xe_vm_range_rebind(9)

✨ xe_vm_range_rebind(9) ✨

okay so bestie, welcome to the kernel API manual for xe_vm_range_rebind — yes that's the manual name, it's giving "9" as in section 9, we love a numbering moment 💅 This whole page is about one lil function that (re)binds an SVM range, and I'm gonna walk you through it fr fr.

---

🎀 NAME 🎀

xe_vm_range_rebind — VM range (re)bind. that's literally it, that's the whole vibe: we're rebinding a range in a VM. iconic, short, slays 🌸

---

🦋 SYNOPSIS 🦋

struct dma_fence * xe_vm_range_rebind (struct xe_vm *vm , struct xe_vma
*vma , struct xe_svm_range *range , u8 tile_mask );

so it hands you back a struct dma_fence * and it takes four things: a vm, a vma, a range, and a tile_mask. keep that in your notes bestie, we're about to go through each one 🫶

---

💖 ARGUMENTS 💖

• vm — the VM which the range belongs to. like, which VM owns this range? that one. 🏠
• vma — the VMA which the range belongs to. yes, both a VM and a VMA are in the mix, don't mix them up, they're different besties 💀
• range — the SVM range to rebind. this is the actual thing we're here for 💫
• tile_mask — the tile mask to bind the range to. it's a u8 mask, and it's telling us which tiles we're binding to ✨

---

🔥 DESCRIPTION 🔥

(re)bind SVM range setting up GPU page tables for the range.

that's the entire description and honestly it ate — we're setting up the GPU page tables for the range, and the "(re)" in parens means it's a bind or a rebind situation, no cap. 🖥️💜

---

⭐ RETURN ⭐

on success you get a dma fence for the rebind to signal completion on (so like, watch the fence, it'll tell you when it's done 🚦). on failure you get an ERR_PTR 😭 — so check it, bestie, don't just vibe past it.

---

👑 SEE ALSO 👑

the whole squad, in order, and yes I'm listing every single one because that's what we do here 💖:

• Kernel file ./drivers/gpu/drm/xe/xe_vm.c
• xe_vm_drm_exec_lock(9)
• xe_vm_remove_compute_exec_queue(9)
• xe_vm_kill(9)
• xe_vm_validate_rebind(9)
• xe_vm_bump_lru(9)
• xe_vm_resume_rebind_worker(9)
• xe_vm_add_fault_entry_pf(9)
• xe_vm_range_unbind(9)
• xe_vma_mem_attr_copy(9)
• xe_vm_lock_vma(9)
• xe_vm_create_scratch(9)
• xe_vm_find_vma_by_addr(9)
• xe_vma_has_default_mem_attrs(9)
• struct xe_vma_lock_and_validate_flags(9)
• xe_vm_bind_kernel_bo(9)
• xe_vm_lock(9)
• xe_vm_unlock(9)
• xe_vm_invalidate_vma_submit(9)
• xe_vm_invalidate_vma(9)
• xe_vma_need_vram_for_atomic(9)
• xe_vm_alloc_madvise_vma(9)
• xe_vm_find_cpu_addr_mirror_vma_range(9)
• xe_vm_alloc_cpu_addr_mirror_vma(9)
• xe_vm_add_exec_queue(9)
• xe_vm_remove_exec_queue(9)

(and yes, struct xe_vma_lock_and_validate_flags(9) is written exactly like that in the original — struct and all — so I'm keeping it exactly like that 🫡)

---

ngl if you're skimming this page for the answer: bind → set up GPU page tables → get a fence on success or an ERR_PTR on failure. that's the tea ☕ and the rest is just the vibes.

---

🌈 alright bestie, that's the whole page! this came out of the ./drivers/gpu/drm/xe source tree 🧚 and it's dated October 2026 ✨ — and honestly? she's serving kernel documentation realness. periodt. 💅🎀