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

Re: [Xen-devel] [PATCH] x86/hvm/viridian: set shutdown_code in response to CrashNotify



> -----Original Message-----
> From: Andrew Cooper
> Sent: 16 August 2018 10:40
> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>; xen-devel@xxxxxxxxxxxxxxxxxxxx
> Cc: Jan Beulich <jbeulich@xxxxxxxx>
> Subject: Re: [PATCH] x86/hvm/viridian: set shutdown_code in response to
> CrashNotify
> 
> On 10/08/18 16:59, Paul Durrant wrote:
> >> -----Original Message-----
> >> From: Andrew Cooper
> >> Sent: 10 August 2018 16:57
> >> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>; xen-
> devel@xxxxxxxxxxxxxxxxxxxx
> >> Cc: Jan Beulich <jbeulich@xxxxxxxx>
> >> Subject: Re: [PATCH] x86/hvm/viridian: set shutdown_code in response
> to
> >> CrashNotify
> >>
> >> On 10/08/18 16:43, Paul Durrant wrote:
> >>> When Windows writes the CrashNotify bit in the CRASH_CTL MSR then
> we
> >> know
> >>> it is crashing, so set the domain shutdown code appropriately.
> >>>
> >>> Signed-off-by: Paul Durrant <paul.durrant@xxxxxxxxxx>
> >>> ---
> >>> Cc: Jan Beulich <jbeulich@xxxxxxxx>
> >>> Cc: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> >>> ---
> >>>  xen/arch/x86/hvm/viridian.c | 4 ++++
> >>>  1 file changed, 4 insertions(+)
> >>>
> >>> diff --git a/xen/arch/x86/hvm/viridian.c b/xen/arch/x86/hvm/viridian.c
> >>> index 486065182c..294cf486cc 100644
> >>> --- a/xen/arch/x86/hvm/viridian.c
> >>> +++ b/xen/arch/x86/hvm/viridian.c
> >>> @@ -645,6 +645,10 @@ int wrmsr_viridian_regs(uint32_t idx, uint64_t
> val)
> >>>          if ( !ctl.u.CrashNotify )
> >>>              break;
> >>>
> >>> +        spin_lock(&d->shutdown_lock);
> >>> +        d->shutdown_code = SHUTDOWN_crash;
> >>> +        spin_unlock(&d->shutdown_lock);
> >> How does the domain eventually shut down?
> > I assume it shuts down when the guest writes to the PIIX.
> >
> >>   It feels slightly odd to have
> >> a shutdown code before the domain has finished executing code.
> >>
> > That's the norm. The PV drivers (if they are installed) set it via a 
> > schedop.
> This just makes sure it is set even if the PV drivers aren't there.
> 
> What happens if the user has configured windows to reboot after a crash?
> 

That's exactly what this is trying to fix. Without the patch a VM without PV 
drivers will appear to just shut down (because nothing will set the shutdown 
code to anything else) so the actions-after-crash will not take effect.

  Paul

> ~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

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