Removed payload size dependency and option

[ethouris]

Fixes #876

[maxsharabayko]

Please provide a description on the changes this PR introduces, and estimate the impact of those changes on the behavior of SRT in live mode.
E.g. is the average payload size calculated correctly?
Which payload size is used to calculate byte statistics for missing packets: default or average?
Is is still possible to limit the payload size to 1316?

[ethouris]

In general:

The SRTO_PAYLOADSIZE option is completely removed, this setting is ignored, the option itself only remains as deprecated.

There's still an internally calculated payload size that is needed for various calculations in congestion control module. This is then calculated as starting from the maximum possible value.

Maximum possible value is simply MSS minus UDP and SRT header size, although it also undergoes correction with the extra FEC header in case when filter is on. Also, when payloadsize is set to a nonzero value, which is required for correct functioning of the live CC, it's still checked if a single sending function doesn't try to send more at once.

New functions:

calcMaxPayloadSize() - simply to calculate maximum payload size in a packet, regarding the MSS size, SRT header and possibly FEC. This option may only be used in the connecting and configuring phase; during connection it should rely on the precalculated - and no longer changeable - value stored in m_zUserPayloadSize.
userPayloadSize() - the cache of the above value per packet in m_zUserPayloadSize, available externally (mainly for CC).

[mbakholdina]

@maxsharabayko Please review the PR, consider the issue #1322 while reviewing and let @ethouris know if any additional tests should be done on his side with regards to #1322.

[maxsharabayko]

Is it still possible to limit the payload size to 1316?

@ethouris It feels like it will not be possible to limit to e.g. 1316 if needed then?

[ethouris]

Yes. But I don't think anyone uses this limit for anything. At best you can limit MSS.