[issue21131] test_faulthandler.test_register_chain fails on 64bit ppc/arm with kernel >= 3.10

[STINNER Victor]

STINNER Victor <vstinner at redhat.com> added the comment:

> Its interesting to read the comment on the IA64 definition for SIGSTKSZ:

Python doesn't support IA64 architecture.

Note: I'm not sure that this architecture is going to survive in the long term... More and more systems stopped to support it.


> Well, one argument for the dynamic approach is that existing python
binaries can adjust without needing to be respun for new CPUs.

PR 13649 gets the default thread stack size, it doesn't get SIGSTKSZ. I expect a thread stack size to be way larger than SIGSTKSZ.

For on my Fedora 30 (Linux kernel 5.1.19-300.fc30.x86_64, glibc-2.29-15.fc30.x86_64) on x86-64, SIGSTKSZ is just 8 KiB (8192 bytes) whereas pthread_attr_getstacksize() returns 8 MiB (8388608 bytes).

As I already wrote, using 8 MiB instead of 8 KiB looks like a waste of memory: faulthandler is only useful for stack overflow, which is a very unlikely bug.

----------

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue21131>
_______________________________________