[vortex] Re: Query: Any solution to 3c905B problem?

[Smoke]

On Sunday, Jan  6, 2002 at 12:18, Bogdan Costescu wrote:
> This seems like a different problem: incorrect media settings. If you send 
> the large amount of data, what hangs: the transfer itself or network 
> communication in general (a ping at the same time) ? Only transfer hanging 
> would most likely be a media mismatch, in this case you should also see 
> some error messages in the logs. If it's the network hanging, would it be 
> possible to run the diag tools (mii-diag and vortex-diag) while the 
> network is in this state ? 

It is definitely the whole interface that is hanging. I did post the
various error messages to the list a while back, but here they are
again:

-=-=-=-=-=-=-
> The symptoms of this "hanging" are that the session freezes, and
> nothing else can use the ethernet interface (pings all fail,
> etc.). the only solution is to take down the interface, bring it back
> up, and add the necessary routing information again. then i have to

That does indicate that the interface is hanging.

> $ mii-diag -w
> Using the default interface 'eth0'.
> Basic registers of MII PHY #24:  3000 786d 0000 0000 01e1 0020 0004
> 2001.
>  Basic mode control register 0x3000: Auto-negotiation enabled.
>  You have link beat, and everything is working OK.
>  Your link partner is generating 10baseT link beat  (no autonegotiation).
>    End of basic transceiver information.
> 
> Monitoring the MII transceiver status.
> 18:29:35.888  Baseline value of MII BMSR (basic mode status register) is 786d.
> 18:31:18.231  MII BMSR now 7849:   no link, NWay busy, No Jabber (0000).
> 18:31:20.141  MII BMSR now 7869:   no link, NWay done, No Jabber (0020).
>    New link partner capability is 0020 0004: 10baseT.
> 18:31:20.153  MII BMSR now 786d: Good link, NWay done, No Jabber (0020).
> 
> after the first time stamp nothing happens until i start an X client
> and cause a hang. then the next 4 lines appear and the connection is
> hung.

Hmmm, it's difficult to tell if the driver is resetting the transceiver,
or if the link partner (the repeater) is going bad.

> Dec 20 19:11:03 noname kernel: eth0: transmit timed out, tx_status 00 status 0000.

Definitely a problem, but there are no obvious errors here.

> Dec 20 19:11:03 noname kernel:   diagnostics: net 0012 media 8880 dma 00000092.
> Dec 20 19:11:03 noname kernel:   Flags; bus-master 1, dirty 63(15) current 79(15)
> Dec 20 19:11:03 noname kernel:   Transmit list 013823c0 vs. c13825c0.
..
> Dec 20 19:11:03 noname kernel:   6: @c1382380  length 800003be status 000103be
> Dec 20 19:11:03 noname kernel:   7: @c13823c0  length 800005ea status 000005ea

Hmmm, this packet hasn't been marked as having been transmitted.
Something is blocking transmits from occuring.

> What driver version?
> What is the detection message?

Oops, sorry, I meant to include that. from dmseg:

PCI: Found IRQ 11 for device 00:11.0
PCI: Sharing IRQ 11 with 00:07.2
3c59x: Donald Becker and others. www.scyld.com/network/vortex.html
00:11.0: 3Com PCI 3c905B Cyclone 100baseTx at 0xdc00. Vers LK1.1.16

This is on Linux kernel 2.4.16.

> Hmmm, it's difficult to tell if the driver is resetting the transceiver,
> or if the link partner (the repeater) is going bad.

Are there perhaps some diagnostics I could try to determine that?

> Hmmm, this packet hasn't been marked as having been transmitted.
> Something is blocking transmits from occuring.

These messages continue at about every 10s until the machine eventually completely hangs.

Is there any other information i can dig up that might be useful?
-=-=-=-=-=-=-

Thanks,
-Conrad.