Segmentation fault although no optimization is done
wackazong opened this issue · comments
wackazong commented
I have a wasm file which causes a segementation fault in wasm-opt, even if not using any optimization. What can I do to resolve this?
writeInt8: 0 (at 9855286)
writeInt8: 3 (at 9855287)
writeInt8: 0 (at 9855288)
writeInt8: 0 (at 9855289)
writeInt8: 0 (at 9855290)
writeInt8: 0 (at 9855291)
writeInt8: 0 (at 9855292)
writeInt8: 0 (at 9855293)
writeInt8: 0 (at 9855294)
writeInt8: 0 (at 9855295)
writeInt8: 0 (at 9855296)
writeInt8: 0 (at 9855297)
writeInt8: 0 (at 9855298)
writeInt8: 0 (at 9855299)
writeInt8: 0 (at 9855300)
writeInt8: 0 (at 9855301)
writeInt8: 0 (at 9855302)
writeInt8: 0 (at 9855303)
writeInt8: 0 (at 9855304)
writeInt8: 0 (at 9855305)
writeInt8: 0 (at 9855306)
writeInt8: 0 (at 9855307)
writeInt8: 0 (at 9855308)
writeInt8: 0 (at 9855309)
writeInt8: 0 (at 9855310)
writeInt8: 2 (at 9855311)
writeInt8: 0 (at 9855312)
writeInt8: 0 (at 9855313)
writeInt8: 0 (at 9855314)
writeAtU32LEB: 8373064 (at 1482246)
[1] 48856 segmentation fault wasm-opt App_bg.wasm -o test.wasm --debug
Thomas Lively commented
Are you able to share the wasm file so we can investigate and fix the bug?
wackazong commented
Also, wasm-gc gives me this error:
thread 'main' panicked at /.../wasm-gc-0.1.6/src/main.rs:45:36:
failed to parse wasm module: Error(UnknownOpcode(195))
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
[1] 49797 abort wasm-gc app.wasm
Alon Zakai commented
Stack trace shows it crashes in DWARF support code. Our DWARF support is not very robust, and not the focus of current work. To work around this issue, pass --strip-dwarf
(which will remove it before writing).