Strange networking problem on SPARC64
Marius Strobl
marius at alchemy.franken.de
Mon Jul 12 16:03:27 PDT 2004
On Mon, Jul 12, 2004 at 08:18:07PM +0900, Pyun YongHyeon wrote:
> On Mon, Jul 12, 2004 at 07:04:44PM +0800, Xin LI wrote:
> > Hi, folks,
> >
> > My friend at Nankai University has a problem with Networking subsystem on
> > FreeBSD/SPARC64. With a heavy FTP load, ipkterr raises rapidly.
> >
> > As I don't have sparc64 based hardware at hand, I send this mail on behalf
> > of him with a hope that someone can give him some hints to figure out what
> > is happening. Please fell free if I have missed some information and I will
> > forward these to him.
> >
> > Hope we have provided useful information :-) fell free to ask us for more if
> > necessary.
> >
> > Upgrading to -CURRENT does not solved the problem. After some experiences
> > we believe that this is not due to a networking hardware/device error. Same
> > hardware works well under other operating system.
> >
> > Here comes some detailed information:
> >
> ...
> > hme0: <Sun HME 10/100 Ethernet> mem 0x8000-0xffff at device 1.1 on pci0
> > hme0: Ethernet address: 08:00:20:e6:8a:f0
> > miibus0: <MII bus> on hme0
> > nsphy0: <DP83840 10/100 media interface> on miibus0
> ^^^^^^^
> > nsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
> ...
> > /var: superblock summary recomputed
> > hme0: invalid packet size 8422; dropping
> > hme0: invalid packet size 7; dropping
> > hme0: invalid packet size 7; dropping
> > hme0: invalid packet size 7; dropping
> > hme0: invalid packet size 7; dropping
> ^^^^^^^^^^^^^^^^^^^^^^
> See the commit log(if_hme.c rev 1.27) of marius.
> Or see
> http://marc.theaimsgroup.com/?l=freebsd-sparc64&m=108942881831209&w=2
>
Does reverting to 1.26 of if_hme.c really fix the problem or is it
just a workaround because auto-negotation then establishes a 100mbit
full-duplex link, i.e. is it also gone with 1.26 when using a half-
duplex link?
More information about the freebsd-sparc64
mailing list