Losses (or gains) multiply. This leads to the convenience of using a log measure dB for them, which add.
So yes, 6.1 dB per 100 m is 0.061 dB per metre.
Note that there are additional discontinuity loss terms if the cable and the measurement system are mismatched in impedance, the larger the mismatch, the larger the error. These errors could easily be bigger than 0.06 dB. Only try to estimate the loss of a cable by measurement by measuring a long length, short lengths could be proportionally very wrong.
100m of cable would measure 6.1 +/- discontinuity_loss dB.
200m of cable would measure 12.2 +/- discontinuity_loss dB.
50m of cable would measure 3.05 +/- discontinuity_loss dB.
1m of cable would measure 0.061 +/- discontinuity_loss dB.
... where discontinuity_loss depends on any cable mismatches, connector repeatability, precise length of cable (due to the phase that the two end reflections add up) and the order of length of cable (as the cable loss increases between the two ends, the two ends interact less). It's good when making a measurement of insertion loss to sweep the frequency over sufficient range to alter the electrical length of the cable under test by at least λ/4, this will swing the combination of discontinuity_losses from each end of the cable over its maximum and minimum values, allowing you to assess whether they are significant or not.