This adds an error message if the isSVESizelessBuiltinType like __SVFloat32_t / __SVInt64_t / etc, which provide the backing for the svfloat32_t / svint64_t / etc ACLE types, are used in a function without SVE. The alternative is a crash in the backend, which is not capable of handling scalable vector types.
When SVE is available, either through a -march=..+sve option or via a target("sve") attribute, nothing should change. Without the sve feature, this patch gives an error for any function arguments, return values and variable declarations involving the scalable types. Struct/class members and global variables already give an error. (Let me know if there is anything else that should give an error. This might not be all cases, but should rule out the most obvious.) As this can be based on the current function target attributes, the error sometimes needs to be handled later than would otherwise if it was just based on the global target.
This also adds a basic initFeatureMap method for AArch64 to include a few transient sve target dependencies. This should ideally be based on the existing map of features already present in the backend, not reproduced in the target parser.
VLSTBuiltinType is synonymous with Vector Length Specific BuiltinType, from the docs:
Additionally, this function doesn't include the tuple types (e.g. svint32x2_t), so if you want this to work for SVE types in general you'll need to create a new method for this.