The "Out of mbuf clusters" problem
chrisj at rtems.org
Fri Sep 3 03:53:08 UTC 2004
Phil Torre wrote:
> With the default mbuf and mbuf cluster allocations, and the MPC860's
> caches turned off, it looks like the receive daemon can grab all of
> the available mbuf clusters before the network task has a chance to
> process any of them, resulting in deadlock.
If I assume your receiver cannot keep up so it will need to drop data.
What queue limits do you have for the receive queue ?
My understanding is MBUF queues have a limit count and MBUFs will not be
queued if the queue is full.
So why not tune the receiver's queue to drop the packet if full.
More information about the users