Re: sysbus failed assert for xen_sysdev

[Jason Andryuk <jandryuk@xxxxxxxxx>]

On Tue, Jun 23, 2020 at 7:46 AM Paul Durrant <xadimgnik@xxxxxxxxx> wrote:
>
> > -----Original Message-----
> > From: Markus Armbruster <armbru@xxxxxxxxxx>
> > Sent: 23 June 2020 09:41
> > To: Jason Andryuk <jandryuk@xxxxxxxxx>
> > Cc: Mark Cave-Ayland <mark.cave-ayland@xxxxxxxxxxxx>; Anthony PERARD 
> > <anthony.perard@xxxxxxxxxx>; xen-
> > devel <xen-devel@xxxxxxxxxxxxxxxxxxxx>; Paul Durrant <paul@xxxxxxx>; QEMU 
> > <qemu-devel@xxxxxxxxxx>
> > Subject: Re: sysbus failed assert for xen_sysdev
> >
> > Jason Andryuk <jandryuk@xxxxxxxxx> writes:
> > > Then it gets farther... until
> > > qemu-system-i386: hw/core/qdev.c:439: qdev_assert_realized_properly:
> > > Assertion `dev->realized' failed.
> > >
> > > dev->id is NULL. The failing device is:
> > > (gdb) p *dev.parent_obj.class.type
> > > $12 = {name = 0x555556207770 "cfi.pflash01",
> > >
>
> Having commented out the call to xen_be_init() entirely (and xen_bus_init() 
> for good measure) I also get this assertion failure, so
> I don't think is related.

Yes, this is something different.  pc_pflash_create() calls
qdev_new(TYPE_PFLASH_CFI01), but it is only realized in
pc_system_flash_map()...  and pc_system_flash_map() isn't called for
Xen.

Removing the call to pc_system_flash_create() from pc_machine_initfn()
lets QEMU startup and run a Xen HVM again.  xen_enabled() doesn't work
there since accelerators have not been initialized yes, I guess?

Regards,
Jason