[llvm] [Offload][AMDGPU] accept generic target (PR #118919)

[via llvm-commits]

================
@@ -190,8 +190,9 @@ Error asyncMemCopy(bool UseMultipleSdmaEngines, void *Dst, hsa_agent_t DstAgent,
 #endif
 }
 
-Expected<std::string> getTargetTripleAndFeatures(hsa_agent_t Agent) {
-  std::string Target;
+Expected<StringRef>
+getTargetTripleAndFeatures(hsa_agent_t Agent, SmallVector<StringRef> &Targets) {
+  StringRef SpecificTarget;
----------------
hidekisaito wrote:

Had an offline chat with @shiltian. We agreed that by controlling the image package ordering --- all specific images before any generic images, OpenMP runtime (or HSA) doesn't have to worry about whether the matched ISA is specific or not. With that, I'll gladly drop the code to identify the specific target. Will follow up with the packager side.

https://github.com/llvm/llvm-project/pull/118919