Subject: WANTED: ep(4) testers
To: None <current-users@netbsd.org>
From: Jaromír <jdolecek@netbsd.org>
List: current-users
Date: 05/12/2001 13:01:59
Hi,
I've made some changes to ep(4) driver and would need testers with
appropriate hardware to test them on -current. I had only a chance
to test it with 3c529 (MCA) so far. Everything seems to work okay,
though the thing didn't help performance on my machine as I hoped :(

The changes:
* g/c the mbuf preallocation code and make sure all mbuf allocation
  is done out of splsched() in epget() - the idea is that preallocation code
  doesn't help performance noticably and needed to do some mbuf frobbing
  to parts which are commonly treated as opaque by drivers, thus making
  experiments with mbufs harder; also,
  when the network would be really busy, it would still be possible
  mbufs could have been allocated on splsched(), hence causing a clock
  skew
  - I've changed the epget() routine to do the mbuf allocation similarily
    to other ethernet drivers, and make sure the data area is allocated
    big enough to hold all available data; I've also moved some other
    misc bits out of critical section in epget()
* implement RX Early support - the idea here is to intercept the
  completition of packet receive and do part of the processing before
  the whole packet is received; thus in theory, once whole packet
  is received, the rest of it can be processed faster, thus making the
  FIFO space faster available for next packet and hence improving
  performance
  - this didn't seem to help on my machine (386DX), however might
    help on something less crappy; currently, the RX Early threshold
    is fixed at ETHERMTU - 200, the code could be improved
    to adjust the threshold if too many packet overruns are encountered

The patch is available at:
	ftp://ftp.NetBSD.org/pub/misc/jdolecek/ep_rxearly.diff

If you'd try it out and tell me performance difference (if any)
and report any encountered problems, you'd help me a lot. Unless
there would some problems, I'd like to get this into -current tree
sometime next week.

Thanks,

Jaromir
-- 
Jaromir Dolecek <jdolecek@NetBSD.org>      http://www.ics.muni.cz/~dolecek/
NetBSD - just plain best OS! -=*=- Got spare MCA cards or docs? Hand me them!