Lumping EQ gain into the total gain makes little sense, unless you are happy to undo the EQ curve after sampling,
and you only need those frequencies. Anyway aren't the AVC and bass enhancement only on the output side, not input,
so are irrelevant.
There is 40dB + 22.5dB gain on the mic input, making 62.5dB gain with flat respose, and the datasheet is
_entirely silent_ about the microphone input noise performance. That's a big clue its poor. Probably very poor
as this is a low cost part (likely will be standard CMOS rather than a hybrid analog/digital process).
Basically the factor you seem to be completely ignoring (but which is usually dominant in microphone systems), is the noise floor
of the system, and to keep this low you need to know the source impedance, and thus the input noise figure of the
preamp for that source impedance. If the noise floor is too high adding gain cannot do anything(*). If we are still talking about
0.1mV full scale, the SGTL5000 can only raise the signal to 0.13V, and with an unknown amount of noise.
This is a microphone handling very low signal levels (100µV), and it likely needs a low noise microphone preamp to get good
signal-to-noise ratio.
Once you have a mic preamp boosting the signal to line level you don't have to worry about gain in the SGTL5000.
(*) Once the noise floor is more than a few LSBs you have all the useful information in the sampled signal in most
scenarios.