diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c index c7bb168b0d97..842427ff2b3c 100644 --- a/arch/arm/kernel/smp.c +++ b/arch/arm/kernel/smp.c @@ -667,9 +667,9 @@ static void do_handle_IPI(int ipinr) break; case IPI_CPU_BACKTRACE: - printk_nmi_enter(); + printk_deferred_enter(); nmi_cpu_backtrace(get_irq_regs()); - printk_nmi_exit(); + printk_deferred_exit(); break; default: diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index dfbce527c98e..e103b89234cd 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -171,7 +171,6 @@ extern void panic_flush_kmsg_start(void) extern void panic_flush_kmsg_end(void) { - printk_safe_flush_on_panic(); kmsg_dump(KMSG_DUMP_PANIC); bust_spinlocks(0); debug_locks_off(); diff --git a/arch/powerpc/kernel/watchdog.c b/arch/powerpc/kernel/watchdog.c index a165635fd214..f9ea0e5357f9 100644 --- a/arch/powerpc/kernel/watchdog.c +++ b/arch/powerpc/kernel/watchdog.c @@ -184,11 +184,6 @@ static void watchdog_smp_panic(int cpu, u64 tb) wd_smp_unlock(&flags); - printk_safe_flush(); - /* - * printk_safe_flush() seems to require another print - * before anything actually goes out to console. - */ if (sysctl_hardlockup_all_cpu_backtrace) trigger_allbutself_cpu_backtrace(); diff --git a/arch/powerpc/kexec/crash.c b/arch/powerpc/kexec/crash.c index 10f997e6bb95..22ceeeb705ab 100644 --- a/arch/powerpc/kexec/crash.c +++ b/arch/powerpc/kexec/crash.c @@ -313,7 +313,7 @@ void default_machine_crash_shutdown(struct pt_regs *regs) int (*old_handler)(struct pt_regs *regs); /* Avoid hardlocking with irresponsive CPU holding logbuf_lock */ - printk_nmi_enter(); + printk_deferred_enter(); /* * This function is only called after the system diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 69bc86ea382c..76878b357ffa 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -116,7 +116,6 @@ extern void rcu_nmi_exit(void); do { \ lockdep_off(); \ arch_nmi_enter(); \ - printk_nmi_enter(); \ BUG_ON(in_nmi() == NMI_MASK); \ __preempt_count_add(NMI_OFFSET + HARDIRQ_OFFSET); \ } while (0) @@ -135,7 +134,6 @@ extern void rcu_nmi_exit(void); do { \ BUG_ON(!in_nmi()); \ __preempt_count_sub(NMI_OFFSET + HARDIRQ_OFFSET); \ - printk_nmi_exit(); \ arch_nmi_exit(); \ lockdep_on(); \ } while (0) diff --git a/include/linux/printk.h b/include/linux/printk.h index da7a1de2257a..259af4f97f50 100644 --- a/include/linux/printk.h +++ b/include/linux/printk.h @@ -141,18 +141,6 @@ static inline __printf(1, 2) __cold void early_printk(const char *s, ...) { } #endif -#ifdef CONFIG_PRINTK_NMI -extern void printk_nmi_enter(void); -extern void printk_nmi_exit(void); -extern void printk_nmi_direct_enter(void); -extern void printk_nmi_direct_exit(void); -#else -static inline void printk_nmi_enter(void) { } -static inline void printk_nmi_exit(void) { } -static inline void printk_nmi_direct_enter(void) { } -static inline void printk_nmi_direct_exit(void) { } -#endif /* PRINTK_NMI */ - struct dev_printk_info; #ifdef CONFIG_PRINTK @@ -172,6 +160,16 @@ int _printk(const char *fmt, ...); */ __printf(1, 2) __cold int _printk_deferred(const char *fmt, ...); +extern void __printk_safe_enter(void); +extern void __printk_safe_exit(void); +/* + * The printk_deferred_enter/exit macros are available only as a hack for + * some code paths that need to defer all printk console printing. Interrupts + * must be disabled for the deferred duration. + */ +#define printk_deferred_enter __printk_safe_enter +#define printk_deferred_exit __printk_safe_exit + /* * Please don't use printk_ratelimit(), because it shares ratelimiting state * with all other unrelated printk_ratelimit() callsites. Instead use @@ -200,8 +198,6 @@ void dump_stack_print_info(const char *log_lvl); void show_regs_print_info(const char *log_lvl); extern asmlinkage void dump_stack_lvl(const char *log_lvl) __cold; extern asmlinkage void dump_stack(void) __cold; -extern void printk_safe_flush(void); -extern void printk_safe_flush_on_panic(void); #else static inline __printf(1, 0) int vprintk(const char *s, va_list args) @@ -218,6 +214,15 @@ int _printk_deferred(const char *s, ...) { return 0; } + +static inline void printk_deferred_enter(void) +{ +} + +static inline void printk_deferred_exit(void) +{ +} + static inline int printk_ratelimit(void) { return 0; @@ -269,14 +274,6 @@ static inline void dump_stack_lvl(const char *log_lvl) static inline void dump_stack(void) { } - -static inline void printk_safe_flush(void) -{ -} - -static inline void printk_safe_flush_on_panic(void) -{ -} #endif #ifdef CONFIG_SMP diff --git a/init/Kconfig b/init/Kconfig index ccffa7ae5ccc..a2bbe49893af 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1523,11 +1523,6 @@ config PRINTK very difficult to diagnose system problems, saying N here is strongly discouraged. -config PRINTK_NMI - def_bool y - depends on PRINTK - depends on HAVE_NMI - config BUG bool "BUG() support" if EXPERT default y diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index 4b34a9aa32bc..5a5d192a89ac 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -979,7 +979,6 @@ void crash_kexec(struct pt_regs *regs) old_cpu = atomic_cmpxchg(&panic_cpu, PANIC_CPU_INVALID, this_cpu); if (old_cpu == PANIC_CPU_INVALID) { /* This is the 1st CPU which comes here, so go ahead. */ - printk_safe_flush_on_panic(); __crash_kexec(regs); /* diff --git a/kernel/panic.c b/kernel/panic.c index edad89660a2b..cefd7d82366f 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -248,7 +248,6 @@ void panic(const char *fmt, ...) * Bypass the panic_cpu check and call __crash_kexec directly. */ if (!_crash_kexec_post_notifiers) { - printk_safe_flush_on_panic(); __crash_kexec(NULL); /* @@ -272,8 +271,6 @@ void panic(const char *fmt, ...) */ atomic_notifier_call_chain(&panic_notifier_list, 0, buf); - /* Call flush even twice. It tries harder with a single online CPU */ - printk_safe_flush_on_panic(); kmsg_dump(KMSG_DUMP_PANIC); /* diff --git a/kernel/printk/internal.h b/kernel/printk/internal.h index 1596e2837318..9f3ed2fdb721 100644 --- a/kernel/printk/internal.h +++ b/kernel/printk/internal.h @@ -6,12 +6,6 @@ #ifdef CONFIG_PRINTK -#define PRINTK_SAFE_CONTEXT_MASK 0x007ffffff -#define PRINTK_NMI_DIRECT_CONTEXT_MASK 0x008000000 -#define PRINTK_NMI_CONTEXT_MASK 0xff0000000 - -#define PRINTK_NMI_CONTEXT_OFFSET 0x010000000 - /* Flags for a single printk record. */ enum printk_info_flags { LOG_NEWLINE = 2, /* text ended with a newline */ @@ -25,10 +19,7 @@ int vprintk_store(int facility, int level, __printf(1, 0) int vprintk_default(const char *fmt, va_list args); __printf(1, 0) int vprintk_deferred(const char *fmt, va_list args); -void __printk_safe_enter(void); -void __printk_safe_exit(void); -void printk_safe_init(void); bool printk_percpu_data_ready(void); #define printk_safe_enter_irqsave(flags) \ @@ -43,18 +34,6 @@ bool printk_percpu_data_ready(void); local_irq_restore(flags); \ } while (0) -#define printk_safe_enter_irq() \ - do { \ - local_irq_disable(); \ - __printk_safe_enter(); \ - } while (0) - -#define printk_safe_exit_irq() \ - do { \ - __printk_safe_exit(); \ - local_irq_enable(); \ - } while (0) - void defer_console_output(void); u16 printk_parse_prefix(const char *text, int *level, @@ -69,9 +48,5 @@ u16 printk_parse_prefix(const char *text, int *level, #define printk_safe_enter_irqsave(flags) local_irq_save(flags) #define printk_safe_exit_irqrestore(flags) local_irq_restore(flags) -#define printk_safe_enter_irq() local_irq_disable() -#define printk_safe_exit_irq() local_irq_enable() - -static inline void printk_safe_init(void) { } static inline bool printk_percpu_data_ready(void) { return false; } #endif /* CONFIG_PRINTK */ diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 776e3565f013..825277e1e742 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -351,7 +351,7 @@ static int console_msg_format = MSG_FORMAT_DEFAULT; */ /* syslog_lock protects syslog_* variables and write access to clear_seq. */ -static DEFINE_RAW_SPINLOCK(syslog_lock); +static DEFINE_MUTEX(syslog_lock); #ifdef CONFIG_PRINTK DECLARE_WAIT_QUEUE_HEAD(log_wait); @@ -727,27 +727,22 @@ static ssize_t devkmsg_read(struct file *file, char __user *buf, if (ret) return ret; - printk_safe_enter_irq(); if (!prb_read_valid(prb, atomic64_read(&user->seq), r)) { if (file->f_flags & O_NONBLOCK) { ret = -EAGAIN; - printk_safe_exit_irq(); goto out; } - printk_safe_exit_irq(); ret = wait_event_interruptible(log_wait, prb_read_valid(prb, atomic64_read(&user->seq), r)); if (ret) goto out; - printk_safe_enter_irq(); } if (r->info->seq != atomic64_read(&user->seq)) { /* our last seen message is gone, return error and reset */ atomic64_set(&user->seq, r->info->seq); ret = -EPIPE; - printk_safe_exit_irq(); goto out; } @@ -757,7 +752,6 @@ static ssize_t devkmsg_read(struct file *file, char __user *buf, &r->info->dev_info); atomic64_set(&user->seq, r->info->seq + 1); - printk_safe_exit_irq(); if (len > count) { ret = -EINVAL; @@ -792,7 +786,6 @@ static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence) if (offset) return -ESPIPE; - printk_safe_enter_irq(); switch (whence) { case SEEK_SET: /* the first record */ @@ -813,7 +806,6 @@ static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence) default: ret = -EINVAL; } - printk_safe_exit_irq(); return ret; } @@ -828,7 +820,6 @@ static __poll_t devkmsg_poll(struct file *file, poll_table *wait) poll_wait(file, &log_wait, wait); - printk_safe_enter_irq(); if (prb_read_valid_info(prb, atomic64_read(&user->seq), &info, NULL)) { /* return error when data has vanished underneath us */ if (info.seq != atomic64_read(&user->seq)) @@ -836,7 +827,6 @@ static __poll_t devkmsg_poll(struct file *file, poll_table *wait) else ret = EPOLLIN|EPOLLRDNORM; } - printk_safe_exit_irq(); return ret; } @@ -869,9 +859,7 @@ static int devkmsg_open(struct inode *inode, struct file *file) prb_rec_init_rd(&user->record, &user->info, &user->text_buf[0], sizeof(user->text_buf)); - printk_safe_enter_irq(); atomic64_set(&user->seq, prb_first_valid_seq(prb)); - printk_safe_exit_irq(); file->private_data = user; return 0; @@ -1037,9 +1025,6 @@ static inline void log_buf_add_cpu(void) {} static void __init set_percpu_data_ready(void) { - printk_safe_init(); - /* Make sure we set this flag only after printk_safe() init is done */ - barrier(); __printk_percpu_data_ready = true; } @@ -1077,6 +1062,7 @@ void __init setup_log_buf(int early) struct prb_desc *new_descs; struct printk_info info; struct printk_record r; + unsigned int text_size; size_t new_descs_size; size_t new_infos_size; unsigned long flags; @@ -1137,24 +1123,37 @@ void __init setup_log_buf(int early) new_descs, ilog2(new_descs_count), new_infos); - printk_safe_enter_irqsave(flags); + local_irq_save(flags); log_buf_len = new_log_buf_len; log_buf = new_log_buf; new_log_buf_len = 0; free = __LOG_BUF_LEN; - prb_for_each_record(0, &printk_rb_static, seq, &r) - free -= add_to_rb(&printk_rb_dynamic, &r); + prb_for_each_record(0, &printk_rb_static, seq, &r) { + text_size = add_to_rb(&printk_rb_dynamic, &r); + if (text_size > free) + free = 0; + else + free -= text_size; + } - /* - * This is early enough that everything is still running on the - * boot CPU and interrupts are disabled. So no new messages will - * appear during the transition to the dynamic buffer. - */ prb = &printk_rb_dynamic; - printk_safe_exit_irqrestore(flags); + local_irq_restore(flags); + + /* + * Copy any remaining messages that might have appeared from + * NMI context after copying but before switching to the + * dynamic buffer. + */ + prb_for_each_record(seq, &printk_rb_static, seq, &r) { + text_size = add_to_rb(&printk_rb_dynamic, &r); + if (text_size > free) + free = 0; + else + free -= text_size; + } if (seq != prb_next_seq(&printk_rb_static)) { pr_err("dropped %llu messages\n", @@ -1476,12 +1475,14 @@ static u64 find_first_fitting_seq(u64 start_seq, u64 max_seq, size_t size, return seq; } +/* The caller is responsible for making sure @size is greater than 0. */ static int syslog_print(char __user *buf, int size) { struct printk_info info; struct printk_record r; char *text; int len = 0; + u64 seq; text = kmalloc(CONSOLE_LOG_MAX, GFP_KERNEL); if (!text) @@ -1489,17 +1490,35 @@ static int syslog_print(char __user *buf, int size) prb_rec_init_rd(&r, &info, text, CONSOLE_LOG_MAX); - while (size > 0) { + mutex_lock(&syslog_lock); + + /* + * Wait for the @syslog_seq record to be available. @syslog_seq may + * change while waiting. + */ + do { + seq = syslog_seq; + + mutex_unlock(&syslog_lock); + len = wait_event_interruptible(log_wait, prb_read_valid(prb, seq, NULL)); + mutex_lock(&syslog_lock); + + if (len) + goto out; + } while (syslog_seq != seq); + + /* + * Copy records that fit into the buffer. The above cycle makes sure + * that the first record is always available. + */ + do { size_t n; size_t skip; + int err; - printk_safe_enter_irq(); - raw_spin_lock(&syslog_lock); - if (!prb_read_valid(prb, syslog_seq, &r)) { - raw_spin_unlock(&syslog_lock); - printk_safe_exit_irq(); + if (!prb_read_valid(prb, syslog_seq, &r)) break; - } + if (r.info->seq != syslog_seq) { /* message is gone, move to next valid one */ syslog_seq = r.info->seq; @@ -1526,13 +1545,15 @@ static int syslog_print(char __user *buf, int size) syslog_partial += n; } else n = 0; - raw_spin_unlock(&syslog_lock); - printk_safe_exit_irq(); if (!n) break; - if (copy_to_user(buf, text + skip, n)) { + mutex_unlock(&syslog_lock); + err = copy_to_user(buf, text + skip, n); + mutex_lock(&syslog_lock); + + if (err) { if (!len) len = -EFAULT; break; @@ -1541,8 +1562,9 @@ static int syslog_print(char __user *buf, int size) len += n; size -= n; buf += n; - } - + } while (size); +out: + mutex_unlock(&syslog_lock); kfree(text); return len; } @@ -1561,7 +1583,6 @@ static int syslog_print_all(char __user *buf, int size, bool clear) return -ENOMEM; time = printk_time; - printk_safe_enter_irq(); /* * Find first record that fits, including all following records, * into the user-provided buffer for this dump. @@ -1582,23 +1603,20 @@ static int syslog_print_all(char __user *buf, int size, bool clear) break; } - printk_safe_exit_irq(); if (copy_to_user(buf + len, text, textlen)) len = -EFAULT; else len += textlen; - printk_safe_enter_irq(); if (len < 0) break; } if (clear) { - raw_spin_lock(&syslog_lock); + mutex_lock(&syslog_lock); latched_seq_write(&clear_seq, seq); - raw_spin_unlock(&syslog_lock); + mutex_unlock(&syslog_lock); } - printk_safe_exit_irq(); kfree(text); return len; @@ -1606,23 +1624,9 @@ static int syslog_print_all(char __user *buf, int size, bool clear) static void syslog_clear(void) { - printk_safe_enter_irq(); - raw_spin_lock(&syslog_lock); + mutex_lock(&syslog_lock); latched_seq_write(&clear_seq, prb_next_seq(prb)); - raw_spin_unlock(&syslog_lock); - printk_safe_exit_irq(); -} - -/* Return a consistent copy of @syslog_seq. */ -static u64 read_syslog_seq_irq(void) -{ - u64 seq; - - raw_spin_lock_irq(&syslog_lock); - seq = syslog_seq; - raw_spin_unlock_irq(&syslog_lock); - - return seq; + mutex_unlock(&syslog_lock); } int do_syslog(int type, char __user *buf, int len, int source) @@ -1648,11 +1652,6 @@ int do_syslog(int type, char __user *buf, int len, int source) return 0; if (!access_ok(buf, len)) return -EFAULT; - - error = wait_event_interruptible(log_wait, - prb_read_valid(prb, read_syslog_seq_irq(), NULL)); - if (error) - return error; error = syslog_print(buf, len); break; /* Read/clear last kernel messages */ @@ -1698,12 +1697,10 @@ int do_syslog(int type, char __user *buf, int len, int source) break; /* Number of chars in the log buffer */ case SYSLOG_ACTION_SIZE_UNREAD: - printk_safe_enter_irq(); - raw_spin_lock(&syslog_lock); + mutex_lock(&syslog_lock); if (!prb_read_valid_info(prb, syslog_seq, &info, NULL)) { /* No unread messages. */ - raw_spin_unlock(&syslog_lock); - printk_safe_exit_irq(); + mutex_unlock(&syslog_lock); return 0; } if (info.seq != syslog_seq) { @@ -1731,8 +1728,7 @@ int do_syslog(int type, char __user *buf, int len, int source) } error -= syslog_partial; } - raw_spin_unlock(&syslog_lock); - printk_safe_exit_irq(); + mutex_unlock(&syslog_lock); break; /* Size of the log buffer */ case SYSLOG_ACTION_SIZE_BUFFER: @@ -1935,6 +1931,76 @@ static void call_console_drivers(const char *ext_text, size_t ext_len, } } +/* + * Recursion is tracked separately on each CPU. If NMIs are supported, an + * additional NMI context per CPU is also separately tracked. Until per-CPU + * is available, a separate "early tracking" is performed. + */ +static DEFINE_PER_CPU(u8, printk_count); +static u8 printk_count_early; +#ifdef CONFIG_HAVE_NMI +static DEFINE_PER_CPU(u8, printk_count_nmi); +static u8 printk_count_nmi_early; +#endif + +/* + * Recursion is limited to keep the output sane. printk() should not require + * more than 1 level of recursion (allowing, for example, printk() to trigger + * a WARN), but a higher value is used in case some printk-internal errors + * exist, such as the ringbuffer validation checks failing. + */ +#define PRINTK_MAX_RECURSION 3 + +/* + * Return a pointer to the dedicated counter for the CPU+context of the + * caller. + */ +static u8 *__printk_recursion_counter(void) +{ +#ifdef CONFIG_HAVE_NMI + if (in_nmi()) { + if (printk_percpu_data_ready()) + return this_cpu_ptr(&printk_count_nmi); + return &printk_count_nmi_early; + } +#endif + if (printk_percpu_data_ready()) + return this_cpu_ptr(&printk_count); + return &printk_count_early; +} + +/* + * Enter recursion tracking. Interrupts are disabled to simplify tracking. + * The caller must check the boolean return value to see if the recursion is + * allowed. On failure, interrupts are not disabled. + * + * @recursion_ptr must be a variable of type (u8 *) and is the same variable + * that is passed to printk_exit_irqrestore(). + */ +#define printk_enter_irqsave(recursion_ptr, flags) \ +({ \ + bool success = true; \ + \ + typecheck(u8 *, recursion_ptr); \ + local_irq_save(flags); \ + (recursion_ptr) = __printk_recursion_counter(); \ + if (*(recursion_ptr) > PRINTK_MAX_RECURSION) { \ + local_irq_restore(flags); \ + success = false; \ + } else { \ + (*(recursion_ptr))++; \ + } \ + success; \ +}) + +/* Exit recursion tracking, restoring interrupts. */ +#define printk_exit_irqrestore(recursion_ptr, flags) \ + do { \ + typecheck(u8 *, recursion_ptr); \ + (*(recursion_ptr))--; \ + local_irq_restore(flags); \ + } while (0) + int printk_delay_msec __read_mostly; static inline void printk_delay(void) @@ -2037,11 +2103,14 @@ int vprintk_store(int facility, int level, struct prb_reserved_entry e; enum printk_info_flags flags = 0; struct printk_record r; + unsigned long irqflags; u16 trunc_msg_len = 0; char prefix_buf[8]; + u8 *recursion_ptr; u16 reserve_size; va_list args2; u16 text_len; + int ret = 0; u64 ts_nsec; /* @@ -2052,6 +2121,9 @@ int vprintk_store(int facility, int level, */ ts_nsec = local_clock(); + if (!printk_enter_irqsave(recursion_ptr, irqflags)) + return 0; + /* * The sprintf needs to come first since the syslog prefix might be * passed in as a parameter. An extra byte must be reserved so that @@ -2089,7 +2161,8 @@ int vprintk_store(int facility, int level, prb_commit(&e); } - return text_len; + ret = text_len; + goto out; } } @@ -2105,7 +2178,7 @@ int vprintk_store(int facility, int level, prb_rec_init_wr(&r, reserve_size + trunc_msg_len); if (!prb_reserve(&e, prb, &r)) - return 0; + goto out; } /* fill message */ @@ -2127,7 +2200,10 @@ int vprintk_store(int facility, int level, else prb_final_commit(&e); - return (text_len + trunc_msg_len); + ret = text_len + trunc_msg_len; +out: + printk_exit_irqrestore(recursion_ptr, irqflags); + return ret; } asmlinkage int vprintk_emit(int facility, int level, @@ -2136,7 +2212,6 @@ asmlinkage int vprintk_emit(int facility, int level, { int printed_len; bool in_sched = false; - unsigned long flags; /* Suppress unimportant messages after panic happens */ if (unlikely(suppress_printk)) @@ -2150,9 +2225,7 @@ asmlinkage int vprintk_emit(int facility, int level, boot_delay_msec(level); printk_delay(); - printk_safe_enter_irqsave(flags); printed_len = vprintk_store(facility, level, dev_info, fmt, args); - printk_safe_exit_irqrestore(flags); /* If called from the scheduler, we can not call up(). */ if (!in_sched) { @@ -2573,9 +2646,9 @@ again: for (;;) { size_t ext_len = 0; + int handover; size_t len; - printk_safe_enter_irqsave(flags); skip: if (!prb_read_valid(prb, console_seq, &r)) break; @@ -2625,19 +2698,22 @@ skip: * were to occur on another CPU, it may wait for this one to * finish. This task can not be preempted if there is a * waiter waiting to take over. + * + * Interrupts are disabled because the hand over to a waiter + * must not be interrupted until the hand over is completed + * (@console_waiter is cleared). */ + printk_safe_enter_irqsave(flags); console_lock_spinning_enable(); stop_critical_timings(); /* don't trace print latency */ call_console_drivers(ext_text, ext_len, text, len); start_critical_timings(); - if (console_lock_spinning_disable_and_check()) { - printk_safe_exit_irqrestore(flags); - return; - } - + handover = console_lock_spinning_disable_and_check(); printk_safe_exit_irqrestore(flags); + if (handover) + return; if (do_cond_resched) cond_resched(); @@ -2656,8 +2732,6 @@ skip: * flush, no worries. */ retry = prb_read_valid(prb, next_seq, NULL); - printk_safe_exit_irqrestore(flags); - if (retry && console_trylock()) goto again; } @@ -2719,13 +2793,8 @@ void console_flush_on_panic(enum con_flush_mode mode) console_trylock(); console_may_schedule = 0; - if (mode == CONSOLE_REPLAY_ALL) { - unsigned long flags; - - printk_safe_enter_irqsave(flags); + if (mode == CONSOLE_REPLAY_ALL) console_seq = prb_first_valid_seq(prb); - printk_safe_exit_irqrestore(flags); - } console_unlock(); } @@ -2860,7 +2929,6 @@ static int try_enable_new_console(struct console *newcon, bool user_specified) */ void register_console(struct console *newcon) { - unsigned long flags; struct console *bcon = NULL; int err; @@ -2965,9 +3033,9 @@ void register_console(struct console *newcon) exclusive_console_stop_seq = console_seq; /* Get a consistent copy of @syslog_seq. */ - raw_spin_lock_irqsave(&syslog_lock, flags); + mutex_lock(&syslog_lock); console_seq = syslog_seq; - raw_spin_unlock_irqrestore(&syslog_lock, flags); + mutex_unlock(&syslog_lock); } console_unlock(); console_sysfs_notify(); @@ -3377,14 +3445,12 @@ bool kmsg_dump_get_line(struct kmsg_dump_iter *iter, bool syslog, struct printk_info info; unsigned int line_count; struct printk_record r; - unsigned long flags; size_t l = 0; bool ret = false; if (iter->cur_seq < min_seq) iter->cur_seq = min_seq; - printk_safe_enter_irqsave(flags); prb_rec_init_rd(&r, &info, line, size); /* Read text or count text lines? */ @@ -3405,7 +3471,6 @@ bool kmsg_dump_get_line(struct kmsg_dump_iter *iter, bool syslog, iter->cur_seq = r.info->seq + 1; ret = true; out: - printk_safe_exit_irqrestore(flags); if (len) *len = l; return ret; @@ -3437,7 +3502,6 @@ bool kmsg_dump_get_buffer(struct kmsg_dump_iter *iter, bool syslog, u64 min_seq = latched_seq_read_nolock(&clear_seq); struct printk_info info; struct printk_record r; - unsigned long flags; u64 seq; u64 next_seq; size_t len = 0; @@ -3450,7 +3514,6 @@ bool kmsg_dump_get_buffer(struct kmsg_dump_iter *iter, bool syslog, if (iter->cur_seq < min_seq) iter->cur_seq = min_seq; - printk_safe_enter_irqsave(flags); if (prb_read_valid_info(prb, iter->cur_seq, &info, NULL)) { if (info.seq != iter->cur_seq) { /* messages are gone, move to first available one */ @@ -3459,10 +3522,8 @@ bool kmsg_dump_get_buffer(struct kmsg_dump_iter *iter, bool syslog, } /* last entry */ - if (iter->cur_seq >= iter->next_seq) { - printk_safe_exit_irqrestore(flags); + if (iter->cur_seq >= iter->next_seq) goto out; - } /* * Find first record that fits, including all following records, @@ -3494,7 +3555,6 @@ bool kmsg_dump_get_buffer(struct kmsg_dump_iter *iter, bool syslog, iter->next_seq = next_seq; ret = true; - printk_safe_exit_irqrestore(flags); out: if (len_out) *len_out = len; @@ -3512,12 +3572,8 @@ EXPORT_SYMBOL_GPL(kmsg_dump_get_buffer); */ void kmsg_dump_rewind(struct kmsg_dump_iter *iter) { - unsigned long flags; - - printk_safe_enter_irqsave(flags); iter->cur_seq = latched_seq_read_nolock(&clear_seq); iter->next_seq = prb_next_seq(prb); - printk_safe_exit_irqrestore(flags); } EXPORT_SYMBOL_GPL(kmsg_dump_rewind); diff --git a/kernel/printk/printk_safe.c b/kernel/printk/printk_safe.c index 94232186fccb..ef0f9a2044da 100644 --- a/kernel/printk/printk_safe.c +++ b/kernel/printk/printk_safe.c @@ -4,347 +4,16 @@ */ #include -#include -#include #include #include #include -#include #include #include #include "internal.h" -/* - * In NMI and safe mode, printk() avoids taking locks. Instead, - * it uses an alternative implementation that temporary stores - * the strings into a per-CPU buffer. The content of the buffer - * is later flushed into the main ring buffer via IRQ work. - * - * The alternative implementation is chosen transparently - * by examining current printk() context mask stored in @printk_context - * per-CPU variable. - * - * The implementation allows to flush the strings also from another CPU. - * There are situations when we want to make sure that all buffers - * were handled or when IRQs are blocked. - */ - -#define SAFE_LOG_BUF_LEN ((1 << CONFIG_PRINTK_SAFE_LOG_BUF_SHIFT) - \ - sizeof(atomic_t) - \ - sizeof(atomic_t) - \ - sizeof(struct irq_work)) - -struct printk_safe_seq_buf { - atomic_t len; /* length of written data */ - atomic_t message_lost; - struct irq_work work; /* IRQ work that flushes the buffer */ - unsigned char buffer[SAFE_LOG_BUF_LEN]; -}; - -static DEFINE_PER_CPU(struct printk_safe_seq_buf, safe_print_seq); static DEFINE_PER_CPU(int, printk_context); -static DEFINE_RAW_SPINLOCK(safe_read_lock); - -#ifdef CONFIG_PRINTK_NMI -static DEFINE_PER_CPU(struct printk_safe_seq_buf, nmi_print_seq); -#endif - -/* Get flushed in a more safe context. */ -static void queue_flush_work(struct printk_safe_seq_buf *s) -{ - if (printk_percpu_data_ready()) - irq_work_queue(&s->work); -} - -/* - * Add a message to per-CPU context-dependent buffer. NMI and printk-safe - * have dedicated buffers, because otherwise printk-safe preempted by - * NMI-printk would have overwritten the NMI messages. - * - * The messages are flushed from irq work (or from panic()), possibly, - * from other CPU, concurrently with printk_safe_log_store(). Should this - * happen, printk_safe_log_store() will notice the buffer->len mismatch - * and repeat the write. - */ -static __printf(2, 0) int printk_safe_log_store(struct printk_safe_seq_buf *s, - const char *fmt, va_list args) -{ - int add; - size_t len; - va_list ap; - -again: - len = atomic_read(&s->len); - - /* The trailing '\0' is not counted into len. */ - if (len >= sizeof(s->buffer) - 1) { - atomic_inc(&s->message_lost); - queue_flush_work(s); - return 0; - } - - /* - * Make sure that all old data have been read before the buffer - * was reset. This is not needed when we just append data. - */ - if (!len) - smp_rmb(); - - va_copy(ap, args); - add = vscnprintf(s->buffer + len, sizeof(s->buffer) - len, fmt, ap); - va_end(ap); - if (!add) - return 0; - - /* - * Do it once again if the buffer has been flushed in the meantime. - * Note that atomic_cmpxchg() is an implicit memory barrier that - * makes sure that the data were written before updating s->len. - */ - if (atomic_cmpxchg(&s->len, len, len + add) != len) - goto again; - - queue_flush_work(s); - return add; -} - -static inline void printk_safe_flush_line(const char *text, int len) -{ - /* - * Avoid any console drivers calls from here, because we may be - * in NMI or printk_safe context (when in panic). The messages - * must go only into the ring buffer at this stage. Consoles will - * get explicitly called later when a crashdump is not generated. - */ - printk_deferred("%.