unclear about the ifm->channel[] use

@dcobas I saw that we had different a-priori about the use of the ifm->channel[256] configuration:

to me:

• .channel[256] with .nchannels and .channel_id suggest that we can have channel_id ordered this way: [35,36,78,58, X,X,X], nchannels = 4.

but I can find other types:

• for( i = 0; i < 256; i++) with
• continue on .type == FSI_CH_NONE
• or continue on .disabled == 1
• or assert (i == cfg[i].channel_id).

I think this is linked with your static preference, so I'd like to have your prefered way so that I can fix my implementation