Commit 5752674e authored by Ingo Molnar's avatar Ingo Molnar

Documentation/ftrace.txt: update

- fix typos/grammos and clarify the text
- prettify the document some more

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
parent 985ec20a
...@@ -15,31 +15,31 @@ Introduction ...@@ -15,31 +15,31 @@ Introduction
Ftrace is an internal tracer designed to help out developers and Ftrace is an internal tracer designed to help out developers and
designers of systems to find what is going on inside the kernel. designers of systems to find what is going on inside the kernel.
It can be used for debugging or analyzing latencies and performance It can be used for debugging or analyzing latencies and
issues that take place outside of user-space. performance issues that take place outside of user-space.
Although ftrace is the function tracer, it also includes an Although ftrace is the function tracer, it also includes an
infrastructure that allows for other types of tracing. Some of the infrastructure that allows for other types of tracing. Some of
tracers that are currently in ftrace include a tracer to trace the tracers that are currently in ftrace include a tracer to
context switches, the time it takes for a high priority task to trace context switches, the time it takes for a high priority
run after it was woken up, the time interrupts are disabled, and task to run after it was woken up, the time interrupts are
more (ftrace allows for tracer plugins, which means that the list of disabled, and more (ftrace allows for tracer plugins, which
tracers can always grow). means that the list of tracers can always grow).
The File System The File System
--------------- ---------------
Ftrace uses the debugfs file system to hold the control files as well Ftrace uses the debugfs file system to hold the control files as
as the files to display output. well as the files to display output.
To mount the debugfs system: To mount the debugfs system:
# mkdir /debug # mkdir /debug
# mount -t debugfs nodev /debug # mount -t debugfs nodev /debug
(Note: it is more common to mount at /sys/kernel/debug, but for simplicity ( Note: it is more common to mount at /sys/kernel/debug, but for
this document will use /debug) simplicity this document will use /debug)
That's it! (assuming that you have ftrace configured into your kernel) That's it! (assuming that you have ftrace configured into your kernel)
...@@ -50,94 +50,124 @@ of ftrace. Here is a list of some of the key files: ...@@ -50,94 +50,124 @@ of ftrace. Here is a list of some of the key files:
Note: all time values are in microseconds. Note: all time values are in microseconds.
current_tracer: This is used to set or display the current tracer current_tracer:
that is configured.
This is used to set or display the current tracer
available_tracers: This holds the different types of tracers that that is configured.
have been compiled into the kernel. The tracers
listed here can be configured by echoing their name available_tracers:
into current_tracer.
This holds the different types of tracers that
tracing_enabled: This sets or displays whether the current_tracer have been compiled into the kernel. The
is activated and tracing or not. Echo 0 into this tracers listed here can be configured by
file to disable the tracer or 1 to enable it. echoing their name into current_tracer.
trace: This file holds the output of the trace in a human readable tracing_enabled:
format (described below).
This sets or displays whether the current_tracer
latency_trace: This file shows the same trace but the information is activated and tracing or not. Echo 0 into this
is organized more to display possible latencies file to disable the tracer or 1 to enable it.
in the system (described below).
trace:
trace_pipe: The output is the same as the "trace" file but this
file is meant to be streamed with live tracing. This file holds the output of the trace in a human
Reads from this file will block until new data readable format (described below).
is retrieved. Unlike the "trace" and "latency_trace"
files, this file is a consumer. This means reading latency_trace:
from this file causes sequential reads to display
more current data. Once data is read from this This file shows the same trace but the information
file, it is consumed, and will not be read is organized more to display possible latencies
again with a sequential read. The "trace" and in the system (described below).
"latency_trace" files are static, and if the
tracer is not adding more data, they will display trace_pipe:
the same information every time they are read.
The output is the same as the "trace" file but this
trace_options: This file lets the user control the amount of data file is meant to be streamed with live tracing.
that is displayed in one of the above output Reads from this file will block until new data
files. is retrieved. Unlike the "trace" and "latency_trace"
files, this file is a consumer. This means reading
trace_max_latency: Some of the tracers record the max latency. from this file causes sequential reads to display
For example, the time interrupts are disabled. more current data. Once data is read from this
This time is saved in this file. The max trace file, it is consumed, and will not be read
will also be stored, and displayed by either again with a sequential read. The "trace" and
"trace" or "latency_trace". A new max trace will "latency_trace" files are static, and if the
only be recorded if the latency is greater than tracer is not adding more data, they will display
the value in this file. (in microseconds) the same information every time they are read.
buffer_size_kb: This sets or displays the number of kilobytes each CPU trace_options:
buffer can hold. The tracer buffers are the same size
for each CPU. The displayed number is the size of the This file lets the user control the amount of data
CPU buffer and not total size of all buffers. The that is displayed in one of the above output
trace buffers are allocated in pages (blocks of memory files.
that the kernel uses for allocation, usually 4 KB in size).
If the last page allocated has room for more bytes trace_max_latency:
than requested, the rest of the page will be used,
making the actual allocation bigger than requested. Some of the tracers record the max latency.
(Note, the size may not be a multiple of the page size due For example, the time interrupts are disabled.
to buffer managment overhead.) This time is saved in this file. The max trace
will also be stored, and displayed by either
This can only be updated when the current_tracer "trace" or "latency_trace". A new max trace will
is set to "nop". only be recorded if the latency is greater than
the value in this file. (in microseconds)
tracing_cpumask: This is a mask that lets the user only trace
on specified CPUS. The format is a hex string buffer_size_kb:
representing the CPUS.
This sets or displays the number of kilobytes each CPU
set_ftrace_filter: When dynamic ftrace is configured in (see the buffer can hold. The tracer buffers are the same size
section below "dynamic ftrace"), the code is dynamically for each CPU. The displayed number is the size of the
modified (code text rewrite) to disable calling of the CPU buffer and not total size of all buffers. The
function profiler (mcount). This lets tracing be configured trace buffers are allocated in pages (blocks of memory
in with practically no overhead in performance. This also that the kernel uses for allocation, usually 4 KB in size).
has a side effect of enabling or disabling specific functions If the last page allocated has room for more bytes
to be traced. Echoing names of functions into this file than requested, the rest of the page will be used,
will limit the trace to only those functions. making the actual allocation bigger than requested.
( Note, the size may not be a multiple of the page size
set_ftrace_notrace: This has an effect opposite to that of due to buffer managment overhead. )
set_ftrace_filter. Any function that is added here will not
be traced. If a function exists in both set_ftrace_filter This can only be updated when the current_tracer
and set_ftrace_notrace, the function will _not_ be traced. is set to "nop".
set_ftrace_pid: Have the function tracer only trace a single thread. tracing_cpumask:
set_graph_function: Select the function where the trace have to start This is a mask that lets the user only trace
with the function graph tracer (See the section on specified CPUS. The format is a hex string
"dynamic ftrace" for more details). representing the CPUS.
available_filter_functions: This lists the functions that ftrace set_ftrace_filter:
has processed and can trace. These are the function
names that you can pass to "set_ftrace_filter" or When dynamic ftrace is configured in (see the
"set_ftrace_notrace". (See the section "dynamic ftrace" section below "dynamic ftrace"), the code is dynamically
below for more details.) modified (code text rewrite) to disable calling of the
function profiler (mcount). This lets tracing be configured
in with practically no overhead in performance. This also
has a side effect of enabling or disabling specific functions
to be traced. Echoing names of functions into this file
will limit the trace to only those functions.
set_ftrace_notrace:
This has an effect opposite to that of
set_ftrace_filter. Any function that is added here will not
be traced. If a function exists in both set_ftrace_filter
and set_ftrace_notrace, the function will _not_ be traced.
set_ftrace_pid:
Have the function tracer only trace a single thread.
set_graph_function:
Set a "trigger" function where tracing should start
with the function graph tracer (See the section
"dynamic ftrace" for more details).
available_filter_functions:
This lists the functions that ftrace
has processed and can trace. These are the function
names that you can pass to "set_ftrace_filter" or
"set_ftrace_notrace". (See the section "dynamic ftrace"
below for more details.)
The Tracers The Tracers
...@@ -145,44 +175,66 @@ The Tracers ...@@ -145,44 +175,66 @@ The Tracers
Here is the list of current tracers that may be configured. Here is the list of current tracers that may be configured.
function - function tracer that uses mcount to trace all functions. "function"
Function call tracer to trace all kernel functions.
"function_graph_tracer"
Similar to the function tracer except that the
function tracer probes the functions on their entry
whereas the function graph tracer traces on both entry
and exit of the functions. It then provides the ability
to draw a graph of function calls similar to C code
source.
"sched_switch"
Traces the context switches and wakeups between tasks.
"irqsoff"
Traces the areas that disable interrupts and saves
the trace with the longest max latency.
See tracing_max_latency. When a new max is recorded,
it replaces the old trace. It is best to view this
trace via the latency_trace file.
function_graph_tracer - similar to the function tracer except that the "preemptoff"
function tracer probes the functions on their entry whereas the
function graph tracer traces on both entry and exit of the
functions. It then provides the ability to draw a graph of
function calls like a primitive C code source.
sched_switch - traces the context switches between tasks. Similar to irqsoff but traces and records the amount of
time for which preemption is disabled.
irqsoff - traces the areas that disable interrupts and saves "preemptirqsoff"
the trace with the longest max latency.
See tracing_max_latency. When a new max is recorded,
it replaces the old trace. It is best to view this
trace via the latency_trace file.
preemptoff - Similar to irqsoff but traces and records the amount of Similar to irqsoff and preemptoff, but traces and
time for which preemption is disabled. records the largest time for which irqs and/or preemption
is disabled.
preemptirqsoff - Similar to irqsoff and preemptoff, but traces and "wakeup"
records the largest time for which irqs and/or preemption
is disabled.
wakeup - Traces and records the max latency that it takes for Traces and records the max latency that it takes for
the highest priority task to get scheduled after the highest priority task to get scheduled after
it has been woken up. it has been woken up.
nop - This is not a tracer. To remove all tracers from tracing "hw-branch-tracer"
simply echo "nop" into current_tracer.
hw-branch-tracer - traces branches on all cpu's in a circular buffer. Uses the BTS CPU feature on x86 CPUs to traces all
branches executed.
"nop"
This is the "trace nothing" tracer. To remove all
tracers from tracing simply echo "nop" into
current_tracer.
Examples of using the tracer Examples of using the tracer
---------------------------- ----------------------------
Here are typical examples of using the tracers when controlling them only Here are typical examples of using the tracers when controlling
with the debugfs interface (without using any user-land utilities). them only with the debugfs interface (without using any
user-land utilities).
Output format: Output format:
-------------- --------------
...@@ -199,16 +251,16 @@ Here is an example of the output format of the file "trace" ...@@ -199,16 +251,16 @@ Here is an example of the output format of the file "trace"
bash-4251 [01] 10152.583855: _atomic_dec_and_lock <-dput bash-4251 [01] 10152.583855: _atomic_dec_and_lock <-dput
-------- --------
A header is printed with the tracer name that is represented by the trace. A header is printed with the tracer name that is represented by
In this case the tracer is "function". Then a header showing the format. Task the trace. In this case the tracer is "function". Then a header
name "bash", the task PID "4251", the CPU that it was running on showing the format. Task name "bash", the task PID "4251", the
"01", the timestamp in <secs>.<usecs> format, the function name that was CPU that it was running on "01", the timestamp in <secs>.<usecs>
traced "path_put" and the parent function that called this function format, the function name that was traced "path_put" and the
"path_walk". The timestamp is the time at which the function was parent function that called this function "path_walk". The
entered. timestamp is the time at which the function was entered.
The sched_switch tracer also includes tracing of task wakeups and The sched_switch tracer also includes tracing of task wakeups
context switches. and context switches.
ksoftirqd/1-7 [01] 1453.070013: 7:115:R + 2916:115:S ksoftirqd/1-7 [01] 1453.070013: 7:115:R + 2916:115:S
ksoftirqd/1-7 [01] 1453.070013: 7:115:R + 10:115:S ksoftirqd/1-7 [01] 1453.070013: 7:115:R + 10:115:S
...@@ -217,8 +269,8 @@ context switches. ...@@ -217,8 +269,8 @@ context switches.
kondemand/1-2916 [01] 1453.070013: 2916:115:S ==> 7:115:R kondemand/1-2916 [01] 1453.070013: 2916:115:S ==> 7:115:R
ksoftirqd/1-7 [01] 1453.070013: 7:115:S ==> 0:140:R ksoftirqd/1-7 [01] 1453.070013: 7:115:S ==> 0:140:R
Wake ups are represented by a "+" and the context switches are shown as Wake ups are represented by a "+" and the context switches are
"==>". The format is: shown as "==>". The format is:
Context switches: Context switches:
...@@ -232,19 +284,20 @@ Wake ups are represented by a "+" and the context switches are shown as ...@@ -232,19 +284,20 @@ Wake ups are represented by a "+" and the context switches are shown as
<pid>:<prio>:<state> + <pid>:<prio>:<state> <pid>:<prio>:<state> + <pid>:<prio>:<state>
The prio is the internal kernel priority, which is the inverse of the The prio is the internal kernel priority, which is the inverse
priority that is usually displayed by user-space tools. Zero represents of the priority that is usually displayed by user-space tools.
the highest priority (99). Prio 100 starts the "nice" priorities with Zero represents the highest priority (99). Prio 100 starts the
100 being equal to nice -20 and 139 being nice 19. The prio "140" is "nice" priorities with 100 being equal to nice -20 and 139 being
reserved for the idle task which is the lowest priority thread (pid 0). nice 19. The prio "140" is reserved for the idle task which is
the lowest priority thread (pid 0).
Latency trace format Latency trace format
-------------------- --------------------
For traces that display latency times, the latency_trace file gives For traces that display latency times, the latency_trace file
somewhat more information to see why a latency happened. Here is a typical gives somewhat more information to see why a latency happened.
trace. Here is a typical trace.
# tracer: irqsoff # tracer: irqsoff
# #
...@@ -271,20 +324,20 @@ irqsoff latency trace v1.1.5 on 2.6.26-rc8 ...@@ -271,20 +324,20 @@ irqsoff latency trace v1.1.5 on 2.6.26-rc8
<idle>-0 0d.s1 98us : trace_hardirqs_on (do_softirq) <idle>-0 0d.s1 98us : trace_hardirqs_on (do_softirq)
This shows that the current tracer is "irqsoff" tracing the time
for which interrupts were disabled. It gives the trace version
and the version of the kernel upon which this was executed on
(2.6.26-rc8). Then it displays the max latency in microsecs (97
us). The number of trace entries displayed and the total number
recorded (both are three: #3/3). The type of preemption that was
used (PREEMPT). VP, KP, SP, and HP are always zero and are
reserved for later use. #P is the number of online CPUS (#P:2).
