[SOL] [lldb] sbf / sbpf naming confusion

[procdump]

The problem

When receving the target.xml from a gdb-remote it typically contains the architecture name the remote represents. This name is later used for finding the corresponding CoreDefiniton so that lldb would be able to set up the target accordingly:

llvm-project/lldb/source/Utility/ArchSpec.cpp

Line 1093 in e9617fb

const CoreDefinition *core_def = FindCoreDefinition(arch_name);

SBFv0-v4's CoreDefinitons are given name strings starting with sbfv*. Yet when invoking the parseBPFArch function it's the sbpfv* names that are compared against the provided ArchName in the attempt to return the most appropriate Triple:

llvm-project/llvm/lib/TargetParser/Triple.cpp

Line 421 in e9617fb

ArchName == "sbpfv0" || ArchName == "sbpfv1" ||

Unittests for v0-v4 triples are only written for the sbpfv* cases:

llvm-project/llvm/unittests/TargetParser/TripleTest.cpp

Line 773 in e9617fb

T = Triple("sbpfv1-solana-solana");

To make matters worse if remote's target.xml contains, for example, <architecture>sbfv1</architecture> this later breaks lldb when trying to get the output of target dump typesystem.

What's changed

The aim of this patch is to prevent this confusion and align the names to using simply sbpf*. Structures and enums aren't touched.