#4683

Crash on start: thread 'main' panicked attempt to multiply with overflow (ppc)

Openpeppergrayxyzopened 18d ago

building on 32-bit powerpc/musl fails with error:

2026-10-12T18:55:23.619313Z  INFO niri: starting version 26.04 (v26.04-187-g90af637e-modified)
2026-10-12T18:55:23.710813Z DEBUG niri_config: loaded config from "/tmp/niri-minimal.kdl"
2026-10-12T18:55:24.514947Z  INFO niri::backend::tty: using as the render node: "/dev/dri/renderD128"
2026-10-12T18:55:25.353594Z DEBUG niri::backend::tty: adding device: 57856 "/dev/dri/card0"
2026-10-12T18:55:25.355824Z DEBUG niri::backend::tty: this is the primary node
2026-10-12T18:55:25.881912Z DEBUG niri::backend::tty: got render node: renderD128
2026-10-12T18:55:25.882543Z DEBUG niri::backend::tty: initializing the primary renderer
r300 FP: Compiler error: Too many ALU instructions used: 241, max: 64
r300 FP: Use RADEON_DEBUG=dummysh to force dummy shader instead.
2026-10-12T18:55:27.539724Z  WARN niri::render_helpers::shaders: error compiling border shader: ProgramLinkError
r300 FP: Compiler error: Too many ALU instructions used: 161, max: 64
r300 FP: Use RADEON_DEBUG=dummysh to force dummy shader instead.
2026-10-12T18:55:27.949443Z  WARN niri::render_helpers::shaders: error compiling shadow shader: ProgramLinkError
2026-10-12T18:55:28.951596Z DEBUG niri::backend::tty: device changed: 57856
2026-10-12T18:55:29.028583Z DEBUG niri::backend::tty: new connector: LVDS-1 "Apple Computer Inc Color LCD Unknown"
2026-10-12T18:55:29.030043Z DEBUG niri::backend::tty: couldn't find connector CRTC_ID property
2026-10-12T18:55:29.030783Z DEBUG niri::backend::tty: couldn't find connector CRTC_ID property
2026-10-12T18:55:29.033999Z DEBUG niri::backend::tty: connecting connector: LVDS-1
2026-10-12T18:55:29.036826Z DEBUG niri::backend::tty: picking mode: Mode { name: "1440x960", clock: 102960, size: (1440, 960), hsync: (1504, 1536, 1760), vsync: (964, 968, 976), hskew: 0, vscan: 0, vrefresh: 60, mode_type: ModeTypeFlags(PREFERRED | DRIVER) }
2026-10-12T18:55:29.044006Z DEBUG niri::backend::tty: failed to set HDR properties: couldn't find property: "HDR_OUTPUT_METADATA"
2026-10-12T18:55:30.262222Z DEBUG niri::niri: putting output LVDS-1 at x=0 y=0

thread 'main' (4211) panicked at vendor/smithay/src/utils/clock.rs:90:9:
attempt to multiply with overflow
stack backtrace:
   0: __rustc::rust_begin_unwind
   1: core::panicking::panic_fmt
   2: core::panicking::panic_const::panic_const_mul_overflow
   3: smithay::utils::clock::Time<smithay::utils::clock::Monotonic>::as_micros
             at niri/vendor/smithay/src/utils/clock.rs:90:9
   4: smithay::utils::clock::Time<smithay::utils::clock::Monotonic>::as_millis
             at niri/vendor/smithay/src/utils/clock.rs:79:15
   5: smithay::backend::input::InputTime::now
             at niri/vendor/smithay/src/backend/input/mod.rs:33:59
   6: niri::niri::State::move_cursor
             at niri/src/niri.rs:995:23
   7: niri::niri::State::move_cursor_to_output
             at niri/src/niri.rs:1237:14
   8: niri::niri::State::focus_default_monitor
             at niri/src/niri.rs:1092:14
   9: niri::niri::State::new
             at niri/src/niri.rs:782:15
  10: niri::main
             at niri/src/main.rs:186:21
  11: core::ops::function::FnOnce::call_once
             at /usr/lib/rust/1.96.1/lib/rustlib/src/rust/library/core/src/ops/function.rs:250:5
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

due to a bug in rustix: https://github.com/bytecodealliance/rustix/issues/1701 that causes an overflow im Smithay: https://github.com/Smithay/smithay/issues/2211

System Information

Checklist

  • I searched existing issues and discussions and this issue was not already reported.
  • I did not use an LLM to write this issue.

Activity

2
YaLTeR17d ago

Well I guess we're waiting on the aforementioned projects then

peppergrayxyz17d ago

Unfortunately. I thought I leave it here to have it documented, but please let me know if I should avoid creating not-niri type issues.