[llvm] [AMDGPU] Replace dynamic VGPR feature with attribute (PR #133444)

[Shilei Tian via llvm-commits]

================
@@ -714,11 +714,12 @@ void SIFrameLowering::emitEntryFunctionPrologue(MachineFunction &MF,
     assert(hasFP(MF));
     Register FPReg = MFI->getFrameOffsetReg();
     assert(FPReg != AMDGPU::FP_REG);
-    unsigned VGPRSize =
-        llvm::alignTo((ST.getAddressableNumVGPRs() -
-                       AMDGPU::IsaInfo::getVGPRAllocGranule(&ST)) *
-                          4,
-                      FrameInfo.getMaxAlign());
+    unsigned VGPRSize = llvm::alignTo(
----------------
shiltian wrote:

This code format is sooo ugly but I understand it is what clang-format likes…

https://github.com/llvm/llvm-project/pull/133444