Dynamic generic dictionary expansion feature

[fadimounir]

TODO:

• Update JITInterface GUID in corinfo.h after JIT: Optimize "constant_string".Length #1378 (comment)

These changes introduce dynamic size expansion for generic dictionary layouts when we run out of slots.
The original implementation allowed for an expansion, but using a linked list structure, which made it
impossible to use fast lookup slots once we're out of slots in the first bucket.

This new implementation allows for the usage of fast lookup slots always, for all generic lookups.

This also removes the constraint we had on R2R, where we disabled the usage of fast slots all-together.

NOTE:
This is a much simpler implementation than the original commit which got reverted due to failures related to collectible assemblies (dotnet/coreclr#26262). The difference between these changes and the original ones is that dictionary expansion for types and methods is now done lazily, on-demand, whenever we attempt to read a value beyond the size of a given dictionary. This makes uses of the newly added JIT support from #31957. The bug with collectible assemblies have been also fixed, and I added a test to cover various interesting aspects of the feature.

BONUS:
Part of these changes is some cleanup work to really simplify the R2R dictionary access stubs. These stubs are now much more compact with fewer instructions and better registry usage (many #ifdef UNIX_AMD64_ABI blocks removed from the amd64 stub)

cc @dotnet/crossgen-contrib

[jkotas]

Overall, I like this much better than the original version. Thanks!

[fadimounir]

Couple of interesting benchmark numbers I tried:

• Roslyn compiler, compiling Roslyn source code: 1~2% performance gain.
• NewtonSoft serialization benchmark (found something on the web): 4% performance gain.

[jkotas]

Thanks!