Use type parameters' source information for kotlinx.serialization diagnostics

To be more consistent with K1, certain diagnostics should be reported
not on the whole properties' types, but on type arguments inside them.

Note that there is still a difference with K2 because K2 reports on a type argument
including its annotations, while K1 used KtTypeReference.typeElement.

IMO, K2 conveys the same or better meaning here, so I am willing to leave this
difference instead of providing PositioningStrategy.

#KT-53861 Fixed

plugins/kotlinx-serialization/testData/diagnostics/NonSerializable.kt

@@ -1,6 +1,6 @@
 // FIR_DISABLE_LAZY_RESOLVE_CHECKS
 // WITH_STDLIB
-// FIR_DIFFERENCE: KT-53861
+// FIR_IDENTICAL
 // FILE: test.kt
 import kotlinx.serialization.*
 
@@ -12,6 +12,9 @@ class Basic(val foo: <!SERIALIZER_NOT_FOUND("NonSerializable")!>NonSerializable<
 @Serializable
 class Inside(val foo: List<<!SERIALIZER_NOT_FOUND("NonSerializable")!>NonSerializable<!>>)
 
+@Serializable
+class Inside2(val foo: List<List<<!SERIALIZER_NOT_FOUND("NonSerializable")!>NonSerializable<!>>>)
+
 @Serializable
 class WithImplicitType {
     <!SERIALIZER_NOT_FOUND("NonSerializable")!>val foo = NonSerializable()<!>