https://bugs.kde.org/show_bug.cgi?id=138424
Alexander Grund <alexander.gr...@tu-dresden.de> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |alexander.grund@tu-dresden.
| |de
Version|3.2.1 |3.25.0
--- Comment #23 from Alexander Grund <alexander.gr...@tu-dresden.de> ---
I just ran into this (again) debugging a SIGSEGV in a Python library. I could
narrow it down to Python itself:
```
$ valgrind python -c 'print(1)'
valgrind: mmap(0x400000, 4096) failed in UME with error 22 (Invalid argument).
valgrind: this can be caused by executables with very large text, data or bss
segments.
```
It also seems to depend on how Valgrind and potentially Python are optimized.
With the system provided Python I can avoid the failure by not building
Valgrind with `-march=native`. I fail to see how that could cause the syscall
to fail though.
However we build (a newer) Python from source and there it fails independent on
the optimization of Python and Valgrind (although it is possible that some
other library used by Python, which is still optimized with -march=native,
causes the issue
The sizes of the Python binary don't seem to be the issue:
```
$ size -x /usr/bin/python3
text data bss dec hex filename
0x75f 0x2bc 0x4 2591 a1f /usr/bin/python3
$ size -x $(which python)
text data bss dec hex filename
0x643 0x280 0x8 2251 8cb
/software/Python/3.12.3-GCCcore-13.3.0/bin/python
```
Any ideas or anything I could test to narrow this down?
--
You are receiving this mail because:
You are watching all bug changes.
