split out record-buffer patch

3 files changed, 596 insertions, 476 deletions

diff --git a/printk-devkmsg.patch b/printk-devkmsg.patch
new file mode 100644
index 0000000..a7fa795
--- /dev/null
+++ b/printk-devkmsg.patch
@@ -0,0 +1,505 @@
+---
+ drivers/base/core.c    |   52 +++++++-
+ drivers/char/mem.c     |   61 ---------
+ include/linux/printk.h |    2 
+ kernel/printk.c        |  313 +++++++++++++++++++++++++++++++++++++++++++++++++
+ 4 files changed, 365 insertions(+), 63 deletions(-)
+
+--- a/drivers/base/core.c
++++ b/drivers/base/core.c
+@@ -25,6 +25,7 @@
+ #include <linux/mutex.h>
+ #include <linux/async.h>
+ #include <linux/pm_runtime.h>
++#include <linux/netdevice.h>
+ 
+ #include "base.h"
+ #include "power/power.h"
+@@ -1843,15 +1844,60 @@ void device_shutdown(void)
+  */
+ 
+ #ifdef CONFIG_PRINTK
+-
+ int __dev_printk(const char *level, const struct device *dev,
+ 		 struct va_format *vaf)
+ {
++	char dict[128];
++	size_t dictlen = 0;
++	const char *subsys;
++
+ 	if (!dev)
+ 		return printk("%s(NULL device *): %pV", level, vaf);
+ 
+-	return printk("%s%s %s: %pV",
+-		      level, dev_driver_string(dev), dev_name(dev), vaf);
++	if (dev->class)
++		subsys = dev->class->name;
++	else if (dev->bus)
++		subsys = dev->bus->name;
++	else
++		subsys = "(NULL subsystem)";
++
++	dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
++			    "SUBSYSTEM=%s", subsys);
++
++	/*
++	 * Add device identifier DEVICE=:
++	 *   b12:8         block dev_t
++	 *   c127:3        char dev_t
++	 *   n8            netdev ifindex
++	 *   +sound:card0  subsystem:devname
++	 */
++	if (MAJOR(dev->devt)) {
++		char c;
++
++		if (strcmp(subsys, "block") == 0)
++			c = 'b';
++		else
++			c = 'c';
++		dictlen++;
++		dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
++				   "DEVICE=%c%u:%u",
++				   c, MAJOR(dev->devt), MINOR(dev->devt));
++	} else if (strcmp(subsys, "net") == 0) {
++		struct net_device *net = to_net_dev(dev);
++
++		dictlen++;
++		dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
++				    "DEVICE=n%u", net->ifindex);
++	} else {
++		dictlen++;
++		dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
++				    "DEVICE=+%s:%s", subsys, dev_name(dev));
++	}
++
++	return printk_emit(0, level[1] - '0',
++			   dict, dictlen,
++			   "%s %s: %pV",
++			   dev_driver_string(dev), dev_name(dev), vaf);
+ }
+ EXPORT_SYMBOL(__dev_printk);
+ 
+--- a/drivers/char/mem.c
++++ b/drivers/char/mem.c
+@@ -807,65 +807,6 @@ static const struct file_operations oldm
+ };
+ #endif
+ 
+-static ssize_t kmsg_writev(struct kiocb *iocb, const struct iovec *iv,
+-			   unsigned long count, loff_t pos)
+-{
+-	char *buf, *line;
+-	int i;
+-	int level = default_message_loglevel;
+-	int facility = 1;	/* LOG_USER */
+-	size_t len = iov_length(iv, count);
+-	ssize_t ret = len;
+-
+-	if (len > 1024)
+-		return -EINVAL;
+-	buf = kmalloc(len+1, GFP_KERNEL);
+-	if (buf == NULL)
+-		return -ENOMEM;
+-
+-	line = buf;
+-	for (i = 0; i < count; i++) {
+-		if (copy_from_user(line, iv[i].iov_base, iv[i].iov_len))
+-			goto out;
+-		line += iv[i].iov_len;
+-	}
+-
+-	/*
+-	 * Extract and skip the syslog prefix <[0-9]*>. Coming from userspace
+-	 * the decimal value represents 32bit, the lower 3 bit are the log
+-	 * level, the rest are the log facility.
+-	 *
+-	 * If no prefix or no userspace facility is specified, we
+-	 * enforce LOG_USER, to be able to reliably distinguish
+-	 * kernel-generated messages from userspace-injected ones.
+-	 */
+-	line = buf;
+-	if (line[0] == '<') {
+-		char *endp = NULL;
+-
+-		i = simple_strtoul(line+1, &endp, 10);
+-		if (endp && endp[0] == '>') {
+-			level = i & 7;
+-			if (i >> 3)
+-				facility = i >> 3;
+-			endp++;
+-			len -= endp - line;
+-			line = endp;
+-		}
+-	}
+-	line[len] = '\0';
+-
+-	printk_emit(facility, level, NULL, 0, "%s", line);
+-out:
+-	kfree(buf);
+-	return ret;
+-}
+-
+-static const struct file_operations kmsg_fops = {
+-	.aio_write = kmsg_writev,
+-	.llseek = noop_llseek,
+-};
+-
+ static const struct memdev {
+ 	const char *name;
+ 	umode_t mode;
+@@ -884,7 +825,7 @@ static const struct memdev {
+ 	 [7] = { "full", 0666, &full_fops, NULL },
+ 	 [8] = { "random", 0666, &random_fops, NULL },
+ 	 [9] = { "urandom", 0666, &urandom_fops, NULL },
+-	[11] = { "kmsg", 0, &kmsg_fops, NULL },
++	[11] = { "kmsg", 0644, &kmsg_fops, NULL },
+ #ifdef CONFIG_CRASH_DUMP
+ 	[12] = { "oldmem", 0, &oldmem_fops, NULL },
+ #endif
+--- a/include/linux/printk.h
++++ b/include/linux/printk.h
+@@ -300,6 +300,8 @@ extern void dump_stack(void) __cold;
+ 	no_printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)
+ #endif
+ 
++extern const struct file_operations kmsg_fops;
++
+ enum {
+ 	DUMP_PREFIX_NONE,
+ 	DUMP_PREFIX_ADDRESS,
+--- a/kernel/printk.c
++++ b/kernel/printk.c
+@@ -41,6 +41,7 @@
+ #include <linux/cpu.h>
+ #include <linux/notifier.h>
+ #include <linux/rculist.h>
++#include <linux/poll.h>
+ 
+ #include <asm/uaccess.h>
+ 
+@@ -149,6 +150,48 @@ static int console_may_schedule;
+  * length of the message text is stored in the header, the stored message
+  * is not terminated.
+  *
++ * Optionally, a message can carry a dictionary of properties (key/value pairs),
++ * to provide userspace with a machine-readable message context.
++ *
++ * Examples for well-defined, commonly used property names are:
++ *   DEVICE=b12:8               device identifier
++ *                                b12:8         block dev_t
++ *                                c127:3        char dev_t
++ *                                n8            netdev ifindex
++ *                                +sound:card0  subsystem:devname
++ *   SUBSYSTEM=pci              driver-core subsystem name
++ *
++ * Valid characters in property names are [a-zA-Z0-9.-_]. The plain text value
++ * follows directly after a '=' character. Every property is terminated by
++ * a '\0' character. The last property is not terminated.
++ *
++ * Example of a message structure:
++ *   0000  ff 8f 00 00 00 00 00 00      monotonic time in nsec
++ *   0008  34 00                        record is 52 bytes long
++ *   000a        0b 00                  text is 11 bytes long
++ *   000c              1f 00            dictionary is 23 bytes long
++ *   000e                    03 00      LOG_KERN (facility) LOG_ERR (level)
++ *   0010  69 74 27 73 20 61 20 6c      "it's a l"
++ *         69 6e 65                     "ine"
++ *   001b           44 45 56 49 43      "DEVIC"
++ *         45 3d 62 38 3a 32 00 44      "E=b8:2\0D"
++ *         52 49 56 45 52 3d 62 75      "RIVER=bu"
++ *         67                           "g"
++ *   0032     00 00 00                  padding to next message header
++ *
++ * The 'struct log' buffer header must never be directly exported to
++ * userspace, it is a kernel-private implementation detail that might
++ * need to be changed in the future, when the requirements change.
++ *
++ * /dev/kmsg exports the structured data in the following line format:
++ *   "level,sequnum,timestamp;<message text>\n"
++ *
++ * The optional key/value pairs are attached as continuation lines starting
++ * with a space character and terminated by a newline. All possible
++ * non-prinatable characters are escaped in the "\xff" notation.
++ *
++ * Users of the export format should ignore possible additional values
++ * separated by ',', and find the message after the ';' character.
+  */
+ 
+ struct log {
+@@ -297,6 +340,276 @@ static void log_store(int facility, int
+ 	log_next_seq++;
+ }
+ 
++/* /dev/kmsg - userspace message inject/listen interface */
++struct devkmsg_user {
++	u64 seq;
++	u32 idx;
++	struct mutex lock;
++	char buf[8192];
++};
++
++static ssize_t devkmsg_writev(struct kiocb *iocb, const struct iovec *iv,
++			      unsigned long count, loff_t pos)
++{
++	char *buf, *line;
++	int i;
++	int level = default_message_loglevel;
++	int facility = 1;	/* LOG_USER */
++	size_t len = iov_length(iv, count);
++	ssize_t ret = len;
++
++	if (len > LOG_LINE_MAX)
++		return -EINVAL;
++	buf = kmalloc(len+1, GFP_KERNEL);
++	if (buf == NULL)
++		return -ENOMEM;
++
++	line = buf;
++	for (i = 0; i < count; i++) {
++		if (copy_from_user(line, iv[i].iov_base, iv[i].iov_len))
++			goto out;
++		line += iv[i].iov_len;
++	}
++
++	/*
++	 * Extract and skip the syslog prefix <[0-9]*>. Coming from userspace
++	 * the decimal value represents 32bit, the lower 3 bit are the log
++	 * level, the rest are the log facility.
++	 *
++	 * If no prefix or no userspace facility is specified, we
++	 * enforce LOG_USER, to be able to reliably distinguish
++	 * kernel-generated messages from userspace-injected ones.
++	 */
++	line = buf;
++	if (line[0] == '<') {
++		char *endp = NULL;
++
++		i = simple_strtoul(line+1, &endp, 10);
++		if (endp && endp[0] == '>') {
++			level = i & 7;
++			if (i >> 3)
++				facility = i >> 3;
++			endp++;
++			len -= endp - line;
++			line = endp;
++		}
++	}
++	line[len] = '\0';
++
++	printk_emit(facility, level, NULL, 0, "%s", line);
++out:
++	kfree(buf);
++	return ret;
++}
++
++static ssize_t devkmsg_read(struct file *file, char __user *buf,
++			    size_t count, loff_t *ppos)
++{
++	struct devkmsg_user *user = file->private_data;
++	struct log *msg;
++	size_t i;
++	size_t len;
++	ssize_t ret;
++
++	if (!user)
++		return -EBADF;
++
++	mutex_lock(&user->lock);
++	raw_spin_lock(&logbuf_lock);
++	while (user->seq == log_next_seq) {
++		if (file->f_flags & O_NONBLOCK) {
++			ret = -EAGAIN;
++			raw_spin_unlock(&logbuf_lock);
++			goto out;
++		}
++
++		raw_spin_unlock(&logbuf_lock);
++		ret = wait_event_interruptible(log_wait,
++					       user->seq != log_next_seq);
