>>> On 13.06.17 at 20:39, wrote:
> On 06/13/2017 04:19 AM, Jan Beulich wrote:
> On 12.06.17 at 23:28, wrote:
>>> On Mon, 2017-06-12 at 13:01 -0400, Boris Ostrovsky wrote:
On 06/12/2017 04:08 AM, Jan Beulich wrote:
On 19.05.17 at 17:50, wrote:
>> Instead of scrubbing pages d
>>> On 13.06.17 at 20:20, wrote:
> +if ( node_need_scrub[node] )
> +{
> +if ( !get_node )
> +return node;
> +
> +dist = __node_distance(local_node, node);
> +if ( dist < shortest || closest == NUMA_NO_
On 06/13/2017 04:36 PM, Dario Faggioli wrote:
> On Tue, 2017-06-13 at 14:39 -0400, Boris Ostrovsky wrote:
>> On 06/13/2017 04:19 AM, Jan Beulich wrote:
>> On 12.06.17 at 23:28, wrote:
If some vcpu context work is scheduled, SCHEDULE_SOFTIRQ is
raised.
do_softirq() will call the
On Tue, 2017-06-13 at 14:39 -0400, Boris Ostrovsky wrote:
> On 06/13/2017 04:19 AM, Jan Beulich wrote:
> > > > > On 12.06.17 at 23:28, wrote:
> > > If some vcpu context work is scheduled, SCHEDULE_SOFTIRQ is
> > > raised.
> > > do_softirq() will call the scheduler, which will see that there
> > >
On 06/13/2017 04:19 AM, Jan Beulich wrote:
On 12.06.17 at 23:28, wrote:
>> On Mon, 2017-06-12 at 13:01 -0400, Boris Ostrovsky wrote:
>>> On 06/12/2017 04:08 AM, Jan Beulich wrote:
>>> On 19.05.17 at 17:50, wrote:
> Instead of scrubbing pages during guest destruction (from
> free_
+if ( node_need_scrub[node] )
+{
+if ( !get_node )
+return node;
+
+dist = __node_distance(local_node, node);
+if ( dist < shortest || closest == NUMA_NO_NODE )
+{
+
>>> On 12.06.17 at 23:28, wrote:
> On Mon, 2017-06-12 at 13:01 -0400, Boris Ostrovsky wrote:
>> On 06/12/2017 04:08 AM, Jan Beulich wrote:
>> > > > > On 19.05.17 at 17:50, wrote:
>> > >
>> > > Instead of scrubbing pages during guest destruction (from
>> > > free_heap_pages()) do this opportunist
>>> On 12.06.17 at 19:01, wrote:
> On 06/12/2017 04:08 AM, Jan Beulich wrote:
> On 19.05.17 at 17:50, wrote:
>>> Instead of scrubbing pages during guest destruction (from
>>> free_heap_pages()) do this opportunistically, from the idle loop.
>> This is too brief for my taste. In particular the
On Mon, 2017-06-12 at 13:01 -0400, Boris Ostrovsky wrote:
> On 06/12/2017 04:08 AM, Jan Beulich wrote:
> > > > > On 19.05.17 at 17:50, wrote:
> > >
> > > Instead of scrubbing pages during guest destruction (from
> > > free_heap_pages()) do this opportunistically, from the idle loop.
> >
> > This
On 06/12/2017 04:08 AM, Jan Beulich wrote:
On 19.05.17 at 17:50, wrote:
>> Instead of scrubbing pages during guest destruction (from
>> free_heap_pages()) do this opportunistically, from the idle loop.
> This is too brief for my taste. In particular the re-ordering ...
>
>> --- a/xen/arch/x86
>>> On 19.05.17 at 17:50, wrote:
> Instead of scrubbing pages during guest destruction (from
> free_heap_pages()) do this opportunistically, from the idle loop.
This is too brief for my taste. In particular the re-ordering ...
> --- a/xen/arch/x86/domain.c
> +++ b/xen/arch/x86/domain.c
> @@ -118
Instead of scrubbing pages during guest destruction (from
free_heap_pages()) do this opportunistically, from the idle loop.
Signed-off-by: Boris Ostrovsky
---
Changes in v4:
* Be careful with tasklets in idle_loop()
* Use per-cpu mapcache override
* Update node_to_scrub() algorithm to select clos
12 matches
Mail list logo
