[llvm] [RISCV] Eliminate getVLENFactoredAmount and expose muladd [nfc] (PR #87881)

[Philip Reames via llvm-commits]

================
@@ -195,7 +195,16 @@ void RISCVRegisterInfo::adjustReg(MachineBasicBlock &MBB,
     Register ScratchReg = DestReg;
     if (DestReg == SrcReg)
       ScratchReg = MRI.createVirtualRegister(&RISCV::GPRRegClass);
-    TII->getVLENFactoredAmount(MF, MBB, II, DL, ScratchReg, ScalableValue, Flag);
+
+    assert(ScalableValue > 0 && "There is no need to get VLEN scaled value.");
+    assert(ScalableValue % 8 == 0 &&
+           "Reserve the stack by the multiple of one vector size.");
+    assert(isInt<32>(ScalableValue / 8) &&
----------------
preames wrote:

Unchanged from the original code.  It arguably should be changed, but I'd prefer to revisit the cornercases of extremely large offsets here in a separate change.  (Remember, we're talking about more than 2^31 VLEN sized slots here, that's... unlikely.)

https://github.com/llvm/llvm-project/pull/87881