[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] #599161: Xen debug patch for the "clock shifts by 50 minutes" bug.



>>> On 24.11.12 at 13:06, Mauro <mrsanna1@xxxxxxxxx> wrote:
> (XEN) XXX plt_overflow: plt_now=569ed0c10fa3 plt_wrap=569ed0bd2ac3
> now=569ed0bf15a4 old_stamp=36f62 new_stamp=800377ce
> plt_stamp64=13d800377ce plt_mask=ffffffff tsc=cfdf69d72c82
> tsc_stamp=cfdf69ccc91d
> (XEN) XXX pts=13d80036988/13d80036988 sps=569ed0bd2ac3/569ed0bd2ac3
> (XEN) XXX ps=7:8baebcb8/7:8baebcb8

I.e. plt_wrap == stime_platform_stamp, which would imply
plt_stamp64 + plt_mask + 1 == platform_timer_stamp, which
visibly isn't the case.

Having to consider mis-compilation of some sort at this point,
can you please attach the xen-syms matching the xen.gz from
the run above?

The only other alternative would appear to be register or stack
corruption (which of the two might also be visible from the
xen-syms, so it would be needed for that determination too). As
the code in question runs with interrupts disabled, the only
sources for that would be NMI, MCE, or SMI. Adding logging of
instances of the former two would be possible, but otoh the
entry/exit code of them is uniform with other exceptions and
hence can pretty much be excluded to corrupt anything.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.