Resguardo inicial

linux-7.2.3/kernel/printk/.kunitconfig

@@ -0,0 +1,3 @@
+CONFIG_KUNIT=y
+CONFIG_PRINTK=y
+CONFIG_PRINTK_RINGBUFFER_KUNIT_TEST=y

linux-7.2.3/kernel/printk/Makefile

@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: GPL-2.0-only
+obj-y	= printk.o
+obj-$(CONFIG_PRINTK)	+= printk_safe.o nbcon.o
+obj-$(CONFIG_A11Y_BRAILLE_CONSOLE)	+= braille.o
+obj-$(CONFIG_PRINTK_INDEX)	+= index.o
+
+obj-$(CONFIG_PRINTK)                 += printk_support.o
+printk_support-y	             := printk_ringbuffer.o
+printk_support-$(CONFIG_SYSCTL)	     += sysctl.o
+
+obj-$(CONFIG_PRINTK_RINGBUFFER_KUNIT_TEST) += printk_ringbuffer_kunit_test.o

linux-7.2.3/kernel/printk/braille.c

@@ -0,0 +1,58 @@
+// SPDX-License-Identifier: GPL-2.0
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/kernel.h>
+#include <linux/console.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+
+#include "console_cmdline.h"
+#include "braille.h"
+
+int _braille_console_setup(char **str, char **brl_options)
+{
+	size_t len;
+
+	len = str_has_prefix(*str, "brl,");
+	if (len) {
+		*brl_options = "";
+		*str += len;
+		return 0;
+	}
+
+	len = str_has_prefix(*str, "brl=");
+	if (len) {
+		*brl_options = *str + len;
+		*str = strchr(*brl_options, ',');
+		if (!*str) {
+			pr_err("need port name after brl=\n");
+			return -EINVAL;
+		}
+		*((*str)++) = 0;
+	}
+
+	return 0;
+}
+
+int
+_braille_register_console(struct console *console, struct console_cmdline *c)
+{
+	int rtn = 0;
+
+	if (c->brl_options) {
+		console->flags |= CON_BRL;
+		rtn = braille_register_console(console, c->index, c->options,
+					       c->brl_options);
+	}
+
+	return rtn;
+}
+
+int
+_braille_unregister_console(struct console *console)
+{
+	if (console->flags & CON_BRL)
+		return braille_unregister_console(console);
+
+	return 0;
+}

linux-7.2.3/kernel/printk/braille.h

@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _PRINTK_BRAILLE_H
+#define _PRINTK_BRAILLE_H
+
+#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
+
+static inline void
+braille_set_options(struct console_cmdline *c, char *brl_options)
+{
+	c->brl_options = brl_options;
+}
+
+/*
+ * Setup console according to braille options.
+ * Return -EINVAL on syntax error, 0 on success (or no braille option was
+ * actually given).
+ * Modifies str to point to the serial options
+ * Sets brl_options to the parsed braille options.
+ */
+int
+_braille_console_setup(char **str, char **brl_options);
+
+int
+_braille_register_console(struct console *console, struct console_cmdline *c);
+
+int
+_braille_unregister_console(struct console *console);
+
+#else
+
+static inline void
+braille_set_options(struct console_cmdline *c, char *brl_options)
+{
+}
+
+static inline int
+_braille_console_setup(char **str, char **brl_options)
+{
+	return 0;
+}
+
+static inline int
+_braille_register_console(struct console *console, struct console_cmdline *c)
+{
+	return 0;
+}
+
+static inline int
+_braille_unregister_console(struct console *console)
+{
+	return 0;
+}
+
+#endif
+
+#endif

linux-7.2.3/kernel/printk/console_cmdline.h

@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _CONSOLE_CMDLINE_H
+#define _CONSOLE_CMDLINE_H
+
+struct console_cmdline
+{
+	char	name[16];			/* Name of the driver	    */
+	int	index;				/* Minor dev. to use	    */
+	char	devname[32];			/* DEVNAME:0.0 style device name */
+	bool	user_specified;			/* Specified by command line vs. platform */
+	char	*options;			/* Options for the driver   */
+#ifdef CONFIG_A11Y_BRAILLE_CONSOLE
+	char	*brl_options;			/* Options for braille driver */
+#endif
+};
+
+#endif

linux-7.2.3/kernel/printk/index.c

@@ -0,0 +1,194 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Userspace indexing of printk formats
+ */
+
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/printk.h>
+#include <linux/slab.h>
+#include <linux/string_helpers.h>
+
+#include "internal.h"
+
+extern struct pi_entry *__start_printk_index[];
+extern struct pi_entry *__stop_printk_index[];
+
+/* The base dir for module formats, typically debugfs/printk/index/ */
+static struct dentry *dfs_index;
+
+static struct pi_entry *pi_get_entry(const struct module *mod, loff_t pos)
+{
+	struct pi_entry **entries;
+	unsigned int nr_entries;
+
+#ifdef CONFIG_MODULES
+	if (mod) {
+		entries = mod->printk_index_start;
+		nr_entries = mod->printk_index_size;
+	} else
+#endif
+	{
+		/* vmlinux, comes from linker symbols */
+		entries = __start_printk_index;
+		nr_entries = __stop_printk_index - __start_printk_index;
+	}
+
+	if (pos >= nr_entries)
+		return NULL;
+
+	return entries[pos];
+}
+
+static void *pi_next(struct seq_file *s, void *v, loff_t *pos)
+{
+	const struct module *mod = s->file->f_inode->i_private;
+	struct pi_entry *entry = pi_get_entry(mod, *pos);
+
+	(*pos)++;
+
+	return entry;
+}
+
+static void *pi_start(struct seq_file *s, loff_t *pos)
+{
+	/*
+	 * Make show() print the header line. Do not update *pos because
+	 * pi_next() still has to return the entry at index 0 later.
+	 */
+	if (*pos == 0)
+		return SEQ_START_TOKEN;
+
+	return pi_next(s, NULL, pos);
+}
+
+/*
+ * We need both ESCAPE_ANY and explicit characters from ESCAPE_SPECIAL in @only
+ * because otherwise ESCAPE_NAP will cause double quotes and backslashes to be
+ * ignored for quoting.
+ */
+#define seq_escape_printf_format(s, src) \
+	seq_escape_str(s, src, ESCAPE_ANY | ESCAPE_NAP | ESCAPE_APPEND, "\"\\")
+
+static int pi_show(struct seq_file *s, void *v)
+{
+	const struct pi_entry *entry = v;
+	int level = LOGLEVEL_DEFAULT;
+	enum printk_info_flags flags = 0;
+	u16 prefix_len = 0;
+
+	if (v == SEQ_START_TOKEN) {
+		seq_puts(s, "# <level/flags> filename:line function \"format\"\n");
+		return 0;
+	}
+
+	if (!entry->fmt)
+		return 0;
+
+	if (entry->level)
+		printk_parse_prefix(entry->level, &level, &flags);
+	else
+		prefix_len = printk_parse_prefix(entry->fmt, &level, &flags);
+
+
+	if (flags & LOG_CONT) {
+		/*
+		 * LOGLEVEL_DEFAULT here means "use the same level as the
+		 * message we're continuing from", not the default message
+		 * loglevel, so don't display it as such.
+		 */
+		if (level == LOGLEVEL_DEFAULT)
+			seq_puts(s, "<c>");
+		else
+			seq_printf(s, "<%d,c>", level);
+	} else
+		seq_printf(s, "<%d>", level);
+
+	seq_printf(s, " %s:%d %s \"", entry->file, entry->line, entry->func);
+	if (entry->subsys_fmt_prefix)
+		seq_escape_printf_format(s, entry->subsys_fmt_prefix);
+	seq_escape_printf_format(s, entry->fmt + prefix_len);
+	seq_puts(s, "\"\n");
+
+	return 0;
+}
+
+static void pi_stop(struct seq_file *p, void *v) { }
+
+static const struct seq_operations dfs_index_sops = {
+	.start = pi_start,
+	.next  = pi_next,
+	.show  = pi_show,
+	.stop  = pi_stop,
+};
+
+DEFINE_SEQ_ATTRIBUTE(dfs_index);
+
+#ifdef CONFIG_MODULES
+static const char *pi_get_module_name(struct module *mod)
+{
+	return mod ? mod->name : "vmlinux";
+}
+#else
+static const char *pi_get_module_name(struct module *mod)
+{
+	return "vmlinux";
+}
+#endif
+
+static void pi_create_file(struct module *mod)
+{
+	debugfs_create_file(pi_get_module_name(mod), 0444, dfs_index,
+				       mod, &dfs_index_fops);
+}
+
+#ifdef CONFIG_MODULES
+static void pi_remove_file(struct module *mod)
+{
+	debugfs_lookup_and_remove(pi_get_module_name(mod), dfs_index);
+}
+
+static int pi_module_notify(struct notifier_block *nb, unsigned long op,
+			    void *data)
+{
+	struct module *mod = data;
+
+	switch (op) {
+	case MODULE_STATE_COMING:
+		pi_create_file(mod);
+		break;
+	case MODULE_STATE_GOING:
+		pi_remove_file(mod);
+		break;
+	default: /* we don't care about other module states */
+		break;
+	}
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block module_printk_fmts_nb = {
+	.notifier_call = pi_module_notify,
+};
+
+static void __init pi_setup_module_notifier(void)
+{
+	register_module_notifier(&module_printk_fmts_nb);
+}
+#else
+static inline void __init pi_setup_module_notifier(void) { }
+#endif
+
+static int __init pi_init(void)
+{
+	struct dentry *dfs_root = debugfs_create_dir("printk", NULL);
+
+	dfs_index = debugfs_create_dir("index", dfs_root);
+	pi_setup_module_notifier();
+	pi_create_file(NULL);
+
+	return 0;
+}
+
+/* debugfs comes up on core and must be initialised first */
+postcore_initcall(pi_init);

