[vortex-bug] 3c905B dead after kernel upgrade

Donald Becker becker@scyld.com
Tue Aug 27 19:01:01 2002


On 27 Aug 2002, Stephan Herrmann wrote:

> I have been using the 3c59x driver for years now, and last
> uptime was 231 days without any problems. Now I decided to
> upgrade my kernel from 2.4.0 to 2.4.19. The result:
>
> No single packet gets delivered through the card, but
> also no error message. Same computer, no hardware change. 

What does
  cat /proc/net/dev
report?

> 3c59x.c:LK1.1.11 13 Nov 2000  Donald Becker and others. http://www.scyld.com/network/vortex.html $Revision: 1.102.2.46 $
> See Documentation/networking/vortex.txt
> eth0: 3Com PCI 3c905B Cyclone 100baseTx at 0xd000, PCI: Found IRQ 11 for device 00:0c.0
>  00:50:04:43:7a:b5, IRQ 11
>   8K byte-wide RAM 5:3 Rx:Tx split, autoselect/Autonegotiate interface.
>   MII transceiver found at address 24, status 782d.
...
> 3c59x.c:v0.99Xc 6/27/2002 Donald Becker, becker@scyld.com
>   http://www.scyld.com/network/vortex.html
> eth0: 3Com 3c905B Cyclone 100baseTx at 0xd000,  00:50:04:43:7a:b5, IRQ 11
>   Internal config register is 01800000, transceivers 0xa.
>   8K buffer 5:3 Rx:Tx split, autoselect/Autonegotiate interface.
>   MII transceiver found at address 24, status 782d.
...
> eth0: Initial media type Autonegotiate half-duplex.
> eth0: MII #24 status 782d, link partner capability 41e1, setting full-duplex.
> eth0: vortex_open() irq 11 media status 8880.

Looks good so far.

> Using vortex-diag -aa I see this difference:
> (is the interrupt enable value significant?)

> --- vortex.aa	Tue Aug 27 03:04:24 2002
> +++ vortex.aa-new	Tue Aug 27 05:14:33 2002
> @@ -1,22 +1,22 @@
>  vortex-diag.c:v2.06 4/18/2002 Donald Becker (becker@scyld.com)
>   http://www.scyld.com/diag/index.html
> Index #1: Found a 3c905B Cyclone 100baseTx adapter at 0xd000.
>  The Vortex chip may be active, so FIFO registers will not be read.
>  To see all register values use the '-f' flag.
> -Initial window 7, registers values by window:
> +Initial window 4, registers values by window:

Hmmm, I didn't expect to see the register window is changed, but it
shouldn't be a problem.   Window 4 is used when reading the MII status
registers.

> -  Window 5: 1ffc 0000 0000 0600 0807 06ce 06c6 a000.
> +  Window 5: 1ffc 0000 0000 0600 0807 06de 06c6 a000.

> -  Window 6: 0000 0000 0000 fb01 0100 c2ff 01c2 c000.
> +  Window 6: 0000 0000 0000 2301 0100 bb3c 0000 c000.

Window 6 contains the statistics registers, so we expect differences.
Both versions are reporting only a single Tx packet, and multiple Rx
packets.  Reads clear these registers, so a single Tx packet might be
normal, although the "new" entry lists 0 bytes transmitted.

>    Window 7: 0000 0000 0000 0000 0000 0000 0000 e000.
>  Vortex chip registers at 0xd000
> -  0xD010: **FIFO** 00000000 0000000a *STATUS*
> +  0xD010: **FIFO** 00000000 00000002 *STATUS*

Only a difference in packet ID...

> -  0xD020: 00000020 00000000 00080000 00000004
> +  0xD020: 00000020 01864a20 00080000 00001404

The down list pointer is non-zero, which means that there are packets
queued waiting for the Tx unit.

> -  0xD030: 00000000 2f38d0c8 2ffd3180 00080004
> +  0xD030: 00000000 d7ba2846 018649f0 00080004

Just timer counts and buffer address differences here.

> - Indication enable is 06c6, interrupt enable is 06ce.
> + Indication enable is 06c6, interrupt enable is 06de.

Hmmm, this is enabling the RxComplete interrupt bit.
That shouldn't have an impact on operation.

> Basic registers of MII PHY #24:  3000 782d 0040 6120 01e1 41e1 0003 0000.
...
> I'm not a network guru, but our network docs say, I'm connected
> to our ATM backbone via a 10Base5-Segment.

Well, perhaps at some distant point.  Your link partner is a 100baseTx
switch with full duplex but not flow control.

Bottom line: I don't see what would be causing a problem.  You'll have
to try loading the driver with a higher debug setting.  With the correct
driver and 'mii-diag' version you can turn on all messages with
    mii-diag --msg_level 0xffffffff

-- 
Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993