8.1-RC2 - PCI fatal error or MCE triggered by USB/ehci on Sun
X4100M2?
John Baldwin
jhb at freebsd.org
Mon Jul 12 12:51:56 UTC 2010
On Monday, July 12, 2010 8:25:54 am Markus Gebert wrote:
>
> On 10.07.2010, at 19:37, Alan Cox wrote:
>
> > On Fri, Jul 9, 2010 at 6:53 PM, Markus Gebert <markus.gebert at hostpoint.ch>
wrote:
> > [snip]
> >
> > Yes, this hardware comes from Sun directly, but getting Sun (/Oracle)
support for this issue is gonna be tough. FreeBSD is unsupported, and in a
short test we couldn't reproduce the problem with a Linux kernel. While I
agree that a hardware issue has always been and still is a possibility to be
considered, the fact that we tested this on two machines remains as well as
the fact that 6.x, 7.x do not show the behavior. Another possibility is of
course, that the X4100 is prone to such issues and somehow 6.x and 7.x have
workarounds we're not aware of or just do something different in way so that
this issue does not get triggered.
> >
> >
> > 8.1 is our first release to have the driver for configuring and reporting
machine check exceptions enabled by default. Prior to 8.1, you had to
explicitly enable the driver at boot time.
>
>
> I was aware of that, but I don't think that it might be the cause. Disabling
MCA just makes the reporting go away, but the MCE and subsequent fatal trap
remain. With default BIOS settings, the OS does not even get a chance to
panic, the system just forces a reset before the OS could do anything. And, as
far as I can tell, that did not happen on previous stable branches.
Hmm with mca disabled in the loader you should not be getting any MCE's at all
as we don't enable the MCE interrupt in the CPU in that case. Are you
disabling it in the BIOS rather than loader.conf?
> Don't know though wether MCA changes the situation even when disabled in
loader.conf (hw.mca.enabled=0). I just checked our 7.2 setup, and MCA does not
seem to be in an 7.2 kernel, so I guess this was added to 8.0 and activated by
default in 8.1. To be honest, we did not check, wether 8.0 shows the same
behavior, but I guess running 8.1 with hw.mca.enabled=0 should pretty much
give the same situation as far as MCA is concerned.
7.3 has MCA support, but disabled by default.
> Is there a way to get rid of MCA completely? (as opposed to just "turning it
off" via loader.conf)
Turning it off in loader.conf does get rid of it completely as it prevents us
from initializing the MSRs.
--
John Baldwin
More information about the freebsd-stable
mailing list