summaryrefslogtreecommitdiffstats
path: root/lib
diff options
context:
space:
mode:
authorTetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>2018-12-17 22:05:04 +0100
committerPetr Mladek <pmladek@suse.com>2018-12-18 10:53:14 +0100
commit15ff2069cb7f967dae6a8f8c176ba51447c75f00 (patch)
treeb3e4f0437c240668ad73e32b194d4e827669dbaa /lib
parentprintk: Remove print_prefix() calls with NULL buffer. (diff)
downloadlinux-15ff2069cb7f967dae6a8f8c176ba51447c75f00.tar.xz
linux-15ff2069cb7f967dae6a8f8c176ba51447c75f00.zip
printk: Add caller information to printk() output.
Sometimes we want to print a series of printk() messages to consoles without being disturbed by concurrent printk() from interrupts and/or other threads. But we can't enforce printk() callers to use their local buffers because we need to ask them to make too much changes. Also, even buffering up to one line inside printk() might cause failing to emit an important clue under critical situation. Therefore, instead of trying to help buffering, let's try to help reconstructing messages by saving caller information as of calling log_store() and adding it as "[T$thread_id]" or "[C$processor_id]" upon printing to consoles. Some examples for console output: [ 1.222773][ T1] x86: Booting SMP configuration: [ 2.779635][ T1] pci 0000:00:01.0: PCI bridge to [bus 01] [ 5.069193][ T268] Fusion MPT base driver 3.04.20 [ 9.316504][ C2] random: fast init done [ 13.413336][ T3355] Initialized host personality Some examples for /dev/kmsg output: 6,496,1222773,-,caller=T1;x86: Booting SMP configuration: 6,968,2779635,-,caller=T1;pci 0000:00:01.0: PCI bridge to [bus 01] SUBSYSTEM=pci DEVICE=+pci:0000:00:01.0 6,1353,5069193,-,caller=T268;Fusion MPT base driver 3.04.20 5,1526,9316504,-,caller=C2;random: fast init done 6,1575,13413336,-,caller=T3355;Initialized host personality Note that this patch changes max length of messages which can be printed by printk() or written to /dev/kmsg interface from 992 bytes to 976 bytes, based on an assumption that userspace won't try to write messages hitting that border line to /dev/kmsg interface. Link: http://lkml.kernel.org/r/93f19e57-5051-c67d-9af4-b17624062d44@i-love.sakura.ne.jp Cc: Dmitry Vyukov <dvyukov@google.com> Cc: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: LKML <linux-kernel@vger.kernel.org> Cc: syzkaller <syzkaller@googlegroups.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Signed-off-by: Petr Mladek <pmladek@suse.com>
Diffstat (limited to 'lib')
-rw-r--r--lib/Kconfig.debug17
1 files changed, 17 insertions, 0 deletions
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 8d24f4ed66fd..89df6750b365 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -17,6 +17,23 @@ config PRINTK_TIME
The behavior is also controlled by the kernel command line
parameter printk.time=1. See Documentation/admin-guide/kernel-parameters.rst
+config PRINTK_CALLER
+ bool "Show caller information on printks"
+ depends on PRINTK
+ help
+ Selecting this option causes printk() to add a caller "thread id" (if
+ in task context) or a caller "processor id" (if not in task context)
+ to every message.
+
+ This option is intended for environments where multiple threads
+ concurrently call printk() for many times, for it is difficult to
+ interpret without knowing where these lines (or sometimes individual
+ line which was divided into multiple lines due to race) came from.
+
+ Since toggling after boot makes the code racy, currently there is
+ no option to enable/disable at the kernel command line parameter or
+ sysfs interface.
+
config CONSOLE_LOGLEVEL_DEFAULT
int "Default console loglevel (1-15)"
range 1 15