Base64.Decode: fixed latent bug for non-ASCII inputs (#76795)

dotnet · Oct 10, 2022 · ef6bc67 · ef6bc67
1 parent 505c27e
commit ef6bc67
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 1 deletion.
diff --git a/src/libraries/System.Memory/tests/Base64/Base64DecoderUnitTests.cs b/src/libraries/System.Memory/tests/Base64/Base64DecoderUnitTests.cs
@@ -272,6 +272,23 @@ public void BasicDecodingWithFinalBlockTrueKnownInputInvalid(string inputString,
             Assert.True(Base64TestHelper.VerifyDecodingCorrectness(expectedConsumed, decodedBytes.Length, source, decodedBytes));
         }
 
+        [Theory]
+        [InlineData("\u00ecz/T", 0, 0)]                                              // scalar code-path
+        [InlineData("z/Ta123\u00ec", 4, 3)]
+        [InlineData("\u00ecz/TpH7sqEkerqMweH1uSw==", 0, 0)]                          // Vector128 code-path
+        [InlineData("z/TpH7sqEkerqMweH1uSw\u00ec==", 20, 15)]
+        [InlineData("\u00ecz/TpH7sqEkerqMweH1uSw1a5ebaAF9xa8B0ze1wet4epo==", 0, 0)]  // Vector256 / AVX code-path
+        [InlineData("z/TpH7sqEkerqMweH1uSw1a5ebaAF9xa8B0ze1wet4epo\u00ec==", 44, 33)]
+        public void BasicDecodingNonAsciiInputInvalid(string inputString, int expectedConsumed, int expectedWritten)
+        {
+            Span<byte> source = Encoding.UTF8.GetBytes(inputString);
+            Span<byte> decodedBytes = new byte[Base64.GetMaxDecodedFromUtf8Length(source.Length)];
+
+            Assert.Equal(OperationStatus.InvalidData, Base64.DecodeFromUtf8(source, decodedBytes, out int consumed, out int decodedByteCount));
+            Assert.Equal(expectedConsumed, consumed);
+            Assert.Equal(expectedWritten, decodedByteCount);
+        }
+
         [Theory]
         [InlineData("AQID", 3)]
         [InlineData("AQIDBAUG", 6)]

diff --git a/src/libraries/System.Private.CoreLib/src/System/Buffers/Text/Base64Decoder.cs b/src/libraries/System.Private.CoreLib/src/System/Buffers/Text/Base64Decoder.cs
@@ -592,8 +592,9 @@ private static unsafe void Vector128Decode(ref byte* srcBytes, ref byte* destByt
 
                 // lookup
                 Vector128<byte> hiNibbles = Vector128.ShiftRightLogical(str.AsInt32(), 4).AsByte() & mask2F;
+                Vector128<byte> loNibbles = str & mask2F;
                 Vector128<byte> hi = SimdShuffle(lutHi, hiNibbles, mask8F);
-                Vector128<byte> lo = SimdShuffle(lutLo, str, mask8F);
+                Vector128<byte> lo = SimdShuffle(lutLo, loNibbles, mask8F);
 
                 // Check for invalid input: if any "and" values from lo and hi are not zero,
                 // fall back on bytewise code to do error checking and reporting: