Correct nacl_helper SI count on Linux

This is to green up the Google Chrome Linux bot [1].

NaCl now has the same SIs as Chrome [2].  They are:
WelsThreadPool.cpp
atomicops_internals_x86.cc
debugallocation_shim.cc
iostream.cpp
spinlock.cc

[1] https://ci.chromium.org/buildbot/chromium.chrome/Google%20Chrome%20Linux%20x64/36988
[2] https://logs.chromium.org/logs/chromium/bb/chromium.chrome/Google_Chrome_Linux_x64/36988/+/recipes/steps/sizes/0/stdout

BUG=896171
R=thakis

Change-Id: Ibcb553b9790b08e46e46b6eb343183f4de60b888
Reviewed-on: https://chromium-review.googlesource.com/c/1296049
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#602054}

infra/scripts/legacy/scripts/slave/chromium/sizes.py

@@ -31,9 +31,13 @@ SRC_DIR = os.path.abspath(
 
 # If something adds a static initializer, revert it, don't increase these
 # numbers. We don't accept regressions in static initializers.
+#
+# Note: Counts for chrome and nacl_helper are one higher in branded builds
+# compared to release builds.  This is due to a static initializer in
+# WelsThreadPool.cpp (https://crbug.com/893594).
 EXPECTED_LINUX_SI_COUNTS = {
   'chrome': 5,
-  'nacl_helper': 4,
+  'nacl_helper': 5,
   'nacl_helper_bootstrap': 0,
 }