nx2: get title 4 past the difficulty confirm

- MME: implement MME_DMA_READ / MME_DMA_READ_FIFOED (Turing macros read query
  results into their load FIFO; the conditional-render macro at 0x142 needs it).
- vi2 present waits up to 4 s for the frame fence (was 250 ms) so shader
  compile stalls don't let the game lap its equal-value gate semaphores.
- Native QMD: fire the third release slot at word 52, which NVN2 patches for
  queue-to-queue syncs (compute queue releasing a gate the 3D queue waits on).
- Compute class SET_REPORT_SEMAPHORE (0x6c0-0x6c3) releases.
- Ampere lowering: AOFFI, any LOP3 truth table, STG RZ, LDG 0x981, IABS,
  I2FP rounding, IMAD negated addend, SHFL 0x989 (all 77 title-4 shaders lower).
- Debug: EDEN_NX2_SEM_WATCH=<addr> traces acquires, releases, method
  arguments and compute macro emits for one semaphore block;
  EDEN_NX2_KEY_SCRIPT scripted key presses; per-submit logs moved to TRACE.

1 files changed, 20 insertions, 2 deletions

diff --git a/src/video_core/engines/puller.cpp b/src/video_core/engines/puller.cpp
index 97a58903fe..047e383b85 100644
--- a/src/video_core/engines/puller.cpp
+++ b/src/video_core/engines/puller.cpp
@@ -68,9 +68,16 @@ bool Puller::CanResumeAcquire(DmaPusher& dma_pusher) {
     if (!pending_acquire->reported &&
         std::chrono::steady_clock::now() - pending_acquire->since > std::chrono::seconds{2}) {
         pending_acquire->reported = true;
-        LOG_WARNING(HW_GPU, "Channel {} waits on semaphore {:#x} (op {}, {:#x}, now {:#x}) for 2 s",
+        const auto device = dma_pusher.memory_manager.GpuToCpuAddress(pending_acquire->address);
+        LOG_WARNING(HW_GPU, "Channel {} waits on semaphore {:#x} (op {}, {:#x}, now {:#x}) for 2 s; "
+                    "device {:#x}, buffer cache holds GPU data: {}, words {:08x} {:08x} {:08x} {:08x}",
                     dma_pusher.channel_state.bind_id, pending_acquire->address,
-                    pending_acquire->operation, pending_acquire->value, current);
+                    pending_acquire->operation, pending_acquire->value, current, device.value_or(0),
+                    device && dma_pusher.rasterizer->MustFlushRegion(*device, 16, VideoCommon::CacheType::All),
+                    dma_pusher.memory_manager.Read<u32>(pending_acquire->address & ~GPUVAddr{0xf}),
+                    dma_pusher.memory_manager.Read<u32>((pending_acquire->address & ~GPUVAddr{0xf}) + 4),
+                    dma_pusher.memory_manager.Read<u32>((pending_acquire->address & ~GPUVAddr{0xf}) + 8),
+                    dma_pusher.memory_manager.Read<u32>((pending_acquire->address & ~GPUVAddr{0xf}) + 12));
     }
     return false;
 }
@@ -157,6 +164,17 @@ void Puller::ProcessFenceActionMethod(DmaPusher& dma_pusher) {
 void Puller::ProcessSemaphoreTriggerMethod(DmaPusher& dma_pusher) {
     const auto semaphoreOperationMask = 0xF;
     const auto op = GpuSemaphoreOperation(regs.semaphore_trigger & semaphoreOperationMask);
+    // EDEN_NX2_SEM_WATCH=<gpu address>: log every release/acquire on the 16 bytes there.
+    static const GPUVAddr watch = [] {
+        const char* value = std::getenv("EDEN_NX2_SEM_WATCH");
+        return value ? GPUVAddr{std::strtoull(value, nullptr, 16)} : GPUVAddr{};
+    }();
+    if (watch != 0 && (regs.semaphore_address.SemaphoreAddress() & ~GPUVAddr{0xf}) == watch) {
+        LOG_INFO(HW_GPU, "Semaphore watch: channel {} op {} {:#x} value {:#x} (now {:#x})",
+                 dma_pusher.channel_state.bind_id, static_cast<u32>(op),
+                 regs.semaphore_address.SemaphoreAddress(), regs.semaphore_sequence,
+                 dma_pusher.memory_manager.Read<u32>(regs.semaphore_address.SemaphoreAddress()));
+    }
     if (op == GpuSemaphoreOperation::WriteLong) {
         const GPUVAddr sequence_address{regs.semaphore_address.SemaphoreAddress()};
         const u32 payload = regs.semaphore_sequence;