This shows that the current tracer is "irqsoff" tracing the time for which The task is the process that was running when the latency
interrupts were disabled. It gives the trace version and the version occurred. (swapper pid: 0).
of the kernel upon which this was executed on (2.6.26-rc8). Then it displays
the max latency in microsecs (97 us). The number of trace entries displayed
and the total number recorded (both are three: #3/3). The type of
preemption that was used (PREEMPT). VP, KP, SP, and HP are always zero
and are reserved for later use. #P is the number of online CPUS (#P:2).
The task is the process that was running when the latency occurred.
(swapper pid: 0).
The start and stop (the functions in which the interrupts were disabled and The start and stop (the functions in which the interrupts were
enabled respectively) that caused the latencies: disabled and enabled respectively) that caused the latencies:
apic_timer_interrupt is where the interrupts were disabled. apic_timer_interrupt is where the interrupts were disabled.
do_softirq is where they were enabled again. do_softirq is where they were enabled again.
...@@ -320,12 +373,12 @@ The above is mostly meaningful for kernel developers. ...@@ -320,12 +373,12 @@ The above is mostly meaningful for kernel developers.
latency_trace file is relative to the start of the trace. latency_trace file is relative to the start of the trace.
delay: This is just to help catch your eye a bit better. And delay: This is just to help catch your eye a bit better. And
needs to be fixed to be only relative to the same CPU. needs to be fixed to be only relative to the same CPU.
The marks are determined by the difference between this The marks are determined by the difference between this
current trace and the next trace. current trace and the next trace.
'!' - greater than preempt_mark_thresh (default 100) '!' - greater than preempt_mark_thresh (default 100)
'+' - greater than 1 microsecond '+' - greater than 1 microsecond
' ' - less than or equal to 1 microsecond. ' ' - less than or equal to 1 microsecond.
The rest is the same as the 'trace' file. The rest is the same as the 'trace' file.
...@@ -333,14 +386,15 @@ The above is mostly meaningful for kernel developers. ...@@ -333,14 +386,15 @@ The above is mostly meaningful for kernel developers.
trace_options trace_options
------------- -------------
The trace_options file is used to control what gets printed in the trace The trace_options file is used to control what gets printed in
output. To see what is available, simply cat the file: the trace output. To see what is available, simply cat the file:
cat /debug/tracing/trace_options cat /debug/tracing/trace_options
print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \ print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj
To disable one of the options, echo in the option prepended with "no". To disable one of the options, echo in the option prepended with
"no".
echo noprint-parent > /debug/tracing/trace_options echo noprint-parent > /debug/tracing/trace_options
...@@ -350,8 +404,8 @@ To enable an option, leave off the "no". ...@@ -350,8 +404,8 @@ To enable an option, leave off the "no".
Here are the available options: Here are the available options:
print-parent - On function traces, display the calling function print-parent - On function traces, display the calling (parent)
as well as the function being traced. function as well as the function being traced.
print-parent: print-parent:
bash-4000 [01] 1477.606694: simple_strtoul <-strict_strtoul bash-4000 [01] 1477.606694: simple_strtoul <-strict_strtoul
...@@ -360,15 +414,16 @@ Here are the available options: ...@@ -360,15 +414,16 @@ Here are the available options:
bash-4000 [01] 1477.606694: simple_strtoul bash-4000 [01] 1477.606694: simple_strtoul
sym-offset - Display not only the function name, but also the offset sym-offset - Display not only the function name, but also the
in the function. For example, instead of seeing just offset in the function. For example, instead of
"ktime_get", you will see "ktime_get+0xb/0x20". seeing just "ktime_get", you will see
"ktime_get+0xb/0x20".
sym-offset: sym-offset:
bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0 bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0
sym-addr - this will also display the function address as well as sym-addr - this will also display the function address as well
the function name. as the function name.
sym-addr: sym-addr:
bash-4000 [01] 1477.606694: simple_strtoul <c0339346> bash-4000 [01] 1477.606694: simple_strtoul <c0339346>
...@@ -378,35 +433,41 @@ Here are the available options: ...@@ -378,35 +433,41 @@ Here are the available options:
bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \ bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \
(+0.000ms): simple_strtoul (strict_strtoul) (+0.000ms): simple_strtoul (strict_strtoul)
raw - This will display raw numbers. This option is best for use with raw - This will display raw numbers. This option is best for
user applications that can translate the raw numbers better than use with user applications that can translate the raw
having it done in the kernel. numbers better than having it done in the kernel.
hex - Similar to raw, but the numbers will be in a hexadecimal format. hex - Similar to raw, but the numbers will be in a hexadecimal
format.
bin - This will print out the formats in raw binary. bin - This will print out the formats in raw binary.
block - TBD (needs update) block - TBD (needs update)
stacktrace - This is one of the options that changes the trace itself. stacktrace - This is one of the options that changes the trace
When a trace is recorded, so is the stack of functions. itself. When a trace is recorded, so is the stack
This allows for back traces of trace sites. of functions. This allows for back traces of
trace sites.
userstacktrace - This option changes the trace. userstacktrace - This option changes the trace. It records a
It records a stacktrace of the current userspace thread. stacktrace of the current userspace thread.
sym-userobj - when user stacktrace are enabled, look up which object the sym-userobj - when user stacktrace are enabled, look up which
address belongs to, and print a relative address object the address belongs to, and print a
This is especially useful when ASLR is on, otherwise you don't relative address. This is especially useful when
get a chance to resolve the address to object/file/line after the app is no ASLR is on, otherwise you don't get a chance to
longer running resolve the address to object/file/line after
the app is no longer running
The lookup is performed when you read trace,trace_pipe,latency_trace. Example: The lookup is performed when you read
trace,trace_pipe,latency_trace. Example:
a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0 a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0
x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6] x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6]
sched-tree - TBD (any users??) sched-tree - trace all tasks that are on the runqueue, at
every scheduling event. Will add overhead if
there's a lot of tasks running at once.
sched_switch sched_switch
...@@ -443,18 +504,19 @@ of how to use it. ...@@ -443,18 +504,19 @@ of how to use it.
[...] [...]
As we have discussed previously about this format, the header shows As we have discussed previously about this format, the header
the name of the trace and points to the options. The "FUNCTION" shows the name of the trace and points to the options. The
is a misnomer since here it represents the wake ups and context "FUNCTION" is a misnomer since here it represents the wake ups
switches. and context switches.
The sched_switch file only lists the wake ups (represented with '+') The sched_switch file only lists the wake ups (represented with
and context switches ('==>') with the previous task or current task '+') and context switches ('==>') with the previous task or
first followed by the next task or task waking up. The format for both current task first followed by the next task or task waking up.
of these is PID:KERNEL-PRIO:TASK-STATE. Remember that the KERNEL-PRIO The format for both of these is PID:KERNEL-PRIO:TASK-STATE.
is the inverse of the actual priority with zero (0) being the highest Remember that the KERNEL-PRIO is the inverse of the actual
priority and the nice values starting at 100 (nice -20). Below is priority with zero (0) being the highest priority and the nice
a quick chart to map the kernel priority to user land priorities. values starting at 100 (nice -20). Below is a quick chart to map
the kernel priority to user land priorities.
