Fix incorrect usage of segment_size instead of segment_bitmask

[redawl]

I noticed the docs for gwsettings had what I believe is a mistake in the explanation for how segment_bitmask is used in the firmware.

[jclehner]

Thanks. What I'm also noticing though is a discrepancy between the footer format described here, with segment_bitmask being a 16 bit integer, vs 32 bit in gwsettings.cc.

I'll investigate that further.

[redawl]

I've been taking a look at this.

From the docs, the previous 2 bytes before the segment_bitmask are padding, and are always 0xFFFF. Because of that, parsing the last 4 bytes as an int32_t or parsing the last two bytes as an int16_t are equivalent.

Using the example from the docs, this c code shows they are equivalent:

int a =  -1 * (int32_t)(0xfffffffc);
int b = -1 * (int16_t)(0xfffc);

printf("Are these equivalent? %d = %d\n", a, b);

outputs:

Are these equivalent? 4 = 4