diff --git a/src/url/url_canon_icu.cc b/src/url/url_canon_icu.cc
index 5181442..98d23e4 100644
--- a/src/url/url_canon_icu.cc
+++ b/src/url/url_canon_icu.cc
@@ -12,10 +12,8 @@
 #include "third_party/icu/include/unicode/char16ptr.h"
 #include "third_party/icu/include/unicode/ucnv.h"
 #include "third_party/icu/include/unicode/ucnv_cb.h"
-#include "third_party/icu/include/unicode/uidna.h"
 #include "url/url_canon_icu.h"
 #include "url/url_canon_internal.h"  // for _itoa_s
-#include "util/gtl/lazy_static_ptr.h"
 
 namespace url {
 
@@ -74,42 +72,6 @@
   const void* old_context_;
 };
 
-// A wrapper to use LazyInstance<>::Leaky with ICU's UIDNA, a C pointer to
-// a UTS46/IDNA 2008 handling object opened with uidna_openUTS46().
-//
-// We use UTS46 with BiDiCheck to migrate from IDNA 2003 (with unassigned
-// code points allowed) to IDNA 2008 with
-// the backward compatibility in mind. What it does:
-//
-// 1. Use the up-to-date Unicode data.
-// 2. Define a case folding/mapping with the up-to-date Unicode data as
-//    in IDNA 2003.
-// 3. Use transitional mechanism for 4 deviation characters (sharp-s,
-//    final sigma, ZWJ and ZWNJ) for now.
-// 4. Continue to allow symbols and punctuations.
-// 5. Apply new BiDi check rules more permissive than the IDNA 2003 BiDI rules.
-// 6. Do not apply STD3 rules
-// 7. Do not allow unassigned code points.
-//
-// It also closely matches what IE 10 does except for the BiDi check (
-// http://goo.gl/3XBhqw ).
-// See http://http://unicode.org/reports/tr46/ and references therein
-// for more details.
-struct UIDNAWrapper {
-  UIDNAWrapper() {
-    UErrorCode err = U_ZERO_ERROR;
-    // TODO(jungshik): Change options as different parties (browsers,
-    // registrars, search engines) converge toward a consensus.
-    value = uidna_openUTS46(UIDNA_CHECK_BIDI, &err);
-    if (U_FAILURE(err)) {
-      LOG(FATAL) << "failed to open UTS46 data with error: " << err;
-      value = NULL;
-    }
-  }
-
-  UIDNA* value;
-};
-
 }  // namespace
 
 ICUCharsetConverter::ICUCharsetConverter(UConverter* converter)
@@ -147,47 +109,4 @@
   } while (true);
 }
 
-static gtl::LazyStaticPtr<UIDNAWrapper> g_uidna;
-
-// Converts the Unicode input representing a hostname to ASCII using IDN rules.
-// The output must be ASCII, but is represented as wide characters.
-//
-// On success, the output will be filled with the ASCII host name and it will
-// return true. Unlike most other canonicalization functions, this assumes that
-// the output is empty. The beginning of the host will be at offset 0, and
-// the length of the output will be set to the length of the new host name.
-//
-// On error, this will return false. The output in this case is undefined.
-// TODO(jungshik): use UTF-8/ASCII version of nameToASCII.
-// Change the function signature and callers accordingly to avoid unnecessary
-// conversions in our code. In addition, consider using icu::IDNA's UTF-8/ASCII
-// version with StringByteSink. That way, we can avoid C wrappers and additional
-// string conversion.
-bool IDNToASCII(const base::char16* src, int src_len, CanonOutputW* output) {
-  DCHECK(output->length() == 0);  // Output buffer is assumed empty.
-
-  UIDNA* uidna = g_uidna->value;
-  DCHECK(uidna != NULL);
-  while (true) {
-    UErrorCode err = U_ZERO_ERROR;
-    UIDNAInfo info = UIDNA_INFO_INITIALIZER;
-    int output_length = uidna_nameToASCII(uidna,
-                                          icu::ConstChar16Ptr(src), src_len,
-                                          icu::Char16Ptr(output->data()),
-                                          output->capacity(), &info, &err);
-    if (U_SUCCESS(err) && info.errors == 0) {
-      output->set_length(output_length);
-      return true;
-    }
-
-    // TODO(jungshik): Look at info.errors to handle them case-by-case basis
-    // if necessary.
-    if (err != U_BUFFER_OVERFLOW_ERROR || info.errors != 0)
-      return false;  // Unknown error, give up.
-
-    // Not enough room in our buffer, expand.
-    output->Resize(output_length);
-  }
-}
-
 }  // namespace url
diff --git a/src/url/url_idna_icu.cc b/src/url/url_idna_icu.cc
new file mode 100644
index 0000000..3fe68c8
--- /dev/null
+++ b/src/url/url_idna_icu.cc
@@ -0,0 +1,90 @@
+#include "third_party/icu/include/unicode/char16ptr.h"
+#include "third_party/icu/include/unicode/uidna.h"
+#include "url/url_canon_internal.h"  // for _itoa_s
+#include "util/gtl/lazy_static_ptr.h"
+
+namespace url {
+
+namespace {
+// A wrapper to use LazyInstance<>::Leaky with ICU's UIDNA, a C pointer to
+// a UTS46/IDNA 2008 handling object opened with uidna_openUTS46().
+//
+// We use UTS46 with BiDiCheck to migrate from IDNA 2003 (with unassigned
+// code points allowed) to IDNA 2008 with
+// the backward compatibility in mind. What it does:
+//
+// 1. Use the up-to-date Unicode data.
+// 2. Define a case folding/mapping with the up-to-date Unicode data as
+//    in IDNA 2003.
+// 3. Use transitional mechanism for 4 deviation characters (sharp-s,
+//    final sigma, ZWJ and ZWNJ) for now.
+// 4. Continue to allow symbols and punctuations.
+// 5. Apply new BiDi check rules more permissive than the IDNA 2003 BiDI rules.
+// 6. Do not apply STD3 rules
+// 7. Do not allow unassigned code points.
+//
+// It also closely matches what IE 10 does except for the BiDi check (
+// http://goo.gl/3XBhqw ).
+// See http://http://unicode.org/reports/tr46/ and references therein
+// for more details.
+struct UIDNAWrapper {
+  UIDNAWrapper() {
+    UErrorCode err = U_ZERO_ERROR;
+    // TODO(jungshik): Change options as different parties (browsers,
+    // registrars, search engines) converge toward a consensus.
+    value = uidna_openUTS46(UIDNA_CHECK_BIDI, &err);
+    if (U_FAILURE(err)) {
+      LOG(FATAL) << "failed to open UTS46 data with error: " << err;
+      value = NULL;
+    }
+  }
+
+  UIDNA* value;
+};
+
+}  // namespace
+
+static gtl::LazyStaticPtr<UIDNAWrapper> g_uidna;
+
+// Converts the Unicode input representing a hostname to ASCII using IDN rules.
+// The output must be ASCII, but is represented as wide characters.
+//
+// On success, the output will be filled with the ASCII host name and it will
+// return true. Unlike most other canonicalization functions, this assumes that
+// the output is empty. The beginning of the host will be at offset 0, and
+// the length of the output will be set to the length of the new host name.
+//
+// On error, this will return false. The output in this case is undefined.
+// TODO(jungshik): use UTF-8/ASCII version of nameToASCII.
+// Change the function signature and callers accordingly to avoid unnecessary
+// conversions in our code. In addition, consider using icu::IDNA's UTF-8/ASCII
+// version with StringByteSink. That way, we can avoid C wrappers and additional
+// string conversion.
+bool IDNToASCII(const base::char16* src, int src_len, CanonOutputW* output) {
+  DCHECK(output->length() == 0);  // Output buffer is assumed empty.
+
+  UIDNA* uidna = g_uidna->value;
+  DCHECK(uidna != NULL);
+  while (true) {
+    UErrorCode err = U_ZERO_ERROR;
+    UIDNAInfo info = UIDNA_INFO_INITIALIZER;
+    int output_length = uidna_nameToASCII(uidna,
+                                          icu::ConstChar16Ptr(src), src_len,
+                                          icu::Char16Ptr(output->data()),
+                                          output->capacity(), &info, &err);
+    if (U_SUCCESS(err) && info.errors == 0) {
+      output->set_length(output_length);
+      return true;
+    }
+
+    // TODO(jungshik): Look at info.errors to handle them case-by-case basis
+    // if necessary.
+    if (err != U_BUFFER_OVERFLOW_ERROR || info.errors != 0)
+      return false;  // Unknown error, give up.
+
+    // Not enough room in our buffer, expand.
+    output->Resize(output_length);
+  }
+}
+
+}  // namespace url