dma_resv_replace_fences(9)

📦 version 7.3.1, in nixpkgs unstable

💅 dma_resv_replace_fences(9) — like, the whole glow-up

okay so bestie, welcome to the Kernel API Manual entry for dma_resv_replace_fences(9) — yes the function name is literally the same as the page name, that's just how kernel docs roll, iykyk 🎀

✨ NAME ✨

dma_resv_replace_fences — swap out fences inside the dma_resv obj, no big deal 💅

🌸 SYNOPSIS 🌸

void dma_resv_replace_fences (struct dma_resv *obj , uint64_t context ,
       struct dma_fence *replacement , enum dma_resv_usage usage );

like, she returns void, she takes four things, she does the thing. periodt. 🦋

🎀 ARGUMENTS 🎀

  • obj — the reservation object. the main character of this whole situation fr.
  • context — the context of the fences you wanna replace. like, which ones are getting swapped out 👀
  • replacement — the new fence to use instead. the understudy. the glow-up. the swap 💖
  • usage — how the new fence is used — go peep enum dma_resv_usage for the deets.

🔥 DESCRIPTION 🔥

so what she does: she replaces fences that have a specified context with a new fence. simple, clean, iconic. ✨

⚠️ but bestie, listen — this is only valid if the operation represented by the original fence no longer has access to the resources represented by the dma_resv object when the new fence completes. read that twice. no cap. 😭

and an example (the manpage says "And example", which, mood, we're keeping it) for using this is replacing a preemption fence with a page table update fence which makes the resource inaccessible. see? the old one loses access, the new one takes over. main character energy 🔥👑

🫶 SEE ALSO 🫶

like, the whole reservation-object friend group, honestly. go say hi 🌈

and that's the page!! this lil slice of docs lives in ./drivers/dma-buf and it's stamped October 2026, page name dma_resv_replace_fences(9). mwah, slay, bye bestie 💜🍓🎀

📦 where this page comes from