From 9f4f01f76e8556941c0c0c3f57edc1d600ed23ea Mon Sep 17 00:00:00 2001
From: Sssnowyyy <78863999+Sssnowyyy@users.noreply.github.com>
Date: Mon, 2 Nov 2026 13:18:13 -0500
Subject: nx2: map GPU memory before later command lists are read

837133792d queues the GMMU maps on the GPU thread (GPU::RunInOrder) so the
service thread never waits for the texture cache. With the GPU split
(Low GPU accuracy) the GPU thread defers queued functions to its back, but
its front reads the command lists: it read the lists submitted after a map
through GPU addresses that weren't mapped yet. Pokopia stopped after its
first frame.

Add GPU::MapInOrder for the GMMU maps and the sparse tile remaps. It runs
in order like RunInOrder, and with the split it runs on the front after
draining the back, before the front reads further command lists. The
service thread still never takes the texture cache lock, and the back is
idle while the map changes the caches. Unmaps and the other queued work
still run on the back.

Assisted-by: Claude Opus 5.5 (bisect, fix, testing and commit message)
---
 src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'src/core')

diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
index 843d31caf1..c5b23bef3e 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
@@ -176,7 +176,7 @@ void nvhost_as_gpu::Nx2Map(u64 gpu_address, u64 cpu_address, u64 size) {
     Nx2GpuMemoryRegistry::Get().RecordDeviceRange(cpu_address, size, device_address);
     // On the GPU thread, like unmaps: a GMMU map takes the texture cache lock, and the callers hold
     // nx2:gpu locks that the GPU thread takes under that lock (sparse blocks backed on write).
-    system.GPU().RunInOrder([gmmu = gmmu.get(), gpu_address, device_address, size] {
+    system.GPU().MapInOrder([gmmu = gmmu.get(), gpu_address, device_address, size] {
         gmmu->Map(gpu_address, device_address, size, Tegra::PTEKind::PITCH, false);
     });
     LOG_DEBUG(Service_NVDRV, "Switch 2 GPU mapping {:#x}+{:#x} -> CPU {:#x}", gpu_address, size,
@@ -253,7 +253,7 @@ void nvhost_as_gpu::Nx2Sparse(u64 gpu_address, u64 size, bool create) {
     }
     nx2_sparse_ranges[gpu_address] = {size, device_address};
     // On the GPU thread, as in Nx2Map.
-    system.GPU().RunInOrder([gmmu = gmmu.get(), gpu_address, device_address, size] {
+    system.GPU().MapInOrder([gmmu = gmmu.get(), gpu_address, device_address, size] {
         gmmu->Map(gpu_address, device_address, size, Tegra::PTEKind::PITCH, false);
     });
     LOG_DEBUG(Service_NVDRV, "Switch 2 sparse region {:#x}+{:#x} -> device {:#x}", gpu_address,
@@ -295,7 +295,7 @@ void nvhost_as_gpu::Nx2SparseTile(u64 gpu_address, u64 cpu_address, u64 size) {
     // whose contents are now what the game's memory holds there.
     const DAddr device = device_address + (gpu_address - start);
     auto& smmu = system.Host1x().MemoryManager();
-    system.GPU().RunInOrder([&smmu, device, cpu_address, size, asid = session->asid] {
+    system.GPU().MapInOrder([&smmu, device, cpu_address, size, asid = session->asid] {
         smmu.Unmap(device, size);
         if (cpu_address != 0) {
             smmu.Map(device, cpu_address, size, asid, true);
-- 
cgit v1.2.3