Re: tg3 appears to be sick in 2.6.33

["Matt Carlson" <mcarlson@broadcom.com>]

On Thu, Jan 07, 2010 at 02:45:43PM -0800, Dmitry Torokhov wrote:
> On Thursday 07 January 2010 02:24:44 pm Eric Dumazet wrote:
> > Le 07/01/2010 21:48, Matt Carlson a ?crit :
> > > Nothing jumps to mind.  Let me see if I can repro this here.
> > >
> > > On Thu, Jan 07, 2010 at 11:56:31AM -0800, Dmitry Torokhov wrote:
> > >> Hi,
> > >>
> > >> Ever since I moved post 2.6.32 tg3 seems to be very sick in my Latitude
> > >> D630, it discards most of the packages for some reason:
> > >>
> > >> [root@dtor-d630 ~]# ethtool -S eth0 | grep rx_
> > >>      rx_octets: 35886
> > >>      rx_fragments: 0
> > >>      rx_ucast_packets: 9
> > >>      rx_mcast_packets: 93
> > >>      rx_bcast_packets: 237
> > >>      rx_fcs_errors: 0
> > >>      rx_align_errors: 0
> > >>      rx_xon_pause_rcvd: 0
> > >>      rx_xoff_pause_rcvd: 0
> > >>      rx_mac_ctrl_rcvd: 0
> > >>      rx_xoff_entered: 0
> > >>      rx_frame_too_long_errors: 0
> > >>      rx_jabbers: 0
> > >>      rx_undersize_packets: 0
> > >>      rx_in_length_errors: 0
> > >>      rx_out_length_errors: 0
> > >>      rx_64_or_less_octet_packets: 0
> > >>      rx_65_to_127_octet_packets: 0
> > >>      rx_128_to_255_octet_packets: 0
> > >>      rx_256_to_511_octet_packets: 0
> > >>      rx_512_to_1023_octet_packets: 0
> > >>      rx_1024_to_1522_octet_packets: 0
> > >>      rx_1523_to_2047_octet_packets: 0
> > >>      rx_2048_to_4095_octet_packets: 0
> > >>      rx_4096_to_8191_octet_packets: 0
> > >>      rx_8192_to_9022_octet_packets: 0
> > >>      rx_discards: 304
> > >>      rx_errors: 0
> > >>      rx_threshold_hit: 0
> > >>
> > >> The above on last night pull from Linux (so 2.6.33+).
> > >>
> > >> Everything works fine if I use wireless or another wired card (Realtek
> > >> in cardbus slot):
> > >>
> > >> [root@dtor-d630 ~]# lspci | grep -ri net
> > >> 04:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd.
> > >> RTL-8139/8139C/8139C+ (rev 10)
> > >> 09:00.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5755M
> > >> Gigabit Ethernet PCI Express (rev 02)
> > >> 0c:00.0 Network controller: Intel Corporation PRO/Wireless 3945ABG
> > >> [Golan] Network Connection (rev 02)
> > >>
> > >> [root@dtor-d630 ~]# ethtool eth0
> > >> Settings for eth0:
> > >>         Supported ports: [ TP ]
> > >>         Supported link modes:   10baseT/Half 10baseT/Full
> > >>                                 100baseT/Half 100baseT/Full
> > >>                                 1000baseT/Half 1000baseT/Full
> > >>         Supports auto-negotiation: Yes
> > >>         Advertised link modes:  10baseT/Half 10baseT/Full
> > >>                                 100baseT/Half 100baseT/Full
> > >>                                 1000baseT/Half 1000baseT/Full
> > >>         Advertised auto-negotiation: Yes
> > >>         Speed: 1000Mb/s
> > >>         Duplex: Full
> > >>         Port: Twisted Pair
> > >>         PHYAD: 1
> > >>         Transceiver: internal
> > >>         Auto-negotiation: on
> > >>         Supports Wake-on: g
> > >>         Wake-on: g
> > >>         Current message level: 0x000000ff (255)
> > >>         Link detected: yes
> > >>
> > >> Any ideas? Thanks!
> > 
> > My laptop is a D630, and running 2.6.33-rc3 with no special problems on tg3
> >
> 
> This is wierd... It all works well with Fedora 12 kernel and .32 so hardware 
> is fine here. I think it also worked when I plugged it into my home router 
> (will try it again tonight). But in the office it just does not want to work. 
> It's like it just does not like the addresses that we use here...

Is the problem related to link speed?  I see that Eric was running at
100Mbps.  Your ethtool output above shows that you are running at 1Gbps.