sugov_iowait_boost(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… sugov_iowait_boost(9) β€” like, the Kernel API Manual bestie ✨

okay so picture this: it's giving kernel scheduler, it's giving cpufreq, and this one little function is out here updating the IO boost status of a CPU. that's it, that's the vibe. πŸ¦‹


✨ NAME ✨

sugov_iowait_boost β€” Updates the IO boost status of a CPU. no cap, that's literally the whole job. πŸŽ€


🌸 SYNOPSIS 🌸

void sugov_iowait_boost (struct sugov_cpu *sg_cpu , u64 time , unsigned
int flags );

🫢 ARGUMENTS 🫢

  • sg_cpu β€” the sugov data for the CPU to boost. like, the actual little data bundle for whichever CPU we're hyping up πŸ’–
  • time β€” the update time, handed to us by the caller. respectfully, we did not pick this timestamp, they did ⭐
  • flags β€” SCHED_CPUFREQ_IOWAIT if the task is waking up after an IO wait. that's the flag that says "bestie just came back from IO" πŸ‘€

πŸ”₯ DESCRIPTION πŸ”₯

okay so here's the lore. every single time a task wakes up after an IO operation, the CPU utilization can be boosted to a certain utilization. and the way that certain utilization doubles at each "frequent and successive" wakeup from IO?? iconic behaviour honestly. ✨

and the range on this thing? it goes from IOWAIT_BOOST_MIN all the way up to the utilization of the maximum OPP. so it starts cute and small and can climb to the tippy top πŸ‘‘

but here's the catch bestie, listen β€” to keep that doubling going, an IO boost has to be requested at least once per tick. miss a tick and it's over, we restart from the utilization of the minimum OPP. no doubling, no drama, just back to the bottom 😭 that's the rule, periodt.

🌈 SEE ALSO 🌈

go read those too, they're all in the same friend group 🧚


aaaand that's the page!! mwah πŸ’œ this little drop of kernel wisdom came from ./kernel/sched, dated October 2026 β€” so yeah, take that timestamp and remember us by it πŸ“πŸŒΈ

πŸ“¦ where this page comes from