bcc-softirqs - Man Page

Measure soft IRQ (soft interrupt) event time. Uses Linux eBPF/bcc.

Synopsis

softirqs [-h] [-T] [-N] [-C] [-d] [-c CPU] [interval] [count]

Description

This summarizes the time spent servicing soft IRQs (soft interrupts), and can show this time as either totals or histogram distributions. A system-wide summary of this time is shown by the %soft column of mpstat(1), and soft IRQ event counts (but not times) are available in /proc/softirqs.

This tool uses the irq:softirq_enter and irq:softirq_exit kernel tracepoints, which is a stable tracing mechanism. BPF programs can attach to tracepoints from Linux 4.7 only. An older version of this tool is available in tools/old, and uses kprobes instead of tracepoints.

Since this uses BPF, only the root user can use this tool.

Requirements

CONFIG_BPF and bcc.

Options

-h

Print usage message.

-T

Include timestamps on output.

-N

Output in nanoseconds.

-C

Show the number of soft irq events.

-d

Show IRQ time distribution as histograms.

-c CPU

Trace on this CPU only.

Examples

Sum soft IRQ event time until Ctrl-C:

# softirqs

Show the number of soft irq events:

# softirqs -C

Show soft IRQ event time as histograms:

# softirqs -d

Print 1 second summaries, 10 times:

# softirqs 1 10

1 second summaries, printed in nanoseconds, with timestamps:

# softirqs -NT 1

Sum soft IRQ event time on CPU 1 until Ctrl-C:

# softirqs -c 1

Fields

SOFTIRQ

The kernel function name that performs the soft IRQ action.

TOTAL_usecs

Total time spent in this soft IRQ function in microseconds.

TOTAL_nsecs

Total time spent in this soft IRQ function in nanoseconds.

usecs

Range of microseconds for this bucket.

nsecs

Range of nanoseconds for this bucket.

count

Number of soft IRQs in this time range.

distribution

ASCII representation of the distribution (the count column).

Overhead

This traces kernel functions and maintains in-kernel counts, which are asynchronously copied to user-space. While the rate of interrupts be very high (>1M/sec), this is a relatively efficient way to trace these events, and so the overhead is expected to be small for normal workloads, but could become noticeable for heavy workloads. Measure in a test environment before use.

Source

This is from bcc.

https://github.com/iovisor/bcc

Also look in the bcc distribution for a companion _examples.txt file containing example usage, output, and commentary for this tool.

OS

Linux

Stability

Unstable - in development.

Authors

Brendan Gregg, Sasha Goldshtein, Rocky Xing

See Also

hardirqs(8)

Info

2015-10-20 USER COMMANDS