iommu_unmap_fast(9)

📦 version 7.3.1, in nixpkgs unstable

💅 iommu_unmap_fast ✨

Okay bestie, welcome to the Kernel API Manual 🎀 — this one's iommu_unmap_fast(9), and no cap, it's kind of a niche icon but she knows what she's doing.

🌸 NAME 🌸

iommu_unmap_fast — remove mappings from a range of IOVA without the IOTLB sync, periodt. 💅

✨ SYNOPSIS ✨

size_t iommu_unmap_fast (struct iommu_domain *domain , unsigned long
iova , size_t size , struct iommu_iotlb_gather *iotlb_gather );

🦋 ARGUMENTS 🦋

  • domain — the domain you're messing with. your main character 🫶
  • iova — the IO virtual address where the range starts, like, the beginning of the story ✨
  • size — length of the range starting from iova. how much we're doing today 💖
  • iotlb_gather — range info for a pending IOTLB flush. she's holding the deets for later 🔥

👑 DESCRIPTION 👑

So iommu_unmap_fast will remove a translation that was created by iommu_map. literally undoing the thing 💀

Important vibe check: she can't subdivide a mapping created by iommu_map, so you should call her with IOVA ranges that match what was passed to iommu_map. The range can aggregate contiguous iommu_map calls — but only so long as no individual range is split. no splitting, bestie, that's the rule 🙈

Basically, iommu_unmap_fast is the same as iommu_unmap but for callers who manage the IOTLB flushing externally to perform a batched sync. it's giving "I'll handle the follow-up, you just do the unmapping" ✨

🌈 RETURN 🌈

Number of bytes of IOVA unmapped. and like, iova + res will be the point where unmapping stopped — so you always know exactly where she left off 🎀

🧚 SEE ALSO 🧚

Kernel file ./drivers/iommu/iommu.c

and that's the whole page, bestie — mwah 💋 this lil number lives over in ./drivers/iommu and she's dated October 2026, so like, mark your calendar 🗓️💜 slay, go unmap responsibly ✨🦋🎀

📦 where this page comes from