Re[2]: weird dump on console

Michiel Brandenburg michielb <michielb@stack.nl>
Thu, 4 Oct 2001 16:22:29 +0200


Hello Bart,

<SNIP>
>>Sep 16 05:27:28 sec kernel:   Flags; bus-master 1, dirty 378136(8) current
>>378136(8)
>>Sep 16 05:27:28 sec kernel:   Transmit list 00000000 vs. c6eec400.
>>Sep 16 05:27:28 sec kernel:   0: @c6eec200  length 8000003a status 0001003a
>>Sep 16 05:27:28 sec kernel:   1: @c6eec240  length 80000036 status 00010036
>>Sep 16 05:27:28 sec kernel:   2: @c6eec280  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   3: @c6eec2c0  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   4: @c6eec300  length 800004a5 status 000104a5
>>Sep 16 05:27:28 sec kernel:   5: @c6eec340  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   6: @c6eec380  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   7: @c6eec3c0  length 800005ba status 800105ba
>>Sep 16 05:27:28 sec kernel:   8: @c6eec400  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   9: @c6eec440  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   10: @c6eec480  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   11: @c6eec4c0  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   12: @c6eec500  length 80000579 status 00010579
>>Sep 16 05:27:28 sec kernel:   13: @c6eec540  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   14: @c6eec580  length 800005ba status 000105ba
>>Sep 16 05:27:28 sec kernel:   15: @c6eec5c0  length 80000435 status 00010435
<SNIP>
BT> firewall uses iptables and runs mandrake7.2, the proxy runs redhat7.0.
BT> It happens when I try to move a lot of data from the proxy server over 
BT> the firewall to another pc. Both firewall and proxy server generate 
BT> those messages.
<SNIP>
BT> All network cards are 3Com. (3c905c, 3c905b)

I had the same problems as the one above.  I rememeber that
there is something not realy ok in the driver of the 3com905 series
cards.  This could be the reason that both machines are producing them
messages under heavy network load.  As they then might start to
busmaster. I will try to reproduct them errors my getting my
other machine to push a lot of data.  It's a cyrix 120Mhz so I hope
the cpu can handle it.

This problem was solved in my case by using Intel Pro 100 cards in all
my machines.

-- 
Best regards,
 Michiel                            mailto:michielb@stack.nl