Re: [PATCH] ns16550c: avoid crash in ns16550_endboot in PV shim mode

[David Woodhouse <dwmw2@xxxxxxxxxxxxx>]

On Fri, 2023-10-20 at 15:29 +0200, Roger Pau Monné wrote:
> On Fri, Oct 20, 2023 at 02:25:35PM +0100, Andrew Cooper wrote:
> > On 20/10/2023 11:29 am, David Woodhouse wrote:
> > > On Fri, 2023-10-20 at 11:14 +0100, Andrew Cooper wrote:
> > > > On 19/10/2023 5:21 pm, David Woodhouse wrote:
> > > > > From: David Woodhouse <dwmw@xxxxxxxxxxxx>
> > > > > 
> > > > > In shim mode there is no hardware_domain. Dereferencing the pointer
> > > > > doesn't end well.
> > > > > 
> > > > > Signed-off-by: David Woodhouse <dwmw@xxxxxxxxxxxx>
> > > > > ---
> > > > > This is about as far as I got in my abortive attempt to use the PV 
> > > > > shim
> > > > > without an actual PV console being provided by the HVM hosting
> > > > > environment. It still doesn't pass the guest's console through to
> > > > > serial; that only seems to shim to an actual PV console.
> > > > There's no such thing as a Xen VM without a PV console.
> > > Huh? There are literally millions of them.
> > 
> > I'm very prepared to believe there are millions which don't overtly
> > malfunction when you don't fill in the HVM Params. 
> > 
> > Which is very different from saying "there's a way in the Xen guest ABI
> > to express 'you don't have a PV console' ".
> 
> FWIW, Linux assumes that either the console page or the event channel
> being 0 implies no console available [0], so I guess that's the ABI
> now.

Or if the HVMOP_get_param call returns an error.

> Roger.
> 
> [0] 
> https://elixir.bootlin.com/linux/latest/source/drivers/tty/hvc/hvc_xen.c#L258

I'm not convinced I believe what the comment says there about evtchn 0
being theoretically valid. I don't believe zero is a valid evtchn#, is
it?

Attachment: smime.p7s
Description: S/MIME cryptographic signature