Comment 9 for bug 1162160

The problem is that cpufreqd uses MAX_PATH_LEN (defined as 512 in cpufreqd.h) to set buffer sizes when it should really use PATH_MAX (defined as 4096 in /usr/include/linux/limits.h). The cpufreqd package in quantal was built on maverick using an old glibc that didn't validate the buffer size. The cpufreqd package in raring was built on raring using a new glibc that does validate the buffer size, which is why we now see this error.

Looks like the issue was fixed in upstream git long ago, but upstream git hasn't been released as a new version since then:
http://git.kamineko.org/cgi-bin/gitweb.cgi?p=cpufreqd.git;a=commitdiff;h=b5b23525edcc09898288360c48e92b4a6c9cb0ee