Re: [Qemu-devel] [PATCH] xen: handle inbound migration of VMs without ioreq server pages

[Paul Durrant <Paul.Durrant@citrix.com>]

> -----Original Message-----
> From: Stefano Stabellini [mailto:sstabellini@kernel.org]
> Sent: 11 August 2016 21:07
> To: Paul Durrant
> Cc: Anthony Perard; xen-devel@lists.xenproject.org; qemu-
> devel@nongnu.org; Stefano Stabellini
> Subject: RE: [PATCH] xen: handle inbound migration of VMs without ioreq
> server pages
> 
> On Tue, 9 Aug 2016, Paul Durrant wrote:
> > > -----Original Message-----
> > > From: Anthony PERARD [mailto:anthony.perard@citrix.com]
> > > Sent: 09 August 2016 16:19
> > > To: Paul Durrant
> > > Cc: xen-devel@lists.xenproject.org; qemu-devel@nongnu.org; Stefano
> > > Stabellini
> > > Subject: Re: [PATCH] xen: handle inbound migration of VMs without
> ioreq
> > > server pages
> > >
> > > On Mon, Aug 01, 2016 at 10:16:25AM +0100, Paul Durrant wrote:
> > > > VMs created on older versions on Xen will not have been provisioned
> with
> > > > pages to support creation of non-default ioreq servers. In this case
> > > > the ioreq server API is not supported and QEMU's only option is to fall
> > > > back to using the default ioreq server pages as it did prior to
> > > > commit 3996e85c ("Xen: Use the ioreq-server API when available").
> > > >
> > > > This patch therefore changes the code in xen_common.h to stop
> > > considering
> > > > a failure of xc_hvm_create_ioreq_server() as a hard failure but simply
> > > > as an indication that the guest is too old to support the ioreq server
> > > > API. Instead a boolean is set to cause reversion to old behaviour such
> > > > that the default ioreq server is then used.
> > > >
> > > > Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
> > > > Cc: Stefano Stabellini <sstabellini@kernel.org>
> > > > Cc: Anthony Perard <anthony.perard@citrix.com>
> > >
> > > There are just two lines too long:
> > >
> > > WARNING: line over 80 characters
> > > #31: FILE: include/hw/xen/xen_common.h:110:
> > > +static inline int xen_get_default_ioreq_server_info(xc_interface *xc,
> > > domid_t dom,
> > >
> > > WARNING: line over 80 characters
> > > #34: FILE: include/hw/xen/xen_common.h:113:
> > > +                                                    evtchn_port_t *bufioreq_evtchn)
> > >
> > > With that fixed: Acked-by: Anthony PERARD
> <anthony.perard@citrix.com>
> > >
> > > Thanks,
> > >
> >
> > Ok, I'll post v2 with those fixes and your ack. Thanks,
> 
> Thank you for fixing this bug and Thanks Anthony for the review.
> 
> I was about to commit it but my sense of style rebelled: I really don't
> like all the if statements. Too many! Sorry for coming in so late in
> commenting on a patch, I realize that it is unfair.
> 
> Would you be up for rewriting the fix so that instead of introducing
> 
>     if (use_default_ioreq_server) {
>         return;
>     }
> 
> in many functions, we turn xc_hvm_* calls into function pointers calls
> that get set to the right behavior depending on the success of
> xc_hvm_create_ioreq_server?
> 
> 
> The call sites would be something like:
> 
>     ioreq_server->unmap_io_range_from_ioreq_server(xc, dom, ioservid, 0,
>                                             start_addr, end_addr);
> 
> At boot time, if xc_hvm_create_ioreq_server returns error:
> 
>     ioreq_server = empty_stubs_functions;
> 
> else
> 
>     ioreq_server = useful_functions;
> 
> 
> What do you guys think?

Personally I don't think it's worth it. This is not a performance critical codepath but if you insist I will re-factor the code.

  Paul