dieharder - results interpretation

I'm playing around with dieharder on some of my systems and have gotten some worrisome results from OS X 10.10.2 (14C109).

This is running dieharder -a -g 500 so it should be testing /dev/random. I didn't let it finish, but the FAILED ones are worrisome. diehard_birthdays| 0| 100| 100|0.99104587| PASSED diehard_operm5| 0| 1000000| 100|0.61968542| PASSED diehard_rank_32x32| 0| 40000| 100|0.73898121| PASSED diehard_rank_6x8| 0| 100000| 100|0.99601339| WEAK diehard_bitstream| 0| 2097152| 100|0.47126864| PASSED diehard_opso| 0| 2097152| 100|0.64959894| PASSED diehard_oqso| 0| 2097152| 100|0.24846701| PASSED diehard_dna| 0| 2097152| 100|0.08482001| PASSED diehard_count_1s_str| 0| 256000| 100|0.41125667| PASSED diehard_count_1s_byt| 0| 256000| 100|0.62675323| PASSED diehard_parking_lot| 0| 12000| 100|0.00000000| FAILED diehard_2dsphere| 2| 8000| 100|0.00000000| FAILED diehard_3dsphere| 3| 4000| 100|0.00000000| FAILED diehard_squeeze| 0| 100000| 100|0.00000000| FAILED diehard_sums| 0| 100| 100|0.00000000| FAILED diehard_runs| 0| 100000| 100|0.12981577| PASSED diehard_runs| 0| 100000| 100|0.18812720| PASSED

I read up some and learned that OS X, like BSD, uses Yarrow PRNG for /dev/random with entropy gathered by securityd. I thought that maybe something was wrong there, so I tried piping in some entropy from my Arch Linux machine (which I have passing dieharder with havaged and rngd + the intel drng providing entropy).

Essentially the same results:

diehard_parking_lot| 0| 12000| 100|0.00000000| FAILED diehard_2dsphere| 2| 8000| 100|0.00000000| FAILED diehard_3dsphere| 3| 4000| 100|0.00000000| FAILED diehard_squeeze| 0| 100000| 100|0.00000000| FAILED diehard_sums| 0| 100| 100|0.00000000| FAILED

So, is this anything to be worried about? Maybe its a bug in dieharder? Those 0 p-values seem suspect to me...