dma_sync_sgtable_for_device(9)

💅 dma_sync_sgtable_for_device(9) 💅

okay bestie, welcome to the Kernel API Manual — this one's a kernel-function page and it's genuinely iconic once you get the vibe, i promise 🫶

---

✨ NAME ✨

dma_sync_sgtable_for_device — synchronize the given buffer for DMA 🌈

---

✨ SYNOPSIS ✨

void dma_sync_sgtable_for_device (struct device *dev , struct sg_table
*sgt , enum dma_data_direction dir );

no cap, that's the whole signature — void, so nothing comes back to you, it just does the thing 💅

---

✨ ARGUMENTS ✨

• dev — the device you're doing the DMA operation for 🎀 like, the actual hardware main character here.
• sgt — the sg_table object that describes the buffer 👑 that's the map of what memory we're talking about.
• dir — the DMA direction 🦋 (an enum dma_data_direction, if you're keeping track).

---

✨ DESCRIPTION ✨

so what it does is: it performs the cache synchronization that's needed, and then it moves the ownership of the buffer back over to the DMA domain 💫 which means — and this is the important part — it's now safe to actually perform the DMA operation. slay.

ngl the ownership handoff is the whole vibe check here. it's giving "here DMA, this is yours now, don't embarrass me" 😭

and once you're finished with all that? you have to call dma_sync_sgtable_for_cpu or dma_unmap_sgtable. like, that's not optional bestie, don't just leave it hanging 🙈

---

✨ SEE ALSO ✨

• Kernel file ./include/linux/dma-mapping.h 📁
• dma_use_iova(9)
• dma_unmap_sgtable(9)
• dma_sync_sgtable_for_cpu(9)
• dma_get_seg_boundary_nr_pages(9)

---

and that's the page fr fr 💖 this beauty lives over in ./include/linux and it's dated October 2026 — so go forth, sync those sg_tables, and remember: hand the buffer back when you're done. periodt 🎀✨💜