dma_iova_try_alloc(9)

📦 version 7.3.1, in nixpkgs unstable

✨ dma_iova_try_alloc(9) ✨

okay so bestie, welcome to the dma_iova_try_alloc page from the Kernel API Manual — yes that's the manual name, and yes we're covering it, because this little function is literally the vibes check for whether your device can do the whole IOVA-based DMA thing or whether you gotta slum it with the regular DMA API. 🎀

🌸 NAME 🌸

dma_iova_try_alloc — bestie's honest attempt at allocating an IOVA space 💅

🦋 SYNOPSIS 🦋

bool dma_iova_try_alloc (struct device *dev , struct dma_iova_state
*state , phys_addr_t phys , size_t size );

like, that's the whole signature, no cap. returns a bool and everything 🔥

🌈 ARGUMENTS 🌈

  • dev — the device you're allocating the IOVA space for. it's the main character here 👑
  • state — the IOVA state. yes, it's just called state. slay 🫶
  • phys — a physical address. hold that thought, there's a note about it below 👀
  • size — the IOVA size. how much space we're claiming, periodt ✨

💖 DESCRIPTION 💖

so what this does is: it checks if dev actually supports the IOVA-based DMA API, and if it does — icon behaviour — it allocates IOVA space for the given base address and size. that's it, that's the function 🧚

⭐ NOTE ⭐

heads up bestie: phys is only used to calculate the IOVA alignment. so if your callers always do PAGE_SIZE aligned transfers, you can safely pass 0 here. no stress 💜

and the return value, because this matters: it returns true if the IOVA-based DMA API can be used and IOVA space has been allocated — or false if you should use the regular DMA API instead. basically a glow-up vs. staying home situation 😭

🍓 SEE ALSO 🍓

kernel file ./drivers/iommu/dma-iommu.c, plus the whole friend group:

okay that's the whole page, ily, go allocate something 💅✨

source: ./drivers/iommu · version/date: October 2026 · that's the sign-off bestie, mwah 🫶🎀

📦 where this page comes from