Merge pull request #121 from Lowell137/fix/consistent-mac-gpu-report

Report the profile GPU consistently in system_profiler

docs/PLATFORM_IDENTITY.md

@@ -164,3 +164,12 @@ keeps no identity of its own, so it reports the same values as IOKit, including
 override saved by this tool. A field the guest does not provide is left out.
 `python3 -m unittest tests.test_system_profiler` checks the output with stub
 commands.
+
+`SPDisplaysDataType` reads the host DRM cards. With `DARLING_METAL_MAC_GPU_NAME`
+set (the launcher default), an AMD or NVIDIA host card is reported as the
+`MacBookPro16,1` discrete GPU: `AMD Radeon Pro 5500M`, vendor `0x1002`, device
+`0x7340`, bus PCIe and VRAM from `DARLING_METAL_VRAM_MB` (4096 by default; a
+value that is not a positive decimal keeps the default). An Intel card is
+reported as `Intel UHD Graphics 630` on bus Built-In with shared memory, as
+`MTLDevice.name` does. This keeps the display report consistent with Metal;
+without the variable the host card is reported as-is.

docs/SETUP.md

@@ -196,7 +196,7 @@ The hand-off experiments (`metal-present-dispatch-wait`, `metal-deferred-blit` a
 | `/tmp/metal-skip-cpu-devices` | software Vulkan devices (lavapipe) are skipped while a hardware device exists |
 | `/tmp/metal-gpu` | diagnostic: selects the Vulkan device used for Metal by index or case-insensitive name substring instead of the loader's first device. The OpenGL context that presents does not follow it; see [hybrid laptops](TROUBLESHOOTING.md#hybrid-graphics-laptops-amdintel-igpu--nvidia). `METAL_INDIUM_GPU` is the environment equivalent; the file wins |
 | `/tmp/metal-all-devices` | a `VkDevice` is created for every supported Vulkan device, as before 2026-10-16. Without it only the selected device gets one; on a machine with lavapipe installed this also drops lavapipe's idle rasterizer threads |
-| `/tmp/metal-no-drawable-vram` | drawables of a layer with `framebufferOnly` off (League sets it off) go back to host-visible system RAM, as before 2026-10-16. Without the file they are allocated in GPU memory like `framebufferOnly` drawables; nothing maps drawable memory from the CPU. On the RX 6600 this gave about 20 FPS more in a practice-tool fight and the same FPS when idle, not yet measured in an ARAM. `/tmp/metal-device.log` shows `drawable-vram=on` and `storage=vram` |
+| `/tmp/metal-no-drawable-vram` | drawables of a layer with `framebufferOnly` off (League sets it off) go back to host-visible system RAM, as before 2026-10-16. Without the file they are allocated in GPU memory like `framebufferOnly` drawables; nothing maps drawable memory from the CPU. On the RX 6600 this gave about 20 FPS more in a practice-tool fight and the same FPS when idle, not yet measured in an ARAM. `/tmp/metal-device.log` (with `DARLING_METAL_DEVICE_LOG=1`) shows `drawable-vram=on` and `storage=vram` |
 | `/tmp/metal-no-lazy-wakeup` | every completion registration signals the event loop's wake-up semaphore from the CPU again, as before 2026-10-18. The handshake is per device. Without the file a registration signals only while the poller has announced a blocking wait (a spinning poller re-reads the registrations on its next pass, and its spin ends when one arrives). On the NVIDIA proprietary driver each `vkSignalSemaphore` is an RM control ioctl that waits on a driver-wide lock, about 0.3 ms each and ~25 per League frame, and it was the largest kernel cost of the submit worker. Re-read every 256 registrations; the wait log counts `wakeupSignalled` and `wakeupSkipped` |
