[llvm-branch-commits] [llvm] [InlineSpiller][AMDGPU] Implement subreg reload during RA spill (PR #175002)

[Christudasan Devadasan via llvm-branch-commits]

================
@@ -1248,18 +1249,62 @@ void InlineSpiller::spillAroundUses(Register Reg) {
 
     // Create a new virtual register for spill/fill.
     // FIXME: Infer regclass from instruction alone.
-    Register NewVReg = Edit->createFrom(Reg);
+
+    unsigned SubReg = 0;
+    LaneBitmask CoveringLanes = LaneBitmask::getNone();
+    // If the subreg liveness is enabled, identify the subreg use(s) to try
+    // subreg reload. Skip if the instruction also defines the register.
+    // For copy bundles, get the covering lane masks.
+    if (MRI.subRegLivenessEnabled() && !RI.Writes) {
+      for (auto [MI, OpIdx] : Ops) {
+        const MachineOperand &MO = MI->getOperand(OpIdx);
+        assert(MO.isReg() && MO.getReg() == Reg);
+        if (MO.isUse()) {
+          SubReg = MO.getSubReg();
+          if (SubReg)
+            CoveringLanes |= TRI.getSubRegIndexLaneMask(SubReg);
+        }
+      }
+    }
+
+    if (MI.isBundled() && CoveringLanes.any()) {
+      CoveringLanes = LaneBitmask(bit_ceil(CoveringLanes.getAsInteger()) - 1);
+      // Obtain the covering subregister index, including any missing indices
+      // within the identified small range. Although this may be suboptimal due
+      // to gaps in the subregisters that are not part of the copy bundle, it is
+      // benificial when components outside this range of the original tuple can
+      // be completely skipped from the reload.
+      SubReg = TRI.getSubRegIdxFromLaneMask(CoveringLanes);
+    }
+
+    // If the target doesn't support subreg reload, fallback to restoring the
+    // full tuple.
+    if (SubReg && !TRI.shouldEnableSubRegReload(SubReg))
----------------
cdevadas wrote:

The check `MRI.subRegLivenessEnabled()` added above skips it for the unintended targets. This is mainly to restrict when targets further want to constrain this optimization for the identified SubReg. For instance, we don't want this optimization for 16-bit subreg accesses. I think the comment I added here is misleading. I need to refine that to include the point I just mentioned. Will do.

https://github.com/llvm/llvm-project/pull/175002