Re: [Xen-devel] [PATCH] tools: hvmloader: attempt to SHUTDOWN_crash on BUG

[Ian Campbell <Ian.Campbell@xxxxxxxxxxxxx>]

Keir already applied this, but...

On Wed, 2011-04-20 at 18:24 +0100, Ian Jackson wrote:
> Ian Campbell writes ("[Xen-devel] [PATCH] tools: hvmloader: attempt to 
> SHUTDOWN_crash on BUG"):
> > tools: hvmloader: attempt to SHUTDOWN_crash on BUG
> > 
> > If we have got as far as having enabled hypercalls then this signals to the
> > toolstack that something went wrong. Otherwise they tend to assume the guest
> > has either shutdown or rebooted which can lead to simply trying again
> > repeatedly.
> 
> Surely in principle this is wrong.  Anything other than a deliberate
> decision by hvmloader to exit cleanly (how would that ever happen?)
> should be treated as a crash.

The toolstack can't tell the difference between hvmloader and the
subsequent bootloader then OS doing stuff and they may not be PV'd up.

In this case, the triple fault resulting from an hvmloader crash could
also be a legitimate attempt by a guest to reboot itself, since it's one
of the ways to reset the processor.

Ian.


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel