On Fri, Sep 06, 2024 at 12:54:37PM +0200, Juraj Marcin wrote:
> Hi Peter,
>
> On Thu, Sep 5, 2024 at 6:00 PM Peter Xu wrote:
> >
> > On Thu, Sep 05, 2024 at 05:32:46PM +0200, Juraj Marcin wrote:
> > > Hi Peter,
> >
> > Hi, Juraj,
> >
> > [...]
> >
> > > > unsigned int kvm_get_max_memslots(void)
Hi Peter,
On Thu, Sep 5, 2024 at 6:00 PM Peter Xu wrote:
>
> On Thu, Sep 05, 2024 at 05:32:46PM +0200, Juraj Marcin wrote:
> > Hi Peter,
>
> Hi, Juraj,
>
> [...]
>
> > > unsigned int kvm_get_max_memslots(void)
> > > {
> > > KVMState *s = KVM_STATE(current_accel());
> > > @@ -193,15 +247,20
On Thu, Sep 05, 2024 at 05:32:46PM +0200, Juraj Marcin wrote:
> Hi Peter,
Hi, Juraj,
[...]
> > unsigned int kvm_get_max_memslots(void)
> > {
> > KVMState *s = KVM_STATE(current_accel());
> > @@ -193,15 +247,20 @@ unsigned int kvm_get_free_memslots(void)
> > /* Called with KVMMemoryListen
Hi Peter,
On Thu, Sep 5, 2024 at 12:35 AM Peter Xu wrote:
>
> Zhiyi reported an infinite loop issue in VFIO use case. The cause of that
> was a separate discussion, however during that I found a regression of
> dirty sync slowness when profiling.
>
> Each KVMMemoryListerner maintains an array of
Zhiyi reported an infinite loop issue in VFIO use case. The cause of that
was a separate discussion, however during that I found a regression of
dirty sync slowness when profiling.
Each KVMMemoryListerner maintains an array of kvm memslots. Currently it's
statically allocated to be the max suppo
