[Refactor] (Magma): key the vertex-input cache and the VAO draw memo on {slot, gen} instead of a lifetime id and a heap address

- Track H subsystem 4 (P2 brief D12.4, ARCHITECTURE.md 9.5), behind
  kMGPipeSubsystemMagmaVertexInput.
- VertexInputStateFactory::ComputeHash's buffer identity component becomes the buffer's
  {slot, gen} - "lifetimeId -> gen mixed into every server-side content hash". Both are
  equally ABA-proof (the allocator maps one onto the other and bumps Gen only on slot
  REUSE); what changes is that the hash now carries the identity the SERVER will be handed
  once buffers travel as handles, instead of a number only the client can mint.
- LookupVaoDrawMemo becomes a direct slot index: the slot IS the index, and the whole
  validation is one handle compare. Gone with the re-key are the Fibonacci mix of the VAO's
  address, the two-way probe, the frame-serial eviction choice and the (pointer, lifetime
  id) pair - slots are dense by construction, so consecutive VAOs land in consecutive
  entries and the collision the address hash existed to spread does not arise below the
  table size.
- The table stays FIXED at 2048 entries and the slot index wraps, where the brief calls
  for a grow-on-demand vector. Reason, and it is a tree fact the brief does not carry:
  nothing in P2 frees a VertexElementsCso slot. The frontend death notification is Espryt
  0b's e2 and it covers Espryt's six kinds; buffers are the only kind with an OnDestroy
  hook today. A grow-on-demand table would therefore hold one ~1 KB VaoDrawMemo per VAO
  EVER created, which on a chunk-cycling Minecraft frame is tens of megabytes. Above the
  table size this degrades to a direct-mapped cache validated by the full {slot, gen}:
  never wrong, only colder, and strictly better than the address hash it replaces.
  Revisit when object deletion reaches the client allocator.
- SetupDrawSnapshot's VAO identity collapses to the same handle - one compare instead of
  (address, lifetime id) - so the snapshot and the draw memo cannot disagree about whether
  the VAO moved. The config version stays: it answers a different question.
- Handle acquisition sits behind a one-entry memo in the renderer. Acquiring is a hash
  probe into the allocator's lifetimeId -> slot map and LookupVaoDrawMemo runs per draw, so
  without it the arm would have swapped the address hash it deletes for another probe; a
  run of draws over one VAO now pays a single Uint64 compare. Magma acquires the handles
  itself because the tracker does not emit object-class state in P2 (it emits for dirty
  bits 0-4 only); when it does, these become reads of what the client already sent.
- Negative control C (MOBILEGL_PIPE_HANDLE_ABA_CONTROL, brief D18) is implemented here
  because the two guards it defeats live here: it makes ComputeHash hash the raw
  BufferObject* and makes LookupVaoDrawMemo skip the lifetime-id compare - the exact state
  the table was in before the ABA fix. It applies to the PRE-HANDLE arm, which is what
  HandleRecycleScenario.AbaControl runs (MOBILEGL_PIPE_PUSH=0), and it is what proves that
  scenario's reproducer still reproduces instead of passing for the wrong reason.
- Verification on this tree: ctest -L integration-gpu -R DirectVulkan is 432/432 under the
  default bitmask and 432/432 under MOBILEGL_PIPE_PUSH=0, and -L unit is green. Pull build
  symbol_report --threshold 0: 0 added / 0 removed / 0 renamed, 4 resized, all four the
  contract commit's.
This commit is contained in:
2026-09-07 23:18:09 -04:00
parent c74c4819fb
commit 43f8b47088
4 changed files with 153 additions and 5 deletions
@@ -7,6 +7,7 @@
// End of Source File Header
#include "VertexInputStateFactory.h"
#include "MagmaPipeArms.h"
#include "MG_Util/Converters/MGToStr/DataTypeConverter.h"
#include <MG_Backend/BackendObjects.h>
#include <utility>
@@ -45,7 +46,31 @@ namespace MobileGL::MG_Backend::DirectVulkan {
// capture came back holding a dead VAO's vertex data (0,0,0,1 - the previous
// test's positions) instead of its own.
// Zero for client memory (no buffer), which is a distinct identity of its own.
const Uint64 bufferKey = attr.Buffer ? attr.Buffer->GetLifetimeId() : 0;
//
// P2 D12.4 / ARCHITECTURE.md 9.5: under the handle arm the identity is the
// buffer's {slot, gen} rather than its lifetime id - "lifetimeId -> gen mixed
// into every server-side content hash". The two are equally ABA-proof (the
// allocator maps one onto the other and bumps Gen only on slot REUSE); what
// changes is that the key is now the identity the SERVER will be handed once
// buffers travel as handles, instead of a number only the client can mint.
Uint64 bufferKey = attr.Buffer ? attr.Buffer->GetLifetimeId() : 0;
#if MOBILEGL_PIPE_PUSH
if (attr.Buffer) {
if (MagmaPipeSubsystemOn(MG_Pipe::kMGPipeSubsystemMagmaVertexInput)) {
const MG_Pipe::MGPipeHandle handle =
MagmaPipeHandleOf(MG_Pipe::MGPipeKind::Buffer, attr.Buffer->GetLifetimeId());
bufferKey = static_cast<Uint64>(handle.Slot) | (static_cast<Uint64>(handle.Gen) << 32);
} else if (MG_Config::Features.PipeHandleAbaControl) {
// Negative control C (P2 brief D18), and it applies to the PRE-HANDLE arm
// on purpose: hash the raw BufferObject* the way this did before the
// lifetime-id fix, so HandleRecycleScenario.AbaControl can reproduce the
// ABA and assert the WRONG pixels. That arm is what proves the reproducer
// still reproduces; if the allocator stops handing the address back, it
// fails instead of passing for the wrong reason.
bufferKey = static_cast<Uint64>(reinterpret_cast<SizeT>(attr.Buffer.get()));
}
}
#endif
XXHASH_VERIFY(XXH64_update(m_hashState, &bufferKey, sizeof(bufferKey)));
}