Michal Privoznik <mpriv...@redhat.com> writes: > On 09/05/2017 10:36 PM, Eric Blake wrote: >> On 09/05/2017 08:22 AM, Michal Privoznik wrote: >>> Currently, the only time that users can set watchdog action is at >>> the start as all we expose is this -watchdog-action command line >>> argument. This is suboptimal when users want to plug the device >>> later via monitor. Alternatively, they might want to change the >>> action for already existing device on the fly. >>> >>> At the same time, drop local redefinition of the actions eum in >> >> s/eum/enum/ >> >>> watchdog.h in favour of the ones defined in schema. >>> >>> Inspired by: https://bugzilla.redhat.com/show_bug.cgi?id=1447169 >>> >>> Signed-off-by: Michal Privoznik <mpriv...@redhat.com> [...] >>> +++ b/qapi-schema.json >>> @@ -6539,3 +6539,12 @@ >>> # Since 2.9 >>> ## >>> { 'command': 'query-vm-generation-id', 'returns': 'GuidInfo' } >>> + >>> +## >>> +# @watchdog-set-action: >>> +# >>> +# Set watchdog action >>> +# >>> +# Since 2.11 >>> +## >>> +{ 'command': 'watchdog-set-action', 'data' : {'action': >>> 'WatchdogExpirationAction'} } >> >> Can a machine have more than one watchdog device, in which case you'd >> want a device name to select which watchdog gets which action? But the >> command-line version that you are replacing seems to be global, so I >> guess you're okay with this interface. >> > > Yeah, I don't think that a guest can have more than one watchdog: > > /qemu.git $ ./x86_64-softmmu/qemu-system-x86_64 \ > -watchdog ib700 -watchdog i6300esb > qemu-system-x86_64: -watchdog i6300esb: only one watchdog option may be > given
-watchdog is a thin wrapper around -device, and all it adds is this error. It's quite redundant. I recommend using -device instead. "-device i6300esb -device i6300esb" works. It's just a PCI device after all. > Also, would it make sense? I mean, what would be the benefit of having > two or more watchdogs? None. All it would accomplish is complicating things.
