Re: [WIP][PATCHES] Network xmit batching

[Evgeniy Polyakov <johnpol@2ka.mipt.ru>]

On Fri, Jun 08, 2007 at 07:31:07AM -0400, jamal (hadi@cyberus.ca) wrote:
> On Fri, 2007-08-06 at 12:38 +0400, Evgeniy Polyakov wrote:
> > On Thu, Jun 07, 2007 at 06:23:16PM -0400, jamal (hadi@cyberus.ca) wrote:
> 
> > > I believe both are called with no lock. The idea is to avoid the lock
> > > entirely when unneeded. That code may end up finding that the packet
> [..]
> > +	netif_tx_lock_bh(odev);
> > +	if (!netif_queue_stopped(odev)) {
> > +
> > +		idle_start = getCurUs();
> > +		pkt_dev->tx_entered++;
> > +		ret = odev->hard_batch_xmit(&odev->blist, odev);
> 
> [..]
> > The same applies to *_gso case.
> > 
> 
> You missed an important piece which is grabbing of
> __LINK_STATE_QDISC_RUNNING

But lock is still being hold - or there was no intention to reduce lock
usage? As far as I read Krishna's mail, lock usage was not an issue, so
that hunk probably should be dropped from the analysis.
 
> > Without lock that would be wrong - it accesses hardware.
> 
> We are achieving the goal of only a single CPU entering that path. Are
> you saying that is not good enough?

Then why essentially the same code (current batch_xmit callback)
previously was always called with disabled interrupts? Aren't there
some watchdog/link/poll/whatever issues present?

> > and i also do not know, what service demand is :)
> 
> From the explanation seems to be how much cpu was used while sending. Do
> you have any suggestions for computing cpu use?
> in pktgen i added code to count how many microsecs were used in
> transmitting.

Something, that anyone can understand :)
For example /proc stats, although it is not very accurate, but it is
really usable parameter from userspace point ov view.

> > Result looks good, but I still do not understand how it appeared, that
> > is why I'm not that excited about idea - I just do not know it in
> > details.
> 
> To add to KKs explanation on other email:
> Essentially the value is in amortizing the cost of barriers and IO per
> packet. For example the queue lock is held/released only once per X
> packets. DMA kicking which includes both a PCI IO write and mbs is done
> only once per X packets. There are still a lot of room for improvement
> of such IO;

Btw, what is the size of the packet in pktgen in your tests? Likely it
is small, since result is that good. That can explain alot.

> cheers,
> jamal

-- 
	Evgeniy Polyakov