[llvm] [AMDGPU] Support preloading hidden kernel arguments (PR #98861)

[Matt Arsenault via llvm-commits]

================
@@ -281,6 +468,14 @@ static bool lowerKernelArguments(Function &F, const TargetMachine &TM) {
   KernArgSegment->addRetAttr(
       Attribute::getWithAlignment(Ctx, std::max(KernArgBaseAlign, MaxAlign)));
 
+  if (InPreloadSequence) {
+    uint64_t ImplicitArgsBaseOffset =
+        alignTo(ExplicitArgOffset, ST.getAlignmentForImplicitArgPtr()) +
+        BaseOffset;
+    PreloadInfo.tryAllocImplicitArgPreloadSGPRs(ImplicitArgsBaseOffset,
+                                                Builder);
----------------
arsenm wrote:

Should this break if the alloc failed

https://github.com/llvm/llvm-project/pull/98861