8.1-RC2 - PCI fatal error or MCE triggered by USB/ehci on Sun X4100M2?

[Markus Gebert]

On 12.07.2010, at 14:41, Markus Gebert wrote:

> In the meantime, I'll try harder to reproduce the MCE on current...


Well, I can't. It's been running the test load for almost 24 hours without any sign of problems. The kernel config is the same as under 8.1, GENERIC+ipfw (USB not excluded!). Our 8.1 test config includes the CURRENT debug options, so as far as I can tell we should be testing under the same conditions. Yet I don't know wether CURRENT includes any additional debug magic I'm not aware of, which could slow the system down an prevent the issue from occurring.

Another difference to our primary 8.1 test machine is that we didn't use ZFS on CURRENT. But since we are able to crash 8.1 with UFS, I don't think this will matter. Still I'm going to try to get together a CURRENT ZFS setup on the second machine.

But, so far, bottom line seems to be:

- no problems on 6.x, 7.x and CURRENT
- no problems on Linux (that was only a short test though)
- crash within a minute on 8.1 (with ehci in the kernel)

Unfortunately, I have not been able to get anything useful out the svn commit logs, which could explain this. Maybe someone else has an idea what could have changed between 7 and 8 to break it, and again between 8 and CURRENT to magically fix it again.


Markus