problem with bus-master driver for 3Com 'Boomerang'

Harald Koenig (koenig@ceres.tat.physik.uni-tuebingen.de)
Mon, 24 Feb 1997 20:43:58 +0100 (MEZ)


Hi,

we're using Donald Becker's "3c900.c:v0.37 12/20/96 becker@cesdis.gsfc.nasa.gov\n"
driver with Linux-2.0.29 for 10 PPro200 PCs with 3Com Boomerang cards.

performace is great (esp. compared with the non-busmaster driver)
but now we got problems for the 2nd time on two of the PCs (all are in the
same network; connected to the same hub) which started to show up the
following messages:

-------------------------------------------------------------------------------
Feb 24 09:51:37 alamak kernel: eth0: transmit timed out, tx_status 88 status e000.
Feb 24 09:51:37 alamak kernel: eth0: Transmitter encountered 16 collisions -- network network cable problem?
Feb 24 09:51:37 alamak kernel: Flags; bus-master 1, full 1; dirty 2882564 current 2882580.
Feb 24 09:51:37 alamak kernel: Down list 00ff1168 vs. 00ff1128.
Feb 24 09:51:37 alamak kernel: 0: 00ff1128 length 80000052 status 00000052
Feb 24 09:51:37 alamak kernel: 1: 00ff1138 length 80000066 status 00000066
Feb 24 09:51:37 alamak kernel: 2: 00ff1148 length 800000b6 status 800000b6
Feb 24 09:51:37 alamak kernel: 3: 00ff1158 length 80000052 status 80000052
Feb 24 09:51:37 alamak kernel: 4: 00ff1168 length 8000002a status 0000002a
Feb 24 09:51:37 alamak kernel: 5: 00ff1178 length 8000002a status 0000002a
Feb 24 09:51:37 alamak kernel: 6: 00ff1188 length 8000008e status 0000008e
Feb 24 09:51:37 alamak kernel: 7: 00ff1198 length 8000002a status 0000002a
Feb 24 09:51:37 alamak kernel: 8: 00ff11a8 length 8000009e status 0000009e
Feb 24 09:51:37 alamak kernel: 9: 00ff11b8 length 8000009e status 0000009e
Feb 24 09:51:37 alamak kernel: 10: 00ff11c8 length 800000b6 status 000000b6
Feb 24 09:51:37 alamak kernel: 11: 00ff11d8 length 80000052 status 00000052
Feb 24 09:51:37 alamak kernel: 12: 00ff11e8 length 8000005a status 0000005a
Feb 24 09:51:37 alamak kernel: 13: 00ff11f8 length 8000004d status 0000004d
Feb 24 09:51:37 alamak kernel: 14: 00ff1208 length 80000052 status 00000052
Feb 24 09:51:37 alamak kernel: 15: 00ff1218 length 800000b6 status 000000b6
Feb 24 09:51:41 alamak kernel: eth0: transmit timed out, tx_status 00 status e000.
Feb 24 09:51:41 alamak kernel: Flags; bus-master 1, full 1; dirty 2882564 current 2882580.
Feb 24 09:51:41 alamak kernel: Down list 00000000 vs. 00ff1128.
Feb 24 09:51:41 alamak kernel: 0: 00ff1128 length 80000052 status 00000052
Feb 24 09:51:41 alamak kernel: 1: 00ff1138 length 80000066 status 00000066
Feb 24 09:51:41 alamak kernel: 2: 00ff1148 length 800000b6 status 800000b6
Feb 24 09:51:41 alamak kernel: 3: 00ff1158 length 80000052 status 80000052
Feb 24 09:51:41 alamak kernel: 4: 00ff1168 length 8000002a status 0000002a
Feb 24 09:51:41 alamak kernel: 5: 00ff1178 length 8000002a status 0000002a
Feb 24 09:51:41 alamak kernel: 6: 00ff1188 length 8000008e status 0000008e
Feb 24 09:51:41 alamak kernel: 7: 00ff1198 length 8000002a status 0000002a
Feb 24 09:51:41 alamak kernel: 8: 00ff11a8 length 8000009e status 0000009e
Feb 24 09:51:41 alamak kernel: 9: 00ff11b8 length 8000009e status 0000009e
Feb 24 09:51:41 alamak kernel: 10: 00ff11c8 length 800000b6 status 000000b6
Feb 24 09:51:41 alamak kernel: 11: 00ff11d8 length 80000052 status 00000052
Feb 24 09:51:41 alamak kernel: 12: 00ff11e8 length 8000005a status 0000005a
Feb 24 09:51:41 alamak kernel: 13: 00ff11f8 length 8000004d status 0000004d
Feb 24 09:51:41 alamak kernel: 14: 00ff1208 length 80000052 status 00000052
Feb 24 09:51:41 alamak kernel: 15: 00ff1218 length 800000b6 status 000000b6
------------------------------------------------------------------------------
Feb 24 09:52:04 alkyone kernel: eth0: transmit timed out, tx_status 88 status e000.
Feb 24 09:52:04 alkyone kernel: eth0: Transmitter encountered 16 collisions -- network network cab
le problem?
Feb 24 09:52:04 alkyone kernel: Flags; bus-master 1, full 1; dirty 560482 current 560498.
Feb 24 09:52:04 alkyone kernel: Down list 07ffe940 vs. 07ffe920.
Feb 24 09:52:04 alkyone kernel: 0: 07ffe920 length 8000002a status 8000002a
Feb 24 09:52:04 alkyone kernel: 1: 07ffe930 length 80000052 status 80000052
Feb 24 09:52:04 alkyone kernel: 2: 07ffe940 length 8000009e status 0000009e
Feb 24 09:52:04 alkyone kernel: 3: 07ffe950 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 4: 07ffe960 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 5: 07ffe970 length 80000052 status 00000052
Feb 24 09:52:04 alkyone kernel: 6: 07ffe980 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 7: 07ffe990 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 8: 07ffe9a0 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 9: 07ffe9b0 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 10: 07ffe9c0 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 11: 07ffe9d0 length 8000002a status 0000002a
Feb 24 09:52:04 alkyone kernel: 12: 07ffe9e0 length 8000004e status 0000004e
Feb 24 09:52:04 alkyone kernel: 13: 07ffe9f0 length 80000052 status 00000052
Feb 24 09:52:04 alkyone kernel: 14: 07ffea00 length 8000005a status 0000005a
Feb 24 09:52:04 alkyone kernel: 15: 07ffea10 length 8000004e status 0000004e
Feb 24 09:52:08 alkyone kernel: eth0: transmit timed out, tx_status 00 status e000.
Feb 24 09:52:08 alkyone kernel: Flags; bus-master 1, full 1; dirty 560482 current 560498.
Feb 24 09:52:08 alkyone kernel: Down list 00000000 vs. 07ffe920.
Feb 24 09:52:08 alkyone kernel: 0: 07ffe920 length 8000002a status 8000002a
Feb 24 09:52:08 alkyone kernel: 1: 07ffe930 length 80000052 status 80000052
Feb 24 09:52:08 alkyone kernel: 2: 07ffe940 length 8000009e status 0000009e
Feb 24 09:52:08 alkyone kernel: 3: 07ffe950 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 4: 07ffe960 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 5: 07ffe970 length 80000052 status 00000052
Feb 24 09:52:08 alkyone kernel: 6: 07ffe980 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 7: 07ffe990 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 8: 07ffe9a0 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 9: 07ffe9b0 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 10: 07ffe9c0 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 11: 07ffe9d0 length 8000002a status 0000002a
Feb 24 09:52:08 alkyone kernel: 12: 07ffe9e0 length 8000004e status 0000004e
Feb 24 09:52:08 alkyone kernel: 13: 07ffe9f0 length 80000052 status 00000052
Feb 24 09:52:08 alkyone kernel: 14: 07ffea00 length 8000005a status 0000005a
Feb 24 09:52:08 alkyone kernel: 15: 07ffea10 length 8000004e status 0000004e
-------------------------------------------------------------------------------

the "Transmitter encountered 16 collisions" showed up only once and then the
"transmit timed out" etc. message repeated every few seconds filling up the
logfile until rebooting (doesn't respond anymore to "ping" etc).

note that

- this happend almost at the same time for both PCs (might be an external trigger?)
but the other 8 machines had no problems

- we had this problem once before for 2 other PCs: again starting almost at the same
time, once showing the "encountered 16 collisions" and then only "transmit timed out"...

- all the PCs have been remotely rebooted at the same time (+/- a few seconds),
so the "same time" might be an effect of "same uptime" instead of
an external trigger too.

- on one other PC (#5) which never showed network problems so far we got

Feb 16 19:17:19 maya kernel: floppy0: Unable to allocate DMA memory
Feb 16 19:19:44 maya kernel: floppy0: Unable to allocate DMA memory

one time after we installed the bus-master driver (never happend with the old driver
and never happend again so far; but OTOH we're using floppy only *very* rarely).

any idea what to try or test ? is there a newer version of this driver available?

btw: 8 of 10 PCs are single-PPro200 with 128MB, two are dual-PPro200 with 256MB
(running SMP kernel) and we have the same problems for both types pf PCs

in the mean time 2 more machines suffered this problem once and one machine
twice now (but no more coincidences in time) :-(

thanks,

Harald

-- 
All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \
koenig@tat.physik.uni-tuebingen.de                     ^^^^^       ^^^^^