From a41eebab7537890409ea9dfe0fcda9b5fbdb090d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 14 Jul 2008 16:41:12 -0400 Subject: [PATCH] ftrace: document updates The following updates were recommended by Elias Oltmanns and Randy Dunlap. [ updates based on Andrew Morton's comments are still to come. ] Signed-off-by: Steven Rostedt Signed-off-by: Linus Torvalds --- Documentation/ftrace.txt | 134 +++++++++++++++++++++------------------ 1 file changed, 71 insertions(+), 63 deletions(-) diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 13e4bf054c38..77d3faa1a611 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -2,8 +2,11 @@ ======================== Copyright 2008 Red Hat Inc. -Author: Steven Rostedt + Author: Steven Rostedt + License: The GNU Free Documentation License, Version 1.2 +Reviewers: Elias Oltmanns and Randy Dunlap +Writen for: 2.6.26-rc8 linux-2.6-tip.git tip/tracing/ftrace branch Introduction ------------ @@ -46,7 +49,7 @@ of ftrace. Here is a list of some of the key files: that is configured. available_tracers : This holds the different types of tracers that - has been compiled into the kernel. The tracers + have been compiled into the kernel. The tracers listed here can be configured by echoing in their name into current_tracer. @@ -90,11 +93,13 @@ of ftrace. Here is a list of some of the key files: trace_entries : This sets or displays the number of trace entries each CPU buffer can hold. The tracer buffers are the same size for each CPU, so care must be - taken when modifying the trace_entries. The number - of actually entries will be the number given - times the number of possible CPUS. The buffers - are saved as individual pages, and the actual entries - will always be rounded up to entries per page. + taken when modifying the trace_entries. The trace + buffers are allocated in pages (blocks of memory that + the kernel uses for allocation, usually 4 KB in size). + Since each entry is smaller than a page, if the last + allocated page has room for more entries than were + requested, the rest of the page is used to allocate + entries. This can only be updated when the current_tracer is set to "none". @@ -114,13 +119,13 @@ of ftrace. Here is a list of some of the key files: in performance. This also has a side effect of enabling or disabling specific functions to be traced. Echoing in names of functions into this - file will limit the trace to only those files. + file will limit the trace to only these functions. set_ftrace_notrace: This has the opposite effect that set_ftrace_filter has. 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_ bet traced. + in both set_ftrace_filter and set_ftrace_notrace, + the function will _not_ be traced. available_filter_functions : When a function is encountered the first time by the dynamic tracer, it is recorded and @@ -138,7 +143,7 @@ Here are the list of current tracers that can be configured. ftrace - function tracer that uses mcount to trace all functions. It is possible to filter out which functions that are - traced when dynamic ftrace is configured in. + to be traced when dynamic ftrace is configured in. sched_switch - traces the context switches between tasks. @@ -297,13 +302,13 @@ explains which is which. The above is mostly meaningful for kernel developers. - time: This differs from the trace output where as the trace output - contained a absolute timestamp. This timestamp is relative - to the start of the first entry in the the trace. + time: This differs from the trace file output. The trace file output + included an absolute timestamp. The timestamp used by the + latency_trace file is relative to the start of the trace. 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. - The marks is determined by the difference between this + The marks are determined by the difference between this current trace and the next trace. '!' - greater than preempt_mark_thresh (default 100) '+' - greater than 1 microsecond @@ -322,13 +327,13 @@ output. To see what is available, simply cat the file: print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \ noblock nostacktrace nosched-tree -To disable one of the options, echo in the option appended with "no". +To disable one of the options, echo in the option prepended with "no". echo noprint-parent > /debug/tracing/iter_ctrl To enable an option, leave off the "no". - echo sym-offest > /debug/tracing/iter_ctrl + echo sym-offset > /debug/tracing/iter_ctrl Here are the available options: @@ -344,7 +349,7 @@ Here are the available options: sym-offset - Display not only the function name, but also the offset in the function. For example, instead of seeing just - "ktime_get" you will see "ktime_get+0xb/0x20" + "ktime_get", you will see "ktime_get+0xb/0x20". sym-offset: bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0 @@ -364,7 +369,7 @@ Here are the available options: user applications that can translate the raw 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. @@ -381,7 +386,7 @@ sched_switch ------------ This tracer simply records schedule switches. Here's an example -on how to implement it. +of how to use it. # echo sched_switch > /debug/tracing/current_tracer # echo 1 > /debug/tracing/tracing_enabled @@ -470,7 +475,7 @@ interrupt from triggering or the mouse interrupt from letting the kernel know of a new mouse event. The result is a latency with the reaction time. -The irqsoff tracer tracks the time interrupts are disabled and when +The irqsoff tracer tracks the time interrupts are disabled to the time they are re-enabled. When a new maximum latency is hit, it saves off the trace so that it may be retrieved at a later time. Every time a new maximum in reached, the old saved trace is discarded and the new @@ -519,7 +524,7 @@ The difference between the 6 and the displayed timestamp 7us is because the clock must have incremented between the time of recording the max latency and recording the function that had that latency. -Note the above had ftrace_enabled not set. If we set the ftrace_enabled +Note the above had ftrace_enabled not set. If we set the ftrace_enabled, we get a much larger output: # tracer: irqsoff @@ -570,21 +575,21 @@ vim:ft=help Here we traced a 50 microsecond latency. But we also see all the -functions that were called during that time. Note that enabling -function tracing we endure an added overhead. This overhead may -extend the latency times. But never the less, this trace has provided -some very helpful debugging. +functions that were called during that time. Note that by enabling +function tracing, we endure an added overhead. This overhead may +extend the latency times. But nevertheless, this trace has provided +some very helpful debugging information. preemptoff ---------- -When preemption is disabled we may be able to receive interrupts but -the task can not be preempted and a higher priority task must wait +When preemption is disabled, we may be able to receive interrupts but +the task cannot be preempted and a higher priority task must wait for preemption to be enabled again before it can preempt a lower priority task. -The preemptoff tracer traces the places that disables preemption. +The preemptoff tracer traces the places that disable preemption. Like the irqsoff, it records the maximum latency that preemption was disabled. The control of preemptoff is much like the irqsoff. @@ -696,7 +701,7 @@ Notice that the __do_softirq when called doesn't have a preempt_count. It may seem that we missed a preempt enabled. What really happened is that the preempt count is held on the threads stack and we switched to the softirq stack (4K stacks in effect). The code -does not copy the preempt count, but because interrupts are disabled +does not copy the preempt count, but because interrupts are disabled, we don't need to worry about it. Having a tracer like this is good to let people know what really happens inside the kernel. @@ -732,7 +737,7 @@ To record this time, use the preemptirqsoff tracer. Again, using this trace is much like the irqsoff and preemptoff tracers. - # echo preemptoff > /debug/tracing/current_tracer + # echo preemptirqsoff > /debug/tracing/current_tracer # echo 0 > /debug/tracing/tracing_max_latency # echo 1 > /debug/tracing/tracing_enabled # ls -ltr @@ -862,9 +867,9 @@ This is a very interesting trace. It started with the preemption of the ls task. We see that the task had the "need_resched" bit set with the 'N' in the trace. Interrupts are disabled in the spin_lock and the trace started. We see that a schedule took place to run -sshd. When the interrupts were enabled we took an interrupt. -On return of the interrupt the softirq ran. We took another interrupt -while running the softirq as we see with the capital 'H'. +sshd. When the interrupts were enabled, we took an interrupt. +On return from the interrupt handler, the softirq ran. We took another +interrupt while running the softirq as we see with the capital 'H'. wakeup @@ -876,9 +881,9 @@ time it executes. This is also known as "schedule latency". I stress the point that this is about RT tasks. It is also important to know the scheduling latency of non-RT tasks, but the average schedule latency is better for non-RT tasks. Tools like -LatencyTop is more appropriate for such measurements. +LatencyTop are more appropriate for such measurements. -Real-Time environments is 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 not the average. We can have a very fast scheduler that may only have a large latency once in a while, but that would not work well @@ -889,8 +894,8 @@ tasks that are unpredictable will overwrite the worst case latency of RT tasks. Since this tracer only deals with RT tasks, we will run this slightly -different than we did with the previous tracers. Instead of performing -an 'ls' we will run 'sleep 1' under 'chrt' which changes the +differently than we did with the previous tracers. Instead of performing +an 'ls', we will run 'sleep 1' under 'chrt' which changes the priority of the task. # echo wakeup > /debug/tracing/current_tracer @@ -924,9 +929,9 @@ wakeup latency trace v1.1.5 on 2.6.26-rc8 vim:ft=help -Running this on an idle system we see that it only took 4 microseconds +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 +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. @@ -992,12 +997,15 @@ ksoftirq-7 1d..4 50us : schedule (__cond_resched) The interrupt went off while running ksoftirqd. This task