++		if (ret)
++			goto out;
++		raw_spin_lock(&logbuf_lock);
++	}
++
++	if (user->seq < log_first_seq) {
++		/* our last seen message is gone, return error and reset */
++		user->idx = log_first_idx;
++		user->seq = log_first_seq;
++		ret = -EPIPE;
++		raw_spin_unlock(&logbuf_lock);
++		goto out;
++	}
++
++	msg = log_from_idx(user->idx);
++	len = sprintf(user->buf, "%u,%llu,%llu;",
++		      msg->level, user->seq, msg->ts_nsec / 1000);
++
++	/* escape non-printable characters */
++	for (i = 0; i < msg->text_len; i++) {
++		char c = log_text(msg)[i];
++
++		if (c < ' ' || c >= 128)
++			len += sprintf(user->buf + len, "\\x%02x", c);
++		else
++			user->buf[len++] = c;
++	}
++	user->buf[len++] = '\n';
++
++	if (msg->dict_len) {
++		bool line = true;
++
++		for (i = 0; i < msg->dict_len; i++) {
++			char c = log_dict(msg)[i];
++
++			if (line) {
++				user->buf[len++] = ' ';
++				line = false;
++			}
++
++			if (c == '\0') {
++				user->buf[len++] = '\n';
++				line = true;
++				continue;
++			}
++
++			if (c < ' ' || c >= 128) {
++				len += sprintf(user->buf + len, "\\x%02x", c);
++				continue;
++			}
++
++			user->buf[len++] = c;
++		}
++		user->buf[len++] = '\n';
++	}
++
++	user->idx = log_next(user->idx);
++	user->seq++;
++	raw_spin_unlock(&logbuf_lock);
++
++	if (len > count) {
++		ret = -EINVAL;
++		goto out;
++	}
++
++	if (copy_to_user(buf, user->buf, len)) {
++		ret = -EFAULT;
++		goto out;
++	}
++	ret = len;
++out:
++	mutex_unlock(&user->lock);
++	return ret;
++}
++
++static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence)
++{
++	struct devkmsg_user *user = file->private_data;
++	loff_t ret = 0;
++
++	if (!user)
++		return -EBADF;
++	if (offset)
++		return -ESPIPE;
++
++	raw_spin_lock(&logbuf_lock);
++	switch (whence) {
++	case SEEK_SET:
++		/* the first record */
++		user->idx = log_first_idx;
++		user->seq = log_first_seq;
++		break;
++	case SEEK_DATA:
++		/*
++		 * The first record after the last SYSLOG_ACTION_CLEAR,
++		 * like issued by 'dmesg -c'. Reading /dev/kmsg itself
++		 * changes no global state, and does not clear anything.
++		 */
++		user->idx = clear_idx;
++		user->seq = clear_seq;
++		break;
++	case SEEK_END:
++		/* after the last record */
++		user->idx = log_next_idx;
++		user->seq = log_next_seq;
++		break;
++	default:
++		ret = -EINVAL;
++	}
++	raw_spin_unlock(&logbuf_lock);
++	return ret;
++}
++
++static unsigned int devkmsg_poll(struct file *file, poll_table *wait)
++{
++	struct devkmsg_user *user = file->private_data;
++	int ret = 0;
++
++	if (!user)
++		return POLLERR|POLLNVAL;
++
++	poll_wait(file, &log_wait, wait);
++
++	raw_spin_lock(&logbuf_lock);
++	if (user->seq < log_next_seq) {
++		/* return error when data has vanished underneath us */
++		if (user->seq < log_first_seq)
++			ret = POLLIN|POLLRDNORM|POLLERR|POLLPRI;
++		ret = POLLIN|POLLRDNORM;
++	}
++	raw_spin_unlock(&logbuf_lock);
++
++	return ret;
++}
++
++static int devkmsg_open(struct inode *inode, struct file *file)
++{
++	struct devkmsg_user *user;
++	int err;
++
++	/* write-only does not need any file context */
++	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
++		return 0;
++
++	err = security_syslog(SYSLOG_ACTION_READ_ALL);
++	if (err)
++		return err;
++
++	user = kmalloc(sizeof(struct devkmsg_user), GFP_KERNEL);
++	if (!user)
++		return -ENOMEM;
++
++	mutex_init(&user->lock);
++
++	raw_spin_lock(&logbuf_lock);
++	user->idx = log_first_idx;
++	user->seq = log_first_seq;
++	raw_spin_unlock(&logbuf_lock);
++
++	file->private_data = user;
++	return 0;
++}
++
++static int devkmsg_release(struct inode *inode, struct file *file)
++{
++	struct devkmsg_user *user = file->private_data;
++
++	if (!user)
++		return 0;
++
++	mutex_destroy(&user->lock);
++	kfree(user);
++	return 0;
++}
++
++const struct file_operations kmsg_fops = {
++	.open = devkmsg_open,
++	.read = devkmsg_read,
++	.aio_write = devkmsg_writev,
++	.llseek = devkmsg_llseek,
++	.poll = devkmsg_poll,
++	.release = devkmsg_release,
++};
++
+ #ifdef CONFIG_KEXEC
+ /*
+  * This appends the listed symbols to /proc/vmcoreinfo
diff --git a/printk.patch b/printk-records.patch
index 0938cea..50c0f50 100644
--- a/printk.patch
+++ b/printk-records.patch
@@ -100,142 +100,85 @@ Various features of this patch:
 Tested-by: William Douglas <william.douglas@intel.com>
 Signed-off-by: Kay Sievers <kay@vrfy.org>
 ---
- drivers/base/core.c    |   52 +
- drivers/char/mem.c     |   40 -
- include/linux/printk.h |   13 
- kernel/printk.c        | 1333 +++++++++++++++++++++++++++++++++----------------
- 4 files changed, 969 insertions(+), 469 deletions(-)
+ drivers/char/mem.c     |   55 +-
+ include/linux/printk.h |   11 
+ kernel/printk.c        | 1020 ++++++++++++++++++++++++++++---------------------
+ 3 files changed, 642 insertions(+), 444 deletions(-)
 
