On Tue, 19 Sep 2000, Donald Becker wrote:
> This is one of the reasons for drivers to keep their Tx queues to a
> reasonable length. Two years ago an almost-always-sufficient Tx queue
> length was between 6 and 10, so most of my bus-master drivers use a queue
> length of 10. Ten 1500 byte packets at 10Mbps is still not too long, but
> ten 60 byte packets (approx. 96 byte periods on the wire) isn't very long at
> 100Mbps.
>
> If you care *that* *much* about priority and scheduling, you are probably
> wrong about what you want. We did some work on utilizing the real-time
> transmit scheduling in the 3Com hardware, and found out that it made almost
> zero difference to [[the multimedia application we were testing]].
Donald,
The Tulip, for example, can have its DMA buffers stashed in a linked list
instead of a ring structure.
Maybe not the most efficient scheme but could you could play with
priority queueing structures/algorithms for the DMA buffers in this
case and for example stick a higher priority packet to the head of the
queue? I have not implemented any such thing, just thinking out
loudly.
cheers,
jamal
|