runs at SCHED_OTHER. Why didn't we see the 'N' set early? This may be -a harmless bug with x86_32 and 4K stacks. The need_reched() function -that tests if we need to reschedule looks on the actual stack. -Where as the setting of the NEED_RESCHED bit happens on the -task's stack. But because we are in a hard interrupt, the test -is with the interrupts stack which has that to be false. We don't -see the 'N' until we switch back to the task's stack. +a harmless bug with x86_32 and 4K stacks. On x86_32 with 4K stacks +configured, the interrupt and softirq runs with their own stack. +Some information is held on the top of the task's stack (need_resched +and preempt_count are both stored there). The setting of the NEED_RESCHED +bit is done directly to the task's stack, but the reading of the +NEED_RESCHED is done by looking at the current stack, which in this case +is the stack for the hard interrupt. This hides the fact that NEED_RESCHED +has been set. We don't see the 'N' until we switch back to the task's +assigned stack. ftrace ------ @@ -1067,10 +1075,10 @@ this works is the mcount function call (placed at the start of every kernel function, produced by the -pg switch in gcc), starts of pointing to a simple return. -When dynamic ftrace is initialized, it calls kstop_machine to make it -act like a uniprocessor so that it can freely modify code without -worrying about other processors executing that same code. At -initialization, the mcount calls are change to call a "record_ip" +When dynamic ftrace is initialized, it calls kstop_machine to make +the machine act like a uniprocessor so that it can freely modify code +without worrying about other processors executing that same code. At +initialization, the mcount calls are changed to call a "record_ip" function. After this, the first time a kernel function is called, it has the calling address saved in a hash table. @@ -1085,8 +1093,8 @@ traced, is that we can now selectively choose which functions we want to trace and which ones we want the mcount calls to remain as nops. -Two files that contain to the enabling and disabling of recorded -functions are: +Two files are used, one for enabling and one for disabling the tracing +of recorded functions. They are: set_ftrace_filter @@ -1094,7 +1102,7 @@ and set_ftrace_notrace -A list of available functions that you can add to this files is listed +A list of available functions that you can add to these files is listed in: available_filter_functions @@ -1133,9 +1141,9 @@ sys_nanosleep Perhaps this isn't enough. The filters also allow simple wild cards. -Only the following is currently available +Only the following are currently available - * - will match functions that begins with + * - will match functions that begin with * - will match functions that end with ** - will match functions that have in it @@ -1187,7 +1195,7 @@ This is because the '>' and '>>' act just like they do in bash. To rewrite 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 # cat /debug/tracing/set_ftrace_filter @@ -1246,8 +1254,8 @@ ftraced As mentioned above, when dynamic ftrace is configured in, a kernel thread wakes up once a second and checks to see if there are mcount -calls that need to be converted into nops. If there is not, then -it simply goes back to sleep. But if there is, it will call +calls that need to be converted into nops. If there are not any, then +it simply goes back to sleep. But if there are some, it will call kstop_machine to convert the calls to nops. There may be a case that you do not want this added latency. @@ -1262,8 +1270,8 @@ mcount calls to nops. Remember that there's a large overhead to calling mcount. Without this kernel thread, that overhead will exist. -Any write to the ftraced_enabled file will cause the kstop_machine -to run if there are recorded calls to mcount. This means that a +If there are recorded calls to mcount, any write to the ftraced_enabled +file will cause the kstop_machine to run. This means that a user can manually perform the updates when they want to by simply echoing a '0' into the ftraced_enabled file. @@ -1315,7 +1323,7 @@ trace entries Having too much or not enough data can be troublesome in diagnosing some issue in the kernel. The file trace_entries is used to modify -the size of the internal trace buffers. The numbers listed +the size of the internal trace buffers. The number listed is the number of entries that can be recorded per CPU. To know the full size, multiply the number of possible CPUS with the number of entries. @@ -1323,7 +1331,7 @@ number of entries. # cat /debug/tracing/trace_entries 65620 -Note, to modify this you must have tracing fulling disabled. To do that, +Note, to modify this, you must have tracing completely disabled. To do that, echo "none" into the current_tracer. # echo none > /debug/tracing/current_tracer @@ -1344,7 +1352,7 @@ it will add them. This shows us that 85 entries can fit on a single page. The number of pages that will be allocated is a percentage of available -memory. Allocating too much will produces an error. +memory. Allocating too much will produce an error. # echo 1000000000000 > /debug/tracing/trace_entries -bash: echo: write error: Cannot allocate memory