radv dynamic rendering is broken after af2a96bb
Description
I'm currently trying to implement support for KHR_dynamic_rendering in Wine. I have an implementation that passes validation, works on stock Mesa 23.1.7, and work with git Mesa with lavapipe, but fails with radv. I was able to bisect the failure to af2a96bb.
I'm not sure of the exact symptoms, but it seems to affect almost all rendering. For the one test I pulled out at random, we do a clear and then a draw, and in that case the framebuffer contains only the clear color, as if the draw never happened.
Steps to reproduce
This would need a custom build of Wine including my local patches, which would be an involved process. If this isn't going to be easily debuggable from a capture, I can try to provide that.
System information
whatsit@camazotz:~/git/wine64/dlls/d3d10core/tests$ inxi -GSC -xx
System:
Host: camazotz Kernel: 6.4.0-4-amd64 arch: x86_64 bits: 64 compiler: gcc
v: 13.2.0 Desktop: KDE Plasma v: 5.27.7 tk: Qt v: 5.15.10 wm: kwin_x11
dm: SDDM Distro: Debian GNU/Linux trixie/sid
CPU:
Info: 8-core model: AMD Ryzen 7 5800X bits: 64 type: MT MCP arch: Zen 3+
rev: 0 cache: L1: 512 KiB L2: 4 MiB L3: 32 MiB
Speed (MHz): avg: 3351 high: 4044 min/max: 2200/4850 boost: enabled cores:
1: 3599 2: 4044 3: 2880 4: 2880 5: 3181 6: 3665 7: 2200 8: 3417 9: 3697
10: 3736 11: 3517 12: 3636 13: 3850 14: 3629 15: 2200 16: 3493
bogomips: 121602
Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm
Graphics:
Device-1: AMD Navi 22 [Radeon RX 6700/6700 XT/6750 XT / 6800M/6850M XT]
vendor: Micro-Star MSI driver: amdgpu v: kernel arch: RDNA-2 pcie:
speed: 16 GT/s lanes: 16 ports: active: HDMI-A-1 empty: DP-1,DP-2,DP-3
bus-ID: 2d:00.0 chip-ID: 1002:73df
Display: x11 server: X.Org v: 1.21.1.8 with: Xwayland v: 23.2.0
compositor: kwin_x11 driver: X: loaded: amdgpu
unloaded: fbdev,modesetting,radeon,vesa dri: radeonsi gpu: amdgpu
display-ID: :0 screens: 1
Screen-1: 0 s-res: 1920x1080 s-dpi: 98
Monitor-1: HDMI-A-1 mapped: HDMI-A-0 model: Asus VS278 res: 1920x1080
dpi: 82 diag: 686mm (27")
API: OpenGL v: 4.6 Mesa 23.1.7-1 renderer: AMD Radeon RX 6700 XT (navi22
LLVM 15.0.7 DRM 3.52 6.4.0-4-amd64) direct-render: Yes
If applicable
- Xserver version: 1.21.1.8
- Wine/Proton version: Custom build of Wine.
Regression
Yes. I bisected this to af2a96bb.
API captures (if applicable, optional)
Further information (optional)
Does the issue reproduce with the LLVM backend (RADV_DEBUG=llvm
) or on the AMDGPU-PRO drivers?
- It does reproduce with the LLVM backend. I have not tested with amdgpu-pro, since that'll be a relatively involved process and I suspect it won't be an interesting test here, but please let me know if I should do that anyway.
Does your environment set any of the variables ACO_DEBUG
, RADV_DEBUG
, and RADV_PERFTEST
?
- No.