Cyclone 3x905b Corruption/problems

Michael Thompson shadowr@ipass.net
Thu Mar 18 10:25:28 1999


I have a DEC 486-66 box running redhat 5.2 with a 2.2.3 kernel.  The box
has two NICs, a 3c509 ISA card and a cyclone.  The cyclone is connected
to an ADSL line and the 509 to an small thinwire ethernet LAN. First the
cyclone only works if the machine has been cold booted.  Warm boots
usually leave the card in a state were all it will do is cause tons
of error messages in the logs.  The cyclone driver is compiled as a
module and removing/reloading it appears to reset it.  When it is
working correctly it runs fine routing packets.  HTML, ICQ, etc work,
but any attempts to download large files (be it HTTP, FTP, samba, etc)
lead to corruption.  It normally crops up in large files (20+ meg)
but corrupts smaller ones randomly.  Zip files pop up with CRC errors,
and other files just get trashed.  No errors, etc crop up in
the logs and ifconfig doesn't show any errors or extreme number of
collisions.  Best I can tell the corruption doesn't show up when using
Win98 with the card just linux (but I really really don't want to run
and have to pay for the Win98 masq program for this setup).  The card's
flaky state does show up in Win98 proving the thing has at least some
kind of hardware problem.


Logs of cyclone in screwy state:
Mar 17 19:05:28 server kernel: CPU0: Intel 486 DX/2 stepping 05
Mar 17 19:05:28 server kernel: SMP motherboard not detected. Using dummy
APIC em
ulation.
Mar 17 19:05:28 server kernel: PCI: PCI BIOS revision 2.00 entry at
0xf0100
Mar 17 19:05:28 server kernel: PCI: Using configuration type 2
Mar 17 19:05:28 server kernel: PCI: Probing PCI hardware
Mar 17 19:05:28 server kernel: eth0: 3c509 at 0x300 tag 1, BNC port,
address  00
 20 af 6e eb 11, IRQ 10.
Mar 17 19:05:28 server kernel: 3c509.c:1.12 6/4/97
becker@cesdis.gsfc.nasa.gov
Mar 17 19:05:28 server kernel: Partition check:
Mar 17 19:05:28 server kernel:  sda: sda1
Mar 17 19:05:28 server kernel:  sdb: sdb1 sdb2
Mar 17 19:05:28 server kernel: VFS: Mounted root (ext2 filesystem)
readonly.
Mar 17 19:05:28 server kernel: Freeing unused kernel memory: 72k freed
Mar 17 19:05:28 server kernel: Adding Swap: 36228k swap-space (priority
-1)
Mar 17 19:05:28 server kernel: 3c59x.c:v0.99H 11/17/98 Donald Becker
http://cesd
is.gsfc.nasa.gov/linux/drivers/vortex.html
Mar 17 19:05:28 server kernel:   The PCI BIOS has not enabled the device
at 0/64
. Updating PCI command 0113->0117.
Mar 17 19:05:28 server kernel: eth1: 3Com 3c905B Cyclone 100baseTx at
0xd200,  0
0:10:5a:ab:fa:69, IRQ 5
Mar 17 19:05:28 server kernel:   8K byte-wide RAM 5:3 Rx:Tx split,
autoselect/Au
tonegotiate interface.
Mar 17 19:05:28 server kernel:   MII transceiver found at address 24,
status 786
d.
Mar 17 19:05:28 server kernel:   MII transceiver found at address 0,
status 786d
.
Mar 17 19:05:28 server kernel:   Enabling bus-master transmits and
whole-frame r
eceives.


ar 17 19:10:32 server kernel: eth1: MII #24 status 786d, link partner
capabilit
y 0020, setting half-duplex.
Mar 17 19:10:32 server kernel: eth1: Host error, FIFO diagnostic
register 0000.
Mar 17 19:10:32 server last message repeated 20 times
Mar 17 19:10:32 server kernel: eth1: Too much work in interrupt, status
8003.  T
emporarily disabling functions (7ffc).
Mar 17 19:10:35 server kernel: eth1: Too much work in interrupt, status
8011.  T
emporarily disabling functions (7fee).
Mar 17 19:12:10 server kernel: eth1: transmit timed out, tx_status 00
status e00
3.
Mar 17 19:12:10 server kernel: eth1: Interrupt posted but not delivered
-- IRQ b
locked by another device?
Mar 17 19:12:10 server kernel: eth1: Host error, FIFO diagnostic
register 0000.
Mar 17 19:12:10 server last message repeated 20 times
Mar 17 19:12:10 server kernel: eth1: Too much work in interrupt, status
8003.  T
emporarily disabling functions (7ffc).
Mar 17 19:12:13 server modprobe: can't locate module net-pf-5
Mar 17 19:12:15 server kernel: eth1: transmit timed out, tx_status 00
status e01
1.
Mar 17 19:12:15 server kernel: eth1: Interrupt posted but not delivered
-- IRQ b
locked by another device?
Mar 17 19:12:15 server kernel: eth1: Too much work in interrupt, status
e011.  T
emporarily disabling functions (7fee).


I also have the more in depth debug info from dmesg if it will help.


-- 
+----------------------------+--------------------------+
Mike Thompson aka Shadow RAM : Made from 100% all natural
shadowr@ipass.net            :          electrons!
mathomp3@eos.ncsu.edu        :
mithomps@nyx10.nyx.net       : (UL) Listed, FCC Approved.
+----------------------------+--------------------------+
:      http://www.geocities.com/SiliconValley/5789      :
+-------------------------------------------------------+
GCS d- s+:- a? C++++ UBAHOLU++++ P++ L++ E---- W+++ N+ o?
K? w++ O--- M-- V-- PS+ PE+ Y+ PGP+ t+ 5- X++ R tv++ b++ 
DI+++ D+++ G e h! r-- y?