linux-7.2.3/kernel/printk/internal.h

@@ -0,0 +1,312 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * internal.h - printk internal definitions
+ */
+#include <linux/console.h>
+#include <linux/types.h>
+#include <linux/sysctl.h>
+
+#if defined(CONFIG_PRINTK) && defined(CONFIG_SYSCTL)
+void __init printk_sysctl_init(void);
+int devkmsg_sysctl_set_loglvl(const struct ctl_table *table, int write,
+			      void *buffer, size_t *lenp, loff_t *ppos);
+#else
+#define printk_sysctl_init() do { } while (0)
+#endif
+
+#define con_printk(lvl, con, fmt, ...)				\
+	printk(lvl pr_fmt("%s%sconsole [%s%d] " fmt),		\
+		(con->flags & CON_NBCON) ? "" : "legacy ",	\
+		(con->flags & CON_BOOT) ? "boot" : "",		\
+		con->name, con->index, ##__VA_ARGS__)
+
+/*
+ * Identify if legacy printing is forced in a dedicated kthread. If
+ * true, all printing via console lock occurs within a dedicated
+ * legacy printer thread. The only exception is on panic, after the
+ * nbcon consoles have had their chance to print the panic messages
+ * first.
+ */
+#ifdef CONFIG_PREEMPT_RT
+# define force_legacy_kthread()	(true)
+#else
+# define force_legacy_kthread()	(false)
+#endif
+
+#ifdef CONFIG_PRINTK
+
+#ifdef CONFIG_PRINTK_CALLER
+#define PRINTK_PREFIX_MAX	48
+#else
+#define PRINTK_PREFIX_MAX	32
+#endif
+
+/*
+ * the maximum size of a formatted record (i.e. with prefix added
+ * per line and dropped messages or in extended message format)
+ */
+#define PRINTK_MESSAGE_MAX	2048
+
+/* the maximum size allowed to be reserved for a record */
+#define PRINTKRB_RECORD_MAX	1024
+
+/* Flags for a single printk record. */
+enum printk_info_flags {
+	/* always show on console, ignore console_loglevel */
+	LOG_FORCE_CON	= 1,
+	LOG_NEWLINE	= 2,	/* text ended with a newline */
+	LOG_CONT	= 8,	/* text is a fragment of a continuation line */
+};
+
+struct printk_ringbuffer;
+struct dev_printk_info;
+
+extern struct printk_ringbuffer *prb;
+extern bool printk_kthreads_running;
+extern bool printk_kthreads_ready;
+extern bool debug_non_panic_cpus;
+
+__printf(4, 0)
+int vprintk_store(int facility, int level,
+		  const struct dev_printk_info *dev_info,
+		  const char *fmt, va_list args);
+
+__printf(1, 0) int vprintk_default(const char *fmt, va_list args);
+
+void __printk_safe_enter(void);
+void __printk_safe_exit(void);
+
+bool printk_percpu_data_ready(void);
+
+#define printk_safe_enter_irqsave(flags)	\
+	do {					\
+		local_irq_save(flags);		\
+		__printk_safe_enter();		\
+	} while (0)
+
+#define printk_safe_exit_irqrestore(flags)	\
+	do {					\
+		__printk_safe_exit();		\
+		local_irq_restore(flags);	\
+	} while (0)
+
+void defer_console_output(void);
+bool is_printk_legacy_deferred(void);
+bool is_printk_force_console(void);
+
+u16 printk_parse_prefix(const char *text, int *level,
+			enum printk_info_flags *flags);
+void console_lock_spinning_enable(void);
+int console_lock_spinning_disable_and_check(int cookie);
+
+u64 nbcon_seq_read(struct console *con);
+void nbcon_seq_force(struct console *con, u64 seq);
+bool nbcon_alloc(struct console *con);
+void nbcon_free(struct console *con);
+enum nbcon_prio nbcon_get_default_prio(void);
+void nbcon_atomic_flush_pending(void);
+bool nbcon_legacy_emit_next_record(struct console *con, bool *handover,
+				   int cookie, bool use_atomic);
+bool nbcon_kthread_create(struct console *con);
+void nbcon_kthread_stop(struct console *con);
+void nbcon_kthreads_wake(void);
+
+/**
+ * nbcon_kthread_wake - Wake up a console printing thread
+ * @con:	Console to operate on
+ */
+static inline void nbcon_kthread_wake(struct console *con)
+{
+	/*
+	 * Guarantee any new records can be seen by tasks preparing to wait
+	 * before this context checks if the rcuwait is empty.
+	 *
+	 * The full memory barrier in rcuwait_wake_up() pairs with the full
+	 * memory barrier within set_current_state() of
+	 * ___rcuwait_wait_event(), which is called after prepare_to_rcuwait()
+	 * adds the waiter but before it has checked the wait condition.
+	 *
+	 * This pairs with nbcon_kthread_func:A.
+	 */
+	rcuwait_wake_up(&con->rcuwait); /* LMM(nbcon_kthread_wake:A) */
+}
+
+#else
+
+#define PRINTK_PREFIX_MAX	0
+#define PRINTK_MESSAGE_MAX	0
+#define PRINTKRB_RECORD_MAX	0
+
+#define printk_kthreads_running (false)
+#define printk_kthreads_ready (false)
+
+/*
+ * In !PRINTK builds we still export console_sem
+ * semaphore and some of console functions (console_unlock()/etc.), so
+ * printk-safe must preserve the existing local IRQ guarantees.
+ */
+#define printk_safe_enter_irqsave(flags) local_irq_save(flags)
+#define printk_safe_exit_irqrestore(flags) local_irq_restore(flags)
+
+static inline bool printk_percpu_data_ready(void) { return false; }
+static inline void defer_console_output(void) { }
+static inline bool is_printk_legacy_deferred(void) { return false; }
+static inline u64 nbcon_seq_read(struct console *con) { return 0; }
+static inline void nbcon_seq_force(struct console *con, u64 seq) { }
+static inline bool nbcon_alloc(struct console *con) { return false; }
+static inline void nbcon_free(struct console *con) { }
+static inline enum nbcon_prio nbcon_get_default_prio(void) { return NBCON_PRIO_NONE; }
+static inline void nbcon_atomic_flush_pending(void) { }
+static inline bool nbcon_legacy_emit_next_record(struct console *con, bool *handover,
+						 int cookie, bool use_atomic) { return false; }
+static inline void nbcon_kthread_wake(struct console *con) { }
+static inline void nbcon_kthreads_wake(void) { }
+
+#endif /* CONFIG_PRINTK */
+
+extern bool have_boot_console;
+extern bool have_nbcon_console;
+extern bool have_legacy_console;
+extern bool legacy_allow_panic_sync;
+
+/**
+ * struct console_flush_type - Define available console flush methods
+ * @nbcon_atomic:	Flush directly using nbcon_atomic() callback
+ * @nbcon_offload:	Offload flush to printer thread
+ * @legacy_direct:	Call the legacy loop in this context
+ * @legacy_offload:	Offload the legacy loop into IRQ or legacy thread
+ *
+ * Note that the legacy loop also flushes the nbcon consoles.
+ */
+struct console_flush_type {
+	bool	nbcon_atomic;
+	bool	nbcon_offload;
+	bool	legacy_direct;
+	bool	legacy_offload;
+};
+
+extern bool console_irqwork_blocked;
+
+/*
+ * Identify which console flushing methods should be used in the context of
+ * the caller.
+ */
+static inline void printk_get_console_flush_type(struct console_flush_type *ft)
+{
+	memset(ft, 0, sizeof(*ft));
+
+	switch (nbcon_get_default_prio()) {
+	case NBCON_PRIO_NORMAL:
+		if (have_nbcon_console && !have_boot_console) {
+			if (printk_kthreads_running && !console_irqwork_blocked)
+				ft->nbcon_offload = true;
+			else
+				ft->nbcon_atomic = true;
+		}
+
+		/* Legacy consoles are flushed directly when possible. */
+		if (have_legacy_console || have_boot_console) {
+			if (!is_printk_legacy_deferred())
+				ft->legacy_direct = true;
+			else if (!console_irqwork_blocked)
+				ft->legacy_offload = true;
+		}
+		break;
+
+	case NBCON_PRIO_EMERGENCY:
+		if (have_nbcon_console && !have_boot_console)
+			ft->nbcon_atomic = true;
+
+		/* Legacy consoles are flushed directly when possible. */
+		if (have_legacy_console || have_boot_console) {
+			if (!is_printk_legacy_deferred())
+				ft->legacy_direct = true;
+			else if (!console_irqwork_blocked)
+				ft->legacy_offload = true;
+		}
+		break;
+
+	case NBCON_PRIO_PANIC:
+		/*
+		 * In panic, the nbcon consoles will directly print. But
+		 * only allowed if there are no boot consoles.
+		 */
+		if (have_nbcon_console && !have_boot_console)
+			ft->nbcon_atomic = true;
+
+		if (have_legacy_console || have_boot_console) {
+			/*
+			 * This is the same decision as NBCON_PRIO_NORMAL
+			 * except that offloading never occurs in panic.
+			 *
+			 * Note that console_flush_on_panic() will flush
+			 * legacy consoles anyway, even if unsafe.
+			 */
+			if (!is_printk_legacy_deferred())
+				ft->legacy_direct = true;
+
+			/*
+			 * In panic, if nbcon atomic printing occurs,
+			 * the legacy consoles must remain silent until
+			 * explicitly allowed.
+			 */
+			if (ft->nbcon_atomic && !legacy_allow_panic_sync)
+				ft->legacy_direct = false;
+		}
+		break;
+
+	default:
+		WARN_ON_ONCE(1);
+		break;
+	}
+}
+
+extern struct printk_buffers printk_shared_pbufs;
+
+/**
+ * struct printk_buffers - Buffers to read/format/output printk messages.
+ * @outbuf:	After formatting, contains text to output.
+ * @scratchbuf:	Used as temporary ringbuffer reading and string-print space.
+ */
+struct printk_buffers {
+	char	outbuf[PRINTK_MESSAGE_MAX];
+	char	scratchbuf[PRINTKRB_RECORD_MAX];
+};
+
+/**
+ * struct printk_message - Container for a prepared printk message.
+ * @pbufs:	printk buffers used to prepare the message.
+ * @outbuf_len:	The length of prepared text in @pbufs->outbuf to output. This
+ *		does not count the terminator. A value of 0 means there is
+ *		nothing to output and this record should be skipped.
+ * @seq:	The sequence number of the record used for @pbufs->outbuf.
+ * @dropped:	The number of dropped records from reading @seq.
+ * @cpu:	CPU on which the message was generated.
+ * @pid:	PID of the task that generated the message
+ * @comm:	Name of the task that generated the message.
+ */
+struct printk_message {
+	struct printk_buffers	*pbufs;
+	unsigned int		outbuf_len;
+	u64			seq;
+	unsigned long		dropped;
+#ifdef CONFIG_PRINTK_EXECUTION_CTX
+	int			cpu;
+	pid_t			pid;
+	char			comm[TASK_COMM_LEN];
+#endif
+};
+
+bool printk_get_next_message(struct printk_message *pmsg, u64 seq,
+			     bool is_extended, bool may_supress);
+
+#ifdef CONFIG_PRINTK
+void console_prepend_dropped(struct printk_message *pmsg, unsigned long dropped);
+void console_prepend_replay(struct printk_message *pmsg);
+#endif
+
+#ifdef CONFIG_SMP
+bool is_printk_cpu_sync_owner(void);
+#else
+static inline bool is_printk_cpu_sync_owner(void) { return false; }
+#endif

linux-7.2.3/kernel/printk/printk_ringbuffer.h

@@ -0,0 +1,442 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef _KERNEL_PRINTK_RINGBUFFER_H
+#define _KERNEL_PRINTK_RINGBUFFER_H
+
+#include <linux/atomic.h>
+#include <linux/bits.h>
+#include <linux/dev_printk.h>
+#include <linux/stddef.h>
+#include <linux/types.h>
+
+/*
+ * Meta information about each stored message.
+ *
+ * All fields are set by the printk code except for @seq, which is
+ * set by the ringbuffer code.
+ */
+struct printk_info {
+	u64	seq;		/* sequence number */
+	u64	ts_nsec;	/* timestamp in nanoseconds */
+	u16	text_len;	/* length of text message */
+	u8	facility;	/* syslog facility */
+	u8	flags:5;	/* internal record flags */
+	u8	level:3;	/* syslog level */
+	u32	caller_id;	/* thread id or processor id */
+#ifdef CONFIG_PRINTK_EXECUTION_CTX
+	u32	caller_id2;	/* caller_id complement */
+	/* name of the task that generated the message */
+	char	comm[TASK_COMM_LEN];
+#endif
+
+	struct dev_printk_info	dev_info;
+};
+
+/*
+ * A structure providing the buffers, used by writers and readers.
+ *
+ * Writers:
+ * Using prb_rec_init_wr(), a writer sets @text_buf_size before calling
+ * prb_reserve(). On success, prb_reserve() sets @info and @text_buf to
+ * buffers reserved for that writer.
+ *
+ * Readers:
+ * Using prb_rec_init_rd(), a reader sets all fields before calling
+ * prb_read_valid(). Note that the reader provides the @info and @text_buf,
+ * buffers. On success, the struct pointed to by @info will be filled and
+ * the char array pointed to by @text_buf will be filled with text data.
+ */
+struct printk_record {
+	struct printk_info	*info;
+	char			*text_buf;
+	unsigned int		text_buf_size;
+};
+
+/* Specifies the logical position and span of a data block. */
+struct prb_data_blk_lpos {
+	unsigned long	begin;
+	unsigned long	next;
+};
+
+/*
+ * A descriptor: the complete meta-data for a record.
+ *
+ * @state_var: A bitwise combination of descriptor ID and descriptor state.
+ */
+struct prb_desc {
+	atomic_long_t			state_var;
+	struct prb_data_blk_lpos	text_blk_lpos;
+};
+
+/* A ringbuffer of "ID + data" elements. */
+struct prb_data_ring {
+	unsigned int	size_bits;
+	char		*data;
+	atomic_long_t	head_lpos;
+	atomic_long_t	tail_lpos;
+};
+
+/* A ringbuffer of "struct prb_desc" elements. */
+struct prb_desc_ring {
+	unsigned int		count_bits;
+	struct prb_desc		*descs;
+	struct printk_info	*infos;
+	atomic_long_t		head_id;
+	atomic_long_t		tail_id;
+	atomic_long_t		last_finalized_seq;
+};
+
+/*
+ * The high level structure representing the printk ringbuffer.
+ *
+ * @fail: Count of failed prb_reserve() calls where not even a data-less
+ *        record was created.
+ */
+struct printk_ringbuffer {
+	struct prb_desc_ring	desc_ring;
+	struct prb_data_ring	text_data_ring;
+	atomic_long_t		fail;
+};
+
+/*
+ * Used by writers as a reserve/commit handle.
+ *
+ * @rb:         Ringbuffer where the entry is reserved.
+ * @irqflags:   Saved irq flags to restore on entry commit.
+ * @id:         ID of the reserved descriptor.
+ * @text_space: Total occupied buffer space in the text data ring, including
+ *              ID, alignment padding, and wrapping data blocks.
+ *
+ * This structure is an opaque handle for writers. Its contents are only
+ * to be used by the ringbuffer implementation.
+ */
+struct prb_reserved_entry {
+	struct printk_ringbuffer	*rb;
+	unsigned long			irqflags;
+	unsigned long			id;
+	unsigned int			text_space;
+};
+
+/* The possible responses of a descriptor state-query. */
+enum desc_state {
+	desc_miss	=  -1,	/* ID mismatch (pseudo state) */
+	desc_reserved	= 0x0,	/* reserved, in use by writer */
+	desc_committed	= 0x1,	/* committed by writer, could get reopened */
+	desc_finalized	= 0x2,	/* committed, no further modification allowed */
+	desc_reusable	= 0x3,	/* free, not yet used by any writer */
+};
+
+#define _DATA_SIZE(sz_bits)	(1UL << (sz_bits))
+#define _DESCS_COUNT(ct_bits)	(1UL << (ct_bits))
+#define DESC_SV_BITS		BITS_PER_LONG
+#define DESC_FLAGS_SHIFT	(DESC_SV_BITS - 2)
+#define DESC_FLAGS_MASK		(3UL << DESC_FLAGS_SHIFT)
+#define DESC_STATE(sv)		(3UL & (sv >> DESC_FLAGS_SHIFT))
+#define DESC_SV(id, state)	(((unsigned long)state << DESC_FLAGS_SHIFT) | id)
+#define DESC_ID_MASK		(~DESC_FLAGS_MASK)
+#define DESC_ID(sv)		((sv) & DESC_ID_MASK)
+
+/*
+ * Special data block logical position values (for fields of
+ * @prb_desc.text_blk_lpos).
+ *
+ * - Bit0 is used to identify if the record has no data block. (Implemented in
+ *   the LPOS_DATALESS() macro.)
+ *
+ * - Bit1 specifies the reason for not having a data block.
+ *
+ * These special values could never be real lpos values because of the
+ * meta data and alignment padding of data blocks. (See to_blk_size() for
+ * details.)
+ */
+#define FAILED_LPOS		0x1
+#define EMPTY_LINE_LPOS		0x3
+
+#define FAILED_BLK_LPOS	\
+{				\
+	.begin	= FAILED_LPOS,	\
+	.next	= FAILED_LPOS,	\
+}
+
+/*
+ * Descriptor Bootstrap
+ *
+ * The descriptor array is minimally initialized to allow immediate usage
+ * by readers and writers. The requirements that the descriptor array
+ * initialization must satisfy:
+ *
+ *   Req1
+ *     The tail must point to an existing (committed or reusable) descriptor.
+ *     This is required by the implementation of prb_first_seq().
+ *
+ *   Req2
+ *     Readers must see that the ringbuffer is initially empty.
+ *
+ *   Req3
+ *     The first record reserved by a writer is assigned sequence number 0.
+ *
+ * To satisfy Req1, the tail initially points to a descriptor that is
+ * minimally initialized (having no data block, i.e. data-less with the
+ * data block's lpos @begin and @next values set to FAILED_LPOS).
+ *
+ * To satisfy Req2, the initial tail descriptor is initialized to the
+ * reusable state. Readers recognize reusable descriptors as existing
+ * records, but skip over them.
+ *
+ * To satisfy Req3, the last descriptor in the array is used as the initial
+ * head (and tail) descriptor. This allows the first record reserved by a
+ * writer (head + 1) to be the first descriptor in the array. (Only the first
+ * descriptor in the array could have a valid sequence number of 0.)
+ *
+ * The first time a descriptor is reserved, it is assigned a sequence number
+ * with the value of the array index. A "first time reserved" descriptor can
+ * be recognized because it has a sequence number of 0 but does not have an
+ * index of 0. (Only the first descriptor in the array could have a valid
+ * sequence number of 0.) After the first reservation, all future reservations
+ * (recycling) simply involve incrementing the sequence number by the array
+ * count.
+ *
+ *   Hack #1
+ *     Only the first descriptor in the array is allowed to have the sequence
+ *     number 0. In this case it is not possible to recognize if it is being
+ *     reserved the first time (set to index value) or has been reserved
+ *     previously (increment by the array count). This is handled by _always_
+ *     incrementing the sequence number by the array count when reserving the
+ *     first descriptor in the array. In order to satisfy Req3, the sequence
+ *     number of the first descriptor in the array is initialized to minus
+ *     the array count. Then, upon the first reservation, it is incremented
+ *     to 0, thus satisfying Req3.
+ *
+ *   Hack #2
+ *     prb_first_seq() can be called at any time by readers to retrieve the
+ *     sequence number of the tail descriptor. However, due to Req2 and Req3,
+ *     initially there are no records to report the sequence number of
+ *     (sequence numbers are u64 and there is nothing less than 0). To handle
+ *     this, the sequence number of the initial tail descriptor is initialized
+ *     to 0. Technically this is incorrect, because there is no record with
+ *     sequence number 0 (yet) and the tail descriptor is not the first
+ *     descriptor in the array. But it allows prb_read_valid() to correctly
+ *     report the existence of a record for _any_ given sequence number at all
+ *     times. Bootstrapping is complete when the tail is pushed the first
+ *     time, thus finally pointing to the first descriptor reserved by a
+ *     writer, which has the assigned sequence number 0.
+ */
+
+/*
+ * Initiating Logical Value Overflows
+ *
+ * Both logical position (lpos) and ID values can be mapped to array indexes
+ * but may experience overflows during the lifetime of the system. To ensure
+ * that printk_ringbuffer can handle the overflows for these types, initial
+ * values are chosen that map to the correct initial array indexes, but will
+ * result in overflows soon.
+ *
+ *   BLK0_LPOS
+ *     The initial @head_lpos and @tail_lpos for data rings. It is at index
+ *     0 and the lpos value is such that it will overflow on the first wrap.
+ *
+ *   DESC0_ID
+ *     The initial @head_id and @tail_id for the desc ring. It is at the last
+ *     index of the descriptor array (see Req3 above) and the ID value is such
+ *     that it will overflow on the second wrap.
+ */
+#define BLK0_LPOS(sz_bits)	(-(_DATA_SIZE(sz_bits)))
+#define DESC0_ID(ct_bits)	DESC_ID(-(_DESCS_COUNT(ct_bits) + 1))
+#define DESC0_SV(ct_bits)	DESC_SV(DESC0_ID(ct_bits), desc_reusable)
+
+/*
+ * Define a ringbuffer with an external text data buffer. The same as
+ * DEFINE_PRINTKRB() but requires specifying an external buffer for the
+ * text data.
+ *
+ * Note: The specified external buffer must be of the size:
+ *       2 ^ (descbits + avgtextbits)
+ */
+#define _DEFINE_PRINTKRB(name, descbits, avgtextbits, text_buf)			\
+static struct prb_desc _##name##_descs[_DESCS_COUNT(descbits)] = {				\
+	/* the initial head and tail */								\
+	[_DESCS_COUNT(descbits) - 1] = {							\
+		/* reusable */									\
+		.state_var	= ATOMIC_INIT(DESC0_SV(descbits)),				\
+		/* no associated data block */							\
+		.text_blk_lpos	= FAILED_BLK_LPOS,						\
+	},											\
+};												\
+static struct printk_info _##name##_infos[_DESCS_COUNT(descbits)] = {				\
+	/* this will be the first record reserved by a writer */				\
+	[0] = {											\
+		/* will be incremented to 0 on the first reservation */				\
+		.seq = -(u64)_DESCS_COUNT(descbits),						\
+	},											\
+	/* the initial head and tail */								\
+	[_DESCS_COUNT(descbits) - 1] = {							\
+		/* reports the first seq value during the bootstrap phase */			\
+		.seq = 0,									\
+	},											\
+};												\
+static struct printk_ringbuffer name = {							\
+	.desc_ring = {										\
+		.count_bits	= descbits,							\
+		.descs		= &_##name##_descs[0],						\
+		.infos		= &_##name##_infos[0],						\
+		.head_id	= ATOMIC_INIT(DESC0_ID(descbits)),				\
+		.tail_id	= ATOMIC_INIT(DESC0_ID(descbits)),				\
+		.last_finalized_seq = ATOMIC_INIT(0),						\
+	},											\
+	.text_data_ring = {									\
+		.size_bits	= (avgtextbits) + (descbits),					\
+		.data		= text_buf,							\
+		.head_lpos	= ATOMIC_LONG_INIT(BLK0_LPOS((avgtextbits) + (descbits))),	\
+		.tail_lpos	= ATOMIC_LONG_INIT(BLK0_LPOS((avgtextbits) + (descbits))),	\
+	},											\
+	.fail			= ATOMIC_LONG_INIT(0),						\
+}
+
+/**
+ * DEFINE_PRINTKRB() - Define a ringbuffer.
+ *
+ * @name:        The name of the ringbuffer variable.
+ * @descbits:    The number of descriptors as a power-of-2 value.
+ * @avgtextbits: The average text data size per record as a power-of-2 value.
+ *
+ * This is a macro for defining a ringbuffer and all internal structures
+ * such that it is ready for immediate use. See _DEFINE_PRINTKRB() for a
+ * variant where the text data buffer can be specified externally.
+ */
+#define DEFINE_PRINTKRB(name, descbits, avgtextbits)				\
+static char _##name##_text[1U << ((avgtextbits) + (descbits))]			\
+			__aligned(__alignof__(unsigned long));			\
+_DEFINE_PRINTKRB(name, descbits, avgtextbits, &_##name##_text[0])
+
+/* Writer Interface */
+
+/**
+ * prb_rec_init_wr() - Initialize a buffer for writing records.
+ *
+ * @r:             The record to initialize.
+ * @text_buf_size: The needed text buffer size.
+ */
+static inline void prb_rec_init_wr(struct printk_record *r,
+				   unsigned int text_buf_size)
+{
+	r->info = NULL;
+	r->text_buf = NULL;
+	r->text_buf_size = text_buf_size;
+}
+
+bool prb_reserve(struct prb_reserved_entry *e, struct printk_ringbuffer *rb,
+		 struct printk_record *r);
+bool prb_reserve_in_last(struct prb_reserved_entry *e, struct printk_ringbuffer *rb,
+			 struct printk_record *r, u32 caller_id, unsigned int max_size);
+void prb_commit(struct prb_reserved_entry *e);
+void prb_final_commit(struct prb_reserved_entry *e);
+
+void prb_init(struct printk_ringbuffer *rb,
+	      char *text_buf, unsigned int text_buf_size,
+	      struct prb_desc *descs, unsigned int descs_count_bits,
+	      struct printk_info *infos);
+unsigned int prb_record_text_space(struct prb_reserved_entry *e);
+
+/* Reader Interface */
+
+/**
+ * prb_rec_init_rd() - Initialize a buffer for reading records.
+ *
+ * @r:             The record to initialize.
+ * @info:          A buffer to store record meta-data.
+ * @text_buf:      A buffer to store text data.
+ * @text_buf_size: The size of @text_buf.
+ *
+ * Initialize all the fields that a reader is interested in. All arguments
+ * (except @r) are optional. Only record data for arguments that are
+ * non-NULL or non-zero will be read.
+ */
+static inline void prb_rec_init_rd(struct printk_record *r,
+				   struct printk_info *info,
+				   char *text_buf, unsigned int text_buf_size)
+{
+	r->info = info;
+	r->text_buf = text_buf;
+	r->text_buf_size = text_buf_size;
+}
+
+/**
+ * prb_for_each_record() - Iterate over the records of a ringbuffer.
+ *
+ * @from: The sequence number to begin with.
+ * @rb:   The ringbuffer to iterate over.
+ * @s:    A u64 to store the sequence number on each iteration.
+ * @r:    A printk_record to store the record on each iteration.
+ *
+ * This is a macro for conveniently iterating over a ringbuffer.
+ * Note that @s may not be the sequence number of the record on each
+ * iteration. For the sequence number, @r->info->seq should be checked.
+ *
+ * Context: Any context.
+ */
+#define prb_for_each_record(from, rb, s, r) \
+for ((s) = from; prb_read_valid(rb, s, r); (s) = (r)->info->seq + 1)
+
+/**
+ * prb_for_each_info() - Iterate over the meta data of a ringbuffer.
+ *
+ * @from: The sequence number to begin with.
+ * @rb:   The ringbuffer to iterate over.
+ * @s:    A u64 to store the sequence number on each iteration.
+ * @i:    A printk_info to store the record meta data on each iteration.
+ * @lc:   An unsigned int to store the text line count of each record.
+ *
+ * This is a macro for conveniently iterating over a ringbuffer.
+ * Note that @s may not be the sequence number of the record on each
+ * iteration. For the sequence number, @i->seq should be checked.
+ *
+ * Context: Any context.
+ */
+#define prb_for_each_info(from, rb, s, i, lc) \
+for ((s) = from; prb_read_valid_info(rb, s, i, lc); (s) = (i)->seq + 1)
+
+bool prb_read_valid(struct printk_ringbuffer *rb, u64 seq,
+		    struct printk_record *r);
+bool prb_read_valid_info(struct printk_ringbuffer *rb, u64 seq,
+			 struct printk_info *info, unsigned int *line_count);
+
+u64 prb_first_seq(struct printk_ringbuffer *rb);
+u64 prb_first_valid_seq(struct printk_ringbuffer *rb);
+u64 prb_next_seq(struct printk_ringbuffer *rb);
+u64 prb_next_reserve_seq(struct printk_ringbuffer *rb);
+
+#ifdef CONFIG_64BIT
+
+#define __u64seq_to_ulseq(u64seq) (u64seq)
+#define __ulseq_to_u64seq(rb, ulseq) (ulseq)
+#define ULSEQ_MAX(rb) (-1)
+
+#else /* CONFIG_64BIT */
+
+#define __u64seq_to_ulseq(u64seq) ((u32)u64seq)
+#define ULSEQ_MAX(rb) __u64seq_to_ulseq(prb_first_seq(rb) + 0x80000000UL)
+
+static inline u64 __ulseq_to_u64seq(struct printk_ringbuffer *rb, u32 ulseq)
+{
+	u64 rb_first_seq = prb_first_seq(rb);
+	u64 seq;
+
+	/*
+	 * The provided sequence is only the lower 32 bits of the ringbuffer
+	 * sequence. It needs to be expanded to 64bit. Get the first sequence
+	 * number from the ringbuffer and fold it.
+	 *
+	 * Having a 32bit representation in the console is sufficient.
+	 * If a console ever gets more than 2^31 records behind
+	 * the ringbuffer then this is the least of the problems.
+	 *
+	 * Also the access to the ring buffer is always safe.
+	 */
+	seq = rb_first_seq - (s32)((u32)rb_first_seq - ulseq);
+
+	return seq;
+}
+
+#endif /* CONFIG_64BIT */
+
+#endif /* _KERNEL_PRINTK_RINGBUFFER_H */

linux-7.2.3/kernel/printk/printk_ringbuffer_kunit_test.c

@@ -0,0 +1,327 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/cpuhplock.h>
+#include <linux/cpumask.h>
+#include <linux/init.h>
+#include <linux/kthread.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/random.h>
+#include <linux/slab.h>
+#include <linux/timer.h>
+#include <linux/wait.h>
+
+#include <kunit/resource.h>
+#include <kunit/test.h>
+
+#include "printk_ringbuffer.h"
+
+/*
+ * This KUnit tests the data integrity of the lockless printk_ringbuffer.
+ * From multiple CPUs it writes messages of varying length and content while
+ * a reader validates the correctness of the messages.
+ *
+ * IMPORTANT: The more CPUs you can use for this KUnit, the better!
+ *
+ * The test works by starting "num_online_cpus() - 1" writer threads, each
+ * pinned to their own CPU. Each writer thread loops, writing data of varying
+ * length into a printk_ringbuffer as fast as possible. The data content is
+ * an embedded data struct followed by string content repeating the byte:
+ *
+ *      'A' + CPUID
+ *
+ * The reader is running on the remaining online CPU, or if there is only one
+ * CPU on the same as the writer.
+ * It ensures that the embedded struct content is consistent with the string
+ * and that the string * is terminated and is composed of the same repeating
+ * byte as its first byte.
+ *
+ * Because the threads are running in such tight loops, they will call
+ * cond_resched() from time to time so the system stays functional.
+ *
+ * If the reader encounters an error, the test is aborted and some
+ * information about the error is reported.
+ * The runtime of the test can be configured with the runtime_ms module parameter.
+ *
+ * Note that the test is performed on a separate printk_ringbuffer instance
+ * and not the instance used by printk().
+ */
+
+static unsigned long runtime_ms = 10 * MSEC_PER_SEC;
+module_param(runtime_ms, ulong, 0400);
+
+/* test data structure */
+struct prbtest_rbdata {
+	unsigned int size;
+	char text[] __counted_by(size);
+};
+
+#define MAX_RBDATA_TEXT_SIZE 0x80
+#define MAX_PRB_RECORD_SIZE (sizeof(struct prbtest_rbdata) + MAX_RBDATA_TEXT_SIZE)
+
+struct prbtest_data {
+	struct kunit *test;
+	struct printk_ringbuffer *ringbuffer;
+	/* used by writers to signal reader of new records */
+	wait_queue_head_t new_record_wait;
+};
+
+struct prbtest_thread_data {
+	unsigned long num;
+	struct prbtest_data *test_data;
+};
+
+static void prbtest_fail_record(struct kunit *test, const struct prbtest_rbdata *dat, u64 seq)
+{
+	unsigned int len;
+
+	len = dat->size - 1;
+
+	KUNIT_FAIL(test, "BAD RECORD: seq=%llu size=%u text=%.*s\n",
+		   seq, dat->size,
+		   len < MAX_RBDATA_TEXT_SIZE ? len : -1,
+		   len < MAX_RBDATA_TEXT_SIZE ? dat->text : "<invalid>");
+}
+
+static bool prbtest_check_data(const struct prbtest_rbdata *dat)
+{
+	unsigned int len;
+
+	/* Sane size? At least one character + trailing '\0' */
+	if (dat->size < 2 || dat->size > MAX_RBDATA_TEXT_SIZE)
+		return false;
+
+	len = dat->size - 1;
+	if (dat->text[len] != '\0')
+		return false;
+
+	/* String repeats with the same character? */
+	while (len--) {
+		if (dat->text[len] != dat->text[0])
+			return false;
+	}
+
+	return true;
+}
+
+static int prbtest_writer(void *data)
+{
+	struct prbtest_thread_data *tr = data;
+	char text_id = 'A' + tr->num;
+	struct prb_reserved_entry e;
+	struct prbtest_rbdata *dat;
+	u32 record_size, text_size;
+	unsigned long count = 0;
+	struct printk_record r;
+
+	kunit_info(tr->test_data->test, "start thread %03lu (writer)\n", tr->num);
+
+	for (;;) {
+		/* ensure at least 1 character + trailing '\0' */
+		text_size = get_random_u32_inclusive(2, MAX_RBDATA_TEXT_SIZE);
+		if (WARN_ON_ONCE(text_size < 2))
+			text_size = 2;
+		if (WARN_ON_ONCE(text_size > MAX_RBDATA_TEXT_SIZE))
+			text_size = MAX_RBDATA_TEXT_SIZE;
+
+		record_size = sizeof(struct prbtest_rbdata) + text_size;
+		WARN_ON_ONCE(record_size > MAX_PRB_RECORD_SIZE);
+
+		/* specify the text sizes for reservation */
+		prb_rec_init_wr(&r, record_size);
+
+		/*
+		 * Reservation can fail if:
+		 *
+		 *      - No free descriptor is available.
+		 *      - The buffer is full, and the oldest record is reserved
+		 *        but not yet committed.
+		 *
+		 * It actually happens in this test because all CPUs are trying
+		 * to write an unbounded number of messages in a tight loop.
+		 * These failures are intentionally ignored because this test
+		 * focuses on races, ringbuffer consistency, and pushing system
+		 * usability limits.
+		 */
+		if (prb_reserve(&e, tr->test_data->ringbuffer, &r)) {
+			r.info->text_len = record_size;
+
+			dat = (struct prbtest_rbdata *)r.text_buf;
+			dat->size = text_size;
+			memset(dat->text, text_id, text_size - 1);
+			dat->text[text_size - 1] = '\0';
+
+			prb_commit(&e);
+
+			wake_up_interruptible(&tr->test_data->new_record_wait);
+		}
+
+		if ((count++ & 0x3fff) == 0)
+			cond_resched();
+
