freebsd-skq/share/man/man4/ktr.4

175 lines
6.0 KiB
Groff

.\" Copyright (c) 2001 John H. Baldwin <jhb@FreeBSD.org>
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd February 16, 2001
.Dt KTR 4
.Os
.Sh NAME
.Nm ktr
.Nd kernel tracing facility
.Sh SYNOPSIS
.Cd options KTR
.Cd options KTR_COMPILE=(KTR_LOCK|KTR_INTR|KTR_PROC)
.Cd options KTR_CPUMASK=0x3
.Cd options KTR_ENTRIES=8192
.Cd options KTR_EXTEND
.Cd options KTR_MASK=(KTR_INTR|KTR_PROC)
.Cd options KTR_VERBOSE
.Sh DESCRIPTION
The
.Nm
facility allows kernel events to be logged while the kernel executes so that
they can be examined later when debugging.
The only mandatory option to enable
.Nm
is
.Dq Li options KTR .
.Pp
The
.Dv KTR_ENTRIES
option sets the size of the buffer of events.
It should be a power of two.
The size of the buffer in the currently running kernel can be found via the
read-only sysctl
.Va debug.ktr.entries .
By default the buffer contains 1024 entries.
.Ss Event Masking
Event levels can be enabled or disabled to trim excessive and overly verbose
logging.
First, a mask of events is specified at compile time via the
.Dv KTR_COMPILE
option to limit which events are actually compiled into the kernel.
The default value for this option is for all events to be enabled.
.Pp
Secondly, the actual events logged while the kernel runs can be further
masked via the run time event mask.
The
.Dv KTR_MASK
option sets the default value of the run time event mask.
The runtime event mask can also be set by the
.Xr loader 8
via the
.Va debug.ktr.mask
environment variable.
It can also be examined and set after booting via the
.Va debug.ktr.mask
sysctl.
By default the run time mask is set to log only
.Dv KTR_GEN
events.
The definitions of the event mask bits can be found in
.Aq Pa sys/ktr.h .
.Ss Extensions
The kernel can be configured to compile with several extensions to the base
functionality via the
.Dv KTR_EXTEND
option.
These extensions can be checked for via the
.Va debug.ktr.extend
read-only sysctl.
It will be set to zero if the extensions are not compiled in and non-zero
if they are compiled in.
If the extensions are compiled in, then each event entry includes the filename
and line number that the event was logged from as well as the CPU on which
the current thread was executing when the event was logged.
.Pp
One extension is a CPU event mask whose default value can be changed via
the
.Dv KTR_CPUMASK
option.
A CPU must have the bit corresponding to its logical id set in this bitmask
for events that occur on it to be logged.
This mask can be set by the
.Xr loader 8
via the
.Va debug.ktr.cpumask
environment variable.
It can also be examined and set after booting via the
.Va debug.ktr.cpumask
sysctl.
By default events on all CPUs are enabled.
.Pp
The log messages go through more processing at log time when the extensions
are enabled as well.
In the basic mode, the format string and arguments are stored directly in
the event entry.
In the extended mode,
.Xr snprintf 9
is invoked and the resulting string is stored directly in the event entry.
This extra processing is more expensive in terms of execution but produces
events that are arguably more readable and easier to parse for some utilities
such as
.Xr ddb 4 .
.Pp
By default, events are only logged to the internal buffer for examination
later, but if the verbose flag is set then they are dumped to the kernel
console as well.
This flag can also be set from the loader via the
.Va debug.ktr.verbose
environment variable, or it can be examined and set after booting via the
.Va debug.ktr.verbose
sysctl.
If the flag is set to zero, which is the default, then verbose output is
disabled.
If the flag is set to one, then the contents of the log message and the CPU
number are printed to the kernel console.
If the flag is greater than one, then the filename and line number of the
event are output to the console in addition to the log message and the CPU
number.
The
.Dv KTR_VERBOSE
option sets the flag to one.
.Ss Examining the Events
The KTR buffer can be examined from within
.Xr ddb 4
via the
.Ic show ktr Op Cm /v
command.
This command displays the contents of the trace buffer one page at a time.
At the
.Dq Li --more--
prompt, the Enter key displays one more entry and prompts again.
The spacebar displays another page of entries.
Any other key quits.
By default the timestamp, filename, and line number are not displayed with
each log entry.
If the
.Cm /v
modifier is specified, then they are displayed in addition to the normal
output.
Note that the events are displayed in reverse chronological order.
That is, the most recent events are displayed first.
.Sh SEE ALSO
.Xr ktr 9
.Sh HISTORY
The KTR kernel tracing facility first appeared in
.Bsx 3.0
and was imported into
.Fx 5.0 .
.Sh BUGS
Currently there is no userland utility outside of a gdb script to extract
the event buffer from a kernel crash dump or a running kernel.