-
-
Notifications
You must be signed in to change notification settings - Fork 33.5k
gh-127271: Remove the PyCell_Get usage for framelocalsproxy #130383
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-127271: Remove the PyCell_Get usage for framelocalsproxy #130383
Conversation
Objects/frameobject.c
Outdated
| PyObject *value = framelocalsproxy_getval(frame->f_frame, co, i); | ||
| assert(value != NULL); |
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.
This doesn't look safe to me. I don't think we can rely on consistency between the earlier framelocalsproxy_getkeyindex call and the framelocalsproxy_getval here.
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.
Do you mean a very specific case where the content of the cell on that index is set to NULL by PyCell_Set from another thread? I think that's probably an outlier. If it's a normal fast variable, can it happen? Multi-threading can't affect the local fast variables on stack right?
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.
Yeah, I'm just thinking about the cell case.
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.
The normal Python operation won't set NULL to a cell right? So we should maybe raise some exception here if the value is NULL? That's the only case I can think of.
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.
I'm not sure about the details here, but in general I think we should avoid this sort of pattern where we perform a "check" separately from a "retrieval" and instead should combine the two. For example, framelocalsproxy_getkeyindex could return the local at the index through an optional output argument.
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.
Yeah that makes sense. I can pass in an pointer to hold the key. I'll work on it.
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.
Okay I made framelocalsproxy_getkeyindex do check & retrieve in a single operation. Tweaked a few other places to make it work.
Co-authored-by: Sam Gross <[email protected]>
nascheme
left a comment
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.
LGTM.
|
@colesbury do you have any other suggestions for the current implementation? |
colesbury
left a comment
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.
One suggestion below, but otherwise the changes to use new references instead of borrowed references LGTM.
Objects/frameobject.c
Outdated
| PyObject *extra = frame->f_extra_locals; | ||
| if (extra != NULL) { | ||
| PyObject *value = PyDict_GetItem(extra, key); | ||
| value = PyDict_GetItem(extra, key); |
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.
Use PyDict_GetItemRef if possible?
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.
Done.
In #127272 a
PyCell_Getusage inframeobject.cwas missed from being replaced for free-threaded safety. I guess the reason was probably the function was labelled to return a borrowed reference and a deeper understanding of the piece of code is needed to make that change. (or maybe I'm just thinking too much and it just slipped away :))I made the change to use
PyCell_GetRefand changed the function to return a new reference (because otherwise it won't be safe in free-threaded build). However, there are a few places in the code where it does not care about the actual value stored, it only cares whether such value exists - so I added a util function to avoid havingPy_XDECREFs all around the code.