This was found during the jsr166 testing, first adoption of @Contended in the wild. It turns out making the static fields @Contended breaks the static oop maps, which is then fails spuriously during the GC, resulting in crashes like these:

#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0x00007f96e8cafeec, pid=23307, tid=140285777536768
#
# JRE version: OpenJDK Runtime Environment (8.0) (build 1.8.0-internal-dl_2013_02_23_13_14-b00)
# Java VM: OpenJDK 64-Bit Server VM (25.0-b19 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# V [libjvm.so+0x507eec] G1ParCopyClosure<false, (G1Barrier)0, false>::copy_to_survivor_space(oopDesc*)+0x3c
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# An error report file with more information is saved as:
# /home/dl/concurrent/src/test/jsr166e/hs_err_pid23307.log
#
# If you would like to submit a bug report, please visit:
# http://bugreport.sun.com/bugreport/crash.jsp
#

Sometimes it dies at:
oopDesc* PSPromotionManager::copy_to_survivor_space<false>(oopDesc*)+0x123

or:
InstanceKlass::oop_push_contents(PSPromotionManager*, oopDesc*)+0x12c

Unfortunately, this also breaks the classes annotated with @Contended, because we pessimistically pad the static fields block as well. This wasn't found during original reviews, it slipped the @Contended-targeted regression testing, and would not manifest unless you use @Contended.

Still, this is the adoption blocker, even though the workaround exists: we can emulate the desired behavior with per-field @Contended annotation.