gh-125604: Move _Py_AuditHookEntry, etc. Out of pycore_runtime.h

[ericsnowcurrently]

This is essentially a cleanup, moving a handful of API declarations to the header files where they fit best, creating new ones when needed.

We do the following:

• add pycore_debugger_utils.h and move _Py_DebugOffsets, etc. there
• inline struct _getargs_runtime_state and struct _gilstate_runtime_state in _PyRuntimeState
• move struct _reftracer_runtime_state to the existing pycore_object_state.h
• add pycore_audit.h and move to it _Py_AuditHookEntry , _PySys_Audit(), and _PySys_ClearAuditHooks
• add audit.h and cpython/audit.h and move the existing audit-related API there
• move the perfmap/trampoline API from cpython/sysmodule.h to cpython/ceval.h, and remove the now-empty cpython/sysmodule.h

@pablogsal, I'd love to know if you're okay with us moving _Py_DebugOffsets like I have. Would it break anything? Would the separate header be inconvenient for external tools or will it make no difference? Also, do you think the name of the new header is okay?

@zooba, are you okay with us moving all the audit API into the new *audit.h header files like I have?

• Issue: Avoid Standalone Sub-structs In pycore_runtime.h #125604

[ericsnowcurrently]

FYI, I was considering adding in the small GENERATE_DEBUG_SECTION() refactor @1st1 does in gh-124640, putting it in the new pycore_debugger_utils.h, but figured we can do that separately (or keep it in gh-124640.

[pablogsal]

@pablogsal, I'd love to know if you're okay with us moving _Py_DebugOffsets like I have. Would it break anything? Would the separate header be inconvenient for external tools or will it make no difference? Also, do you think the name of the new header is okay?

I am fine with moving the definition to a separate header, this is also more convenient for tools to copy-paste. :)

FYI, I was considering adding in the small GENERATE_DEBUG_SECTION() refactor @1st1 does in gh-124640, putting it in the new pycore_debugger_utils.h, but figured we can do that separately (or keep it in gh-124640.

I was planning to do this independently if #124640 (I did that change in the PR 😉 ) takes time to land so I will build on top of this.

[ericsnowcurrently]

this is also more convenient for tools to copy-paste. :)

I had a gut feeling something like that was the case. :)

[zooba]

Provided the audit machinery doesn't gain accessible entry points (which basically means they all exist in the one source unit, via headers is okay), I don't mind where it lives.

If the functions or structure layouts are discoverable at runtime, it makes it much easier to create a bypass. We call an internal-only version of the function so that it's harder to shim it. Just don't make that any more accessible.

[ericsnowcurrently]

I'll merge this tomorrow, unless there are any objections. If this is blocking anyone, feel free to merge it before I get to it.

[vstinner]

I suggest to rename the file pycore_debug_offsets.h instead. "debug utils" is vague.

[ericsnowcurrently]

I was going to point out that there are other debugger-related things that would go there, but your suggestion looks very correct to me. 😄 Thanks! I'll change it.

[ericsnowcurrently]

done

[vstinner]

LGTM