_NoLogging methods in VM

[jasper-d]

Since #68717 there are a bunch of Xyz and Xyz_NoLogging method pairs in VM code.
In most cases Xyz simply calls the _NoLogging variant or has the same implementation, e.g.

runtime/src/coreclr/vm/methodtable.h

Lines 879 to 890 in f5ca995

	inline BOOL IsRestored_NoLogging()
	{
	LIMITED_METHOD_DAC_CONTRACT;
	return !(GetWriteableData_NoLogging()->m_dwFlags & MethodTableWriteableData::enum_flag_Unrestored);
	}
	inline BOOL IsRestored()
	{
	LIMITED_METHOD_DAC_CONTRACT;
	return IsRestored_NoLogging();
	}

I think the only methods that differ in behaviour are MethodTable::GetClass_NoLogging/MethodTable::GetClass.

Could these _NoLogging variants be removed or are they kept for some reason?

[AaronRobinsonMSFT]

Could these _NoLogging variants be removed or are they kept for some reason?

These are an historical artifact for scenarios involving collection of data that were reused for performance related tooling. The "_NoLogging" APIs were lighter-weight in the sense they didn't record accesses as opposed to the "logging" versions that did. All/most of that infrastructure was removed but we kept the "_NoLogging" in case we find a use case for this sort of mechanism again.

I personally would be fine removing them since we are going to add new APIs that probably would need a log/no-log pair but since we don't measure or have an active scenario they won't get written. Which means we are going to need to audit everything again anyways.

@davidwrighton and @jkotas Do either of you have any push back on removing these?

[jkotas]

Do either of you have any push back on removing these?

Sounds good to me.

[AaronRobinsonMSFT]

@jasper-d I look forward to your PR :)