 | `/tmp/metal-submit-batch-us` | the submit batch window in microseconds (500–100000, default 500); re-read every 256 checks, so it can be changed during a match. A presenting commit and every CPU wait still flush the batch at once. On the NVIDIA proprietary driver every `vkQueueSubmit2` is an RM control ioctl that waits on a driver-wide lock (about 0.4 ms each in a League match); with the default window a frame took about 8.5 submits, with `5000` about 2.5 (RTX 4090: 89 → 113 FPS, same match) |
 | `/tmp/metal-submit-batch-max` | command buffers per batch (1–256, default 16), re-read like the window; `64` lets one batch hold a whole frame |

docs/TROUBLESHOOTING.md

@@ -523,7 +523,7 @@ The guest file `/tmp/metal-gpu` (a device index or part of its name, for example
 
 ### Check which GPU rendered
 
-Each process that uses Metal appends to guest `/tmp/metal-device.log` (host `~/.darling/private/tmp/metal-device.log`): every Vulkan device the loader reported, the selected one and why, and the renderer of the presenting OpenGL context with both device UUIDs:
+With `DARLING_METAL_DEVICE_LOG=1` in the launcher environment, each process that uses Metal appends to guest `/tmp/metal-device.log` (host `~/.darling/private/tmp/metal-device.log`): every Vulkan device the loader reported, the selected one and why, and the renderer of the presenting OpenGL context with both device UUIDs. The log is off by default because the guest can read it and it names the host GPU:
 
 ```
 [MetalDevice] pid=… physical name="NVIDIA GeForce RTX 3060 Laptop GPU" type=discrete vendorID=0x10de … supported=yes

patches/cocotron.patch

@@ -10116,12 +10116,13 @@ diff --git a/QuartzCore/CAMetalDrawable.mm b/QuartzCore/CAMetalDrawable.mm
 index d6f86b31..687f05d5 100644
 --- a/QuartzCore/CAMetalDrawable.mm
 +++ b/QuartzCore/CAMetalDrawable.mm
-@@ -31,6 +31,13 @@
+@@ -31,6 +31,14 @@
  #include <OpenGL/glext.h>
  #include <indium/indium.private.hpp>
  
 +#include <atomic>
 +#include <cstdio>
++#include <cstdlib>
 +#include <cstring>
 +#include <mutex>
 +#include <string>
@@ -10130,7 +10131,7 @@ index d6f86b31..687f05d5 100644
  namespace DynamicVK = Indium::DynamicVK;
  
  static void reportGLErrors(void) {
-@@ -50,8 +57,111 @@ static void reportGLErrors(void) {
+@@ -50,8 +57,114 @@
  namespace Indium::DynamicVK {
  	static DynamicFunction<PFN_vkGetSemaphoreFdKHR> vkGetSemaphoreFdKHR("vkGetSemaphoreFdKHR");
  	static DynamicFunction<PFN_vkGetMemoryFdKHR> vkGetMemoryFdKHR("vkGetMemoryFdKHR");
@@ -10154,8 +10155,11 @@ index d6f86b31..687f05d5 100644
 +	}
 +}
 +
-+// Same line format as Indium's device log.
++// Same line format as Indium's device log; off unless DARLING_METAL_DEVICE_LOG is
++// set, so the guest-readable file does not name the host GPU by default.
 +static void appendDeviceLog(const std::string& line) {
++	const char* enabled = getenv("DARLING_METAL_DEVICE_LOG");
++	if (!enabled || !*enabled || (enabled[0] == '0' && !enabled[1])) return;
 +	if (FILE* log = fopen("/tmp/metal-device.log", "a")) {
 +		fprintf(log, "[MetalDevice] pid=%d %s\n", getpid(), line.c_str());
 +		fclose(log);

patches/metal--deps--indium.patch

@@ -2919,7 +2919,7 @@ index 322e186..f712c7a 100644
  #include <indium/instance.private.hpp>
  #include <indium/command-queue.private.hpp>
  #include <indium/render-pipeline.private.hpp>
-@@ -12,39 +17,207 @@
+@@ -12,39 +17,210 @@
  
  #include <iridium/iridium.hpp>
  
@@ -2946,8 +2946,11 @@ index 322e186..f712c7a 100644
 +	// device list is built, plus one when the presenting OpenGL context first
 +	// imports a Vulkan drawable. It exists so a user can tell us which GPU
 +	// actually rendered a match, which the game's own logs and FPS numbers do not
-+	// show.
++	// show. Off by default: the guest can read this file, so enabling it names the
++	// host GPU to anything running in the prefix. Set DARLING_METAL_DEVICE_LOG=1.
 +	void appendDeviceLog(const std::string& line) {
++		const char* enabled = getenv("DARLING_METAL_DEVICE_LOG");
++		if (!enabled || !*enabled || (enabled[0] == '0' && !enabled[1])) return;
 +		if (FILE* log = fopen("/tmp/metal-device.log", "a")) {
 +			fprintf(log, "[MetalDevice] pid=%d %s\n", getpid(), line.c_str());
 +			fclose(log);

tests/test_exec_guest.py

@@ -23,10 +23,12 @@ REQUIRED = {
     '__NV_PRIME_RENDER_OFFLOAD': '1', '__EGL_VENDOR_LIBRARY_FILENAMES': '/usr/share/glvnd/egl_vendor.d/10_nvidia.json',
     'VK_ICD_FILENAMES': '/usr/share/vulkan/icd.d/nvidia_icd.json', 'LIBGL_ALWAYS_SOFTWARE': '0',
     'LD_PRELOAD': 'libgamemodeauto.so.0', 'LD_LIBRARY_PATH': '/run/opengl-driver/lib',
-    'HOME': '/home/u', 'USER': 'u', 'PATH': '/usr/bin:/bin', 'LANG': 'en_US.UTF-8', 'TERM': 'xterm',
+    'HOME': '/home/u', 'USER': 'u', 'PATH': '/usr/bin:/bin', 'LANG': 'en_US.UTF-8',
     'DPREFIX': '/home/u/.darling', 'RCLOL_REFRESH_HZ': '144', 'DYLD_LIBRARY_PATH': '/Users/u/RuntimeLibraries',
 }
 SESSION = {
+    'TERM': 'xterm', 'STARSHIP_SESSION_KEY': 'host-prompt', 'AWS_PROFILE': 'host-profile',
+    'GIT_CONFIG_GLOBAL': '/home/u/.gitconfig', 'EDITOR': '/bin/host-editor',
     'WAYLAND_DISPLAY': 'wayland-0', 'XDG_CURRENT_DESKTOP': 'GNOME', 'XDG_SESSION_TYPE': 'wayland',
     'XDG_SESSION_ID': '2', 'XDG_SEAT': 'seat0', 'DESKTOP_SESSION': 'gnome', 'GDMSESSION': 'gnome',
     'GDM_LANG': 'en_US.UTF-8', 'SESSION_MANAGER': 'local/host:@/tmp/.ICE-unix/1', 'GNOME_SETUP_DISPLAY': ':1',
@@ -45,7 +47,13 @@ class FilterTests(unittest.TestCase):
         environ = {**REQUIRED, **SESSION, 'SHELL': '/bin/fish'}
         removed = exec_guest.filter_host_environment(environ)
         self.assertEqual(set(removed), set(SESSION))
-        self.assertEqual(environ, {**REQUIRED, 'SHELL': '/bin/zsh'})
+        self.assertEqual(environ, {**REQUIRED, 'SHELL': '/bin/zsh', 'PATH': '/usr/bin:/bin:/usr/sbin:/sbin'})
+
+    def test_explicit_keep_preserves_host_path_and_terminal(self):
+        environ = {'PATH': '/custom/bin', 'TERM': 'xterm',
+                   'RCLOL_KEEP_HOST_ENV': 'PATH TERM', 'RCLOL_FILTER_HOST_ENV': '1'}
+        self.assertEqual(exec_guest.filter_host_environment(environ), [])
+        self.assertEqual(environ, {'PATH': '/custom/bin', 'TERM': 'xterm', 'SHELL': '/bin/zsh'})
 
     def test_keep_list(self):
         environ = {'WAYLAND_DISPLAY': 'wayland-0', 'SSH_AUTH_SOCK': '/s', 'SHELL': '/bin/fish',
@@ -53,6 +61,7 @@ class FilterTests(unittest.TestCase):
         self.assertEqual(exec_guest.filter_host_environment(environ), ['SSH_AUTH_SOCK'])
         self.assertEqual(environ['WAYLAND_DISPLAY'], 'wayland-0')
         self.assertEqual(environ['SHELL'], '/bin/fish')
+        self.assertNotIn('RCLOL_KEEP_HOST_ENV', environ)
 
 
 class ExecTests(unittest.TestCase):
@@ -82,8 +91,11 @@ class ExecTests(unittest.TestCase):
         self.assertEqual(child['fds'], [0, 1, 2])
         self.assertFalse(set(SESSION) & set(child['env']))
         for name, value in REQUIRED.items():
-            self.assertEqual(child['env'].get(name), value, name)
+            expected = '/usr/bin:/bin:/usr/sbin:/sbin' if name == 'PATH' else value
+            self.assertEqual(child['env'].get(name), expected, name)
         self.assertEqual(child['env']['SHELL'], '/bin/zsh')
+        self.assertNotIn('RCLOL_FILTER_HOST_ENV', child['env'])
+        self.assertNotIn('RCLOL_KEEP_HOST_ENV', child['env'])
 
 
 if __name__ == '__main__':

tools/darling-system_profiler.sh

@@ -35,7 +35,11 @@ pci_name() { # vendor device (hex without 0x) -> "Vendor Model"
  done
 }
 
-displays() { # one line per GPU: vendor|device|vram MB or empty|name
+displays() { # one line per GPU: vendor|device|vram MB or empty|name|bus
+ # Same switch as the Metal device name: a Mac profile presents the MacBookPro16,1
+ # discrete GPU instead of the host card, so system_profiler and Metal agree.
+ mac_spoof=0
+ case "${DARLING_METAL_MAC_GPU_NAME:-}" in '') ;; 0) ;; *) mac_spoof=1;; esac
  for card in "$host"/sys/class/drm/card[0-9] "$host"/sys/class/drm/card[0-9][0-9]; do
   [ -r "$card/device/vendor" ] || continue
   vendor=$(sed 's/^0x//' "$card/device/vendor"); device=$(sed 's/^0x//' "$card/device/device")
@@ -44,7 +48,23 @@ displays() { # one line per GPU: vendor|device|vram MB or empty|name
    vram=$(awk '{ printf "%d", $1 / 1048576 }' "$card/device/mem_info_vram_total")
   fi
   name=$(pci_name "$vendor" "$device"); [ -n "$name" ] || name="Display"
-  echo "$vendor|$device|$vram|$name"
+  bus=builtin
+  if [ "$mac_spoof" = 1 ]; then
+   case "$vendor" in
+    1002|10de) vendor=1002; device=7340; name='AMD Radeon Pro 5500M'
+               vram=${DARLING_METAL_VRAM_MB:-4096}
+               case "$vram" in ''|*[!0-9]*) vram=4096;; esac
+               # Keep the numeric test overflow-safe: a long digit string must
+               # fall back before the shell compares it.
+               if [ "${#vram}" -gt 7 ] || [ "$vram" -eq 0 ] || [ "$vram" -gt 1048576 ]; then
+                vram=4096
+               fi
+               bus=pcie;;
+    8086) vendor=8086; device=3e9b; name='Intel UHD Graphics 630'
+          vram=; bus=builtin;;
+   esac
+  fi
+  echo "$vendor|$device|$vram|$name|$bus"
  done
 }
 
@@ -144,11 +164,11 @@ if [ "$xml" = 1 ]; then
    SPHardwareDataType) hardware | overview_xml hardware_overview;;
    SPSoftwareDataType) software | overview_xml os_overview;;
    SPDisplaysDataType)
-    displays | while IFS='|' read -r vendor device vram name; do
+    displays | while IFS='|' read -r vendor device vram name bus; do
      name=$(printf '%s' "$name" | escape)
      echo '   <dict>'
      echo "    <key>_name</key><string>$name</string>"
-     echo '    <key>spdisplays_bus</key><string>spdisplays_builtin</string>'
+     echo "    <key>spdisplays_bus</key><string>spdisplays_$bus</string>"
      echo "    <key>spdisplays_device-id</key><string>0x$device</string>"
      echo "    <key>spdisplays_vendor</key><string>${name%% *} (0x$vendor)</string>"
      [ -z "$vram" ] || echo "    <key>spdisplays_vram</key><string>$vram MB</string>"
@@ -190,8 +210,9 @@ for type in $types; do
   SPSoftwareDataType) software | overview_text Software 'System Software Overview';;
   SPDisplaysDataType)
    printf 'Graphics/Displays:\n\n'
-   displays | while IFS='|' read -r vendor device vram name; do
-    printf '    %s:\n\n      Chipset Model: %s\n      Type: GPU\n      Bus: Built-In\n' "$name" "$name"
+   displays | while IFS='|' read -r vendor device vram name bus; do
+    [ "$bus" = pcie ] && bus_text=PCIe || bus_text='Built-In'
+    printf '    %s:\n\n      Chipset Model: %s\n      Type: GPU\n      Bus: %s\n' "$name" "$name" "$bus_text"
     [ -z "$vram" ] || printf '      VRAM (Total): %s MB\n' "$vram"
     printf '      Vendor: %s (0x%s)\n      Device ID: 0x%s\n\n' "${name%% *}" "$vendor" "$device"
    done;;

tools/exec-guest.py

@@ -16,6 +16,12 @@ SESSION_PREFIXES = (
     'XDG_SESSION_', 'XDG_SEAT', 'XDG_VTNR', 'XDG_CURRENT_DESKTOP', 'XDG_MENU_PREFIX',
     'GNOME_', 'KDE_', 'GDM', 'SSH_', 'SYSTEMD_', 'MEMORY_PRESSURE_', 'QT_', 'GTK_', 'GDK_',
     'GIO_', 'GSETTINGS_', 'GJS_', 'FLATPAK_', 'SNAP_', 'KONSOLE_', 'TERM_PROGRAM', 'BASH_FUNC_',
+    # Host shell, prompt and developer-tool metadata. None of it exists on a Mac and
+    # several entries name Linux paths, so the guest must not inherit them.
+    'STARSHIP_', 'ILLOGICAL_IMPULSE_', 'INIR_', 'PNPM_', 'YARN_', 'npm_config_', 'PIP_',
+    'CARGO_', 'GIT_', 'LESS_TERMCAP_', '_P9K_', 'VSSCRIPT_', 'ORCA_', 'CLAUDE', 'TF_',
+    'CUDA_', 'COMPOSER_', 'DEBIAN_', 'FZF_', 'BAT_', 'MYSQL_', 'AWS_', 'LESS',
+    'CLOUDSDK_', 'GH_', 'GLAB_', 'HOMEBREW_', 'ELECTRON_', 'DELTA_', 'HIST', 'GPG_',
 )
 SESSION_NAMES = {
     'WAYLAND_DISPLAY', 'DESKTOP_SESSION', 'DESKTOP_STARTUP_ID', 'SESSION_MANAGER',
@@ -23,6 +29,9 @@ SESSION_NAMES = {
     'LISTEN_FDS', 'LISTEN_PID', 'LISTEN_FDNAMES', 'COLORTERM', 'VTE_VERSION', 'WINDOWID',
     'WINDOWPATH', 'CHROME_DESKTOP', 'NO_AT_BRIDGE', 'MOTD_SHOWN', 'SNAP', 'APPIMAGE', 'APPDIR',
     'OWD', 'VIRTUAL_ENV', 'DEBUGINFOD_URLS', 'LS_COLORS', 'COLORFGBG', 'OLDPWD', 'SHLVL',
+    'AGENT', 'NO_COLOR', 'TERM', 'VISUAL', 'PROMPT_COMMAND', 'HISTFILE', 'HISTSIZE',
+    'SAVEHIST', 'HISTORY_IGNORE', 'PYTHONUNBUFFERED', 'LINES', 'COLUMNS', '_',
+    'CI', 'EDITOR', 'PAGER', 'GIT_PAGER', 'SYSTEMD_PAGER', 'MANPAGER',
 }
 GUEST_SHELL = '/bin/zsh'
 
@@ -40,6 +49,11 @@ def filter_host_environment(environ):
     # The host login shell (for example /bin/fish) does not exist in the guest.
     if 'SHELL' not in keep:
         environ['SHELL'] = GUEST_SHELL
+    # A host PATH names Linux directories; the guest always has the Mac layout.
+    if 'PATH' not in keep:
+        environ['PATH'] = '/usr/bin:/bin:/usr/sbin:/sbin'
+    environ.pop('RCLOL_KEEP_HOST_ENV', None)
+    environ.pop('RCLOL_FILTER_HOST_ENV', None)
     return removed
 
 

tools/run-riot-client.sh

@@ -44,11 +44,19 @@ unset RCLOL_UX_ARG2 RCLOL_UX_ARG2_SLOT
 # Explicit 0 or an empty value selects the CPU fallback for diagnostics.
 # DARLING_METAL_MAC_GPU_NAME reports the MacBookPro16,1 GPU of the host vendor (AMD or Intel) as
 # MTLDevice.name, matching the model IOKit reports; memory figures stay the host's.
+# VRAM clamp: honor an explicit DARLING_METAL_VRAM_MB, otherwise pin 4096 MB so the
+# 5500M identity does not leak the host card's heap size.
+export DARLING_METAL_VRAM_MB="${DARLING_METAL_VRAM_MB:-4096}"
 for option in RCLOL_GPU_DIRECT_PRESENT RCLOL_GPU_DIRECT_ASYNC RCLOL_GPU_DIRECT_VSYNC DARLING_CA_PRESENT_AT_COMMIT DARLING_CA_SKIP_LOCAL_IOSURFACE_RENDER RCLOL_FORWARD_MOUSE_MOVED RCLOL_MOUSE_CONFINEMENT DARLING_METAL_MAC_GPU_NAME; do
  eval 'value=${'"$option"'-1}'
  case "$value" in 0|'') unset "$option";; *) export "$option=$value";; esac
 done
 prefix=$prefix_dir
+# Resolve the loader before the environment filter replaces PATH with the Mac layout.
+darling_bin=$(command -v darling) || {
+ echo 'darling loader not found in PATH' >&2
+ exit 1
+}
 # RIOT_GUEST_HOME is host-side; DYLD paths must be relative to the guest root.
 case "$guest_home" in
  "$prefix"/*) guest_path=/${guest_home#"$prefix"/} ;;
@@ -166,6 +174,6 @@ exec python3 "$runner_dir/exec-guest.py" env -u ELECTRON_RUN_AS_NODE -u DYLD_INS
   DYLD_LIBRARY_PATH="${DYLD_LIBRARY_PATH:-$guest_path/RuntimeLibraries}" \
   DYLD_FRAMEWORK_PATH="${DYLD_FRAMEWORK_PATH:-$guest_path/RuntimeFrameworks}" \
   DARLING_XSHM_LIBRARY="${DARLING_XSHM_LIBRARY:-$guest_path/libDarlingXShm.dylib}" \
-  darling "$A" --region="${RCLOL_REGION:-EUW}" --locale="${RCLOL_LOCALE:-en_GB}" \
+  "$darling_bin" "$A" --region="${RCLOL_REGION:-EUW}" --locale="${RCLOL_LOCALE:-en_GB}" \
     "$@" \
     --log-dir="$guest_path/Library/Logs/Riot Games/Riot Client"