Rpcs3: [Persona 5] Invalid RSX method; Invalid FIFO queue registers

Created on 15 Jul 2017  路  3Comments  路  Source: RPCS3/rpcs3

Game: Persona 5
Renderer: Vulkan

Whilst in dialogue near the start of the game, I encountered:

E {rsx::thread} RSX: Invalid RSX method 0xc080 (arg=0x6041dfa4)
(in file C:\rpcs3\rpcs3\Emu\RSX\rsx_methods.cpp:44)
E {rsx::thread} HLE: Getting the real address failed. (addr=0xc8000000)
E {rsx::thread} RSX: Invalid FIFO queue get/put registers found, get=0xC8000000, put=0x79BD4
S {SPU[0x2000005] Thread (gfdSpursCellSpursKernel5)} SPU: Function detected [0x0e194-0x0e564] (size=0x3d0)

The window's graphics cease to update. Well, admittedly I observed that through some combination of pausing, leaving/entering fullscreen, I managed to get the graphics stuck on the pause menu instead. So maybe not completely stuck.
Music continues playing. If I close the window: RPCS3 does not crash.

system_report.txt
DxDiag.txt
RPCS3.log.zip

OS: Windows 10 Pro (with Creators' Edition update) (10.0, build 15063)
GPU: NVIDIA GeForce GTX 760
GPU driver: NVIDIA 22.21.13.8476 (27/06/2017) (384.76)
CPU: Intel(R) Core(TM) i5-4670K CPU @ 3.40GHz, 3401 Mhz, 4 Core(s), 4 Logical Processor(s)

build: rpcs3-v0.0.2-2017-07-15-25df7676_win64

Used configuration:
Core:
  PPU Decoder: Recompiler (LLVM)
  PPU Threads: 2
  PPU Debug: false
  Save LLVM logs: false
  Use LLVM CPU: ""
  SPU Decoder: Recompiler (ASMJIT)
  Bind SPU threads to secondary cores: false
  Lower SPU thread priority: false
  SPU Debug: false
  Maximum immediate DMA write size: 16384
  Lib Loader: Automatically load required libraries
  Hook static functions: false
  Load libraries:
    []
Video:
  Renderer: Vulkan
  Resolution: 1280x720
  Aspect ratio: 16:9
  Frame limit: Auto
  Write Color Buffers: false
  Write Depth Buffer: false
  Read Color Buffers: false
  Read Depth Buffer: false
  Log shader programs: false
  VSync: false
  Debug output: false
  Debug overlay: false
  Use Legacy OpenGL Buffers: false
  Use GPU texture scaling: true
  Stretch To Display Area: false
  Force High Precision Z buffer: false
  Invalidate Cache Every Frame: false
  Strict Rendering Mode: false
  Batch Instanced Geometry: false
  Vertex Upload Threads: 1
  Enable Frame Skip: false
  Consequtive Frames Drawn: 1
  Consequtive Frames Skept: 1
  D3D12:
    Adapter: NVIDIA GeForce GTX 760
  Vulkan:
    Adapter: GeForce GTX 760

Most helpful comment

That is caused due to inaccuracies in HLE implementation of libgcm.

Use the build that's on the #announcements channel in our Discord server. That contains WIP code that allows for libgcm to be LLE loaded, which solves the issue.

All 3 comments

That is caused due to inaccuracies in HLE implementation of libgcm.

Use the build that's on the #announcements channel in our Discord server. That contains WIP code that allows for libgcm to be LLE loaded, which solves the issue.

I have the same problem...can you please send me a link to that build and if you can can you please talk to ssshadow if he can me red77thunder be back on discord of course ill pay as supporter if i need to i apologize really :)

This is fixed on master since lle-gcm was merged. Closing.

Was this page helpful?
0 / 5 - 0 ratings

Related issues

On1ko picture On1ko  路  3Comments

Xcedf picture Xcedf  路  3Comments

legend800 picture legend800  路  3Comments

LokiGrants picture LokiGrants  路  3Comments

altiereslima picture altiereslima  路  3Comments