Is there any practical difference between an extension method on <T> or on Object?

Is there any practical difference between these two extension methods?

class Extensions
{
    public static void Foo<T>(this T obj) where T : class { ... }
    public static 开发者_运维知识库void Foo(this object obj) { ... }
}

I was poking around in Extension Overflow and I came across the first form, which I haven't used before. Curious what the difference is.

Extension methods on Object will also apply to value types. (And they'll be boxed by the call, reducing performance)

Extension methods on <T> but without where T : class will also work on value types, but will not box them.

In addition, extension methods on <T> can write typeof(T) to get the compile-time type of their invocation.
If you do that, note the difference between

someButton.Extension();
someButton.Extension<Control>();
someButton.Extension<Object>();