Hi Nicolas,

We hit an assertion in our UV bits (bug 6555423). This is caused by my 
changes to the aggr driver. I assumed that only when the aggregation is 
plumbed, that will be any tx/rx traffic going through the aggregation, so 
that I moved the the port->lp_txinfo assignment from aggr_grp_create() to 
aggr_m_start(), and only then, the underlying legacy driver associated with 
the port will be opened, and be ready to handle traffic.

The assumption seems wrong for LACP enabled aggregation though - that 
traffic can go on when the aggregation first created. Is this true?

I can change the code the fix it once I got the confirmation that the above 
assumption is indeed wrong.

Thanks
- Cathy

Reply via email to