Commit Graph
3075 Commits
Author SHA1 Message Date
Henrik Rydgård 8cd7e1b2c0 Delete all support for Qt
Our Qt backend has long been left behind and doesn't even support Vulkan
currently. There would be a lot of work to make it viable, and I don't
think anyone is really interested.

ImGui on SDL will soon fulfill the need for a more classic user interface
with a menu bar on Linux, and on Mac we already have a native UI.
2026-08-08 18:18:28 +02:00
Henrik RydgårdandClaude Sonnet 5 3500293402 SymbolMap: route remaining GUI-thread callers, remove internal locking
Windows/MainWindowMenu.cpp's Load/Save/Clear Symbol Map menu actions
mutated g_symbolMap directly from the WinMain thread; route them through
Core_RunOnCPUThread like the rest of the debugger. Windows/MainWindow.cpp's
WM_USER_GET_MODULE_INFO handler read it directly too; guard with
g_frameMutex. Windows/main.cpp's SortSymbols() calls turn out to already be
safe as-is - both notifications that trigger them (BOOT_DONE,
SYMBOL_MAP_UPDATED) are only ever fired from the CPU/NativeFrame thread -
so just fix the stale comment claiming reliance on the (now removed)
internal lock.

With those covered, every remaining caller is either on the CPU/NativeFrame
thread already or routes through Core_RunOnCPUThread/g_frameMutex, so
SymbolMap's internal recursive_mutex is redundant - remove it and all ~40
lock_guard call sites. Qt's mainwindow.cpp still pokes at g_symbolMap
directly and unguarded (a pre-existing issue, out of scope - Qt isn't a
maintained backend), but removing the lock doesn't change its public API,
so it still builds.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hqm11k99viLfbJm2MkH4BH
2026-08-08 17:22:12 +02:00
Henrik RydgårdandClaude Sonnet 5 f72709feb0 Legacy Win32 debugger: fix the painting problem with a frame-scoped mutex
Debugger windows (register list, disassembly view, memory view, breakpoint/
thread/module/stack lists, watch list) read CPU-thread-owned state directly
from the GUI thread's WM_PAINT/list-fill handlers, racing against the CPU
thread. Routing every read through Core_RunOnCPUThread would be too slow for
something invoked continuously on paint/list-refresh.

Add g_frameMutex (Core.h/Core.cpp), held by NativeFrame() only across the
span where it actually touches that state (running the CPU, processing
breakpoints, running the ImGui debugger) - not across input handling or the
present/frame-pacing waits. Debugger windows now hold the same mutex while
reading, giving synchronized reads without the round-trip cost of queuing
to the CPU thread.

CtrlRegisterList::onPaint() goes back to always reading live values (now
safe under the lock) and grays them out by color alone while the core is
running, rather than the earlier snapshot-caching approach.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hqm11k99viLfbJm2MkH4BH
2026-08-08 17:22:12 +02:00
Henrik RydgårdandClaude Sonnet 5 25c3c90c94 Legacy Win32 debugger: route mutations through the CPU thread
The Win32 debugger dialogs mutate CPU-thread-owned state (breakpoints,
symbol map, registers, memory, kernel threads) directly from the GUI
thread with no synchronization, same problem the WebSocket debugger had.
Route all of these through Core_RunOnCPUThread instead, following the
same pattern used there.

Also drops two forced-pause (Core_Break/Core_WaitInactive/Core_Resume)
dances in CtrlMemView::onChar and DumpMemoryWindow's dump handler, now
unnecessary since Core_RunOnCPUThread works whether the CPU is running
or already stepping.

Modal dialogs (MessageBox, InputBox_GetString, BreakpointWindow::exec,
etc.) are kept outside any queued callback so they never block the CPU
thread on user input. Pure reads/painting (disassembly formatting,
search, list reloads, onPaint) are intentionally left as-is for now.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hqm11k99viLfbJm2MkH4BH
2026-08-08 17:22:12 +02:00
Henrik RydgårdandClaude Sonnet 5 b1f0112cef Debugger: Remove opcode-fusion display and fix cpu step size units
DisassemblyManager used to fuse lui+addiu/load/store into single pseudo-
instructions ("li", fused loads/stores) for display. This only applied to a
handful of opcodes, complicated DisassemblyManager, and was the root cause of
a stepping bug: Core_PerformCPUStep's Into/Over cases treated stepSize as a
byte count, while the WebSocket cpu.stepInto handler computed it as an
instruction count (needed to step over a whole fused macro in one go) - so a
plain, non-fused stepInto silently executed zero instructions.

Removed the fusion logic entirely (DisassemblyMacro, DISTYPE_MACRO) - every
disassembly line is now exactly one 4-byte instruction. With that,
"how many instructions does this line span" is always 1, so the
getInstructionSizeAt() byte-size queries in the legacy Windows and ImGui
debuggers are gone too; step requests just pass 1. Core_RequestCPUStep's
stepSize is now consistently in instructions everywhere.

Also fixes the PPSSPPHeadless build, broken since 0ed1f3e added
OpenWebDebugger() (which calls System_LaunchUrl) without a headless stub.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hqm11k99viLfbJm2MkH4BH
2026-08-08 11:16:36 +02:00
Henrik Rydgård 4160414b31 Vulkan: remove libretro's global function-pointer wrapper hack
libretro/libretro_vulkan.cpp got PPSSPP's libretro core working with
RetroArch's Vulkan integration by globally monkey-patching PPSSPP's Vulkan
loader function pointers (vkCreateInstance, vkCreateDevice,
vkCreateSwapchainKHR, vkAcquireNextImageKHR, vkQueuePresentKHR,
vkQueueSubmit, etc.) so the unmodified VulkanContext class would end up
wrapping RetroArch's already-existing VkInstance/VkDevice instead of
creating its own, and so a fake VkSwapchainKHR (a self-managed array of
images synced against RetroArch's retro_hw_render_interface_vulkan
callbacks) could stand in for the real swapchain that libretro's Vulkan
model doesn't have. Flagged in-code as "a wacky wrapper".

Replaces that with first-class support in VulkanContext for the two things
libretro actually needs:

- Adopting an externally-created instance/device instead of faking
  vkCreateInstance/vkCreateDevice: VulkanContext::CreateInstanceExternal()
  adopts RetroArch's VkInstance; CreateDevice() gained optional
  extraDeviceExtensions/extraRequiredFeatures params so RetroArch's
  requirements get merged into a real vkCreateDevice() call;
  ownsInstance_/ownsDevice_ flags (the latter set via
  SetDeviceExternallyOwned()) mean DestroyInstance()/DestroyDevice() skip
  the real vkDestroy* calls when something else owns the object, without
  needing to intercept anything. VulkanLoader gained
  VulkanLoadFromGetInstanceProcAddr() for bootstrapping from a
  host-supplied proc-addr getter instead of dlopen/dlsym-ing the loader
  ourselves - vkGetDeviceProcAddr is resolved via the real instance handle
  (not NULL), since per the Vulkan spec it's not one of the handful of
  commands queryable with a NULL instance.
- A pluggable presentation backend (Common/GPU/Vulkan/VulkanPresentation.h)
  for hosts with no real VK_KHR_swapchain, replacing the fake-swapchain-
  handle trick. VulkanContext::GetPresentation() is null by default, so
  every existing platform's real-swapchain code path is untouched;
  libretro/LibretroVulkanPresentation implements this interface directly
  against retro_hw_render_interface_vulkan, as real class state instead of
  file-scope globals. Several pieces of state that are normally only
  populated as a side effect of ReinitSurface()/InitSwapchain() - the
  graphics queue/queue family index (ChooseQueue() is entangled with
  real-surface presentation-support checks), the swapchain format, and
  the available present modes - needed presentation-aware fallbacks since
  libretro never calls that real-surface path at all.

libretro/LibretroVulkanContext.cpp now drives VulkanContext's real, public
API directly - no more hijacked function pointers, no more fake surface or
swapchain. libretro/libretro_vulkan.cpp is deleted.

Verified with a full build+run in RetroArch (not just compile-time
checks): the libretro Makefile doesn't track header dependencies
(cl.exe doesn't support -MMD/-MP, and Makefile.common never sets up an
equivalent), so a `make clean` full rebuild is required after any header
change to avoid linking stale object code from before the change - several
of the fixes above were initially masked by exactly that.
2026-08-06 17:07:05 +02:00
Henrik Rydgård 551e4cd0ab Headless: Make all GPU backends work correctly on Windows 2026-08-05 00:15:56 +02:00
Henrik Rydgård 25cf52176f Misc headless work 2026-08-05 00:15:56 +02:00
Henrik Rydgård 5eb4772644 Minor EmuThread refactor 2026-08-05 00:15:56 +02:00
Henrik Rydgård be922eca62 Make OpenGL render thread shutdown more robust. 2026-08-04 16:44:55 +02:00
Henrik Rydgård 971b5eca13 WASAPI: Use device's suggested format when available
When IsFormatSupported returns S_FALSE with a closestMatch format,
evaluate it and use it if it's acceptable (stereo float). This can
improve compatibility with devices that don't support our exact
requested format but have a similar one.

Previously we just discarded closestMatch and forced manual conversion,
which was wasteful when the device provided a perfectly usable
alternative format.
2026-08-04 00:37:55 +02:00
Henrik Rydgård 32f3fbd65a WASAPI: Fix channel mapping for 5.1 audi
CRITICAL BUG FIX: The previous code assumed incorrect channel ordering
for multichannel audio, breaking 5.1 (6-channel) configurations.

Problem:
- Code assumed channel order: FL, FR, RL, RR, C, LFE
o systems- Actual 5.1 standard order: FL, FR, C, LFE, RL, RR
- This caused center/LFE to receive rear audio and vice versa
- Users had to force stereo mode to get working audio

Solution:
- Parse dwChannelMask from WAVEFORMATEXTENSIBLE to determine actual
  channel positions
- Map stereo input to correct output channels based on the mask
- Properly handles 5.1, 7.1, and other multichannel configurations
- Safely handles non-standard channel layouts

The channel mapping now works correctly for all standard speaker
configurations by querying the device's actual channel layout instead
of assuming a fixed order.
2026-08-04 00:37:55 +02:00
Henrik Rydgård 11d4107230 WASAPI: Replace magic numbers with named constants
Define constants for channel upmixing attenuation factors to improve
code readability and maintainability:
- SURROUND_ATTENUATION (0.7f) for rear/side channels
- CENTER_MIX_ATTENUATION (0.7f) for center channel
- LFE_MIX_ATTENUATION (0.5f) for LFE channel
2026-08-04 00:37:55 +02:00
Henrik Rydgård fd1e48ecc5 WASAPI: Add error check for RegisterEndpointNotificationCallback
Log a warning if registering for device change notifications fails.
This is non-fatal but means automatic device switching won't work.
2026-08-04 00:37:55 +02:00
Henrik Rydgård 95e09324a7 WASAPI: Fix ClampFloatToS16 to use full s16 range
Changed minimum clamp from -32767 to -32768 to match the full
range of signed 16-bit integers.
2026-08-04 00:37:55 +02:00
Henrik Rydgård d97f42c225 WASAPI: Cache channel count to avoid unsafe pointer access from audio thread
Store format_->nChannels in atomic curChannels_ member to avoid
unsynchronized access to format_ pointer from the audio thread.
The format_ pointer can be freed/reallocated during Stop() while
the audio thread might still be reading it.
2026-08-04 00:37:55 +02:00
Henrik Rydgård 456cfb0f5a RAIntegration crashfix 2026-07-27 18:37:28 +02:00
Henrik Rydgård 5ef736f6a7 More command line improvements 2026-07-27 18:37:28 +02:00
Henrik Rydgård ccff34b431 Reuse the AndroidJavaGLContext as the new generic OpenGLGraphicsContext 2026-07-26 20:07:14 +02:00
Henrik Rydgård cb623f725c Get rid of the SDL-specific Vulkan graphics context 2026-07-26 19:10:46 +02:00
Henrik Rydgård 3784bbe0e6 Delete the AndroidVulkanContext, replace with the generic VulkanGraphicsContext 2026-07-26 19:10:46 +02:00
Henrik Rydgård e3d36e2741 Move the VulkanGraphicsContext to its correct location in Common/GPU/Vulkan 2026-07-26 14:15:15 +02:00
Henrik Rydgård 12cbe2143a Rename WindowsVulkanContext to VulkanGraphicsContext, preparing for moving it 2026-07-26 14:01:05 +02:00
Henrik Rydgård 3eb056ad86 Move Common/GraphicsContext.h to Common/GPU/GraphicsContext.h 2026-07-26 13:58:17 +02:00
Henrik Rydgård 4a95d8eeaa Fix exit crash on Windows, oops 2026-07-26 13:37:33 +02:00
Henrik Rydgård 7f218acc53 Change the GraphicsContext interface to be more consistent 2026-07-26 10:22:00 +02:00
Henrik Rydgård c2c0fa3634 Remove the AndroidGraphicsContext middle-level class 2026-07-26 10:21:59 +02:00
Henrik Rydgård 85f8715660 Break the circular dependency caused by putting EmuThread in core using an interface 2026-07-24 15:31:42 +02:00
Henrik Rydgård af7977267e Add NeedsRenderThread property to GraphicsContext 2026-07-24 15:31:41 +02:00
Henrik Rydgård 86c98c2d23 Prepare EmuThread for other users 2026-07-24 15:31:41 +02:00
Henrik Rydgård cb18929f3a Make EmuThread.cpp/h compile outside of Windows. 2026-07-21 13:25:31 +02:00
Henrik Rydgård 32f975ef85 Windows: Move the graphics context creation to main.cpp 2026-07-21 13:13:01 +02:00
Henrik Rydgård cc69070f19 Remove MainThread_Start/Stop to increase visibility 2026-07-21 12:58:45 +02:00
Henrik Rydgård 221aba92da Move EmuThread.cpp/h to Core, in preparation for sharing logic 2026-07-21 12:50:22 +02:00
Henrik Rydgård 16c7ec390f Headless: Start taking logic out of the "HeadlessHost"
Also update some agent files
2026-07-21 11:01:48 +02:00
Henrik Rydgård bd25962cbe Convert more options to the new command line parser 2026-07-20 16:33:03 +02:00
Henrik Rydgård e6875fa339 More cmdline work. 2026-07-20 16:33:03 +02:00
Henrik Rydgård a59637975a Call the new command line parser from all backends 2026-07-20 12:02:07 +02:00
Henrik Rydgård acc0a19786 wip 2026-07-20 12:02:07 +02:00
Henrik Rydgård a73fcc7b4b Windows: Call NativeInit/NativeShutdown outside the EmuThread. 2026-07-20 12:02:07 +02:00
Henrik Rydgård f022328521 Remove the "restart emuthread" facility that isn't really working (and won't be compatible with the next refactorings) 2026-07-20 12:02:07 +02:00
Henrik Rydgård 29836ea74e Move the new incomplete command line parser to Core 2026-07-20 12:02:07 +02:00
Henrik Rydgård a886cfb5ab First step of centralizing command line parsing 2026-07-20 12:02:07 +02:00
Henrik Rydgård f5a8ad172e Fix various issues with the headless build (for running frame dumps) 2026-07-19 00:09:06 +02:00
Henrik Rydgård b469e57735 Some refactoring in Windows/EmuThread.cpp/h 2026-07-18 23:54:32 +02:00
Henrik Rydgård 986eee7ef3 Remove the global variable "useEmuThread", add some VK error checking for compute shader compiles 2026-07-18 23:45:27 +02:00
Henrik Rydgård fdf186e43b Windows: Refactor console positioning, moving the code out of EmuThread.cpp 2026-07-18 20:58:29 +02:00
Henrik Rydgård 993c3a9c67 Avoid accurately computing disk space on iOS and Android (where it can be slow). Add a setting. 2026-07-14 22:22:29 +02:00
Henrik Rydgård 61e1ef8f7a Remove the "Software skinning" option. Now always on. 2026-07-14 17:02:57 +02:00
Henrik Rydgård 1b2f87bf1a Rename GPUgstate to GEState 2026-07-02 20:34:09 +02:00