I915_PPGTT_CREATE AND I915_PPGTT_RELEASE TRACEPOINTS(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… i915_ppgtt_create and i915_ppgtt_release tracepoints πŸ’…

okay bestie, welcome to the tracepoints(9) manual β€” yes, the manual name is folded right in up here, we’re not doing the stiff repeat-the-header thing. vibe check: when full ppgtt is enabled, each process using drm will allocate at least one translation table. slay for them, but somebody has to keep track πŸ‘€.

With these traces it’s possible to keep track of the allocation and of the lifetime of the tables. no cap, that’s the whole point. this can be used during testing/debug to verify that we are not leaking ppgtts β€” because leaks are not iconic, they’re giving β€œwhere did that memory go” energy 😭.

These traces identify the ppgtt through the vm pointer, which is also printed by the i915_vma_bind and i915_vma_unbind tracepoints. so the vm pointer is the little name tag that lets you match things up, fr fr 🫢.


πŸ‘€ SEE ALSO πŸ‘€

  • Kernel file ./drivers/gpu/drm/i915/i915_trace.h β€” also go peek at i915_context_create and i915_context_free tracepoints(9) ✨.

mwah πŸ’‹ this page lives under ./drivers/gpu/drm/i915, is dated October 2026, and answers to i915_ppgtt_c... tracepoints(9) β€” periodt.

πŸ“¦ where this page comes from