[vortex-bug] Tx int. mitigation

Donald Becker becker@scyld.com
Thu, 14 Dec 2000 12:19:06 -0500 (EST)


On Thu, 14 Dec 2000, Giuseppe Ciaccio wrote:

> > The mechanism for Tx mitigation is located at the end of start_xmit().
> > However, the current code (slighlty different in Don's drivers):
..
> My personal opinion is:  we do not really need Tx IRQ.

We need some way to scavenge Tx skbuffs in a timely manner, even if we are
not receiving packets.
The protocol layer uses the skbuff-is-free information.  Leaving already
transmitted skbuffs on the transmit queue will cause mysterious performance
problems.
Additional issue: Debugging is easier if the Tx statistics are correct.

> 	/*  Apparently not "enough" room in Tx ring  */
> 	try to free the tail of Tx ring as much as you can;

The 3c509 driver gathers the previous Tx results in the Tx path, but the
situation is different.  3c509 driver copies the packets to the card and
frees the skbuff immediately.

Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993