbc2f0149 Mon Sep 17 00:00:00 2001 From: Marco Elver Date: Wed, 24 Jun 2020 13:32:46 +0200 Subject: [PATCH 2/4] kcsan: Make KCSAN compatible with new IRQ state tracking The new IRQ state tracking code does not honor lockdep_off(), and as such we should again permit tracing by using non-raw functions in core.c. Update the lockdep_off() comment in report.c, to reflect the fact there is still a potential risk of deadlock due to using printk() from scheduler code. Suggested-by: Peter Zijlstra (Intel) Signed-off-by: Marco Elver Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Ingo Molnar Link: https://lkml.kernel.org/r/20200624113246.GA170324@elver.google.com --- kernel/kcsan/core.c | 5 ++--- kernel/kcsan/report.c | 9 +++++---- 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/kernel/kcsan/core.c b/kernel/kcsan/core.c index 15f67949d11e..732623c30359 100644 --- a/kernel/kcsan/core.c +++ b/kernel/kcsan/core.c @@ -397,8 +397,7 @@ kcsan_setup_watchpoint(const volatile void *ptr, size_t size, int type) } if (!kcsan_interrupt_watcher) - /* Use raw to avoid lockdep recursion via IRQ flags tracing. */ - raw_local_irq_save(irq_flags); + local_irq_save(irq_flags); watchpoint = insert_watchpoint((unsigned long)ptr, size, is_write); if (watchpoint == NULL) { @@ -539,7 +538,7 @@ kcsan_setup_watchpoint(const volatile void *ptr, size_t size, int type) kcsan_counter_dec(KCSAN_COUNTER_USED_WATCHPOINTS); out_unlock: if (!kcsan_interrupt_watcher) - raw_local_irq_restore(irq_flags); + local_irq_restore(irq_flags); out: user_access_restore(ua_flags); } diff --git a/kernel/kcsan/report.c b/kernel/kcsan/report.c index ac5f8345bae9..6b2fb1a6d8cd 100644 --- a/kernel/kcsan/report.c +++ b/kernel/kcsan/report.c @@ -606,10 +606,11 @@ void kcsan_report(const volatile void *ptr, size_t size, int access_type, goto out; /* - * With TRACE_IRQFLAGS, lockdep's IRQ trace state becomes corrupted if - * we do not turn off lockdep here; this could happen due to recursion - * into lockdep via KCSAN if we detect a race in utilities used by - * lockdep. + * Because we may generate reports when we're in scheduler code, the use + * of printk() could deadlock. Until such time that all printing code + * called in print_report() is scheduler-safe, accept the risk, and just + * get our message out. As such, also disable lockdep to hide the + * warning, and avoid disabling lockdep for the rest of the kernel. */ lockdep_off(); From 859d069ee1ddd87862e1d6a356a82ed417dbeb67 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 May 2020 15:00:57 +0200 Subject: [PATCH 3/4] lockdep: Prepare for NMI IRQ state tracking There is no reason not to always, accurately, track IRQ state. This change also makes IRQ state tracking ignore lockdep_off(). Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Ingo Molnar Link: https://lkml.kernel.org/r/20200623083721.155449112@infradead.org --- kernel/locking/lockdep.c | 46 ++++++++++++++++++++++++++++++++++++---- 1 file changed, 42 insertions(+), 4 deletions(-) diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c index 29a8de4c50b9..d595623c4b34 100644 --- a/kernel/locking/lockdep.c +++ b/kernel/locking/lockdep.c @@ -395,7 +395,7 @@ void lockdep_init_task(struct task_struct *task) static __always_inline void lockdep_recursion_finish(void) { - if (WARN_ON_ONCE(--current->lockdep_recursion)) + if (WARN_ON_ONCE((--current->lockdep_recursion) & LOCKDEP_RECURSION_MASK)) current->lockdep_recursion = 0; } @@ -3646,7 +3646,16 @@ static void __trace_hardirqs_on_caller(void) */ void lockdep_hardirqs_on_prepare(unsigned long ip) { - if (unlikely(!debug_locks || current->lockdep_recursion)) + if (unlikely(!debug_locks)) + return; + + /* + * NMIs do not (and cannot) track lock dependencies, nothing to do. + */ + if (unlikely(in_nmi())) + return; + + if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK)) return; if (unlikely(current->hardirqs_enabled)) { @@ -3692,7 +3701,27 @@ void noinstr lockdep_hardirqs_on(unsigned long ip) { struct task_struct *curr = current; - if (unlikely(!debug_locks || curr->lockdep_recursion)) + if (unlikely(!debug_locks)) + return; + + /* + * NMIs can happen in the middle of local_irq_{en,dis}able() where the + * tracking state and hardware state are out of sync. + * + * NMIs must save lockdep_hardirqs_enabled() to restore IRQ state from, + * and not rely on hardware state like normal interrupts. + */ + if (unlikely(in_nmi())) { + /* + * Skip: + * - recursion check, because NMI can hit lockdep; + * - hardware state check, because above; + * - chain_key check, see lockdep_hardirqs_on_prepare(). + */ + goto skip_checks; + } + + if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK)) return; if (curr->hardirqs_enabled) { @@ -3720,6 +3749,7 @@ void noinstr lockdep_hardirqs_on(unsigned long ip) DEBUG_LOCKS_WARN_ON(current->hardirq_chain_key != current->curr_chain_key); +skip_checks: /* we'll do an OFF -> ON transition: */ curr->hardirqs_enabled = 1; curr->hardirq_enable_ip = ip; @@ -3735,7 +3765,15 @@ void noinstr lockdep_hardirqs_off(unsigned long ip) { struct task_struct *curr = current; - if (unlikely(!debug_locks || curr->lockdep_recursion)) + if (unlikely(!debug_locks)) + return; + + /* + * Matching lockdep_hardirqs_on(), allow NMIs in the middle of lockdep; + * they will restore the software state. This ensures the software + * state is consistent inside NMIs as well. + */ + if (unlikely(!in_nmi() && (current->lockdep_recursion & LOCKDEP_RECURSION_MASK))) return; /* From ba1f2b2eaa2a529dba722507c55ff3d761d325dd Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 May 2020 15:50:29 +0200 Subject: [PATCH 4/4] x86/entry: Fix NMI vs IRQ state tracking While the nmi_enter() users did trace_hardirqs_{off_prepare,on_finish}() there was no matching lockdep_hardirqs_*() calls to complete the picture. Introduce idtentry_{enter,exit}_nmi() to enable proper IRQ state tracking across the NMIs. Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Ingo Molnar Link: https://lkml.kernel.org/r/20200623083721.216740948@infradead.org --- arch/x86/entry/common.c | 42 +++++++++++++++++++++++++++++---- arch/x86/include/asm/idtentry.h | 3 +++ arch/x86/kernel/nmi.c | 9 ++++--- arch/x86/kernel/traps.c | 17 +++++-------- include/linux/hardirq.h | 28 +++++++++++++++------- 5 files changed, 70 insertions(+), 29 deletions(-) diff --git a/arch/x86/entry/common.c b/arch/x86/entry/common.c index 0521546022cb..63c607dd6c52 100644 --- a/arch/x86/entry/common.c +++ b/arch/x86/entry/common.c @@ -592,7 +592,7 @@ SYSCALL_DEFINE0(ni_syscall) * The return value must be fed into the state argument of * idtentry_exit(). */ -idtentry_state_t noinstr idtentry_enter(struct pt_regs *regs) +noinstr idtentry_state_t idtentry_enter(struct pt_regs *regs) { idtentry_state_t ret = { .exit_rcu = false, @@ -687,7 +687,7 @@ static void idtentry_exit_cond_resched(struct pt_regs *regs, bool may_sched) * Counterpart to idtentry_enter(). The return value of the entry * function must be fed into the @state argument. */ -void noinstr idtentry_exit(struct pt_regs *regs, idtentry_state_t state) +noinstr void idtentry_exit(struct pt_regs *regs, idtentry_state_t state) { lockdep_assert_irqs_disabled(); @@ -731,7 +731,7 @@ void noinstr idtentry_exit(struct pt_regs *regs, idtentry_state_t state) * Invokes enter_from_user_mode() to establish the proper context for * NOHZ_FULL. Otherwise scheduling on exit would not be possible. */ -void noinstr idtentry_enter_user(struct pt_regs *regs) +noinstr void idtentry_enter_user(struct pt_regs *regs) { check_user_regs(regs); enter_from_user_mode(); @@ -749,13 +749,47 @@ void noinstr idtentry_enter_user(struct pt_regs *regs) * * Counterpart to idtentry_enter_user(). */ -void noinstr idtentry_exit_user(struct pt_regs *regs) +noinstr void idtentry_exit_user(struct pt_regs *regs) { lockdep_assert_irqs_disabled(); prepare_exit_to_usermode(regs); } +noinstr bool idtentry_enter_nmi(struct pt_regs *regs) +{ + bool irq_state = lockdep_hardirqs_enabled(current); + + __nmi_enter(); + lockdep_hardirqs_off(CALLER_ADDR0); + lockdep_hardirq_enter(); + rcu_nmi_enter(); + + instrumentation_begin(); + trace_hardirqs_off_finish(); + ftrace_nmi_enter(); + instrumentation_end(); + + return irq_state; +} + +noinstr void idtentry_exit_nmi(struct pt_regs *regs, bool restore) +{ + instrumentation_begin(); + ftrace_nmi_exit(); + if (restore) { + trace_hardirqs_on_prepare(); + lockdep_hardirqs_on_prepare(CALLER_ADDR0); + } + instrumentation_end(); + + rcu_nmi_exit(); + lockdep_hardirq_exit(); + if (restore) + lockdep_hardirqs_on(CALLER_ADDR0); + __nmi_exit(); +} + #ifdef CONFIG_XEN_PV #ifndef CONFIG_PREEMPTION /* diff --git a/arch/x86/include/asm/idtentry.h b/arch/x86/include/asm/idtentry.h index 7227225cf45d..2b0497486525 100644 --- a/arch/x86/include/asm/idtentry.h +++ b/arch/x86/include/asm/idtentry.h @@ -20,6 +20,9 @@ typedef struct idtentry_state { idtentry_state_t idtentry_enter(struct pt_regs *regs); void idtentry_exit(struct pt_regs *regs, idtentry_state_t state); +bool idtentry_enter_nmi(struct pt_regs *regs); +void idtentry_exit_nmi(struct pt_regs *regs, bool irq_state); + /** * DECLARE_IDTENTRY - Declare functions for simple IDT entry points * No error code pushed by hardware diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index d7c5e44b26f7..4fc9954a9560 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -330,7 +330,6 @@ static noinstr void default_do_nmi(struct pt_regs *regs) __this_cpu_write(last_nmi_rip, regs->ip); instrumentation_begin(); - trace_hardirqs_off_finish(); handled = nmi_handle(NMI_LOCAL, regs); __this_cpu_add(nmi_stats.normal, handled); @@ -417,8 +416,6 @@ static noinstr void default_do_nmi(struct pt_regs *regs) unknown_nmi_error(reason, regs); out: - if (regs->flags & X86_EFLAGS_IF) - trace_hardirqs_on_prepare(); instrumentation_end(); } @@ -478,6 +475,8 @@ static DEFINE_PER_CPU(unsigned long, nmi_dr7); DEFINE_IDTENTRY_RAW(exc_nmi) { + bool