diff --git a/app/src/main/cpp/skyline/gpu.h b/app/src/main/cpp/skyline/gpu.h
index 33959cd4..7e7b11bc 100644
--- a/app/src/main/cpp/skyline/gpu.h
+++ b/app/src/main/cpp/skyline/gpu.h
@@ -57,6 +57,8 @@ namespace skyline::gpu {
         cache::RenderPassCache renderPassCache;
         cache::FramebufferCache framebufferCache;
 
+        std::mutex channelLock;
+
         GPU(const DeviceState &state);
     };
 }
diff --git a/app/src/main/cpp/skyline/soc/gm20b/channel.cpp b/app/src/main/cpp/skyline/soc/gm20b/channel.cpp
index aacc5c43..c6372acc 100644
--- a/app/src/main/cpp/skyline/soc/gm20b/channel.cpp
+++ b/app/src/main/cpp/skyline/soc/gm20b/channel.cpp
@@ -12,5 +12,6 @@ namespace skyline::soc::gm20b {
           maxwellDma(state, *this, executor),
           keplerCompute(state, *this),
           inline2Memory(*this),
-          gpfifo(state, *this, numEntries) {}
+          gpfifo(state, *this, numEntries),
+          globalChannelLock{state.gpu->channelLock} {}
 }
diff --git a/app/src/main/cpp/skyline/soc/gm20b/channel.h b/app/src/main/cpp/skyline/soc/gm20b/channel.h
index 30f72525..b79a60b1 100644
--- a/app/src/main/cpp/skyline/soc/gm20b/channel.h
+++ b/app/src/main/cpp/skyline/soc/gm20b/channel.h
@@ -30,7 +30,18 @@ namespace skyline::soc::gm20b {
         engine::KeplerCompute keplerCompute;
         engine::Inline2Memory inline2Memory;
         ChannelGpfifo gpfifo;
+        std::mutex &globalChannelLock;
 
         ChannelContext(const DeviceState &state, std::shared_ptr<AddressSpaceContext> asCtx, size_t numEntries);
+
+        void Lock() {
+            globalChannelLock.lock();
+            executor.LockPreserve();
+        }
+
+        void Unlock() {
+            executor.UnlockPreserve();
+            globalChannelLock.unlock();
+        }
     };
 }
diff --git a/app/src/main/cpp/skyline/soc/gm20b/engines/gpfifo.cpp b/app/src/main/cpp/skyline/soc/gm20b/engines/gpfifo.cpp
index 7b1db598..f447a990 100644
--- a/app/src/main/cpp/skyline/soc/gm20b/engines/gpfifo.cpp
+++ b/app/src/main/cpp/skyline/soc/gm20b/engines/gpfifo.cpp
@@ -25,7 +25,10 @@ namespace skyline::soc::gm20b::engine {
                     Logger::Debug("Wait syncpoint: {}, thresh: {}", +action.index, registers.syncpoint->payload);
 
                     // Wait forever for another channel to increment
+
+                    channelCtx.Unlock();
                     syncpoints.at(action.index).Wait(registers.syncpoint->payload, std::chrono::steady_clock::duration::max());
+                    channelCtx.Lock();
                 }
             })
 
diff --git a/app/src/main/cpp/skyline/soc/gm20b/gpfifo.cpp b/app/src/main/cpp/skyline/soc/gm20b/gpfifo.cpp
index b7dac13d..5be67c7a 100644
--- a/app/src/main/cpp/skyline/soc/gm20b/gpfifo.cpp
+++ b/app/src/main/cpp/skyline/soc/gm20b/gpfifo.cpp
@@ -340,13 +340,23 @@ namespace skyline::soc::gm20b {
             signal::SetSignalHandler({SIGINT, SIGILL, SIGTRAP, SIGBUS, SIGFPE}, signal::ExceptionalSignalHandler);
             signal::SetSignalHandler({SIGSEGV}, nce::NCE::HostSignalHandler); // We may access NCE trapped memory
 
-            gpEntries.Process([this](GpEntry gpEntry) {
+            bool channelLocked{};
+
+            gpEntries.Process([this, &channelLocked](GpEntry gpEntry) {
                 Logger::Debug("Processing pushbuffer: 0x{:X}, Size: 0x{:X}", gpEntry.Address(), +gpEntry.size);
+
+                if (!channelLocked) {
+                    channelCtx.Lock();
+                    channelLocked = true;
+                }
+
                 Process(gpEntry);
-            }, [this]() {
+            }, [this, &channelLocked]() {
                 // If we run out of GpEntries to process ensure we submit any remaining GPU work before waiting for more to arrive
                 Logger::Debug("Finished processing pushbuffer batch");
                 channelCtx.executor.Submit();
+                channelCtx.Unlock();
+                channelLocked = false;
             });
         } catch (const signal::SignalException &e) {
             if (e.signal != SIGINT) {