Re: [Qemu-devel] [PATCH v6 2/4] target-s390x: Migrate to new NMI interface

[Cornelia Huck]

On Thu, 12 Jun 2014 19:38:40 +1000
Alexey Kardashevskiy <address@hidden> wrote:

> On 06/12/2014 04:31 PM, Cornelia Huck wrote:
> > On Thu, 12 Jun 2014 03:03:01 +1000
> > Alexey Kardashevskiy <address@hidden> wrote:
> > 
> >> This implements an NMI interface for s390 machine.
> >>
> >> This removes #ifdef s390 branch in qmp_inject_nmi so new s390's
> >> nmi_monitor_handler() callback is going to be used for NMI.
> >>
> >> Since nmi_monitor_handler()-calling code is platform independent,
> >> CPUState::cpu_index is used instead of S390CPU::env.cpu_num.
> >> There should not be any change in behaviour as both @cpu_index and
> >> @cpu_num are global CPU numbers.
> >>
> >> Also, s390_cpu_restart() takes care of preforming operations in
> >> the specific CPU thread so no extra measure is required here either.
> >>
> >> Since the only error s390_cpu_restart() can return is ENOSYS, convert
> >> it to QERR_UNSUPPORTED.
> >>
> >> Signed-off-by: Alexey Kardashevskiy <address@hidden>
> >> ---
> >> Changes:
> >> v6:
> >> * supported NMI interface
> >>
> >> v5:
> >> * added ENOSYS -> QERR_UNSUPPORTED, qapi/qmp/qerror.h was added for this
> >>
> >> v4:
> >> * s/\<nmi\>/nmi_monitor_handler/
> >>
> >> v3:
> >> * now contains both old code removal and new code insertion, easier to
> >> track changes
> >>
> >> ---
> >> Is there any good reason to have @cpu_num in addition to @cpu_index?
> >> Just asking :)
> >> ---
> >>  cpus.c                 | 14 --------------
> >>  hw/s390x/s390-virtio.c | 31 +++++++++++++++++++++++++++++++
> >>  target-s390x/cpu.c     |  1 +
> >>  3 files changed, 32 insertions(+), 14 deletions(-)
> >>

> I pushed some version to address@hidden:aik/qemu.git , branch nmi-v7
> Please have a look and give it a go - I do not have s390 kernel/images
> handy. Thanks!

Gave it a try with both of the machines: Triggering 'nmi' from the
monitor still triggers the configured on_restart action, so this seems
to work as well as before.

> 
> It does not look like we really need a new file for NMI now.

Yes, I think the v7 code looks fine.