---- a/drivers/base/core.c
-+++ b/drivers/base/core.c
-@@ -25,6 +25,7 @@
- #include <linux/mutex.h>
- #include <linux/async.h>
- #include <linux/pm_runtime.h>
-+#include <linux/netdevice.h>
- 
- #include "base.h"
- #include "power/power.h"
-@@ -1843,15 +1844,60 @@ void device_shutdown(void)
-  */
- 
- #ifdef CONFIG_PRINTK
--
- int __dev_printk(const char *level, const struct device *dev,
- 		 struct va_format *vaf)
- {
-+	char dict[128];
-+	size_t dictlen = 0;
-+	const char *subsys;
-+
- 	if (!dev)
- 		return printk("%s(NULL device *): %pV", level, vaf);
- 
--	return printk("%s%s %s: %pV",
--		      level, dev_driver_string(dev), dev_name(dev), vaf);
-+	if (dev->class)
-+		subsys = dev->class->name;
-+	else if (dev->bus)
-+		subsys = dev->bus->name;
-+	else
-+		subsys = "(NULL subsystem)";
-+
-+	dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
-+			    "SUBSYSTEM=%s", subsys);
-+
-+	/*
-+	 * Add device identifier DEVICE=:
-+	 *   b12:8         block dev_t
-+	 *   c127:3        char dev_t
-+	 *   n8            netdev ifindex
-+	 *   +sound:card0  subsystem:devname
-+	 */
-+	if (MAJOR(dev->devt)) {
-+		char c;
-+
-+		if (strcmp(subsys, "block") == 0)
-+			c = 'b';
-+		else
-+			c = 'c';
-+		dictlen++;
-+		dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
-+				   "DEVICE=%c%u:%u",
-+				   c, MAJOR(dev->devt), MINOR(dev->devt));
-+	} else if (strcmp(subsys, "net") == 0) {
-+		struct net_device *net = to_net_dev(dev);
-+
-+		dictlen++;
-+		dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
-+				    "DEVICE=n%u", net->ifindex);
-+	} else {
-+		dictlen++;
-+		dictlen += snprintf(dict + dictlen, sizeof(dict) - dictlen,
-+				    "DEVICE=+%s:%s", subsys, dev_name(dev));
-+	}
-+
-+	return printk_emit(0, level[1] - '0',
-+			   dict, dictlen,
-+			   "%s %s: %pV",
-+			   dev_driver_string(dev), dev_name(dev), vaf);
- }
- EXPORT_SYMBOL(__dev_printk);
- 
 --- a/drivers/char/mem.c
 +++ b/drivers/char/mem.c
-@@ -807,44 +807,6 @@ static const struct file_operations oldm
- };
- #endif
- 
--static ssize_t kmsg_writev(struct kiocb *iocb, const struct iovec *iv,
--			   unsigned long count, loff_t pos)
--{
+@@ -810,33 +810,54 @@ static const struct file_operations oldm
+ static ssize_t kmsg_writev(struct kiocb *iocb, const struct iovec *iv,
+ 			   unsigned long count, loff_t pos)
+ {
 -	char *line, *p;
--	int i;
++	char *buf, *line;
+ 	int i;
 -	ssize_t ret = -EFAULT;
--	size_t len = iov_length(iv, count);
--
++	int level = default_message_loglevel;
++	int facility = 1;	/* LOG_USER */
+ 	size_t len = iov_length(iv, count);
++	ssize_t ret = len;
+ 
 -	line = kmalloc(len + 1, GFP_KERNEL);
 -	if (line == NULL)
--		return -ENOMEM;
--
++	if (len > 1024)
++		return -EINVAL;
++	buf = kmalloc(len+1, GFP_KERNEL);
++	if (buf == NULL)
+ 		return -ENOMEM;
+ 
 -	/*
 -	 * copy all vectors into a single string, to ensure we do
 -	 * not interleave our log line with other printk calls
 -	 */
 -	p = line;
--	for (i = 0; i < count; i++) {
++	line = buf;
+ 	for (i = 0; i < count; i++) {
 -		if (copy_from_user(p, iv[i].iov_base, iv[i].iov_len))
--			goto out;
++		if (copy_from_user(line, iv[i].iov_base, iv[i].iov_len))
+ 			goto out;
 -		p += iv[i].iov_len;
--	}
++		line += iv[i].iov_len;
++	}
++
++	/*
++	 * Extract and skip the syslog prefix <[0-9]*>. Coming from userspace
++	 * the decimal value represents 32bit, the lower 3 bit are the log
++	 * level, the rest are the log facility.
++	 *
++	 * If no prefix or no userspace facility is specified, we
++	 * enforce LOG_USER, to be able to reliably distinguish
++	 * kernel-generated messages from userspace-injected ones.
++	 */
++	line = buf;
++	if (line[0] == '<') {
++		char *endp = NULL;
++
++		i = simple_strtoul(line+1, &endp, 10);
++		if (endp && endp[0] == '>') {
++			level = i & 7;
++			if (i >> 3)
++				facility = i >> 3;
++			endp++;
++			len -= endp - line;
++			line = endp;
++		}
+ 	}
 -	p[0] = '\0';
--
++	line[len] = '\0';
+ 
 -	ret = printk("%s", line);
 -	/* printk can add a prefix */
 -	if (ret > len)
 -		ret = len;
--out:
++	printk_emit(facility, level, NULL, 0, "%s", line);
+ out:
 -	kfree(line);
--	return ret;
--}
--
--static const struct file_operations kmsg_fops = {
--	.aio_write = kmsg_writev,
--	.llseek = noop_llseek,
--};
--
- static const struct memdev {
- 	const char *name;
- 	umode_t mode;
-@@ -863,7 +825,7 @@ static const struct memdev {
- 	 [7] = { "full", 0666, &full_fops, NULL },
- 	 [8] = { "random", 0666, &random_fops, NULL },
- 	 [9] = { "urandom", 0666, &urandom_fops, NULL },
--	[11] = { "kmsg", 0, &kmsg_fops, NULL },
-+	[11] = { "kmsg", 0644, &kmsg_fops, NULL },
- #ifdef CONFIG_CRASH_DUMP
- 	[12] = { "oldmem", 0, &oldmem_fops, NULL },
- #endif
++	kfree(buf);
+ 	return ret;
+ }
+ 
 --- a/include/linux/printk.h
 +++ b/include/linux/printk.h
 @@ -95,8 +95,19 @@ extern int printk_needs_cpu(int cpu);
@@ -258,26 +201,9 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  asmlinkage __printf(1, 2) __cold
  int printk(const char *fmt, ...);
  
