-
-
Notifications
You must be signed in to change notification settings - Fork 30.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
gh-83004: Clean up refleaks in ssl, socket initialisation #99044
Conversation
hauntsaninja
commented
Nov 3, 2022
•
edited by bedevere-bot
Loading
edited by bedevere-bot
- Issue: Improve stdlib module initialization error handling. #83004
@@ -5925,8 +5925,7 @@ sslmodule_init_constants(PyObject *m) | |||
#define addbool(m, key, value) \ | |||
do { \ | |||
PyObject *bool_obj = (value) ? Py_True : Py_False; \ | |||
Py_INCREF(bool_obj); \ | |||
PyModule_AddObject((m), (key), bool_obj); \ | |||
PyModule_AddObjectRef((m), (key), bool_obj); \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Isn't this still a leak because we're not checking the return value?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's no longer a leak.
The thing that makes PyModule_AddObject
tricky cause leaks is that you have to decref in the failure case, but not in the success case (because it steals the ref). With PyModule_AddObjectRef
you can treat the failure and success cases the same in terms of what they do to ref counts.
But yes, in general we should probably be propagating failures here, and the other cases below you commented on. I didn't do it in this PR because it's a slightly separate change and there are many other places in this module where we don't check return values.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should I check return values for just the ones I'm touching in this PR or fix all of them?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the explanation! I feel it'd be fine to also fix missing error checks in this PR, since the change remains localized and you're fixing a very similar problem to what the PR already fixes.
return NULL; | ||
|
||
#ifdef ENABLE_IPV6 | ||
has_ipv6 = Py_True; | ||
#else | ||
has_ipv6 = Py_False; | ||
#endif | ||
Py_INCREF(has_ipv6); | ||
PyModule_AddObject(m, "has_ipv6", has_ipv6); | ||
PyModule_AddObjectRef(m, "has_ipv6", has_ipv6); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Check the return value.
@@ -8666,7 +8662,8 @@ PyInit__socket(void) | |||
tmp = PyLong_FromUnsignedLong(codes[i]); | |||
if (tmp == NULL) | |||
return NULL; | |||
PyModule_AddObject(m, names[i], tmp); | |||
PyModule_AddObjectRef(m, names[i], tmp); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Here too.
The socket issues have been resolved in gh-103261. |
Closing, as the issues have been resolved in other PRs. Thanks :) |