[eepro100] Re: Need help correcting the following error

Andrey Savochkin saw@saw.sw.com.sg
Tue, 17 Oct 2000 16:53:15 +0800


Daniel,

This case is more difficult, see below.

On Mon, Oct 16, 2000 at 10:01:34PM +0200, Daniel Veillard wrote:
> 
[snip]
> --------------------------------- dmesg
> eth0: Printing Rx ring (next to receive into 423951902, dirty index 423951902).
> eth0:     0 0000a002.
> eth0:     1 0000a022.
> eth0:     2 0000a022.
> eth0:     3 0000a022.
> eth0:     4 0000a002.
> eth0:     5 0000a022.
> eth0:     6 0000a022.
> eth0:     7 0000a022.
> eth0:     8 0000a002.
> eth0:     9 0000a022.
> eth0:    10 0000a022.
> eth0:    11 0000a022.
> eth0:    12 0000a022.
> eth0:    13 0000a022.
> eth0:    14 0000a002.
> eth0:    15 0000a022.
> eth0:    16 0000a022.
> eth0:    17 0000a002.
> eth0:    18 0000a002.
> eth0:    19 0000a022.
> eth0:    20 0000a022.
> eth0:    21 0000a022.
> eth0:    22 0000a022.
> eth0:    23 0000a002.
> eth0:    24 0000a022.
> eth0:    25 0000a022.
> eth0:    26 0000a002.
> eth0:    27 0000a022.
> eth0:    28 0000a022.
> eth0: l  29 c000a022.
> eth0:  *=30 0000a002.
> eth0:    31 0000a002.

Receiving ring is ok;

> eth0: Transmit timed out: status f048  0000 at 696171120/696171148 command 0001a000.

but the status is what's interesting there.
It means that the hardware receiver is in stopped state because it thinks
that we don't have enough buffers.  It may happen, for example, when the
traffic bursts.
The question is why it didn't recover after that.
I'll think why it may happen.

> eth0: Tx ring dump,  Tx queue 696171148 / 696171120:
> eth0:     0 200ca000.
> eth0:     1 000ca000.
> eth0:     2 000ca000.
> eth0:     3 000ca000.
> eth0:     4 000ca000.
> eth0:     5 000ca000.
> eth0:     6 000ca000.
> eth0:     7 000ca000.
> eth0:     8 200ca000.
> eth0:     9 000ca000.
> eth0:    10 000ca000.
> eth0:    11 400ca000.
> eth0:   =12 000ca000.
> eth0:    13 000ca000.
> eth0:    14 000ca000.
> eth0:    15 400ca000.
> eth0:  * 16 0001a000.
> eth0:    17 0002a000.
> eth0:    18 000ca000.
> eth0:    19 000ca000.
> eth0:    20 000ca000.
> eth0:    21 000ca000.
> eth0:    22 000ca000.
> eth0:    23 000ca000.
> eth0:    24 200ca000.
> eth0:    25 000ca000.
> eth0:    26 000ca000.
> eth0:    27 000ca000.
> eth0:    28 000ca000.
> eth0:    29 000ca000.
> eth0:    30 000ca000.
> eth0:    31 000ca000.

Best regards
					Andrey V.
					Savochkin