Optimization of log levels

[Aartih <aarthit2014@xxxxxxxxx>]

Hi,

We have registered callbacks for fetching the debug/trace from libssh in our
server side interface code, and for configuring the log level we use the API
ssh_set_log_level(SSH_LOG_LEVEL).

When the log level is set to 1 we observe all the warnings would be
displayed. Also because of this #define of SSH_LOG_RARE to SSH_LOG_WARNING
dumps us the logs of SSH_LOG_RARE in level 1 . 

Normally when we establish a SSH session we get around 13 messages when the
log level is set to 1. Further if we enter some commands/do some operation
the count increases, and floods more number of messages which actually
doesnot fall under warning conditions. Few informational messages are also
logged under this level 1, which floods the log books that we are implementing.

For example in functions 
channel_default_bufferize : SSH_LOG(SSH_LOG_RARE,
      "placing %d bytes into channel buffer (stderr=%d)", len, is_stderr);

channel_write_common:  SSH_LOG(SSH_LOG_RARE,
        "channel_write wrote %ld bytes", (long int) effectivelen);

The above log messages occur very frequently when we set the level to 1
establish the session, 

The above informational messages also come under warning, And we feel that
this SSH_LOG_RARE has to be made equivalent to either levels 2 or 3, since
the messages are flooding the log books and we fear the possibility of
missing other important messages.

Kindly let us know if this SSH_LOG_RARE's level can be increased so that we
don't get these informational messages continuously.

Thanks
Aarthi