+		if (kthread_should_stop())
+			break;
+	}
+
+	kunit_info(tr->test_data->test, "end thread %03lu: wrote=%lu\n", tr->num, count);
+
+	return 0;
+}
+
+struct prbtest_wakeup_timer {
+	struct timer_list timer;
+	struct task_struct *task;
+};
+
+static void prbtest_wakeup_callback(struct timer_list *timer)
+{
+	struct prbtest_wakeup_timer *wakeup = timer_container_of(wakeup, timer, timer);
+
+	set_tsk_thread_flag(wakeup->task, TIF_NOTIFY_SIGNAL);
+	wake_up_process(wakeup->task);
+}
+
+static int prbtest_reader(struct prbtest_data *test_data, unsigned long timeout_ms)
+{
+	struct prbtest_wakeup_timer wakeup;
+	char text_buf[MAX_PRB_RECORD_SIZE];
+	unsigned long count = 0;
+	struct printk_info info;
+	struct printk_record r;
+	u64 seq = 0;
+
+	wakeup.task = current;
+	timer_setup_on_stack(&wakeup.timer, prbtest_wakeup_callback, 0);
+	mod_timer(&wakeup.timer, jiffies + msecs_to_jiffies(timeout_ms));
+
+	prb_rec_init_rd(&r, &info, text_buf, sizeof(text_buf));
+
+	kunit_info(test_data->test, "start reader\n");
+
+	while (!wait_event_interruptible(test_data->new_record_wait,
+					 prb_read_valid(test_data->ringbuffer, seq, &r))) {
+		/* check/track the sequence */
+		if (info.seq < seq)
+			KUNIT_FAIL(test_data->test, "BAD SEQ READ: request=%llu read=%llu\n",
+				   seq, info.seq);
+
+		if (!prbtest_check_data((struct prbtest_rbdata *)r.text_buf))
+			prbtest_fail_record(test_data->test,
+					    (struct prbtest_rbdata *)r.text_buf, info.seq);
+
+		if ((count++ & 0x3fff) == 0)
+			cond_resched();
+
+		seq = info.seq + 1;
+	}
+
+	timer_delete_sync(&wakeup.timer);
+	timer_destroy_on_stack(&wakeup.timer);
+
+	kunit_info(test_data->test, "end reader: read=%lu seq=%llu\n", count, info.seq);
+
+	return 0;
+}
+
+KUNIT_DEFINE_ACTION_WRAPPER(prbtest_cpumask_cleanup, free_cpumask_var, struct cpumask *);
+KUNIT_DEFINE_ACTION_WRAPPER(prbtest_kthread_cleanup, kthread_stop, struct task_struct *);
+
+static void prbtest_add_cpumask_cleanup(struct kunit *test, cpumask_var_t mask)
+{
+	int err;
+
+	err = kunit_add_action_or_reset(test, prbtest_cpumask_cleanup, mask);
+	KUNIT_ASSERT_EQ(test, err, 0);
+}
+
+static void prbtest_add_kthread_cleanup(struct kunit *test, struct task_struct *kthread)
+{
+	int err;
+
+	err = kunit_add_action_or_reset(test, prbtest_kthread_cleanup, kthread);
+	KUNIT_ASSERT_EQ(test, err, 0);
+}
+
+static inline void prbtest_prb_reinit(struct printk_ringbuffer *rb)
+{
+	prb_init(rb, rb->text_data_ring.data, rb->text_data_ring.size_bits, rb->desc_ring.descs,
+		 rb->desc_ring.count_bits, rb->desc_ring.infos);
+}
+
+static void test_readerwriter(struct kunit *test)
+{
+	/* Equivalent to CONFIG_LOG_BUF_SHIFT=13 */
+	DEFINE_PRINTKRB(test_rb, 8, 5);
+
+	struct prbtest_thread_data *thread_data;
+	struct prbtest_data *test_data;
+	struct task_struct *thread;
+	cpumask_var_t test_cpus;
+	int cpu, reader_cpu;
+
+	KUNIT_ASSERT_TRUE(test, alloc_cpumask_var(&test_cpus, GFP_KERNEL));
+	prbtest_add_cpumask_cleanup(test, test_cpus);
+
+	cpus_read_lock();
+	/*
+	 * Failure of KUNIT_ASSERT() kills the current task
+	 * so it can not be called while the CPU hotplug lock is held.
+	 * Instead use a snapshot of the online CPUs.
+	 * If they change during test execution it is unfortunate but not a grave error.
+	 */
+	cpumask_copy(test_cpus, cpu_online_mask);
+	cpus_read_unlock();
+
+	/* One CPU is for the reader, all others are writers */
+	reader_cpu = cpumask_first(test_cpus);
+	if (cpumask_weight(test_cpus) == 1)
+		kunit_warn(test, "more than one CPU is recommended");
+	else
+		cpumask_clear_cpu(reader_cpu, test_cpus);
+
+	/* KUnit test can get restarted more times. */
+	prbtest_prb_reinit(&test_rb);
+
+	test_data = kunit_kmalloc(test, sizeof(*test_data), GFP_KERNEL);
+	KUNIT_ASSERT_NOT_NULL(test, test_data);
+	test_data->test = test;
+	test_data->ringbuffer = &test_rb;
+	init_waitqueue_head(&test_data->new_record_wait);
+
+	kunit_info(test, "running for %lu ms\n", runtime_ms);
+
+	for_each_cpu(cpu, test_cpus) {
+		thread_data = kunit_kmalloc(test, sizeof(*thread_data), GFP_KERNEL);
+		KUNIT_ASSERT_NOT_NULL(test, thread_data);
+		thread_data->test_data = test_data;
+		thread_data->num = cpu;
+
+		thread = kthread_run_on_cpu(prbtest_writer, thread_data, cpu,
+					    "prbtest writer %u");
+		KUNIT_ASSERT_NOT_ERR_OR_NULL(test, thread);
+		prbtest_add_kthread_cleanup(test, thread);
+	}
+
+	kunit_info(test, "starting test\n");
+
+	set_cpus_allowed_ptr(current, cpumask_of(reader_cpu));
+	prbtest_reader(test_data, runtime_ms);
+
+	kunit_info(test, "completed test\n");
+}
+
+static struct kunit_case prb_test_cases[] = {
+	KUNIT_CASE_SLOW(test_readerwriter),
+	{}
+};
+
+static struct kunit_suite prb_test_suite = {
+	.name       = "printk-ringbuffer",
+	.test_cases = prb_test_cases,
+};
+kunit_test_suite(prb_test_suite);
+
+MODULE_IMPORT_NS("EXPORTED_FOR_KUNIT_TESTING");
+MODULE_AUTHOR("John Ogness <john.ogness@linutronix.de>");
+MODULE_DESCRIPTION("printk_ringbuffer KUnit test");
+MODULE_LICENSE("GPL");

