I'm comfortable with the "spec1" -> P tag.

Some of the use (from nikk's 小/こ entry)
小島	1207895
小山	1562564
小屋	2782880
小石	287085
小降り	102001
小銭	1434535
小首	54995
(all in jmdict)

Some random stuff I found through google, ninjal-lwp which aren't currently in jmdict:
小返し	1472
小広い	2473
小狡い	1360
小ズルい	244
小ずるい	3099
小ずる賢い	45
小狭い	747
小太い	26
小椅子	37982
小箪笥	2144
小むすび	202

Why wouldn't we count it's use in words that are already entries if we're talking about absolute usage? We have over 400 entries beginning 
with 小/こ. That alone makes it worthy of spec1, I think. If it's that productive, it also seems likely there's probably hundreds or 
thousands of similar words/constructions that aren't included in the dictionary.

not entirely sure i agree with this.  despite it being called a 接頭 in most dictionaries, shinmeikai has it as 造語, and i feel like it is not particularly productive outside of words like 小皿 that we already have entries for.  simply wanting something to "stand out" in comparison to other entries is not a valid reason for the tag, IMO.  it should be about absolute frequency of usage.  how often is this used as a productive suffix that could be added to entries that are not already in the dictionary?

Very common n-pref. Should come above お 
and ぐゎー (both written 小) in search 
results.

@@ -5,0 +6 @@
+<ke_pri>spec1</ke_pri>
@@ -8,0 +10 @@
+<re_pri>spec1</re_pri>

@@ -20 +19,0 @@
-<xref type="see" seq="1324510">弱・じゃく・1</xref>