Use persistent memory for receiving metrics from sub-processes.

tools/metrics/histograms/histograms.xml

Index: tools/metrics/histograms/histograms.xml
diff --git a/tools/metrics/histograms/histograms.xml b/tools/metrics/histograms/histograms.xml
index 0e0b248a7ee91c4b187a624c383f1aaa3848d6d3..7dd0831375599bad65e3814ab18302ed9fd11d47 100644
--- a/tools/metrics/histograms/histograms.xml
+++ b/tools/metrics/histograms/histograms.xml
@@ -61329,6 +61329,24 @@ http://cs/file:chrome/histograms.xml - but prefer this file for new entries.
   </summary>
 </histogram>
 
+<histogram name="UMA.PersistentAllocator.GpuMetrics.Allocs" units="bytes">
+  <owner>bcwhite@chromium.org</owner>
+  <owner>asvitkine@chromium.org</owner>
+  <summary>
+    Size, before padding, of objects allocated from persistent memory. This is

[Alexei Svitkine (slow), 2016/08/12 23:14:37]

For both of these histograms, mention something about the fact that it's for the
GPU process.

[bcwhite, 2016/08/15 19:43:10]

Done.

+    updated with every allocation.
+  </summary>
+</histogram>
+
+<histogram name="UMA.PersistentAllocator.GpuMetrics.UsedPct" units="%">
+  <owner>bcwhite@chromium.org</owner>
+  <owner>asvitkine@chromium.org</owner>
+  <summary>
+    Percentage of persistent memory segment that has been allocated. This is
+    updated only once per reporting cycle.
+  </summary>
+</histogram>
+
 <histogram name="UMA.PersistentAllocator.RendererMetrics.Allocs" units="bytes">
   <owner>bcwhite@chromium.org</owner>
   <owner>asvitkine@chromium.org</owner>