-@@ -289,6 +300,8 @@ extern void dump_stack(void) __cold;
- 	no_printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)
- #endif
- 
-+extern const struct file_operations kmsg_fops;
-+
- enum {
- 	DUMP_PREFIX_NONE,
- 	DUMP_PREFIX_ADDRESS,
 --- a/kernel/printk.c
 +++ b/kernel/printk.c
-@@ -41,6 +41,7 @@
- #include <linux/cpu.h>
- #include <linux/notifier.h>
- #include <linux/rculist.h>
-+#include <linux/poll.h>
- 
- #include <asm/uaccess.h>
- 
-@@ -54,8 +55,6 @@ void asmlinkage __attribute__((weak)) ea
+@@ -54,8 +54,6 @@ void asmlinkage __attribute__((weak)) ea
  {
  }
  
@@ -286,7 +212,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  /* printk's without a loglevel use this.. */
  #define DEFAULT_MESSAGE_LOGLEVEL CONFIG_DEFAULT_MESSAGE_LOGLEVEL
  
-@@ -99,24 +98,6 @@ EXPORT_SYMBOL_GPL(console_drivers);
+@@ -99,24 +97,6 @@ EXPORT_SYMBOL_GPL(console_drivers);
  static int console_locked, console_suspended;
  
  /*
@@ -311,7 +237,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
   * If exclusive_console is non-NULL then only this console is to be printed to.
   */
  static struct console *exclusive_console;
-@@ -146,12 +127,488 @@ EXPORT_SYMBOL(console_set_on_cmdline);
+@@ -146,12 +126,176 @@ EXPORT_SYMBOL(console_set_on_cmdline);
  static int console_may_schedule;
  
  #ifdef CONFIG_PRINTK
@@ -338,48 +264,6 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
 + * length of the message text is stored in the header, the stored message
 + * is not terminated.
 + *
-+ * Optionally, a message can carry a dictionary of properties (key/value pairs),
-+ * to provide userspace with a machine-readable message context.
-+ *
-+ * Examples for well-defined, commonly used property names are:
-+ *   DEVICE=b12:8               device identifier
-+ *                                b12:8         block dev_t
-+ *                                c127:3        char dev_t
-+ *                                n8            netdev ifindex
-+ *                                +sound:card0  subsystem:devname
-+ *   SUBSYSTEM=pci              driver-core subsystem name
-+ *
-+ * Valid characters in property names are [a-zA-Z0-9.-_]. The plain text value
-+ * follows directly after a '=' character. Every property is terminated by
-+ * a '\0' character. The last property is not terminated.
-+ *
-+ * Example of a message structure:
-+ *   0000  ff 8f 00 00 00 00 00 00      monotonic time in nsec
-+ *   0008  34 00                        record is 52 bytes long
-+ *   000a        0b 00                  text is 11 bytes long
-+ *   000c              1f 00            dictionary is 23 bytes long
-+ *   000e                    03 00      LOG_KERN (facility) LOG_ERR (level)
-+ *   0010  69 74 27 73 20 61 20 6c      "it's a l"
-+ *         69 6e 65                     "ine"
-+ *   001b           44 45 56 49 43      "DEVIC"
-+ *         45 3d 62 38 3a 32 00 44      "E=b8:2\0D"
-+ *         52 49 56 45 52 3d 62 75      "RIVER=bu"
-+ *         67                           "g"
-+ *   0032     00 00 00                  padding to next message header
-+ *
-+ * The 'struct log' buffer header must never be directly exported to
-+ * userspace, it is a kernel-private implementation detail that might
-+ * need to be changed in the future, when the requirements change.
-+ *
-+ * /dev/kmsg exports the structured data in the following line format:
-+ *   "level,sequnum,timestamp;<message text>\n"
-+ *
-+ * The optional key/value pairs are attached as continuation lines starting
-+ * with a space character and terminated by a newline. All possible
-+ * non-prinatable characters are escaped in the "\xff" notation.
-+ *
-+ * Users of the export format should ignore possible additional values
-+ * separated by ',', and find the message after the ';' character.
 + */
 +
 +struct log {
@@ -389,7 +273,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
 +	u16 dict_len;		/* length of dictionary buffer */
 +	u16 level;		/* syslog level + facility */
 +};
- 
++
 +/*
 + * The logbuf_lock protects kmsg buffer, indices, counters. It is also
 + * used in interesting ways to provide interlocking in console_unlock();
@@ -400,7 +284,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
 +static volatile unsigned int logbuf_cpu = UINT_MAX;
 +
 +#define LOG_LINE_MAX 1024
-+
+ 
 +/* record buffer */
 +#define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT)
  static char __log_buf[__LOG_BUF_LEN];
@@ -530,280 +414,10 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
 +	log_next_idx += msg->len;
 +	log_next_seq++;
 +}
-+
-+/* /dev/kmsg - userspace message inject/listen interface */
-+struct devkmsg_user {
-+	u64 seq;
-+	u32 idx;
-+	struct mutex lock;
-+	char buf[8192];
-+};
-+
-+static ssize_t devkmsg_writev(struct kiocb *iocb, const struct iovec *iv,
-+			      unsigned long count, loff_t pos)
-+{
-+	char *buf, *line;
-+	int i;
-+	int level = default_message_loglevel;
-+	int facility = 1;	/* LOG_USER */
-+	size_t len = iov_length(iv, count);
-+	ssize_t ret = len;
-+
-+	if (len > LOG_LINE_MAX)
-+		return -EINVAL;
-+	buf = kmalloc(len+1, GFP_KERNEL);
-+	if (buf == NULL)
-+		return -ENOMEM;
-+
-+	line = buf;
-+	for (i = 0; i < count; i++) {
-+		if (copy_from_user(line, iv[i].iov_base, iv[i].iov_len))
-+			goto out;
-+		line += iv[i].iov_len;
-+	}
-+
-+	/*
-+	 * Extract and skip the syslog prefix <[0-9]*>. Coming from userspace
-+	 * the decimal value represents 32bit, the lower 3 bit are the log
-+	 * level, the rest are the log facility.
-+	 *
-+	 * If no prefix or no userspace facility is specified, we
-+	 * enforce LOG_USER, to be able to reliably distinguish
-+	 * kernel-generated messages from userspace-injected ones.
-+	 */
-+	line = buf;
-+	if (line[0] == '<') {
-+		char *endp = NULL;
-+
-+		i = simple_strtoul(line+1, &endp, 10);
-+		if (endp && endp[0] == '>') {
-+			level = i & 7;
-+			if (i >> 3)
-+				facility = i >> 3;
-+			endp++;
-+			len -= endp - line;
-+			line = endp;
-+		}
-+	}
-+	line[len] = '\0';
-+
-+	printk_emit(facility, level, NULL, 0, "%s", line);
-+out:
-+	kfree(buf);
-+	return ret;
-+}
-+
-+static ssize_t devkmsg_read(struct file *file, char __user *buf,
-+			    size_t count, loff_t *ppos)
-+{
-+	struct devkmsg_user *user = file->private_data;
-+	struct log *msg;
-+	size_t i;
-+	size_t len;
-+	ssize_t ret;
-+
-+	if (!user)
-+		return -EBADF;
-+
-+	mutex_lock(&user->lock);
-+	raw_spin_lock(&logbuf_lock);
-+	while (user->seq == log_next_seq) {
-+		if (file->f_flags & O_NONBLOCK) {
-+			ret = -EAGAIN;
-+			raw_spin_unlock(&logbuf_lock);
-+			goto out;
-+		}
-+
-+		raw_spin_unlock(&logbuf_lock);
-+		ret = wait_event_interruptible(log_wait,
-+					       user->seq != log_next_seq);
-+		if (ret)
-+			goto out;
-+		raw_spin_lock(&logbuf_lock);
-+	}
-+
-+	if (user->seq < log_first_seq) {
-+		/* our last seen message is gone, return error and reset */
-+		user->idx = log_first_idx;
-+		user->seq = log_first_seq;
-+		ret = -EPIPE;
-+		raw_spin_unlock(&logbuf_lock);
-+		goto out;
-+	}
-+
-+	msg = log_from_idx(user->idx);
-+	len = sprintf(user->buf, "%u,%llu,%llu;",
-+		      msg->level, user->seq, msg->ts_nsec / 1000);
-+
-+	/* escape non-printable characters */
-+	for (i = 0; i < msg->text_len; i++) {
-+		char c = log_text(msg)[i];
-+
-+		if (c < ' ' || c >= 128)
-+			len += sprintf(user->buf + len, "\\x%02x", c);
-+		else
-+			user->buf[len++] = c;
-+	}
-+	user->buf[len++] = '\n';
-+
-+	if (msg->dict_len) {
-+		bool line = true;
-+
-+		for (i = 0; i < msg->dict_len; i++) {
-+			char c = log_dict(msg)[i];
-+
-+			if (line) {
-+				user->buf[len++] = ' ';
-+				line = false;
-+			}
-+
-+			if (c == '\0') {
-+				user->buf[len++] = '\n';
-+				line = true;
-+				continue;
-+			}
-+
-+			if (c < ' ' || c >= 128) {
-+				len += sprintf(user->buf + len, "\\x%02x", c);
-+				continue;
-+			}
-+
-+			user->buf[len++] = c;
-+		}
-+		user->buf[len++] = '\n';
-+	}
-+
-+	user->idx = log_next(user->idx);
-+	user->seq++;
-+	raw_spin_unlock(&logbuf_lock);
-+
-+	if (len > count) {
-+		ret = -EINVAL;
-+		goto out;
-+	}
-+
-+	if (copy_to_user(buf, user->buf, len)) {
-+		ret = -EFAULT;
-+		goto out;
-+	}
-+	ret = len;
-+out:
-+	mutex_unlock(&user->lock);
-+	return ret;
-+}
-+
-+static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence)
-+{
-+	struct devkmsg_user *user = file->private_data;
-+	loff_t ret = 0;
-+
-+	if (!user)
-+		return -EBADF;
-+	if (offset)
-+		return -ESPIPE;
-+
-+	raw_spin_lock(&logbuf_lock);
-+	switch (whence) {
-+	case SEEK_SET:
-+		/* the first record */
-+		user->idx = log_first_idx;
-+		user->seq = log_first_seq;
-+		break;
-+	case SEEK_DATA:
-+		/*
-+		 * The first record after the last SYSLOG_ACTION_CLEAR,
-+		 * like issued by 'dmesg -c'. Reading /dev/kmsg itself
-+		 * changes no global state, and does not clear anything.
-+		 */
-+		user->idx = clear_idx;
-+		user->seq = clear_seq;
-+		break;
-+	case SEEK_END:
-+		/* after the last record */
-+		user->idx = log_next_idx;
-+		user->seq = log_next_seq;
-+		break;
-+	default:
-+		ret = -EINVAL;
-+	}
-+	raw_spin_unlock(&logbuf_lock);
-+	return ret;
-+}
-+
-+static unsigned int devkmsg_poll(struct file *file, poll_table *wait)
-+{
-+	struct devkmsg_user *user = file->private_data;
-+	int ret = 0;
-+
-+	if (!user)
-+		return POLLERR|POLLNVAL;
-+
-+	poll_wait(file, &log_wait, wait);
-+
-+	raw_spin_lock(&logbuf_lock);
-+	if (user->seq < log_next_seq) {
-+		/* return error when data has vanished underneath us */
-+		if (user->seq < log_first_seq)
-+			ret = POLLIN|POLLRDNORM|POLLERR|POLLPRI;
-+		ret = POLLIN|POLLRDNORM;
-+	}
-+	raw_spin_unlock(&logbuf_lock);
-+
-+	return ret;
-+}
-+
-+static int devkmsg_open(struct inode *inode, struct file *file)
-+{
-+	struct devkmsg_user *user;
-+	int err;
-+
-+	/* write-only does not need any file context */
-+	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
-+		return 0;
-+
-+	err = security_syslog(SYSLOG_ACTION_READ_ALL);
-+	if (err)
-+		return err;
-+
-+	user = kmalloc(sizeof(struct devkmsg_user), GFP_KERNEL);
-+	if (!user)
-+		return -ENOMEM;
-+
-+	mutex_init(&user->lock);
-+
-+	raw_spin_lock(&logbuf_lock);
-+	user->idx = log_first_idx;
-+	user->seq = log_first_seq;
-+	raw_spin_unlock(&logbuf_lock);
-+
-+	file->private_data = user;
-+	return 0;
-+}
-+
-+static int devkmsg_release(struct inode *inode, struct file *file)
-+{
-+	struct devkmsg_user *user = file->private_data;
-+
-+	if (!user)
-+		return 0;
-+
-+	mutex_destroy(&user->lock);
-+	kfree(user);
-+	return 0;
-+}
-+
-+const struct file_operations kmsg_fops = {
-+	.open = devkmsg_open,
-+	.read = devkmsg_read,
-+	.aio_write = devkmsg_writev,
-+	.llseek = devkmsg_llseek,
-+	.poll = devkmsg_poll,
-+	.release = devkmsg_release,
-+};
  