Kernel priority: 0 to 99 ==> user RT priority 99 to 0 Kernel priority: 0 to 99 ==> user RT priority 99 to 0
Kernel priority: 100 to 139 ==> user nice -20 to 19 Kernel priority: 100 to 139 ==> user nice -20 to 19
...@@ -475,10 +537,10 @@ The task states are: ...@@ -475,10 +537,10 @@ The task states are:
ftrace_enabled ftrace_enabled
-------------- --------------
The following tracers (listed below) give different output depending The following tracers (listed below) give different output
on whether or not the sysctl ftrace_enabled is set. To set ftrace_enabled, depending on whether or not the sysctl ftrace_enabled is set. To
one can either use the sysctl function or set it via the proc set ftrace_enabled, one can either use the sysctl function or
file system interface. set it via the proc file system interface.
sysctl kernel.ftrace_enabled=1 sysctl kernel.ftrace_enabled=1
...@@ -486,12 +548,12 @@ file system interface. ...@@ -486,12 +548,12 @@ file system interface.
echo 1 > /proc/sys/kernel/ftrace_enabled echo 1 > /proc/sys/kernel/ftrace_enabled
To disable ftrace_enabled simply replace the '1' with '0' in To disable ftrace_enabled simply replace the '1' with '0' in the
the above commands. above commands.
When ftrace_enabled is set the tracers will also record the functions When ftrace_enabled is set the tracers will also record the
that are within the trace. The descriptions of the tracers functions that are within the trace. The descriptions of the
will also show an example with ftrace enabled. tracers will also show an example with ftrace enabled.
irqsoff irqsoff
...@@ -499,17 +561,18 @@ irqsoff ...@@ -499,17 +561,18 @@ irqsoff
When interrupts are disabled, the CPU can not react to any other When interrupts are disabled, the CPU can not react to any other
external event (besides NMIs and SMIs). This prevents the timer external event (besides NMIs and SMIs). This prevents the timer
interrupt from triggering or the mouse interrupt from letting the interrupt from triggering or the mouse interrupt from letting
kernel know of a new mouse event. The result is a latency with the the kernel know of a new mouse event. The result is a latency
reaction time. with the reaction time.
The irqsoff tracer tracks the time for which interrupts are disabled. The irqsoff tracer tracks the time for which interrupts are
When a new maximum latency is hit, the tracer saves the trace leading up disabled. When a new maximum latency is hit, the tracer saves
to that latency point so that every time a new maximum is reached, the old the trace leading up to that latency point so that every time a
saved trace is discarded and the new trace is saved. new maximum is reached, the old saved trace is discarded and the
new trace is saved.
To reset the maximum, echo 0 into tracing_max_latency. Here is an To reset the maximum, echo 0 into tracing_max_latency. Here is
example: an example:
# echo irqsoff > /debug/tracing/current_tracer # echo irqsoff > /debug/tracing/current_tracer
# echo 0 > /debug/tracing/tracing_max_latency # echo 0 > /debug/tracing/tracing_max_latency
...@@ -544,10 +607,11 @@ irqsoff latency trace v1.1.5 on 2.6.26 ...@@ -544,10 +607,11 @@ irqsoff latency trace v1.1.5 on 2.6.26
Here we see that that we had a latency of 12 microsecs (which is Here we see that that we had a latency of 12 microsecs (which is
very good). The _write_lock_irq in sys_setpgid disabled interrupts. very good). The _write_lock_irq in sys_setpgid disabled
The difference between the 12 and the displayed timestamp 14us occurred interrupts. The difference between the 12 and the displayed
because the clock was incremented between the time of recording the max timestamp 14us occurred because the clock was incremented
latency and the time of recording the function that had that latency. between the time of recording the max latency and the time of
recording the function that had that latency.
Note the above example had ftrace_enabled not set. If we set the Note the above example had ftrace_enabled not set. If we set the
ftrace_enabled, we get a much larger output: ftrace_enabled, we get a much larger output:
...@@ -598,24 +662,24 @@ irqsoff latency trace v1.1.5 on 2.6.26-rc8 ...@@ -598,24 +662,24 @@ irqsoff latency trace v1.1.5 on 2.6.26-rc8
Here we traced a 50 microsecond latency. But we also see all the Here we traced a 50 microsecond latency. But we also see all the
functions that were called during that time. Note that by enabling functions that were called during that time. Note that by
function tracing, we incur an added overhead. This overhead may enabling function tracing, we incur an added overhead. This
extend the latency times. But nevertheless, this trace has provided overhead may extend the latency times. But nevertheless, this
some very helpful debugging information. trace has provided some very helpful debugging information.
preemptoff preemptoff
---------- ----------
When preemption is disabled, we may be able to receive interrupts but When preemption is disabled, we may be able to receive
the task cannot be preempted and a higher priority task must wait interrupts but the task cannot be preempted and a higher
for preemption to be enabled again before it can preempt a lower priority task must wait for preemption to be enabled again
priority task. before it can preempt a lower priority task.
The preemptoff tracer traces the places that disable preemption. The preemptoff tracer traces the places that disable preemption.
Like the irqsoff tracer, it records the maximum latency for which preemption Like the irqsoff tracer, it records the maximum latency for
was disabled. The control of preemptoff tracer is much like the irqsoff which preemption was disabled. The control of preemptoff tracer
tracer. is much like the irqsoff tracer.
# echo preemptoff > /debug/tracing/current_tracer # echo preemptoff > /debug/tracing/current_tracer
# echo 0 > /debug/tracing/tracing_max_latency # echo 0 > /debug/tracing/tracing_max_latency
...@@ -649,11 +713,12 @@ preemptoff latency trace v1.1.5 on 2.6.26-rc8 ...@@ -649,11 +713,12 @@ preemptoff latency trace v1.1.5 on 2.6.26-rc8
sshd-4261 0d.s1 30us : trace_preempt_on (__do_softirq) sshd-4261 0d.s1 30us : trace_preempt_on (__do_softirq)
This has some more changes. Preemption was disabled when an interrupt This has some more changes. Preemption was disabled when an
came in (notice the 'h'), and was enabled while doing a softirq. interrupt came in (notice the 'h'), and was enabled while doing
(notice the 's'). But we also see that interrupts have been disabled a softirq. (notice the 's'). But we also see that interrupts
when entering the preempt off section and leaving it (the 'd'). have been disabled when entering the preempt off section and
We do not know if interrupts were enabled in the mean time. leaving it (the 'd'). We do not know if interrupts were enabled
in the mean time.
# tracer: preemptoff # tracer: preemptoff
# #
...@@ -712,28 +777,30 @@ preemptoff latency trace v1.1.5 on 2.6.26-rc8 ...@@ -712,28 +777,30 @@ preemptoff latency trace v1.1.5 on 2.6.26-rc8
sshd-4261 0d.s1 64us : trace_preempt_on (__do_softirq) sshd-4261 0d.s1 64us : trace_preempt_on (__do_softirq)
The above is an example of the preemptoff trace with ftrace_enabled The above is an example of the preemptoff trace with
set. Here we see that interrupts were disabled the entire time. ftrace_enabled set. Here we see that interrupts were disabled
The irq_enter code lets us know that we entered an interrupt 'h'. the entire time. The irq_enter code lets us know that we entered
Before that, the functions being traced still show that it is not an interrupt 'h'. Before that, the functions being traced still
in an interrupt, but we can see from the functions themselves that show that it is not in an interrupt, but we can see from the
this is not the case. functions themselves that this is not the case.
Notice that __do_softirq when called does not have a preempt_count. Notice that __do_softirq when called does not have a
It may seem that we missed a preempt enabling. What really happened preempt_count. It may seem that we missed a preempt enabling.
is that the preempt count is held on the thread's stack and we What really happened is that the preempt count is held on the
switched to the softirq stack (4K stacks in effect). The code thread's stack and we switched to the softirq stack (4K stacks
does not copy the preempt count, but because interrupts are disabled, in effect). The code does not copy the preempt count, but
we do not need to worry about it. Having a tracer like this is good because interrupts are disabled, we do not need to worry about
for letting people know what really happens inside the kernel. it. Having a tracer like this is good for letting people know
what really happens inside the kernel.
preemptirqsoff preemptirqsoff
-------------- --------------
Knowing the locations that have interrupts disabled or preemption Knowing the locations that have interrupts disabled or
disabled for the longest times is helpful. But sometimes we would preemption disabled for the longest times is helpful. But
like to know when either preemption and/or interrupts are disabled. sometimes we would like to know when either preemption and/or
interrupts are disabled.
Consider the following code: Consider the following code:
...@@ -753,11 +820,13 @@ The preemptoff tracer will record the total length of ...@@ -753,11 +820,13 @@ The preemptoff tracer will record the total length of
call_function_with_irqs_and_preemption_off() and call_function_with_irqs_and_preemption_off() and
call_function_with_preemption_off(). call_function_with_preemption_off().
But neither will trace the time that interrupts and/or preemption But neither will trace the time that interrupts and/or
is disabled. This total time is the time that we can not schedule. preemption is disabled. This total time is the time that we can
To record this time, use the preemptirqsoff tracer. not schedule. To record this time, use the preemptirqsoff
tracer.
Again, using this trace is much like the irqsoff and preemptoff tracers. Again, using this trace is much like the irqsoff and preemptoff
tracers.
# echo preemptirqsoff > /debug/tracing/current_tracer # echo preemptirqsoff > /debug/tracing/current_tracer
# echo 0 > /debug/tracing/tracing_max_latency # echo 0 > /debug/tracing/tracing_max_latency
...@@ -793,9 +862,10 @@ preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8 ...@@ -793,9 +862,10 @@ preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
The trace_hardirqs_off_thunk is called from assembly on x86 when The trace_hardirqs_off_thunk is called from assembly on x86 when
interrupts are disabled in the assembly code. Without the function interrupts are disabled in the assembly code. Without the
tracing, we do not know if interrupts were enabled within the preemption function tracing, we do not know if interrupts were enabled
points. We do see that it started with preemption enabled. within the preemption points. We do see that it started with
preemption enabled.
Here is a trace with ftrace_enabled set: Here is a trace with ftrace_enabled set:
...@@ -883,40 +953,42 @@ preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8 ...@@ -883,40 +953,42 @@ preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
sshd-4261 0d.s1 105us : trace_preempt_on (__do_softirq) sshd-4261 0d.s1 105us : trace_preempt_on (__do_softirq)
This is a very interesting trace. It started with the preemption of This is a very interesting trace. It started with the preemption
the ls task. We see that the task had the "need_resched" bit set of the ls task. We see that the task had the "need_resched" bit
via the 'N' in the trace. Interrupts were disabled before the spin_lock set via the 'N' in the trace. Interrupts were disabled before
at the beginning of the trace. We see that a schedule took place to run the spin_lock at the beginning of the trace. We see that a
sshd. When the interrupts were enabled, we took an interrupt. schedule took place to run sshd. When the interrupts were
On return from the interrupt handler, the softirq ran. We took another enabled, we took an interrupt. On return from the interrupt
interrupt while running the softirq as we see from the capital 'H'. handler, the softirq ran. We took another interrupt while
running the softirq as we see from the capital 'H'.
wakeup wakeup
------ ------
In a Real-Time environment it is very important to know the wakeup In a Real-Time environment it is very important to know the
time it takes for the highest priority task that is woken up to the wakeup time it takes for the highest priority task that is woken
time that it executes. This is also known as "schedule latency". up to the time that it executes. This is also known as "schedule
I stress the point that this is about RT tasks. It is also important latency". I stress the point that this is about RT tasks. It is
to know the scheduling latency of non-RT tasks, but the average also important to know the scheduling latency of non-RT tasks,
schedule latency is better for non-RT tasks. Tools like but the average schedule latency is better for non-RT tasks.
LatencyTop are more appropriate for such measurements. Tools like LatencyTop are more appropriate for such
measurements.
Real-Time environments are interested in the worst case latency. Real-Time environments are interested in the worst case latency.
That is the longest latency it takes for something to happen, and That is the longest latency it takes for something to happen,
not the average. We can have a very fast scheduler that may only and not the average. We can have a very fast scheduler that may
have a large latency once in a while, but that would not work well only have a large latency once in a while, but that would not
with Real-Time tasks. The wakeup tracer was designed to record work well with Real-Time tasks. The wakeup tracer was designed
the worst case wakeups of RT tasks. Non-RT tasks are not recorded to record the worst case wakeups of RT tasks. Non-RT tasks are
because the tracer only records one worst case and tracing non-RT not recorded because the tracer only records one worst case and
tasks that are unpredictable will overwrite the worst case latency tracing non-RT tasks that are unpredictable will overwrite the
of RT tasks. worst case latency of RT tasks.
Since this tracer only deals with RT tasks, we will run this slightly Since this tracer only deals with RT tasks, we will run this
differently than we did with the previous tracers. Instead of performing slightly differently than we did with the previous tracers.
an 'ls', we will run 'sleep 1' under 'chrt' which changes the Instead of performing an 'ls', we will run 'sleep 1' under
priority of the task. 'chrt' which changes the priority of the task.
# echo wakeup > /debug/tracing/current_tracer # echo wakeup > /debug/tracing/current_tracer
# echo 0 > /debug/tracing/tracing_max_latency # echo 0 > /debug/tracing/tracing_max_latency
...@@ -946,17 +1018,16 @@ wakeup latency trace v1.1.5 on 2.6.26-rc8 ...@@ -946,17 +1018,16 @@ wakeup latency trace v1.1.5 on 2.6.26-rc8
<idle>-0 1d..4 4us : schedule (cpu_idle) <idle>-0 1d..4 4us : schedule (cpu_idle)
Running this on an idle system, we see that it only took 4
microseconds to perform the task switch. Note, since the trace
marker in the schedule is before the actual "switch", we stop
the tracing when the recorded task is about to schedule in. This
may change if we add a new marker at the end of the scheduler.
Running this on an idle system, we see that it only took 4 microseconds Notice that the recorded task is 'sleep' with the PID of 4901
to perform the task switch. Note, since the trace marker in the and it has an rt_prio of 5. This priority is user-space priority
schedule is before the actual "switch", we stop the tracing when and not the internal kernel priority. The policy is 1 for
the recorded task is about to schedule in. This may change if SCHED_FIFO and 2 for SCHED_RR.
we add a new marker at the end of the scheduler.
Notice that the recorded task is 'sleep' with the PID of 4901 and it
has an rt_prio of 5. This priority is user-space priority and not
the internal kernel priority. The policy is 1 for SCHED_FIFO and 2
for SCHED_RR.
Doing the same with chrt -r 5 and ftrace_enabled set. Doing the same with chrt -r 5 and ftrace_enabled set.
...@@ -1013,24 +1084,25 @@ ksoftirq-7 1d..6 49us : _spin_unlock (tracing_record_cmdline) ...@@ -1013,24 +1084,25 @@ ksoftirq-7 1d..6 49us : _spin_unlock (tracing_record_cmdline)
ksoftirq-7 1d..6 49us : sub_preempt_count (_spin_unlock) ksoftirq-7 1d..6 49us : sub_preempt_count (_spin_unlock)
ksoftirq-7 1d..4 50us : schedule (__cond_resched) ksoftirq-7 1d..4 50us : schedule (__cond_resched)
The interrupt went off while running ksoftirqd. This task runs at The interrupt went off while running ksoftirqd. This task runs
SCHED_OTHER. Why did not we see the 'N' set early? This may be at SCHED_OTHER. Why did not we see the 'N' set early? This may
a harmless bug with x86_32 and 4K stacks. On x86_32 with 4K stacks be a harmless bug with x86_32 and 4K stacks. On x86_32 with 4K
configured, the interrupt and softirq run with their own stack. stacks configured, the interrupt and softirq run with their own
Some information is held on the top of the task's stack (need_resched stack. Some information is held on the top of the task's stack
and preempt_count are both stored there). The setting of the NEED_RESCHED (need_resched and preempt_count are both stored there). The
bit is done directly to the task's stack, but the reading of the setting of the NEED_RESCHED bit is done directly to the task's
NEED_RESCHED is done by looking at the current stack, which in this case stack, but the reading of the NEED_RESCHED is done by looking at
is the stack for the hard interrupt. This hides the fact that NEED_RESCHED the current stack, which in this case is the stack for the hard
has been set. We do not see the 'N' until we switch back to the task's interrupt. This hides the fact that NEED_RESCHED has been set.
We do not see the 'N' until we switch back to the task's
assigned stack. assigned stack.
function function
-------- --------
This tracer is the function tracer. Enabling the function tracer This tracer is the function tracer. Enabling the function tracer
can be done from the debug file system. Make sure the ftrace_enabled is can be done from the debug file system. Make sure the
set; otherwise this tracer is a nop. ftrace_enabled is set; otherwise this tracer is a nop.
# sysctl kernel.ftrace_enabled=1 # sysctl kernel.ftrace_enabled=1
# echo function > /debug/tracing/current_tracer # echo function > /debug/tracing/current_tracer
...@@ -1060,14 +1132,15 @@ set; otherwise this tracer is a nop. ...@@ -1060,14 +1132,15 @@ set; otherwise this tracer is a nop.
[...] [...]
Note: function tracer uses ring buffers to store the above entries. Note: function tracer uses ring buffers to store the above
The newest data may overwrite the oldest data. Sometimes using echo to entries. The newest data may overwrite the oldest data.
stop the trace is not sufficient because the tracing could have overwritten Sometimes using echo to stop the trace is not sufficient because
the data that you wanted to record. For this reason, it is sometimes better to the tracing could have overwritten the data that you wanted to
disable tracing directly from a program. This allows you to stop the record. For this reason, it is sometimes better to disable
tracing at the point that you hit the part that you are interested in. tracing directly from a program. This allows you to stop the
To disable the tracing directly from a C program, something like following tracing at the point that you hit the part that you are
code snippet can be used: interested in. To disable the tracing directly from a C program,
something like following code snippet can be used:
int trace_fd; int trace_fd;
[...] [...]
...@@ -1082,10 +1155,10 @@ int main(int argc, char *argv[]) { ...@@ -1082,10 +1155,10 @@ int main(int argc, char *argv[]) {
} }
Note: Here we hard coded the path name. The debugfs mount is not Note: Here we hard coded the path name. The debugfs mount is not
guaranteed to be at /debug (and is more commonly at /sys/kernel/debug). guaranteed to be at /debug (and is more commonly at
For simple one time traces, the above is sufficent. For anything else, /sys/kernel/debug). For simple one time traces, the above is
a search through /proc/mounts may be needed to find where the debugfs sufficent. For anything else, a search through /proc/mounts may
file-system is mounted. be needed to find where the debugfs file-system is mounted.
Single thread tracing Single thread tracing
...@@ -1186,10 +1259,11 @@ following format: ...@@ -1186,10 +1259,11 @@ following format:
0 scheduler_tick+0x1b6/0x1bf <- scheduler_tick+0x1aa/0x1bf 0 scheduler_tick+0x1b6/0x1bf <- scheduler_tick+0x1aa/0x1bf
The tracer may be used to dump the trace for the oops'ing cpu on a The tracer may be used to dump the trace for the oops'ing cpu on
kernel oops into the system log. To enable this, ftrace_dump_on_oops a kernel oops into the system log. To enable this,
must be set. To set ftrace_dump_on_oops, one can either use the sysctl ftrace_dump_on_oops must be set. To set ftrace_dump_on_oops, one
function or set it via the proc system interface. can either use the sysctl function or set it via the proc system
interface.
sysctl kernel.ftrace_dump_on_oops=1 sysctl kernel.ftrace_dump_on_oops=1
...@@ -1198,8 +1272,8 @@ or ...@@ -1198,8 +1272,8 @@ or
echo 1 > /proc/sys/kernel/ftrace_dump_on_oops echo 1 > /proc/sys/kernel/ftrace_dump_on_oops
Here's an example of such a dump after a null pointer dereference in a Here's an example of such a dump after a null pointer
kernel module: dereference in a kernel module:
[57848.105921] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000 [57848.105921] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
[57848.106019] IP: [<ffffffffa0000006>] open+0x6/0x14 [oops] [57848.106019] IP: [<ffffffffa0000006>] open+0x6/0x14 [oops]
...@@ -1239,25 +1313,34 @@ kernel module: ...@@ -1239,25 +1313,34 @@ kernel module:
function graph tracer function graph tracer
--------------------------- ---------------------------
This tracer is similar to the function tracer except that it probes This tracer is similar to the function tracer except that it
a function on its entry and its exit. probes a function on its entry and its exit. This is done by
This is done by setting a dynamically allocated stack of return addresses on each using a dynamically allocated stack of return addresses in each
task_struct. Then the tracer overwrites the return address of each function traced task_struct. On function entry the tracer overwrites the return
to set a custom probe. Thus the original return address is stored on the stack of return address of each function traced to set a custom probe. Thus the
address in the task_struct. original return address is stored on the stack of return address
in the task_struct.
Probing on both extremities of a function leads to special features such as Probing on both ends of a function leads to special features
such as:
_ measure of function's time execution - measure of a function's time execution
_ having a reliable call stack to draw function calls graph - having a reliable call stack to draw function calls graph
This tracer is useful in several situations: This tracer is useful in several situations:
_ you want to find the reason of a strange kernel behavior and need to see - you want to find the reason of a strange kernel behavior and
what happens in detail on any areas (or specific ones). need to see what happens in detail on any areas (or specific
_ you are experiencing weird latencies but it's difficult to find its origin. ones).
_ you want to find quickly which path is taken by a specific function
_ you just want to see what happens inside your kernel - you are experiencing weird latencies but it's difficult to
find its origin.
- you want to find quickly which path is taken by a specific
function
- you just want to peek inside a working kernel and want to see
what happens there.