*s", len, text); -} - -/* printk part of the temporary buffer line by line */ -static int printk_safe_flush_buffer(const char *start, size_t len) -{ - const char *c, *end; - bool header; - - c = start; - end = start + len; - header = true; - - /* Print line by line. */ - while (c < end) { - if (*c == '\n') { - printk_safe_flush_line(start, c - start + 1); - start = ++c; - header = true; - continue; - } - - /* Handle continuous lines or missing new line. */ - if ((c + 1 < end) && printk_get_level(c)) { - if (header) { - c = printk_skip_level(c); - continue; - } - - printk_safe_flush_line(start, c - start); - start = c++; - header = true; - continue; - } - - header = false; - c++; - } - - /* Check if there was a partial line. Ignore pure header. */ - if (start < end && !header) { - static const char newline[] = KERN_CONT "\n"; - - printk_safe_flush_line(start, end - start); - printk_safe_flush_line(newline, strlen(newline)); - } - - return len; -} - -static void report_message_lost(struct printk_safe_seq_buf *s) -{ - int lost = atomic_xchg(&s->message_lost, 0); - - if (lost) - printk_deferred("Lost %d message(s)!\n", lost); -} - -/* - * Flush data from the associated per-CPU buffer. The function - * can be called either via IRQ work or independently. - */ -static void __printk_safe_flush(struct irq_work *work) -{ - struct printk_safe_seq_buf *s = - container_of(work, struct printk_safe_seq_buf, work); - unsigned long flags; - size_t len; - int i; - - /* - * The lock has two functions. First, one reader has to flush all - * available message to make the lockless synchronization with - * writers easier. Second, we do not want to mix messages from - * different CPUs. This is especially important when printing - * a backtrace. - */ - raw_spin_lock_irqsave(&safe_read_lock, flags); - - i = 0; -more: - len = atomic_read(&s->len); - - /* - * This is just a paranoid check that nobody has manipulated - * the buffer an unexpected way. If we printed something then - * @len must only increase. Also it should never overflow the - * buffer size. - */ - if ((i && i >= len) || len > sizeof(s->buffer)) { - const char *msg = "printk_safe_flush: internal error\n"; - - printk_safe_flush_line(msg, strlen(msg)); - len = 0; - } - - if (!len) - goto out; /* Someone else has already flushed the buffer. */ - - /* Make sure that data has been written up to the @len */ - smp_rmb(); - i += printk_safe_flush_buffer(s->buffer + i, len - i); - - /* - * Check that nothing has got added in the meantime and truncate - * the buffer. Note that atomic_cmpxchg() is an implicit memory - * barrier that makes sure that the data were copied before - * updating s->len. - */ - if (atomic_cmpxchg(&s->len, len, 0) != len) - goto more; - -out: - report_message_lost(s); - raw_spin_unlock_irqrestore(&safe_read_lock, flags); -} - -/** - * printk_safe_flush - flush all per-cpu nmi buffers. - * - * The buffers are flushed automatically via IRQ work. This function - * is useful only when someone wants to be sure that all buffers have - * been flushed at some point. - */ -void printk_safe_flush(void) -{ - int cpu; - - for_each_possible_cpu(cpu) { -#ifdef CONFIG_PRINTK_NMI - __printk_safe_flush(&per_cpu(nmi_print_seq, cpu).work); -#endif - __printk_safe_flush(&per_cpu(safe_print_seq, cpu).work); - } -} - -/** - * printk_safe_flush_on_panic - flush all per-cpu nmi buffers when the system - * goes down. - * - * Similar to printk_safe_flush() but it can be called even in NMI context when - * the system goes down. It does the best effort to get NMI messages into - * the main ring buffer. - * - * Note that it could try harder when there is only one CPU online. - */ -void printk_safe_flush_on_panic(void) -{ - /* - * Make sure that we could access the safe buffers. - * Do not risk a double release when more CPUs are up. - */ - if (raw_spin_is_locked(&safe_read_lock)) { - if (num_online_cpus() > 1) - return; - - debug_locks_off(); - raw_spin_lock_init(&safe_read_lock); - } - - printk_safe_flush(); -} - -#ifdef CONFIG_PRINTK_NMI -/* - * Safe printk() for NMI context. It uses a per-CPU buffer to - * store the message. NMIs are not nested, so there is always only - * one writer running. But the buffer might get flushed from another - * CPU, so we need to be careful. - */ -static __printf(1, 0) int vprintk_nmi(const char *fmt, va_list args) -{ - struct printk_safe_seq_buf *s = this_cpu_ptr(&nmi_print_seq); - - return printk_safe_log_store(s, fmt, args); -} - -void noinstr printk_nmi_enter(void) -{ - this_cpu_add(printk_context, PRINTK_NMI_CONTEXT_OFFSET); -} - -void noinstr printk_nmi_exit(void) -{ - this_cpu_sub(printk_context, PRINTK_NMI_CONTEXT_OFFSET); -} - -/* - * Marks a code that might produce many messages in NMI context - * and the risk of losing them is more critical than eventual - * reordering. - * - * It has effect only when called in NMI context. Then printk() - * will store the messages into the main logbuf directly. - */ -void printk_nmi_direct_enter(void) -{ - if (this_cpu_read(printk_context) & PRINTK_NMI_CONTEXT_MASK) - this_cpu_or(printk_context, PRINTK_NMI_DIRECT_CONTEXT_MASK); -} - -void printk_nmi_direct_exit(void) -{ - this_cpu_and(printk_context, ~PRINTK_NMI_DIRECT_CONTEXT_MASK); -} - -#else - -static __printf(1, 0) int vprintk_nmi(const char *fmt, va_list args) -{ - return 0; -} - -#endif /* CONFIG_PRINTK_NMI */ - -/* - * Lock-less printk(), to avoid deadlocks should the printk() recurse - * into itself. It uses a per-CPU buffer to store the message, just like - * NMI. - */ -static __printf(1, 0) int vprintk_safe(const char *fmt, va_list args) -{ - struct printk_safe_seq_buf *s = this_cpu_ptr(&safe_print_seq); - - return printk_safe_log_store(s, fmt, args); -} - /* Can be preempted by NMI. */ void __printk_safe_enter(void) { @@ -369,46 +38,15 @@ asmlinkage int vprintk(const char *fmt, va_list args) * Use the main logbuf even in NMI. But avoid calling console * drivers that might have their own locks. */ - if ((this_cpu_read(printk_context) & PRINTK_NMI_DIRECT_CONTEXT_MASK)) { - unsigned long flags; + if (this_cpu_read(printk_context) || in_nmi()) { int len; - printk_safe_enter_irqsave(flags); len = vprintk_store(0, LOGLEVEL_DEFAULT, NULL, fmt, args); - printk_safe_exit_irqrestore(flags); defer_console_output(); return len; } - /* Use extra buffer in NMI. */ - if (this_cpu_read(printk_context) & PRINTK_NMI_CONTEXT_MASK) - return vprintk_nmi(fmt, args); - - /* Use extra buffer to prevent a recursion deadlock in safe mode. */ - if (this_cpu_read(printk_context) & PRINTK_SAFE_CONTEXT_MASK) - return vprintk_safe(fmt, args); - /* No obstacles. */ return vprintk_default(fmt, args); } EXPORT_SYMBOL(vprintk); - -void __init printk_safe_init(void) -{ - int cpu; - - for_each_possible_cpu(cpu) { - struct printk_safe_seq_buf *s; - - s = &per_cpu(safe_print_seq, cpu); - init_irq_work(&s->work, __printk_safe_flush); - -#ifdef CONFIG_PRINTK_NMI - s = &per_cpu(nmi_print_seq, cpu); - init_irq_work(&s->work, __printk_safe_flush); -#endif - } - - /* Flush pending messages that did not have scheduled IRQ works. */ - printk_safe_flush(); -} diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index f8b80b5bab71..da057c13f9fe 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -9797,7 +9797,6 @@ void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) tracing_off(); local_irq_save(flags); - printk_nmi_direct_enter(); /* Simulate the iterator */ trace_init_global_iter(&iter); @@ -9879,7 +9878,6 @@ void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) atomic_dec(&per_cpu_ptr(iter.array_buffer->data, cpu)->disabled); } atomic_dec(&dump_running); - printk_nmi_direct_exit(); local_irq_restore(flags); } EXPORT_SYMBOL_GPL(ftrace_dump); diff --git a/lib/nmi_backtrace.c b/lib/nmi_backtrace.c index 8abe1870dba4..f9e89001b52e 100644 --- a/lib/nmi_backtrace.c +++ b/lib/nmi_backtrace.c @@ -75,12 +75,6 @@ void nmi_trigger_cpumask_backtrace(const cpumask_t *mask, touch_softlockup_watchdog(); } - /* - * Force flush any remote buffers that might be stuck in IRQ context - * and therefore could not run their irq_work. - */ - printk_safe_flush(); - clear_bit_unlock(0, &backtrace_flag); put_cpu(); } @@ -92,8 +86,14 @@ module_param(backtrace_idle, bool, 0644); bool nmi_cpu_backtrace(struct pt_regs *regs) { int cpu = smp_processor_id(); + unsigned long flags; if (cpumask_test_cpu(cpu, to_cpumask(backtrace_mask))) { + /* + * Allow nested NMI backtraces while serializing + * against other CPUs. + */ + printk_cpu_lock_irqsave(flags); if (!READ_ONCE(backtrace_idle) && regs && cpu_in_idle(instruction_pointer(regs))) { pr_warn("NMI backtrace for cpu %d skipped: idling at %pS\n", cpu, (void *)instruction_pointer(regs)); @@ -104,6 +104,7 @@ bool nmi_cpu_backtrace(struct pt_regs *regs) else dump_stack(); } + printk_cpu_unlock_irqrestore(flags); cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask)); return true; }