Make ARM64 detection consistent.

chrome/browser/omaha_query_params/omaha_query_params.cc

Index: chrome/browser/omaha_query_params/omaha_query_params.cc
diff --git a/chrome/browser/omaha_query_params/omaha_query_params.cc b/chrome/browser/omaha_query_params/omaha_query_params.cc
index 78a5317b175d6152a7b13a4b13e966b8af6fc7dd..3f5ecc7f44076c7f5aa8fa7bfa2cea74c0509207 100644
--- a/chrome/browser/omaha_query_params/omaha_query_params.cc
+++ b/chrome/browser/omaha_query_params/omaha_query_params.cc
@@ -40,8 +40,8 @@ const char kArch[] =
     "x86";
 #elif defined(__arm__)
     "arm";
-#elif defined(__arm64__)
-    "arm64";
+#elif defined(__aarch64__)
+    "aarch64";

[stuartmorgan, 2014/03/31 13:26:01]

Is there any reason *not* to call this arm64? Seems odd to make something that's
just for dashboard consumption consistent with the define rather than the "arm"
string; arm64 seems more self-explanatory.

I'm fine either way though if there's a good reason for this.

[Nico, 2014/03/31 16:13:43]

(I agree with this. The gyp target_arch value is also called arm64.)

[Primiano Tucci (use gerrit), 2014/03/31 16:25:29]

Right, me and Stuart had a discussion offline and that (GYP) was  the same point
I had about keeping arm64.
However, it looks it's worth double-checking this since this string is
apparently sent to Omaha, and I don't know whether it would be fine having Linux
64 fall into the same category of iOS 64. (And for sake of clarity, changing
this string as it is in patchset 2, looks wrong in any case to me)
I'll hold this CL until I get a confirmation on this point.

 #elif defined(__mips__)
     "mipsel";
 #else