Problem with nfe stability and throughput

[Pyun YongHyeon]

On Fri, Jan 18, 2008 at 07:10:48PM +0100, Chris Poulsen wrote:
 > Hi,
 > 
 > Pyun YongHyeon wrote:
 > >On Fri, Jan 18, 2008 at 10:01:00AM +0900, To Chris Poulsen wrote:
 > > > On Thu, Jan 17, 2008 at 07:05:11PM +0100, Chris Poulsen wrote:
 > > >  > Hi,
 > > >  > 
 > > >  > Pyun YongHyeon wrote:
 > > >  > >Would you show me the output of "ifconfig nfe0"?
 > > >  > >  
 > > >  > nfe0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 
 > > mtu 1500
 > > >  >        options=48<VLAN_MTU,POLLING>
 > > >  >        ether 00:1d:60:6d:73:ec
 > > >  >        inet 192.168.1.11 netmask 0xffffff00 broadcast 192.168.1.255
 > > >  >        media: Ethernet 100baseTX <full-duplex>
 > > >  >        status: active
 > > > 
 > > > Hmmm, it seems that you've set media type manually without relying
 > > > on automatic media detection. Is there any reason not using auto
 > > > media type?  How about using media type 'auto'?
 > > > #ifconfig nfe0 media auto
 > > > 
 > >
 > >I've updated the experimental driver. Please revert previous patch and
 > >apply the following one.
 > >
 > >http://people.freebsd.org/~yongari/atphy.diff2
 > >
 > >And show me the 'ifconfig nfe0' output again.
 > >  
 > 
 > I tried specifying the media type manually, when I was trying to get nfe 
 > up and running earlier, I've reverted to auto now.
 > 
 > The box in question is currently running with your latest patch, I did 
 > manage to stress it (with a couple of concurrent ftp uploads+some web 
 > browing) into a state where it once again gave me a bunch of the following:
 > 
 > kernel: nfe0: discard frame w/o leading ethernet header (len 4 pkt len 4)
 > kernel: nfe0: discard frame w/o leading ethernet header (len 0 pkt len 0)
 > last message repeated 12 times
 > kernel: nfe0: discard frame w/o leading ethernet header (len 3 pkt len 3)
 > kernel: nfe0: discard frame w/o leading ethernet header (len 11 pkt len 11)
 > kernel: nfe0: discard frame w/o leading ethernet header (len 5 pkt len 5)
 > kernel: nfe0: discard frame w/o leading ethernet header (len 5 pkt len 5)
 > kernel: nfe0: link state changed to DOWN
 > kernel: nfe0: link state changed to UP
 > 
 > However bringing the interface down and up seemed to put it back into a 
 > normal state. (I just noticed that nfe0 went down/up again, without my 

Did you have to bring nfe(4) down and up manually due to network lockups?

 > interaction, if that matters).
 > 

Maybe this would come from atphy(4) bug. atphy(4) does not seem to
reliabily detect an established link.

 > ifconfig nfe0 yields:
 > 
 > nfe0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
 >        options=48<VLAN_MTU,POLLING>
 >        ether 00:1d:60:6d:73:ec
 >        inet 192.168.1.11 netmask 0xffffff00 broadcast 192.168.1.255
 >        media: Ethernet autoselect (100baseTX <full-duplex>)
 >        status: active
 > 

How about the following change?
>From /usr/src/sys/dev/mii/athpy.c:
    288         ssr = PHY_READ(sc, ATPHY_SSR);
    289         if ((((bmcr & BMCR_AUTOEN) != 0) && ((bmsr & BMSR_ACOMP) == 0)) ||
    290             (ssr & ATPHY_SSR_SPD_DPLX_RESOLVED) == 0) {
    291                 /* Erg, still trying, I guess... */
    292                 mii->mii_media_active |= IFM_NONE;
    293                 return;
    294         }

To:
    288         ssr = PHY_READ(sc, ATPHY_SSR);
    289         if ((ssr & ATPHY_SSR_SPD_DPLX_RESOLVED) == 0) {
    290                 /* Erg, still trying, I guess... */
    291                 mii->mii_media_active |= IFM_NONE;
    292                 return;
    293         }

Thanks for your patience and testing.
-- 
Regards,
Pyun YongHyeon