Re: [Qemu-devel] [PATCH 5/6 v5] deal with guest panicked event accoring to -onpanic parameter

[Wen Congyang]

At 07/03/2012 02:36 PM, Jan Kiszka Wrote:
> On 2012-07-03 08:07, Wen Congyang wrote:
>> At 06/28/2012 04:26 PM, Jan Kiszka Wrote:
>>> On 2012-06-28 03:15, Wen Congyang wrote:
>>>> At 06/27/2012 10:39 PM, Jan Kiszka Wrote:
>>>>> On 2012-06-27 09:02, Wen Congyang wrote:
>>>>>> When the guest is panicked, it will write 0x1 to the port KVM_PV_PORT.
>>>>>> So if qemu reads 0x1 from this port, we can do the folloing three
>>>>>> things according to the parameter -onpanic:
>>>>>> 1. emit QEVENT_GUEST_PANICKED only
>>>>>> 2. emit QEVENT_GUEST_PANICKED and pause the guest
>>>>>> 3. emit QEVENT_GUEST_PANICKED and poweroff the guest
>>>>>> 4. emit QEVENT_GUEST_PANICKED and reset the guest
>>>>>>
>>>>>> Note: if we emit QEVENT_GUEST_PANICKED only, and the management
>>>>>> application does not receive this event(the management may not
>>>>>> run when the event is emitted), the management won't know the
>>>>>> guest is panicked.
>>>>>>
>>>>>> Signed-off-by: Wen Congyang <address@hidden>
>>>>>> ---
>>>>>>  kvm-all.c       |  101 
>>>>>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>>>>  kvm-stub.c      |    9 +++++
>>>>>>  kvm.h           |    3 ++
>>>>>>  qemu-options.hx |   15 ++++++++
>>>>>>  vl.c            |   10 +++++
>>>>>>  5 files changed, 138 insertions(+), 0 deletions(-)
>>>>>>
>>>>>> diff --git a/kvm-all.c b/kvm-all.c
>>>>>> index f8e4328..9494dd2 100644
>>>>>> --- a/kvm-all.c
>>>>>> +++ b/kvm-all.c
>>>>>> @@ -19,6 +19,8 @@
>>>>>>  #include <stdarg.h>
>>>>>>  
>>>>>>  #include <linux/kvm.h>
>>>>>> +#include <linux/kvm_para.h>
>>>>>> +#include <asm/kvm_para.h>
>>>>>>  
>>>>>>  #include "qemu-common.h"
>>>>>>  #include "qemu-barrier.h"
>>>>>> @@ -32,6 +34,9 @@
>>>>>>  #include "bswap.h"
>>>>>>  #include "memory.h"
>>>>>>  #include "exec-memory.h"
>>>>>> +#include "iorange.h"
>>>>>> +#include "qemu-objects.h"
>>>>>> +#include "monitor.h"
>>>>>>  
>>>>>>  /* This check must be after config-host.h is included */
>>>>>>  #ifdef CONFIG_EVENTFD
>>>>>> @@ -1931,3 +1936,99 @@ int kvm_on_sigbus(int code, void *addr)
>>>>>>  {
>>>>>>      return kvm_arch_on_sigbus(code, addr);
>>>>>>  }
>>>>>> +
>>>>>> +/* Possible values for action parameter. */
>>>>>> +#define PANICKED_REPORT     1   /* emit QEVENT_GUEST_PANICKED only */
>>>>>> +#define PANICKED_PAUSE      2   /* emit QEVENT_GUEST_PANICKED and pause 
>>>>>> VM */
>>>>>> +#define PANICKED_POWEROFF   3   /* emit QEVENT_GUEST_PANICKED and quit 
>>>>>> VM */
>>>>>> +#define PANICKED_RESET      4   /* emit QEVENT_GUEST_PANICKED and reset 
>>>>>> VM */
>>>>>> +
>>>>>> +static int panicked_action = PANICKED_REPORT;
>>>>>> +
>>>>>> +static void kvm_pv_port_read(IORange *iorange, uint64_t offset, 
>>>>>> unsigned width,
>>>>>> +                             uint64_t *data)
>>>>>> +{
>>>>>> +    *data = (1 << KVM_PV_FEATURE_PANICKED);
>>>>>> +}
>>>>>> +
>>>>>> +static void panicked_mon_event(const char *action)
>>>>>> +{
>>>>>> +    QObject *data;
>>>>>> +
>>>>>> +    data = qobject_from_jsonf("{ 'action': %s }", action);
>>>>>> +    monitor_protocol_event(QEVENT_GUEST_PANICKED, data);
>>>>>> +    qobject_decref(data);
>>>>>> +}
>>>>>> +
>>>>>> +static void panicked_perform_action(void)
>>>>>> +{
>>>>>> +    switch (panicked_action) {
>>>>>> +    case PANICKED_REPORT:
>>>>>> +        panicked_mon_event("report");
>>>>>> +        break;
>>>>>> +
>>>>>> +    case PANICKED_PAUSE:
>>>>>> +        panicked_mon_event("pause");
>>>>>> +        vm_stop(RUN_STATE_GUEST_PANICKED);
>>>>>> +        break;
>>>>>> +
>>>>>> +    case PANICKED_POWEROFF:
>>>>>> +        panicked_mon_event("poweroff");
>>>>>> +        exit(0);
>>>>>> +        break;
>>>>>> +    case PANICKED_RESET:
>>>>>> +        panicked_mon_event("reset");
>>>>>> +        qemu_system_reset_request();
>>>>>> +        break;
>>>>>> +    }
>>>>>> +}
>>>>>> +
>>>>>> +static void kvm_pv_port_write(IORange *iorange, uint64_t offset, 
>>>>>> unsigned width,
>>>>>> +                              uint64_t data)
>>>>>> +{
>>>>>> +    if (data == KVM_PV_PANICKED) {
>>>>>> +        panicked_perform_action();
>>>>>> +    }
>>>>>> +}
>>>>>> +
>>>>>> +static void kvm_pv_port_destructor(IORange *iorange)
>>>>>> +{
>>>>>> +    g_free(iorange);
>>>>>> +}
>>>>>> +
>>>>>> +static IORangeOps pv_io_range_ops = {
>>>>>> +    .read = kvm_pv_port_read,
>>>>>> +    .write = kvm_pv_port_write,
>>>>>> +    .destructor = kvm_pv_port_destructor,
>>>>>> +};
>>>>>> +
>>>>>> +#if defined(KVM_PV_PORT)
>>>>>> +void kvm_pv_port_init(void)
>>>>>> +{
>>>>>> +    IORange *pv_io_range = g_malloc(sizeof(IORange));
>>>>>> +
>>>>>> +    iorange_init(pv_io_range, &pv_io_range_ops, KVM_PV_PORT, 1);
>>>>>> +    ioport_register(pv_io_range);
>>>>>
>>>>> This modeling is still not ok. We don't open-code ports anymore, we
>>>>> introduce devices. And this doesn't belong inro generic code as it is
>>>>
>>>> Do you mean introducing a new device instead of I/O port?
>>>
>>> I mean encapsulating the I/O registration (PIO or MMIO) in a QOM device
>>
>> A QOM device? Do you mean introduce a new device? If so, the guest should
>> have a driver to know such device. Another problem is: we cannot use
>> such device when the kernel is starting(the device's driver is not ready).
>> If we use a new device, I think virtio-serial is enough. The reason why
>> I do not use virtio-serial is: I want the feature can also work when the
>> kernel is starting.
> 
> I'm not talking about changing the interface to the guest, I'm talking
> about how to model it in QEMU. And that difference would be transparent
> to the guest. I pointed you to examples like hw/kvm/clock.c.

OK, I will read the code in hw/kvm/clock.c

Thanks for your help.

Wen Congyang

> 
> Jan
>