[llvm] [X86] shouldReduceLoadWidth - don't split loads if ANY uses are a extract+store or a full width legal binop (PR #129695)

[Simon Pilgrim via llvm-commits]

================
@@ -3278,15 +3285,23 @@ bool X86TargetLowering::shouldReduceLoadWidth(SDNode *Load,
       if (Use.getResNo() != 0)
         continue;
 
-      SDNode *User = Use.getUser();
+      const SDNode *User = PeekThroughOneUserBitcasts(Use.getUser());
 
-      // If this use is not an extract + store, it's probably worth splitting.
-      if (User->getOpcode() != ISD::EXTRACT_SUBVECTOR || !User->hasOneUse() ||
-          User->user_begin()->getOpcode() != ISD::STORE)
-        return true;
+      // If any use is an extract + store, it's probably not worth splitting.
+      if (User->getOpcode() == ISD::EXTRACT_SUBVECTOR &&
+          all_of(User->uses(), [&](const SDUse &U) {
+            const SDNode *Inner = PeekThroughOneUserBitcasts(U.getUser());
+            return Inner->getOpcode() == ISD::STORE;
+          }))
+        return false;
+
+      // If any use is a full width legal/target bin op, then assume its legal
----------------
RKSimon wrote:

Yes, we shouldn't have created a target node with 256/512-bit operands unless its supported.

https://github.com/llvm/llvm-project/pull/129695