xe_exec_queu...st_fence_get(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… xe_exec_queue_tlb_inval_last_fence_get β€” the Kernel API Manual tea πŸ«–

okay bestie, so this is a page out of the Kernel API Manual (yes, that manual, the <italic>9</italic> one) and it's all about one little function with a very long name. stick with me πŸ’–


✨ NAME ✨

xe_exec_queue_tlb_inval_last_fence_get β€” literally just "get the last fence for TLB invalidation." that's the whole vibe. no cap πŸŽ€


🌸 SYNOPSIS 🌸

struct dma_fence * xe_exec_queue_tlb_inval_last_fence_get (struct
xe_exec_queue *q , struct xe_vm *vm , unsigned int type );

three args, one struct dma_fence * coming back at you. slay πŸ’…


πŸ¦‹ ARGUMENTS πŸ¦‹

  • q β€” the exec queue. the main character, honestly πŸ‘‘
  • vm β€” the VM that the engine does a bind for. she's doing bind things, don't worry about it 🌈
  • type β€” either the primary or the media GT. pick your fighter πŸ₯Š

πŸ’œ DESCRIPTION πŸ’œ

so what does it actually do? it gets the last fence. and β€” important detail, bestie β€” it takes a ref πŸ”₯ that's it, that's the whole description, and honestly? iconic. short and confident. periodt.


⭐ RETURN ⭐

what you get back depends on the vibe check:

  • last fence if it's not signaled πŸ’–
  • dma fence stub if it is signaled 😭

🧚 SEE ALSO 🧚

okay so this is the part where the man page goes absolutely feral and links like every sibling this function has. buckle up, we're going down the whole family tree 🌳

the kernel file it lives in is ./drivers/gpu/drm/xe/xe_exec_queue.c β€” that's home base, iykyk 🏠

and then, in order, no skipping (i would never):

like i said. the whole family. ate and left no crumbs πŸ“


fr fr bestie, this page comes straight out of ./drivers/gpu/drm/xe and it's dated October 2026 β€” so yes, we're reading docs from the future, that's just how we roll around here ⏰✨

k byeee, go take that ref responsibly πŸ’…πŸ«Ά

πŸ“¦ where this page comes from