On Wed, Sep 04, 2024 at 05:46:55PM -0400, Peter Xu wrote:
> On Wed, Sep 04, 2024 at 11:38:28PM +0200, David Hildenbrand wrote:
> > On 04.09.24 23:34, Peter Xu wrote:
> > > On Wed, Sep 04, 2024 at 11:23:33PM +0200, David Hildenbrand wrote:
> > > >
> > > > > >
> > > > > > Then, you can remove the p
On Wed, Sep 04, 2024 at 11:38:28PM +0200, David Hildenbrand wrote:
> On 04.09.24 23:34, Peter Xu wrote:
> > On Wed, Sep 04, 2024 at 11:23:33PM +0200, David Hildenbrand wrote:
> > >
> > > > >
> > > > > Then, you can remove the parameter from kvm_slots_grow() completely
> > > > > and just call it
On 04.09.24 23:34, Peter Xu wrote:
On Wed, Sep 04, 2024 at 11:23:33PM +0200, David Hildenbrand wrote:
Then, you can remove the parameter from kvm_slots_grow() completely and just
call it
kvm_slots_double() and simplify a bit:
static bool kvm_slots_double(KVMMemoryListener *kml)
{
unsi
On Wed, Sep 04, 2024 at 11:23:33PM +0200, David Hildenbrand wrote:
>
> > >
> > > Then, you can remove the parameter from kvm_slots_grow() completely and
> > > just call it
> > > kvm_slots_double() and simplify a bit:
> > >
> > > static bool kvm_slots_double(KVMMemoryListener *kml)
> > > {
> > >
Then, you can remove the parameter from kvm_slots_grow() completely and just
call it
kvm_slots_double() and simplify a bit:
static bool kvm_slots_double(KVMMemoryListener *kml)
{
unsigned int i, nr_slots_new, cur = kml->nr_slots_allocated;
KVMSlot *slots;
nr_slots_new = MIN(c
On Wed, Sep 04, 2024 at 11:07:44PM +0200, David Hildenbrand wrote:
> On 04.09.24 21:16, 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.
>
On 04.09.24 21:16, 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 kvm memslots. Currently it's
sta
On Wed, Sep 04, 2024 at 10:43:23PM +0200, David Hildenbrand wrote:
> On 04.09.24 21:16, 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.
>
On 04.09.24 22:43, David Hildenbrand wrote:
On 04.09.24 21:16, 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
On 04.09.24 21:16, 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 kvm memslots. Currently it's
sta
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
11 matches
Mail list logo