addr2line.py doesn't show correct location information about wasm-micro-runtime

[lum1n0us]

#3199

Before running this script, please make sure:
  - the wasm file is compiled with debug info. (like: clang -g) 
  - the call-stack dump is generated by iwasm
  - iwasm is compiled with -DWAMR_BUILD_DUMP_CALL_STACK=1
  - iwasm isn't running under fast-interp mode. -DWAMR_FAST_INTERP=0
  - if using .aot, the aot file is generated with `--enable-dump-call-stack`

[lum1n0us]

Nice catch. It is a bug of addr2line.py.

[lum1n0us]

#3194

[eloparco]

Hi @lum1n0us, thanks for the quick reply! I tried the fix in #3194, now I get the backtrace but it's not correct:

#00: 0x004a - $f3
#01: 0x00a9 - $f4 (FILE:trap.c LINE:   10 COLUMN:  7 FUNC:b)
#02: 0x0070 - $f5 (FILE:trap.c LINE:    3 COLUMN:  0 FUNC:c)
#03: 0x0091 - $f6 (FILE:trap.c LINE:    9 COLUMN:  0 FUNC:b)
#04: 0x0088 - _start (FILE:trap.c LINE:    9 COLUMN:  0 FUNC:b)

I was comparing the script to https://github.com/bytecodealliance/wasm-tools/blob/main/src/bin/wasm-tools/addr2line.rs, but couldn't find what the problem is.

Also, looking at #3194, I guess the addresses printed by iwasm are already relative to the code section, is that why the PR removes the code section subtraction?

[eloparco]

Thanks for the fix, closing the issue.
I have some changes that I want to add to the script, I'll raise a PR once #3199 is merged