linux-7.2.3/kernel/printk/printk_safe.c

@@ -0,0 +1,84 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * printk_safe.c - Safe printk for printk-deadlock-prone contexts
+ */
+
+#include <linux/preempt.h>
+#include <linux/kdb.h>
+#include <linux/smp.h>
+#include <linux/cpumask.h>
+#include <linux/printk.h>
+#include <linux/kprobes.h>
+
+#include "internal.h"
+
+/* Context where printk messages are never suppressed */
+static atomic_t force_con;
+
+void printk_force_console_enter(void)
+{
+	atomic_inc(&force_con);
+}
+
+void printk_force_console_exit(void)
+{
+	atomic_dec(&force_con);
+}
+
+bool is_printk_force_console(void)
+{
+	return atomic_read(&force_con);
+}
+
+static DEFINE_PER_CPU(int, printk_context);
+
+/* Can be preempted by NMI. */
+void __printk_safe_enter(void)
+{
+	this_cpu_inc(printk_context);
+}
+
+/* Can be preempted by NMI. */
+void __printk_safe_exit(void)
+{
+	this_cpu_dec(printk_context);
+}
+
+void __printk_deferred_enter(void)
+{
+	cant_migrate();
+	__printk_safe_enter();
+}
+
+void __printk_deferred_exit(void)
+{
+	cant_migrate();
+	__printk_safe_exit();
+}
+
+bool is_printk_legacy_deferred(void)
+{
+	/*
+	 * The per-CPU variable @printk_context can be read safely in any
+	 * context. CPU migration is always disabled when set.
+	 *
+	 * A context holding the printk_cpu_sync must not spin waiting for
+	 * another CPU. For legacy printing, it could be the console_lock
+	 * or the port lock.
+	 */
+	return (force_legacy_kthread() ||
+		this_cpu_read(printk_context) ||
+		in_nmi() ||
+		is_printk_cpu_sync_owner());
+}
+
+asmlinkage int vprintk(const char *fmt, va_list args)
+{
+#ifdef CONFIG_KGDB_KDB
+	/* Allow to pass printk() to kdb but avoid a recursion. */
+	if (unlikely(kdb_trap_printk && kdb_printf_cpu < 0))
+		return vkdb_printf(KDB_MSGSRC_PRINTK, fmt, args);
+#endif
+	return vprintk_default(fmt, args);
+}
+EXPORT_SYMBOL(vprintk);

linux-7.2.3/kernel/printk/sysctl.c

@@ -0,0 +1,83 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * sysctl.c: General linux system control interface
+ */
+
+#include <linux/printk.h>
+#include <linux/capability.h>
+#include <linux/ratelimit.h>
+#include "internal.h"
+
+static const int ten_thousand = 10000;
+
+static int proc_dointvec_minmax_sysadmin(const struct ctl_table *table, int write,
+				void *buffer, size_t *lenp, loff_t *ppos)
+{
+	if (write && !capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
+}
+
+static const struct ctl_table printk_sysctls[] = {
+	{
+		.procname	= "printk",
+		.data		= &console_loglevel,
+		.maxlen		= 4*sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec,
+	},
+	{
+		.procname	= "printk_ratelimit",
+		.data		= &printk_ratelimit_state.interval,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_jiffies,
+	},
+	{
+		.procname	= "printk_ratelimit_burst",
+		.data		= &printk_ratelimit_state.burst,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec,
+	},
+	{
+		.procname	= "printk_delay",
+		.data		= &printk_delay_msec,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= (void *)&ten_thousand,
+	},
+	{
+		.procname	= "printk_devkmsg",
+		.data		= devkmsg_log_str,
+		.maxlen		= DEVKMSG_STR_MAX_SIZE,
+		.mode		= 0644,
+		.proc_handler	= devkmsg_sysctl_set_loglvl,
+	},
+	{
+		.procname	= "dmesg_restrict",
+		.data		= &dmesg_restrict,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax_sysadmin,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_ONE,
+	},
+	{
+		.procname	= "kptr_restrict",
+		.data		= &kptr_restrict,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax_sysadmin,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_TWO,
+	},
+};
+
+void __init printk_sysctl_init(void)
+{
+	register_sysctl_init("kernel", printk_sysctls);
+}