# tracer: function_graph # tracer: function_graph
# #
...@@ -1282,24 +1365,28 @@ _ you just want to see what happens inside your kernel ...@@ -1282,24 +1365,28 @@ _ you just want to see what happens inside your kernel
0) 0.586 us | _spin_unlock(); 0) 0.586 us | _spin_unlock();
There are several columns that can be dynamically enabled/disabled. There are several columns that can be dynamically
You can use every combination of options you want, depending on your needs. enabled/disabled. You can use every combination of options you
want, depending on your needs.
_ The cpu number on which the function executed is default enabled. - The cpu number on which the function executed is default
It is sometimes better to only trace one cpu (see tracing_cpu_mask file) enabled. It is sometimes better to only trace one cpu (see
or you might sometimes see unordered function calls while cpu tracing switch. tracing_cpu_mask file) or you might sometimes see unordered
function calls while cpu tracing switch.
hide: echo nofuncgraph-cpu > /debug/tracing/trace_options hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
show: echo funcgraph-cpu > /debug/tracing/trace_options show: echo funcgraph-cpu > /debug/tracing/trace_options
_ The duration (function's time of execution) is displayed on the closing bracket - The duration (function's time of execution) is displayed on
line of a function or on the same line than the current function in case of a leaf the closing bracket line of a function or on the same line
one. It is default enabled. than the current function in case of a leaf one. It is default
enabled.
hide: echo nofuncgraph-duration > /debug/tracing/trace_options hide: echo nofuncgraph-duration > /debug/tracing/trace_options
show: echo funcgraph-duration > /debug/tracing/trace_options show: echo funcgraph-duration > /debug/tracing/trace_options
_ The overhead field precedes the duration one in case of reached duration thresholds. - The overhead field precedes the duration field in case of
reached duration thresholds.
hide: echo nofuncgraph-overhead > /debug/tracing/trace_options hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
show: echo funcgraph-overhead > /debug/tracing/trace_options show: echo funcgraph-overhead > /debug/tracing/trace_options
...@@ -1328,8 +1415,8 @@ _ The overhead field precedes the duration one in case of reached duration thres ...@@ -1328,8 +1415,8 @@ _ The overhead field precedes the duration one in case of reached duration thres
! means that the function exceeded 100 usecs. ! means that the function exceeded 100 usecs.
_ The task/pid field displays the thread cmdline and pid which executed the function. - The task/pid field displays the thread cmdline and pid which
It is default disabled. executed the function. It is default disabled.
hide: echo nofuncgraph-proc > /debug/tracing/trace_options hide: echo nofuncgraph-proc > /debug/tracing/trace_options
show: echo funcgraph-proc > /debug/tracing/trace_options show: echo funcgraph-proc > /debug/tracing/trace_options
...@@ -1351,8 +1438,9 @@ _ The task/pid field displays the thread cmdline and pid which executed the func ...@@ -1351,8 +1438,9 @@ _ The task/pid field displays the thread cmdline and pid which executed the func
0) sh-4802 | + 49.370 us | } 0) sh-4802 | + 49.370 us | }
_ The absolute time field is an absolute timestamp given by the clock since - The absolute time field is an absolute timestamp given by the
it started. A snapshot of this time is given on each entry/exit of functions system clock since it started. A snapshot of this time is
given on each entry/exit of functions
hide: echo nofuncgraph-abstime > /debug/tracing/trace_options hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
show: echo funcgraph-abstime > /debug/tracing/trace_options show: echo funcgraph-abstime > /debug/tracing/trace_options
...@@ -1377,9 +1465,10 @@ _ The absolute time field is an absolute timestamp given by the clock since ...@@ -1377,9 +1465,10 @@ _ The absolute time field is an absolute timestamp given by the clock since
360.774530 | 1) 0.594 us | __phys_addr(); 360.774530 | 1) 0.594 us | __phys_addr();
You can put some comments on specific functions by using ftrace_printk() You can put some comments on specific functions by using
For example, if you want to put a comment inside the __might_sleep() function, ftrace_printk() For example, if you want to put a comment inside
you just have to include <linux/ftrace.h> and call ftrace_printk() inside __might_sleep() the __might_sleep() function, you just have to include
<linux/ftrace.h> and call ftrace_printk() inside __might_sleep()
ftrace_printk("I'm a comment!\n") ftrace_printk("I'm a comment!\n")
...@@ -1390,8 +1479,9 @@ will produce: ...@@ -1390,8 +1479,9 @@ will produce:
1) 1.449 us | } 1) 1.449 us | }
You might find other useful features for this tracer on the "dynamic ftrace" You might find other useful features for this tracer in the
section such as tracing only specific functions or tasks. following "dynamic ftrace" section such as tracing only specific
functions or tasks.
dynamic ftrace dynamic ftrace
-------------- --------------
...@@ -1399,43 +1489,45 @@ dynamic ftrace ...@@ -1399,43 +1489,45 @@ dynamic ftrace
If CONFIG_DYNAMIC_FTRACE is set, the system will run with If CONFIG_DYNAMIC_FTRACE is set, the system will run with
virtually no overhead when function tracing is disabled. The way virtually no overhead when function tracing is disabled. The way
this works is the mcount function call (placed at the start of this works is the mcount function call (placed at the start of
every kernel function, produced by the -pg switch in gcc), starts every kernel function, produced by the -pg switch in gcc),
of pointing to a simple return. (Enabling FTRACE will include the starts of pointing to a simple return. (Enabling FTRACE will
-pg switch in the compiling of the kernel.) include the -pg switch in the compiling of the kernel.)
At compile time every C file object is run through the At compile time every C file object is run through the
recordmcount.pl script (located in the scripts directory). This recordmcount.pl script (located in the scripts directory). This
script will process the C object using objdump to find all the script will process the C object using objdump to find all the
locations in the .text section that call mcount. (Note, only locations in the .text section that call mcount. (Note, only the
the .text section is processed, since processing other sections .text section is processed, since processing other sections like
like .init.text may cause races due to those sections being freed). .init.text may cause races due to those sections being freed).
A new section called "__mcount_loc" is created that holds references A new section called "__mcount_loc" is created that holds
to all the mcount call sites in the .text section. This section is references to all the mcount call sites in the .text section.
compiled back into the original object. The final linker will add This section is compiled back into the original object. The
all these references into a single table. final linker will add all these references into a single table.
On boot up, before SMP is initialized, the dynamic ftrace code On boot up, before SMP is initialized, the dynamic ftrace code
scans this table and updates all the locations into nops. It also scans this table and updates all the locations into nops. It
records the locations, which are added to the available_filter_functions also records the locations, which are added to the
list. Modules are processed as they are loaded and before they are available_filter_functions list. Modules are processed as they
executed. When a module is unloaded, it also removes its functions from are loaded and before they are executed. When a module is
the ftrace function list. This is automatic in the module unload unloaded, it also removes its functions from the ftrace function
code, and the module author does not need to worry about it. list. This is automatic in the module unload code, and the
module author does not need to worry about it.
When tracing is enabled, kstop_machine is called to prevent races
with the CPUS executing code being modified (which can cause the When tracing is enabled, kstop_machine is called to prevent
CPU to do undesireable things), and the nops are patched back races with the CPUS executing code being modified (which can
to calls. But this time, they do not call mcount (which is just cause the CPU to do undesireable things), and the nops are
a function stub). They now call into the ftrace infrastructure. patched back to calls. But this time, they do not call mcount
(which is just a function stub). They now call into the ftrace
infrastructure.
One special side-effect to the recording of the functions being One special side-effect to the recording of the functions being
traced is that we can now selectively choose which functions we traced is that we can now selectively choose which functions we
wish to trace and which ones we want the mcount calls to remain as wish to trace and which ones we want the mcount calls to remain
nops. as nops.
