[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

STINNER Victor added the comment: The bug should now be fixed in 3.9 and 3.10 branches. The main branch should not be impacted, it has a different implementation. Thanks for the bug report Duncan Grisby. -- resolution: -> fixed stage: patch review -> resolved status: open -> closed

[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

miss-islington added the comment: New changeset 720aef48b558e68c07937f0cc8d62a60f23dcb3d by Miss Islington (bot) in branch '3.9': bpo-9: faulthandler don't modify frame refcnt (GH-27850) https://github.com/python/cpython/commit/720aef48b558e68c07937f0cc8d62a60f23dcb3d --

[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

Change by miss-islington : -- nosy: +miss-islington nosy_count: 2.0 -> 3.0 pull_requests: +26510 pull_request: https://github.com/python/cpython/pull/28066 ___ Python tracker

[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

STINNER Victor added the comment: New changeset fe997e1a67835a929705c8c305d41c4d7dd326e3 by Victor Stinner in branch '3.10': bpo-9: faulthandler don't modify frame refcnt (GH-27850) https://github.com/python/cpython/commit/fe997e1a67835a929705c8c305d41c4d7dd326e3 --

[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

Change by STINNER Victor : -- keywords: +patch pull_requests: +26310 stage: -> patch review pull_request: https://github.com/python/cpython/pull/27850 ___ Python tracker ___

[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

STINNER Victor added the comment: _Py_DumpTracebackThreads() should not use Py_DECREF(). It's a bug. It must only *read* memory, not *modify* memory since it's called from a signal handler. It's a regression in dump_traceback(). Python 3.9 and 3.10 use: frame =

[issue44449] Segfault in _PyTrash_begin when faulthandler tries to dump thread stacks

New submission from Duncan Grisby : I am using Python 3.9.4 on CentOS 7. faulthandler is registered with SIGUSR1: faulthandler.register(signal.SIGUSR1) Sending SIGUSR1 normally correctly dumps the thread stacks, but occasionally it segfaults from the main thread instead: Thread 1 (Thread