bpo-45953: Statically allocate the main interpreter (and initial thread state). #29883
Conversation
Python/ceval.c
Outdated
| int | ||
| _PyEval_InitState(struct _ceval_state *ceval) | ||
| void | ||
| _PyEval_InitState(struct _ceval_state *ceval, PyThread_type_lock pending_lock) |
There was a problem hiding this comment.
Would this be a good opportunity to remove this function?
The _pending_calls struct belongs on the interpreter state, not on the ceval as only the main interpreter can handle them.
There was a problem hiding this comment.
There's enough going on in this PR that I'd rather not.
|
When you're done making the requested changes, leave the comment: |
Include/internal/pycore_interp.h
Outdated
| }; | ||
|
|
||
| #define _PyInterpreterState_INIT \ |
There was a problem hiding this comment.
This is only used once, please remove it.
There was a problem hiding this comment.
FYI, I brought this back to avoid so much nesting and clutter in _PyRuntimeState_INIT.
| // because on Windows we only get a pointer type. | ||
| // All other pointers in PyThreadState are either | ||
| // populated lazily or change but default to NULL. | ||
| } _preallocated; | ||
| } _PyRuntimeState; | ||
|
|
||
| #define _PyRuntimeState_INIT \ |
There was a problem hiding this comment.
Might as well remove this as well. It is also only used once.
There was a problem hiding this comment.
Should _Py_global_objects_INIT also move? It's pretty big and likely to get much bigger.
There was a problem hiding this comment.
Actually, _PyRuntimeState_INIT is used twice.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
|
With basically all feedback addressed I plan on merging this soon, to unblock other changes I have waiting. If I missed something, I'd be glad to address it in a follow-up PR. |
|
Fine with me.
…On Wed, Jan 12, 2022 at 2:56 PM Eric Snow ***@***.***> wrote:
With basically all feedback addressed I plan on merging this soon, to
unblock other changes I have waiting. If I missed something, I'd be glad to
address it in a follow-up PR.
—
Reply to this email directly, view it on GitHub
<#29883 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAWCWMXNEYDUC6YIRLTJAEDUVYBIXANCNFSM5JFJMYAA>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
--
--Guido van Rossum (python.org/~guido)
|
As of CPython 3.11 (via python/cpython#29883) stdbool.h is now included in Python.h so do attempt to redefine bool/true/false.
As of CPython 3.11 (via python/cpython#29883) stdbool.h is now included in Python.h so do attempt to redefine bool/true/false.
Currently the main interpreter is allocated on the heap during runtime initialization. Here we are instead embedding it into
_PyRuntimeState, which means it is statically allocated as part of the_PyRuntimeglobal. The same goes for the initial thread state (of each interpreter, including the main one). Consequently there are fewer allocations during runtime/interpreter init, fewer possible failures, and better memory locality.FYI, this also helps efforts to consolidate globals, which in turns helps work on subinterpreter isolation.
https://bugs.python.org/issue45953