Re: [Qemu-devel] [PATCH] Guest stop notification

[Jan Kiszka]

On 2011-12-01 18:19, Eric B Munson wrote:
> On Thu, 01 Dec 2011, Jan Kiszka wrote:
> 
>> On 2011-11-29 22:36, Eric B Munson wrote:
>>> Often when a guest is stopped from the qemu console, it will
>>> report spurious soft lockup warnings on resume.  There are
>>> kernel patches being discussed that will give the host the
>>> ability to tell the guest that it is being stopped and should
>>> ignore the soft lockup warning that generates.
>>> 
>>> Signed-off-by: Eric B Munson <address@hidden> Cc:
>>> address@hidden Cc: address@hidden Cc:
>>> address@hidden Cc: address@hidden Cc: address@hidden 
>>> Cc: address@hidden --- target-i386/kvm.c |    6
>>> ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)
>>> 
>>> diff --git a/target-i386/kvm.c b/target-i386/kvm.c index
>>> 5bfc21f..defd364 100644 --- a/target-i386/kvm.c +++
>>> b/target-i386/kvm.c @@ -336,12 +336,18 @@ static int
>>> kvm_inject_mce_oldstyle(CPUState *env) return 0; }
>>> 
>>> +static void kvm_put_guest_paused(CPUState *penv) +{ +
>>> kvm_vcpu_ioctl(penv, KVM_GUEST_PAUSED, 0); +}
>> 
>> I see no need in encapsulating this in a separate function.
>> 
> 
> The encapsulated function was from a previous idea, I will remove
> it for V2.
> 
>>> + static void cpu_update_state(void *opaque, int running,
>>> RunState state) { CPUState *env = opaque;
>>> 
>>> if (running) { env->tsc_valid = false; +
>>> kvm_put_guest_paused(env);
>> 
>> checkpatch.pl would have asked you to remove this tab.
> 
> Will change to spaces for V2.
> 
>> 
>> More general:
>> 
>> Why is this x86-only? If the kernel interface is x86-only, what
>> prevents making it generic right from the beginning?
>> 
>> Why do we need a new IOCTL for this? Was there no space left in
>> the kvm_run structure e.g. to pass this flag down on next vcpu
>> execution? No big deal, just wondering.
> 
> Thanks for your review/feedback.
> 
> When I started looking into this problem, the ioctl was the first
> suggestion I got for how to communicate from qemu to guest kernel.
> I don't see a technical reason that this could not be added to the
> kvm_run structure in one of the bytes currently used as padding.  I
> would prefer to keep the ioctl because I have the corresponding
> kernel patches out to work with this, however, if there is a strong
> preference for using kvm_run, I can rework both sets.

My feeling is that a run field would be more elegant, but I might be
wrong on this as well. In any case: You need KVM_CAP in your kernel
interface to announce the new feature and you have to sync in the new
kernel header into QEMU to make it build.

Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux