funclatency yielding odd results

Hi,
funclatency is giving me bloated numbers, so I tried with an empty function, and funclatency tells me that this function lasts 1200 nsec, which is rather absurd on my 3.50 GHz CPU. Here are all details:
- I'm using a very recent version of bcc, cloned about ten days ago;
- the functions I'm measuring are invoked through pointers (if useful for you to know, I'm measuring the duration of the interface function of my BFQ I/O scheduler)
- to get stable results, I do the following:
cpupower frequency-set -g performance -d 3.50GHz -u 3.50GHz
cpupower idle-set -D 0

(/usr/share/bcc/tools/funclatency $FUNC_TO_MEASURE) &
BPFPID=$! # save pid, to stop funclatency when the test finishes

# do the test during which I want to measure the duration of my function of interest

kill -SIGINT $BPFPID
cpupower frequency-set -g powersave -d 800MHz
cpupower idle-set -E

Just ask if other information is needed.

Start with /usr/share/bcc/tools/funclatency and reproduce the report using the empty function, function-pointer invocation, and CPU settings described. Compare the reported 1200 nsec result with the measurement setup and determine whether the behavior is specific to funclatency or pointer-invoked functions. Done means the discrepancy is explained and the appropriate correction or limitation is established.

Written by the indexing model from the issue text.