SEGFAULT on WSL for simple binary that doesn't SEGFAULT on std. Ubuntu or with valgrind #2546
Comments
- Adding mutable global rather than use a closure for `vec_size`. - Manually granting libraries ability to execute on stack in WSL. See: microsoft/WSL#286 https://stackoverflow.com/q/39136040/1068170 microsoft/WSL#916 https://gist.github.com/dhermes/f448415b9160b785f503a7361fe40d51 microsoft/WSL#2546
- Adding mutable global rather than use a closure for `vec_size`. - Manually granting libraries ability to execute on stack in WSL. See: microsoft/WSL#286 https://stackoverflow.com/q/39136040/1068170 microsoft/WSL#916 https://gist.github.com/dhermes/f448415b9160b785f503a7361fe40d51 microsoft/WSL#2546
dhermes
changed the title
|
@therealkenc Has this been resolved? I noticed the |
|
Ah. In 2017 this likely got skipped because the OP did not have any CLI steps visible, and with no repro visible the likelihood of someone clicking through a link was "small". The closed status was just cleaning up because I did a |
See flang-compiler/flang#298 As of this, the build fails for a different reason: ``` F90-S-0439-An internal subprogram cannot be passed as argument - vec_size (.../bezier/src/fortran/curve.f90: 837) 0 inform, 0 warnings, 1 severes, 0 fatal for compute_length ``` Support for passing the function closure (i.e. the "internal subprogram") has been problematic before when using `gfortran` on Windows Subsystem for Linux (WSL): microsoft/WSL#2546
|
Can it be that the issue is with the (lack of support for) executable stack that is used by gfrortran to pass inner functions (closures) as parameters? gfortran creates a trampoline to call inner function when it is passed as an argument. The trampoline instructions are placed on the stack. Hence, they need executable stack. |
|
Plausible. |
|
@therealkenc Would you like me to generate assembly and print it out? (I can do it in WSL vs. a "real" Ubuntu install to compare?) |
|
What could be done is check whether the use case runs on 19041 WSL2. If it works the issue is sort of destined for fixed-in-wsl2 tag. In WSL1 |
|
The repro does take on WSL2. This one is for all intents dupe #2553.
|
|
Awesome! Thanks. (I am just about to buy a new PC and I think I won't even make an Ubuntu partition on it because of the richness of the WSL experience.) |
|
This issue has been automatically closed since it has not had any activity for the past year. If you're still experiencing this issue please re-file this as a new issue or feature request. Thank you! |
>ver ... Microsoft Windows [Version 10.0.15063]gfortran(see gist). The Fortran uses a closure and it seems there is a bad reference to the procedure with the closure (or to the enclosed reference).valgrind/ what should be happening instead: it should not SEGFAULT (it doesn't on "normal" Ubuntu)Context: This is a minimal reproduction from a more involved error experienced when calling into QUADPACK for a numerical library.
The text was updated successfully, but these errors were encountered: