VMA OFFSET MANAGER(9)

💅 vma offset manager 💅

okay so welcome to the vma offset manager, bestie — this is the kernel API manual page for it, and honestly? it's kind of a vibe. let's get into it 🫶

---

🌸 DESCRIPTION 🌸

so the vma-manager's whole job? it's responsible to map arbitrary driver-dependent memory regions into the linear user address-space. that's it, that's the gig ✨ it provides offsets to the caller which can then be used on the address_space of the drm-device. and it takes care to not overlap regions, to size them appropriately, and to not confuse mm-core by inconsistent fake vm_pgoff fields. no cap, it's very considerate 💅

🚨 real talk though: drivers shouldn't use this for object placement in VMEM. this manager should only be used to manage mappings into linear user-space VMs. periodt.

🧚 the backend situation

we use drm_mm as backend to manage object allocations. but — and this is the important part — it is highly optimized for alloc/free calls, not lookups. hence, we use an rb-tree to speed up offset lookups. iykyk, different data structure for different problems 💖

💀 the big no-no

you must not use multiple offset managers on a single address_space. otherwise, mm-core will be unable to tear down memory mappings as the VM will no longer be linear. like... don't do it. it's giving chaos. 😭

🎀 pages, not bytes (please remember this)

this offset manager works on page-based addresses. that is, every argument and return code (with the exception of drm_vma_node_offset_addr) is given in number of pages, not number of bytes. so object sizes and offsets must always be page-aligned (as usual). if you want to get a valid byte-based user-space address for a given offset, please see drm_vma_node_offset_addr. it's literally right there bestie 🦋

🔥 bonus round: access management

additionally to offset management, the vma offset manager also handles access management. for every open-file context that is allowed to access a given node, you must call drm_vma_node_allow. otherwise, an mmap call on this open-file with the offset of the node will fail with -EACCES. 😤 to revoke access again, use drm_vma_node_revoke. however — important — the caller is responsible for destroying already existing mappings, if required. that part's on you, queen 👑

---

💜 SEE ALSO 💜

• Kernel file ./drivers/gpu/drm/drm_vma_manager.c
• drm_vma_offset_manager_init(9)
• drm_vma_offset_manager_destroy(9)
• drm_vma_offset_lookup_locked(9)
• drm_vma_offset_add(9)
• drm_vma_offset_remove(9)
• drm_vma_node_allow(9)
• drm_vma_node_allow_once(9)
• drm_vma_node_revoke(9)
• drm_vma_node_is_allowed(9)

---

and that's the whole page, bestie!! 🌈 this little icon lives in ./drivers/gpu/drm, it's stamped October 2026, and it goes by the name vma offset manager(9) — go forth and map those regions responsibly 💖✨🎀