  #ifdef CONFIG_KEXEC
  /*
-@@ -165,9 +622,9 @@ static int saved_console_loglevel = -1;
+@@ -165,9 +309,9 @@ static int saved_console_loglevel = -1;
  void log_buf_kexec_setup(void)
  {
  	VMCOREINFO_SYMBOL(log_buf);
@@ -815,7 +429,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  }
  #endif
  
-@@ -191,7 +648,6 @@ early_param("log_buf_len", log_buf_len_s
+@@ -191,7 +335,6 @@ early_param("log_buf_len", log_buf_len_s
  void __init setup_log_buf(int early)
  {
  	unsigned long flags;
@@ -823,7 +437,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	char *new_log_buf;
  	int free;
  
-@@ -219,20 +675,8 @@ void __init setup_log_buf(int early)
+@@ -219,20 +362,8 @@ void __init setup_log_buf(int early)
  	log_buf_len = new_log_buf_len;
  	log_buf = new_log_buf;
  	new_log_buf_len = 0;
@@ -846,7 +460,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	raw_spin_unlock_irqrestore(&logbuf_lock, flags);
  
  	pr_info("log_buf_len: %d\n", log_buf_len);
-@@ -332,11 +776,165 @@ static int check_syslog_permissions(int
+@@ -332,11 +463,165 @@ static int check_syslog_permissions(int
  	return 0;
  }
  
@@ -1015,7 +629,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	int error;
  
  	error = check_syslog_permissions(type, from_file);
-@@ -364,28 +962,14 @@ int do_syslog(int type, char __user *buf
+@@ -364,28 +649,14 @@ int do_syslog(int type, char __user *buf
  			goto out;
  		}
  		error = wait_event_interruptible(log_wait,
@@ -1047,7 +661,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  		/* FALL THRU */
  	/* Read last kernel messages */
  	case SYSLOG_ACTION_READ_ALL:
-@@ -399,52 +983,11 @@ int do_syslog(int type, char __user *buf
+@@ -399,52 +670,11 @@ int do_syslog(int type, char __user *buf
  			error = -EFAULT;
  			goto out;
  		}
@@ -1102,7 +716,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	/* Disable logging to console */
  	case SYSLOG_ACTION_CONSOLE_OFF:
  		if (saved_console_loglevel == -1)
-@@ -472,7 +1015,33 @@ int do_syslog(int type, char __user *buf
+@@ -472,7 +702,33 @@ int do_syslog(int type, char __user *buf
  		break;
  	/* Number of chars in the log buffer */
  	case SYSLOG_ACTION_SIZE_UNREAD:
@@ -1137,7 +751,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  		break;
  	/* Size of the log buffer */
  	case SYSLOG_ACTION_SIZE_BUFFER:
-@@ -501,29 +1070,11 @@ void kdb_syslog_data(char *syslog_data[4
+@@ -501,29 +757,11 @@ void kdb_syslog_data(char *syslog_data[4
  {
  	syslog_data[0] = log_buf;
  	syslog_data[1] = log_buf + log_buf_len;
@@ -1169,7 +783,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  static bool __read_mostly ignore_loglevel;
  
  static int __init ignore_loglevel_setup(char *str)
-@@ -540,142 +1091,33 @@ MODULE_PARM_DESC(ignore_loglevel, "ignor
+@@ -540,142 +778,33 @@ MODULE_PARM_DESC(ignore_loglevel, "ignor
  	"print all kernel messages to the console.");
  
