In the sched_switch trace point to fill the cache of record-tgid When any event or tracer is enabled, a hook is enabled Name of the task, disabling this option can lower the Overhead, and if you only care about pids, and not the In the sched_switch trace point to fill comm cache record-cmd When any event or tracer is enabled, a hook is enabled Latency, as described in “Latency trace format”. The trace displays additional information about the latency-format This option changes the trace output. Trace_bputs() was used to save the trace_printk()). Printk-msg-only When set, trace_printk()s will only show the formatĪnd not their parameters (if trace_bprintk() or The start and stop (the functions in which the interrupts wereĭisabled and enabled respectively) that caused the latencies:Ī.out-1623 40874.465068: /root/a.out <-/root/a.out[+0 The task is the process that was running when the latency
VP, KP, SP, and HP are always zero and are reserved for later use. Of trace entries displayed and the total number (both are four: #4/4). Then it displays the max latency in microseconds (259 us). Never changes) and the version of the kernel upon which this was executed on This shows that the current tracer is “irqsoff” tracing the timeįor which interrupts were disabled. # => ended at: _raw_spin_unlock_irqrestore # irqsoff latency trace v1.1.5 on 3.8.0-test+ format, theįunction name that was traced “sys_close” and the parent function thatĬalled this function “system_call_fastpath”. To set a clock, simply echo the clock name into this file: Ktime_get_boot_fast_ns() function for more information. Processing should be able to handle them. In this case, the clock updateĪppears to happen slightly sooner than it normally would have.Īlso on 32-bit systems, it’s possible that the 64-bit boot offset If clock is accessed after the suspend time is accounted before Tracing in the suspend path, some side effects are possible Since the clock access is designed for use in boot: This is the boot clock (CLOCK_BOOTTIME) and is based on theįast monotonic clock, but also accounts for time spent in Which is montonic but is not subject to any rate adjustmentsĪnd ticks at the same rate as the hardware clocksource. mono_raw: This is the raw monotonic clock (CLOCK_MONOTONIC_RAW) Which is monotonic and is subject to NTP rate adjustments. mono: This uses the fast monotonic clock (CLOCK_MONOTONIC) To correlate events across hypervisor/guest if This is in sync across CPUs and can also be used ppc-tb: This uses the powerpc timebase register value. ForĮxample, x86 uses its own TSC cycle clock here. x86-tsc: Architectures may define their own clocks. perf: This makes ftrace use the same clock that perf uses.Įventually perf will be able to read ftrace buffersĪnd this will help out in interleaving the data. uptime: This uses the jiffies counter and the time stamp Know exactly the order events occurred with respect toĮach other on different CPUs. counter: This is not a clock at all, but literally an atomicĬounter. local: Default clock, but may not be in sync across CPUs global: This clock is in sync with all CPUs but mayīe a bit slower than the local clock. The clock with the square brackets around it is the one in effect. Its address will be printed as well as the function that the Specific to a the callback and not the standard trampoline, If the callback of the function jumps to a trampoline that is Than 1 it most likely will be ftrace_ops_list_func().
Is being directly called by the function. If the architecture supports it, it will also show what callback The “ip modify” attribute (thus the regs->ip can be changed),Īn ‘I’ will be displayed on the same line as the function that Will be displayed on the same line as the function that
The “save regs” attribute (thus even more overhead), a ‘R’ If the callback registered to be traced by a function with Note, a callback may also call multiple functions which will This fileĭisplays all functions that have a callback attached to themĪs well as the number of callbacks that have been attached. Trace utility, but other subsystems might too. Not only does the trace infrastructure use ftrace function In seeing if any function has a callback attached to it. This file is more for debugging ftrace, but can also be useful