Two files are used, one for enabling and one for disabling the tracing Two files are used, one for enabling and one for disabling the
of specified functions. They are: tracing of specified functions. They are:
set_ftrace_filter set_ftrace_filter
...@@ -1443,8 +1535,8 @@ and ...@@ -1443,8 +1535,8 @@ and
set_ftrace_notrace set_ftrace_notrace
A list of available functions that you can add to these files is listed A list of available functions that you can add to these files is
in: listed in:
available_filter_functions available_filter_functions
...@@ -1481,8 +1573,8 @@ hrtimer_interrupt ...@@ -1481,8 +1573,8 @@ hrtimer_interrupt
sys_nanosleep sys_nanosleep
Perhaps this is not enough. The filters also allow simple wild cards. Perhaps this is not enough. The filters also allow simple wild
Only the following are currently available cards. Only the following are currently available
<match>* - will match functions that begin with <match> <match>* - will match functions that begin with <match>
*<match> - will match functions that end with <match> *<match> - will match functions that end with <match>
...@@ -1492,9 +1584,9 @@ These are the only wild cards which are supported. ...@@ -1492,9 +1584,9 @@ These are the only wild cards which are supported.
<match>*<match> will not work. <match>*<match> will not work.
Note: It is better to use quotes to enclose the wild cards, otherwise Note: It is better to use quotes to enclose the wild cards,
the shell may expand the parameters into names of files in the local otherwise the shell may expand the parameters into names
directory. of files in the local directory.
# echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
...@@ -1540,7 +1632,8 @@ This is because the '>' and '>>' act just like they do in bash. ...@@ -1540,7 +1632,8 @@ This is because the '>' and '>>' act just like they do in bash.
To rewrite the filters, use '>' To rewrite the filters, use '>'
To append to the filters, use '>>' To append to the filters, use '>>'
To clear out a filter so that all functions will be recorded again: To clear out a filter so that all functions will be recorded
again:
# echo > /debug/tracing/set_ftrace_filter # echo > /debug/tracing/set_ftrace_filter
# cat /debug/tracing/set_ftrace_filter # cat /debug/tracing/set_ftrace_filter
...@@ -1572,7 +1665,8 @@ hrtimer_get_res ...@@ -1572,7 +1665,8 @@ hrtimer_get_res
hrtimer_init_sleeper hrtimer_init_sleeper
The set_ftrace_notrace prevents those functions from being traced. The set_ftrace_notrace prevents those functions from being
traced.
# echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
...@@ -1595,18 +1689,20 @@ Produces: ...@@ -1595,18 +1689,20 @@ Produces:
We can see that there's no more lock or preempt tracing. We can see that there's no more lock or preempt tracing.
* Dynamic ftrace with the function graph tracer * Dynamic ftrace with the function graph tracer
---------------------------------------------
Although what has been explained above concerns both the
function tracer and the function-graph-tracer, there are some
special features only available in the function-graph tracer.
Although what has been explained above concerns both the function tracer and If you want to trace only one function and all of its children,
the function_graph_tracer, the following concerns only the latter. you just have to echo its name into set_graph_function:
If you want to trace only one function and all of its childs, you just have echo __do_fault > set_graph_function
to echo its name on set_graph_function:
echo __do_fault > set_graph_function will produce the following "expanded" trace of the __do_fault()
function:
will produce the following:
0) | __do_fault() { 0) | __do_fault() {
0) | filemap_fault() { 0) | filemap_fault() {
...@@ -1643,23 +1739,24 @@ will produce the following: ...@@ -1643,23 +1739,24 @@ will produce the following:
0) 2.793 us | } 0) 2.793 us | }
0) + 14.012 us | } 0) + 14.012 us | }
You can also select several functions: You can also expand several functions at once:
echo sys_open > set_graph_function echo sys_open > set_graph_function
echo sys_close >> set_graph_function echo sys_close >> set_graph_function
Now if you want to go back to trace all functions Now if you want to go back to trace all functions you can clear
this special filter via:
echo > set_graph_function echo > set_graph_function
trace_pipe trace_pipe
---------- ----------
The trace_pipe outputs the same content as the trace file, but the effect The trace_pipe outputs the same content as the trace file, but
on the tracing is different. Every read from trace_pipe is consumed. the effect on the tracing is different. Every read from
This means that subsequent reads will be different. The trace trace_pipe is consumed. This means that subsequent reads will be
is live. different. The trace is live.
# echo function > /debug/tracing/current_tracer # echo function > /debug/tracing/current_tracer
# cat /debug/tracing/trace_pipe > /tmp/trace.out & # cat /debug/tracing/trace_pipe > /tmp/trace.out &
...@@ -1687,38 +1784,45 @@ is live. ...@@ -1687,38 +1784,45 @@ is live.
bash-4043 [00] 41.267111: select_task_rq_rt <-try_to_wake_up bash-4043 [00] 41.267111: select_task_rq_rt <-try_to_wake_up
Note, reading the trace_pipe file will block until more input is added. Note, reading the trace_pipe file will block until more input is
By changing the tracer, trace_pipe will issue an EOF. We needed added. By changing the tracer, trace_pipe will issue an EOF. We
to set the function tracer _before_ we "cat" the trace_pipe file. needed to set the function tracer _before_ we "cat" the
trace_pipe file.
trace entries trace entries
------------- -------------
Having too much or not enough data can be troublesome in diagnosing Having too much or not enough data can be troublesome in
an issue in the kernel. The file buffer_size_kb is used to modify diagnosing an issue in the kernel. The file buffer_size_kb is
the size of the internal trace buffers. The number listed used to modify the size of the internal trace buffers. The
is the number of entries that can be recorded per CPU. To know number listed is the number of entries that can be recorded per
the full size, multiply the number of possible CPUS with the CPU. To know the full size, multiply the number of possible CPUS
number of entries. with the number of entries.
# cat /debug/tracing/buffer_size_kb # cat /debug/tracing/buffer_size_kb
1408 (units kilobytes) 1408 (units kilobytes)
Note, to modify this, you must have tracing completely disabled. To do that, Note, to modify this, you must have tracing completely disabled.
echo "nop" into the current_tracer. If the current_tracer is not set To do that, echo "nop" into the current_tracer. If the
to "nop", an EINVAL error will be returned. current_tracer is not set to "nop", an EINVAL error will be
returned.
# echo nop > /debug/tracing/current_tracer # echo nop > /debug/tracing/current_tracer
# echo 10000 > /debug/tracing/buffer_size_kb # echo 10000 > /debug/tracing/buffer_size_kb
# cat /debug/tracing/buffer_size_kb # cat /debug/tracing/buffer_size_kb
10000 (units kilobytes) 10000 (units kilobytes)
The number of pages which will be allocated is limited to a percentage The number of pages which will be allocated is limited to a
of available memory. Allocating too much will produce an error. percentage of available memory. Allocating too much will produce
an error.
# echo 1000000000000 > /debug/tracing/buffer_size_kb # echo 1000000000000 > /debug/tracing/buffer_size_kb
-bash: echo: write error: Cannot allocate memory -bash: echo: write error: Cannot allocate memory
# cat /debug/tracing/buffer_size_kb # cat /debug/tracing/buffer_size_kb
85 85
-----------
More details can be found in the source code, in the
kernel/tracing/*.c files.
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment