Commit Graph
615 Commits
Author SHA1 Message Date
Henrik RydgårdandClaude Sonnet 5 29825c1e61 Debugger: Route the rest of SteppingSubscriber.cpp's handlers through the CPU thread
Extends the cpu.stepInto treatment to cpu.stepOver, cpu.stepOut, cpu.runUntil,
and cpu.nextHLE: each now routes its breakpoint/stepping manipulation through
Core_RunOnCPUThread() instead of touching it directly from the WebSocket
handler thread. cpu.runUntil didn't have an explicit "must be stepping"
guard to begin with; since the CPU-thread queue is now drained unconditionally
at the top of every Core_RunLoopUntil() iteration (not just while stepping),
queuing from it is safe regardless of current core state.

Also corrects a stale comment on Core_RunOnCPUThread() left over from before
the drain point moved from Core_ProcessStepping() to the top of
Core_RunLoopUntil() - it's not limited to the stepping/paused case.

Replaced remaining `auto` locals in this file with concrete types.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hqm11k99viLfbJm2MkH4BH
2026-08-08 11:32:01 +02:00
Henrik RydgårdandClaude Sonnet 5 c6fccefa49 Debugger: Route cpu.stepInto's stepping state through the CPU thread
The WebSocket debugger's cpu.stepInto handler ran entirely on the WebSocket
handler thread, directly manipulating breakpoints and stepping state (via
Core_RequestCPUStep, g_breakpoints.SetSkipFirst, etc.) that's otherwise only
ever touched from the CPU thread (the one that calls Core_RunLoopUntil, and
thus indirectly NativeFrame).

Adds Core_RunOnCPUThread() - queues a function to run on the CPU thread and
blocks the caller until it's done. The queue is drained at the top of
Core_RunLoopUntil()'s loop, so it's reached continuously (in a tight spin)
while the CPU is stepping/paused, and at least once per call even while fully
running.

cpu.stepInto is the first consumer: once the CPU is already stepping, the
breakpoint/stepping manipulation is now routed through Core_RunOnCPUThread
instead of happening directly on the WebSocket thread. The "not currently
stepping" path still calls Core_Break() directly from the WebSocket thread,
since it's already documented free-threaded and is what makes the CPU thread
start reaching the queue-drain point in the first place.

More WebSocket debugger commands can be converted the same way going forward.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hqm11k99viLfbJm2MkH4BH
2026-08-08 11:21:46 +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 0ed1f3eceb Add some easy ways to get into the web debugger 2026-08-08 10:13:09 +02:00
Henrik Rydgård d87a98c4a8 Avoid wasting CPU on counting characters in the memblock management
Unittest and a bugfix by claude.
2026-08-07 08:56:41 +02:00
Henrik RydgårdandClaude Opus 5 074c8ac523 Add memory.search and hle.data.* to the WebSocket debugger
Reverse-engineering workflows need to (1) find where an unknown value lives
in memory and (2) label what's found, neither of which the debugger API
could do before:

- memory.search (MemorySubscriber.cpp): Cheat-Engine-style scan of a memory
  range for a u8/u16/u32/float value, or a byte pattern with an optional
  wildcard mask.
- hle.data.list/add/remove/rename (HLESubscriber.cpp): manage ST_DATA
  symbols (structs, tables, buffers), mirroring the existing hle.func.*
  commands for functions. Needed a new SymbolMap::RemoveData, since only
  RemoveFunction existed - added following the same pattern.

Verified live against a running PPSSPP instance (game.status, cpu.stepping,
memory.search in u32/bytes/masked-bytes modes, and the full
add/list/rename/remove data-symbol lifecycle) via Tools/wsdbg.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XDNwPPuidmNxQGRJxBuRL6
2026-07-26 21:17:09 +02:00
Henrik Rydgård 5c340ccd09 Remove the concept of scaled/offset depth buffers. 2026-06-10 10:41:30 +02:00
Henrik Rydgård a317890c08 Remove the GPUDebugInterface class
Just a pointless extra layer in the class hiearchy, making it
unnecessarily hard to modify the interface.

Might as well hit GPUCommon directly.
2026-06-02 11:15:08 +02:00
Henrik Rydgård f60e27a9b7 Just some refactoring of the GPUStatistics struct, and more use of StringWriter 2026-05-29 14:40:31 +02:00
Henrik Rydgård 8dad5a5c57 Remove asserts in GPUBufferSubscriber
Fixes #21683
2026-05-19 11:09:53 +02:00
Henrik Rydgård 1f4d3b3b04 Minor code cleanup 2026-05-10 21:46:13 +02:00
Henrik Rydgård 7d9539a461 Fix multiple issues with temporary breakpoints. Prevents unnecessary jit clears on step-out. Thanks Nemoumbra. 2026-05-10 21:33:54 +02:00
Nemoumbra 1297fdde1f Fix: vector overrun in GPU stats 2026-05-10 05:38:55 +03:00
Henrik Rydgård 5aeb38412d Fix another valgrind warning 2026-02-19 11:46:14 +01:00
Henrik Rydgård d6ebfed432 Fix some bugs and warnings from a pass of static analysis 2026-01-29 17:41:54 +01:00
Henrik Rydgård 3a7911407b Assorted minor cleanups 2026-01-19 15:47:40 +01:00
刘皓 295ac0e9f6 Merge branch 'master' into libretro-vfs 2026-01-02 12:58:27 -05:00
刘皓 5716cbd41d Use the libretro VFS interface in libretro builds 2026-01-01 00:24:01 -05:00
Henrik Rydgård 5f7a937466 Rename ValidSize to ClampValidSizeAt 2025-12-30 20:31:07 +01:00
Henrik Rydgård c20be71c10 Fix the API for MIPSAssembleOpcode to remove a global 2025-11-03 16:08:07 +01:00
Henrik Rydgård b304fbe0c9 Build armips in libretro builds (requirement for some lua functionality) 2025-10-08 09:45:44 -06:00
Silvris 6d5bf32796 initial tests 2025-09-19 02:40:13 -05:00
Henrik Rydgård 33b87578c5 Revive old alignment utility functions, give them better names. 2025-09-03 20:48:51 +02:00
Henrik Rydgård f2df336f79 Resolve #20746 by checking addresses harshly in MIPSAnalyst
By fixing up badly aligned addresses in HLESubscriber.cpp.

This should help eliminate any bad usage within PPSSPP itself, while
also keeping existing websocket code working.

Additionally, this makes some end addresses exclusive instead of
inclusive, which simplifies address math.
2025-09-03 20:32:19 +02:00
Henrik Rydgård 5751e202a5 Remove redundant .c_str() 2025-08-31 13:37:43 +02:00
Henrik Rydgård 3b3257ad99 Fix crash in debug builds in gpu.stats.get 2025-08-26 10:24:30 +02:00
Henrik Rydgård a7b36ae0ae Websocket: Request resets asynchronously. sync won't work. 2025-08-26 10:12:43 +02:00
Henrik Rydgård 18ccef1bd4 Just some minor code modernization 2025-08-25 10:45:12 +02:00
Henrik Rydgård d9b7e370f3 ImDebugger: Add initial implementation of a watch window 2025-08-15 21:58:32 +02:00
Nemoumbra 846c265972 Fixed the docs + better wording 2025-08-10 17:44:38 +03:00
Henrik Rydgård cc3fbb22e4 Change some stat arrays from double to float (to save mem bandwidth) 2025-06-18 11:00:27 +02:00
Henrik Rydgård ff37cbe184 Minor memory safety fixes 2025-05-14 09:39:09 +02:00
Henrik Rydgård b24ce4bd14 Simplifications 2025-04-05 09:18:56 +02:00
Henrik Rydgård 2bfe327dbd Expose PSPThread in the same manner 2025-03-31 10:24:03 +02:00
Henrik Rydgård 644f5e4e6c Expose PSPModule (so the debugger can access it later) 2025-03-31 09:56:08 +02:00
Henrik Rydgård 088a02bfdb Cleanup a lot of the bootup state management. 2025-03-30 14:02:29 +02:00
Henrik Rydgård c1828f69b9 Add an option to compress .ppsym files when saving or not (was always on before). 2025-03-29 18:17:41 +01:00
Henrik Rydgård 1a93d8a1f9 Some work on symbol export, minor cleanups 2025-03-29 14:23:59 +01:00
Henrik Rydgård acad90a041 More work on the lua console 2025-03-27 14:26:48 +01:00
Henrik Rydgård 4989ec61d9 BreakpointManager: Safer and simpler updates. 2025-03-03 13:59:42 +01:00
oltolm 02e767866a fix compiler warnings 2025-02-22 14:15:15 +01:00
Henrik Rydgård 31cf5771f4 Turn the break reason into an enum, fix some minor issues 2025-02-19 16:01:11 -06:00
oltolm 9566065455 make compile on mingw-w64 2025-01-11 01:25:27 +01:00
Henrik Rydgård 23954f5403 Implement the disassembly mode too 2025-01-09 13:35:29 +01:00
Henrik Rydgård c91169e702 Restore removed <algorithm> includes.
Turns out these were needed after all. For some reason, on Windows and
Mac, <algorithm> gets auto-included by something else so I don't notice
when it's missing, and MSVC's include dependency tracker doesn't see it
either.
2024-12-19 09:53:07 +01:00
Henrik Rydgård 4dd3621fa0 Remove some unnecessary <algorithm> includes. 2024-12-18 17:04:27 +01:00
Henrik Rydgård 45cf6d423e Buildfixes 2024-12-18 13:57:29 +01:00
Henrik Rydgård 3e198c53b2 More include cleanup 2024-12-18 13:57:26 +01:00
Henrik Rydgård 96cab3f267 Rework the log manager to be more robust 2024-12-16 12:33:58 +01:00
Henrik Rydgård 638607d29a Refactor: Make GPUCommon own the framedump "recorder". 2024-12-15 12:07:44 +01:00