irq_state; + if (IS_ENABLED(CONFIG_SMP) && arch_cpu_is_offline(smp_processor_id())) return; @@ -491,14 +490,14 @@ DEFINE_IDTENTRY_RAW(exc_nmi) this_cpu_write(nmi_dr7, local_db_save()); - nmi_enter(); + irq_state = idtentry_enter_nmi(regs); inc_irq_stat(__nmi_count); if (!ignore_nmis) default_do_nmi(regs); - nmi_exit(); + idtentry_exit_nmi(regs, irq_state); local_db_restore(this_cpu_read(nmi_dr7)); diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 4627f826fb57..cdd73829e637 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -403,7 +403,7 @@ DEFINE_IDTENTRY_DF(exc_double_fault) } #endif - nmi_enter(); + idtentry_enter_nmi(regs); instrumentation_begin(); notify_die(DIE_TRAP, str, regs, error_code, X86_TRAP_DF, SIGSEGV); @@ -649,15 +649,12 @@ DEFINE_IDTENTRY_RAW(exc_int3) instrumentation_end(); idtentry_exit_user(regs); } else { - nmi_enter(); + bool irq_state = idtentry_enter_nmi(regs); instrumentation_begin(); - trace_hardirqs_off_finish(); if (!do_int3(regs)) die("int3", regs, 0); - if (regs->flags & X86_EFLAGS_IF) - trace_hardirqs_on_prepare(); instrumentation_end(); - nmi_exit(); + idtentry_exit_nmi(regs, irq_state); } } @@ -865,9 +862,8 @@ static void handle_debug(struct pt_regs *regs, unsigned long dr6, bool user) static __always_inline void exc_debug_kernel(struct pt_regs *regs, unsigned long dr6) { - nmi_enter(); + bool irq_state = idtentry_enter_nmi(regs); instrumentation_begin(); - trace_hardirqs_off_finish(); /* * If something gets miswired and we end up here for a user mode @@ -884,10 +880,8 @@ static __always_inline void exc_debug_kernel(struct pt_regs *regs, handle_debug(regs, dr6, false); - if (regs->flags & X86_EFLAGS_IF) - trace_hardirqs_on_prepare(); instrumentation_end(); - nmi_exit(); + idtentry_exit_nmi(regs, irq_state); } static __always_inline void exc_debug_user(struct pt_regs *regs, @@ -903,6 +897,7 @@ static __always_inline void exc_debug_user(struct pt_regs *regs, instrumentation_begin(); handle_debug(regs, dr6, true); + instrumentation_end(); idtentry_exit_user(regs); } diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 03c9fece7d43..754f67ac4326 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -111,32 +111,42 @@ extern void rcu_nmi_exit(void); /* * nmi_enter() can nest up to 15 times; see NMI_BITS. */ -#define nmi_enter() \ +#define __nmi_enter() \ do { \ + lockdep_off(); \ arch_nmi_enter(); \ printk_nmi_enter(); \ - lockdep_off(); \ BUG_ON(in_nmi() == NMI_MASK); \ __preempt_count_add(NMI_OFFSET + HARDIRQ_OFFSET); \ - rcu_nmi_enter(); \ + } while (0) + +#define nmi_enter() \ + do { \ + __nmi_enter(); \ lockdep_hardirq_enter(); \ + rcu_nmi_enter(); \ instrumentation_begin(); \ ftrace_nmi_enter(); \ instrumentation_end(); \ } while (0) +#define __nmi_exit() \ + do { \ + BUG_ON(!in_nmi()); \ + __preempt_count_sub(NMI_OFFSET + HARDIRQ_OFFSET); \ + printk_nmi_exit(); \ + arch_nmi_exit(); \ + lockdep_on(); \ + } while (0) + #define nmi_exit() \ do { \ instrumentation_begin(); \ ftrace_nmi_exit(); \ instrumentation_end(); \ - lockdep_hardirq_exit(); \ rcu_nmi_exit(); \ - BUG_ON(!in_nmi()); \ - __preempt_count_sub(NMI_OFFSET + HARDIRQ_OFFSET); \ - lockdep_on(); \ - printk_nmi_exit(); \ - arch_nmi_exit(); \ + lockdep_hardirq_exit(); \ + __nmi_exit(); \ } while (0) #endif /* LINUX_HARDIRQ_H */