Many experiments are now dedicated to the measurement of the Cosmic Microwave Background (CMB) polarization B-modes from primordial gravitational waves. In case of a detection, this signal would be the smoking-gun of the existence of an inflationary phase occurred in the early Universe, as well as a unique probe of its properties. The experiments aiming at this detection, from the ground or from stratospheric balloons, propose instrumental setups with increasing sensitivities and control of the systematics. Nevertheless, there is a question that these experiments cannot answer yet: what is the amplitude of B-modes from the polarized emission of dust in our galaxy, which could screen this primordial signal? Indeed, the interstellar dust grains of asymmetrical shape align with the Galactic magnetic field and produce a polarized emission. Very little was known on the statistics of this emission on large fractions of the sky before the Planck satellite. With its unprecedented sensitivity to the Galactic dust polarized emission in terms of frequency bands and sky coverage, Planck recently gave some answers. I will present these answers, based on http://arxiv.org/abs/1409.5738. I will present the first statistical measurements of the dust polarization at the angular scales of interest for the detection of the CMB B-modes. I will show what kind of general properties we were able to derive from these data and how we have used this information to demonstrate that there are no regions on the sky where the CMB B-modes signal could be detected without properly taking into account the Galactic dust emission polarization. Finally, I will present what is the expected amplitude of the dust polarization in the region where the BICEP2 experiment observed the sky, which shows that the signal they measured is not totally of cosmological origin.