hardirqs(8)

🎀 hardirqs(8) 🎀

okay bestie, this is the manual page for hardirqs, the tool that tells you how much time your computer spends handling hard IRQs (hard interrupts) 💅 it's a hardirqs(8) situation, so buckle up ✨

---

✨ NAME ✨

hardirqs - Measure hard IRQ (hard interrupt) event time. Uses Linux eBPF/bcc. 💖

🧠 SYNOPSIS 🧠

hardirqs [-h] [-T] [-N] [-C] [-d] [interval] [outputs]

💫 DESCRIPTION 💫

okay so this one summarizes the time spent servicing hard IRQs (hard interrupts), and it can show that time as either totals or histogram distributions 📊 the whole-system version of this time shows up in the %irq column of mpstat(1), and event counts (but not times) live in /proc/interrupts 🦋

this tool uses the irq:irq_handler_entry and irq:irq_handler_exit kernel tracepoints, which is a stable tracing mechanism 🫶 BPF programs can attach to tracepoints from Linux 4.7 only, so keep that in mind. an older version of this tool is available in tools/old, and it uses kprobes instead of tracepoints 🌸

since this uses BPF, only the root user can use this tool 🙈 no root, no hardirqs, sorry bestie

📋 REQUIREMENTS 📋

CONFIG_BPF and bcc 🔥

🎀 OPTIONS 🎀

• -h — prints the usage message 💅
• -T — includes timestamps on the output ⏰
• -N — outputs in nanoseconds 🔬
• -C — counts events only, no times 🔢
• -d — shows the IRQ time distribution as histograms 📊
• -c CPU — traces on this CPU only 🎯

🧚 EXAMPLES 🧚

okay here are the vibes, ranked by how iconic they are 👑

Sum hard IRQ event time until Ctrl-C:

# hardirqs

Show hard IRQ event time as histograms:

# hardirqs -d

Print 1 second summaries, 10 times:

# hardirqs 1 10

1 second summaries, printed in nanoseconds, with timestamps:

# hardirqs -NT 1

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

# hardirqs -c 1

🍓 FIELDS 🍓

• HARDIRQ — the irq action name for this hard IRQ 🏷️
• TOTAL_usecs — total time spent in this hard IRQ, in microseconds ⏱️
• TOTAL_nsecs — total time spent in this hard IRQ, in nanoseconds ⏱️
• usecs — range of microseconds for this bucket 📏
• nsecs — range of nanoseconds for this bucket 📏
• count — number of hard IRQs in this time range 🔢
• distribution — ASCII representation of the distribution (the count column) 📊

🔥 OVERHEAD 🔥

⚠️ heads up bestie: this traces kernel functions and keeps in-kernel counts, which are asynchronously copied to user-space 🥺 if the rate of interrupts is very high (>1M/sec), it's still a relatively efficient way to trace these events, so the overhead should be small for normal workloads, but it could get 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 with example usage, output, and commentary for this tool 📚

💻 OS 💻

Linux 🐧

🚧 STABILITY 🚧

Unstable - in development. 🙈

👩‍💻 AUTHOR 👩‍💻

Brendan Gregg, Hengqi Chen, Rocky Xing 💜

👀 SEE ALSO 👀

softirqs(8) 🦋

---

bestie, that's the whole page 💖 it's from the USER COMMANDS section, dated 2015-10-20, and the manual page is hardirqs(8) ✨ go forth and measure those interrupts, you literally got this 💅🎀