[llvm] [APInt] Slightly simplify APInt::ashrSlowCase. NFC (PR #111220)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-llvm-support

Author: Craig Topper (topperc)

<details>
<summary>Changes</summary>

Use an arithmetic shift for the last word copy when BitShift!=0. This avoids an explicit sign extend after the shift.

---
Full diff: https://github.com/llvm/llvm-project/pull/111220.diff


1 Files Affected:

- (modified) llvm/lib/Support/APInt.cpp (+3-4) 


``````````diff
diff --git a/llvm/lib/Support/APInt.cpp b/llvm/lib/Support/APInt.cpp
index 2348a4c9b795e1..23e365f16d8fea 100644
--- a/llvm/lib/Support/APInt.cpp
+++ b/llvm/lib/Support/APInt.cpp
@@ -1055,11 +1055,10 @@ void APInt::ashrSlowCase(unsigned ShiftAmt) {
         U.pVal[i] = (U.pVal[i + WordShift] >> BitShift) |
                     (U.pVal[i + WordShift + 1] << (APINT_BITS_PER_WORD - BitShift));
 
-      // Handle the last word which has no high bits to copy.
-      U.pVal[WordsToMove - 1] = U.pVal[WordShift + WordsToMove - 1] >> BitShift;
-      // Sign extend one more time.
+      // Handle the last word which has no high bits to copy. Use an arithmetic
+      // shift to preserve the sign bit.
       U.pVal[WordsToMove - 1] =
-          SignExtend64(U.pVal[WordsToMove - 1], APINT_BITS_PER_WORD - BitShift);
+          (int64_t)U.pVal[WordShift + WordsToMove - 1] >> BitShift;
     }
   }
 

``````````

</details>


https://github.com/llvm/llvm-project/pull/111220