  /*
@@ -1331,7 +945,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  }
  
  /*
-@@ -700,16 +1142,6 @@ static void zap_locks(void)
+@@ -700,16 +829,6 @@ static void zap_locks(void)
  	sema_init(&console_sem, 1);
  }
  
@@ -1348,7 +962,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  /* Check if we have any console registered that can be called early in boot. */
  static int have_callable_console(void)
  {
-@@ -722,51 +1154,6 @@ static int have_callable_console(void)
+@@ -722,51 +841,6 @@ static int have_callable_console(void)
  	return 0;
  }
  
@@ -1400,7 +1014,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  /*
   * Can we actually use the console at this time on this cpu?
   *
-@@ -810,17 +1197,12 @@ static int console_trylock_for_printk(un
+@@ -810,17 +884,12 @@ static int console_trylock_for_printk(un
  			retval = 0;
  		}
  	}
@@ -1419,7 +1033,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  
  int printk_delay_msec __read_mostly;
  
-@@ -836,15 +1218,22 @@ static inline void printk_delay(void)
+@@ -836,15 +905,22 @@ static inline void printk_delay(void)
  	}
  }
  
@@ -1449,7 +1063,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  
  	boot_delay_msec();
  	printk_delay();
-@@ -856,7 +1245,7 @@ asmlinkage int vprintk(const char *fmt,
+@@ -856,7 +932,7 @@ asmlinkage int vprintk(const char *fmt,
  	/*
  	 * Ouch, printk recursed into itself!
  	 */
@@ -1458,7 +1072,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  		/*
  		 * If a crash is occurring during printk() on this CPU,
  		 * then try to get the crash message out but make sure
-@@ -873,97 +1262,92 @@ asmlinkage int vprintk(const char *fmt,
+@@ -873,97 +949,92 @@ asmlinkage int vprintk(const char *fmt,
  
  	lockdep_off();
  	raw_spin_lock(&logbuf_lock);
@@ -1627,7 +1241,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	 */
  	if (console_trylock_for_printk(this_cpu))
  		console_unlock();
-@@ -974,12 +1358,73 @@ out_restore_irqs:
+@@ -974,12 +1045,73 @@ out_restore_irqs:
  
  	return printed_len;
  }
@@ -1703,7 +1317,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  {
  }
  
-@@ -1217,7 +1662,7 @@ int is_console_locked(void)
+@@ -1217,7 +1349,7 @@ int is_console_locked(void)
  }
  
  /*
@@ -1712,7 +1326,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
   */
  #define PRINTK_BUF_SIZE		512
  
-@@ -1253,6 +1698,10 @@ void wake_up_klogd(void)
+@@ -1253,6 +1385,10 @@ void wake_up_klogd(void)
  		this_cpu_or(printk_pending, PRINTK_PENDING_WAKEUP);
  }
  
@@ -1723,7 +1337,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  /**
   * console_unlock - unlock the console system
   *
-@@ -1263,15 +1712,16 @@ void wake_up_klogd(void)
+@@ -1263,15 +1399,16 @@ void wake_up_klogd(void)
   * by printk().  If this is the case, console_unlock(); emits
   * the output prior to releasing the lock.
   *
@@ -1743,7 +1357,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  
  	if (console_suspended) {
  		up(&console_sem);
-@@ -1281,17 +1731,41 @@ void console_unlock(void)
+@@ -1281,17 +1418,41 @@ void console_unlock(void)
  	console_may_schedule = 0;
  
  again:
@@ -1793,7 +1407,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  		start_critical_timings();
  		local_irq_restore(flags);
  	}
-@@ -1312,8 +1786,7 @@ again:
+@@ -1312,8 +1473,7 @@ again:
  	 * flush, no worries.
  	 */
  	raw_spin_lock(&logbuf_lock);
@@ -1803,7 +1417,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	raw_spin_unlock_irqrestore(&logbuf_lock, flags);
  
  	if (retry && console_trylock())
-@@ -1549,7 +2022,8 @@ void register_console(struct console *ne
+@@ -1549,7 +1709,8 @@ void register_console(struct console *ne
  		 * for us.
  		 */
  		raw_spin_lock_irqsave(&logbuf_lock, flags);
@@ -1813,7 +1427,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  		raw_spin_unlock_irqrestore(&logbuf_lock, flags);
  		/*
  		 * We're about to replay the log buffer.  Only do this to the
-@@ -1758,6 +2232,9 @@ int kmsg_dump_unregister(struct kmsg_dum
+@@ -1758,6 +1919,9 @@ int kmsg_dump_unregister(struct kmsg_dum
  }
  EXPORT_SYMBOL_GPL(kmsg_dump_unregister);
  
@@ -1823,7 +1437,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  /**
   * kmsg_dump - dump kernel log to kernel message dumpers.
   * @reason: the reason (oops, panic etc) for dumping
-@@ -1767,8 +2244,7 @@ EXPORT_SYMBOL_GPL(kmsg_dump_unregister);
+@@ -1767,8 +1931,7 @@ EXPORT_SYMBOL_GPL(kmsg_dump_unregister);
   */
  void kmsg_dump(enum kmsg_dump_reason reason)
  {
@@ -1833,7 +1447,7 @@ Signed-off-by: Kay Sievers <kay@vrfy.org>
  	struct kmsg_dumper *dumper;
  	const char *s1, *s2;
  	unsigned long l1, l2;
-@@ -1780,24 +2256,27 @@ void kmsg_dump(enum kmsg_dump_reason rea
+@@ -1780,24 +1943,27 @@ void kmsg_dump(enum kmsg_dump_reason rea
  	/* Theoretically, the log could move on after we do this, but
  	   there's not a lot we can do about that. The new messages
  	   will overwrite the start of what we dump. */
diff --git a/series b/series
index 7627619..1bf71d3 100644
--- a/series
+++ b/series
@@ -1,2 +1,3 @@
-printk.patch
+printk-records.patch
+printk-devkmsg.patch
 kern-cont.patch