iris: regression in yuzu
System information
System: Host: Qykopi Kernel: 5.12.14-arch1-2-vfio x86_64 bits: 64 compiler: gcc v: 11.1.0
Desktop: Xfce 4.16.0 tk: Gtk 3.24.29 wm: xfwm4 dm: LightDM Distro: Arch Linux
CPU: Info: 6-Core model: 11th Gen Intel Core i5-11600K bits: 64 type: MT MCP arch: N/A
rev: 1 cache: L2: 12 MiB
flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 93924
Speed: 4800 MHz min/max: 800/5000 MHz Core speeds (MHz): 1: 4800 2: 4800 3: 4800
4: 4800 5: 4800 6: 4800 7: 4800 8: 4800 9: 4800 10: 4800 11: 4800 12: 4800
Graphics: Device-1: Intel RocketLake-S GT1 [UHD Graphics 750] vendor: ASUSTeK driver: i915
v: kernel bus-ID: 00:02.0 chip-ID: 8086:4c8a
Device-2: AMD Tonga PRO [Radeon R9 285/380] vendor: XFX Pine driver: amdgpu v: kernel
bus-ID: 01:00.0 chip-ID: 1002:6939
Device-3: NVIDIA GM200 [GeForce GTX 980 Ti] vendor: ZOTAC driver: nvidia v: 470.42.01
bus-ID: 04:00.0 chip-ID: 10de:17c8
Display: x11 server: X.Org 1.20.11 compositor: picom driver:
loaded: amdgpu,modesetting,nvidia alternate: ati,fbdev,intel,nouveau,nv,vesa
resolution: 1: 1920x1200~60Hz 2: 2560x1440 s-dpi: 96
OpenGL: renderer: Mesa Intel Graphics (RKL GT1) v: 4.6 Mesa 21.1.3 direct render: Yes
Describe the issue
While playing The Legend of Zelda: Breath of the Wild in the yuzu emulator, terrain may visibly sink or rise, although interacting with the glitched terrain does not affect gameplay. Generally the issue happens every time the game is run. How exactly the terrain changes is not consistent between runs. This happens within the first few moments of gameplay.
Regression
According to git bisect, caused by c85ea824
Screenshots/video files
mesa-21.1.3: 2021-07-05_21-34-06
Latest Mesa: 2021-07-05_21-35-46
Note how the mountain behind the Shrine has a hole, the hill is recessed until walking closer to it, and the somewhat nearby terrain rises into the sky.