Messages in this thread | | | Date | Tue, 12 Jan 1999 15:29:25 -0700 | From | Al Youngwerth <> | Subject | Re: Update: 2.0 kernels, tulip driver, crashes and reboots (long) |
| |
At 12:48 PM 1/12/99 +0000, you wrote: >> I think we're getting close on this. Your ideas and comments on the above >> are welcome and appreciated. > >One more suggestion. Drop the 2.0.31 tulip.c into 2.0.36 and see what that >one does. That will give clues if its related to driver revisions (although if >its the driver, the board,... is another matter of course). >
I believe the stock 2.0.31 kernel uses 0.87 tulip (please correct me if I'm wrong). We have tried this with 2.0.36 and those systems do reboot. However, your inquiry caused me to check the 2.0.31 kernel we're testing and it turns out we were running .90f on these systems (my technician and I had a misunderstanding when he setup the units).
.90f with 2.0.36 locks and reboots, with 2.0.31 it seems to be stable. I'm testing .89K/PNIC/2.0.31 at 100mbit now with the netperf/hdparm -t test and it has been running for about 4 hours without a failure. .89K/PNIC/2.0.36 running at 100mbit fails within 30 minutes reliably. Interestingly enough .89K/PNIC/2.0.36 running at 10mbit has been running for 3+ hours now. I will also setup a .90f/PNIC/2.0.36 test running at 100mbit.
In other news, ne2000/2.0.36 running netperf/hdparm -t test rebooted once overnight. The Intel TX systems running .89K/PNIC/2.0.36 running the netperf/hdparm -t test have not failed. Two systems running just the hdparm -t test on 2.0.36 kernels (one ne2000, one .89K/PNIC) did not fail overnight.
The big common thread here that is emerging for me is that all combinations of VPX motherboards, network cards and 2.0.35/36 lockup or reboot. I have not been able to get any combination of motherboards, network cards and a 2.0.31 kernel to reboot, yet.
Anybody care to suggest patch areas between 2.0.31 and 2.0.36 to look for this sort of problem?
Thanks,
Al Youngwerth alberty@apexxtech.com
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |