On 2015/04/02 10:36, Gu Zheng wrote:
Hi Kame, TJ,
On 04/01/2015 04:30 PM, Kamezawa Hiroyuki wrote:
On 2015/04/01 12:02, Tejun Heo wrote:
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
Now, hot-added cpus will have the lowest free cpu id.
Because of this, in most of
Hi Kame, TJ,
On 04/01/2015 04:30 PM, Kamezawa Hiroyuki wrote:
> On 2015/04/01 12:02, Tejun Heo wrote:
>> On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
>>> Now, hot-added cpus will have the lowest free cpu id.
>>>
>>> Because of this, in most of systems which has only
On 2015/04/01 12:02, Tejun Heo wrote:
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
Now, hot-added cpus will have the lowest free cpu id.
Because of this, in most of systems which has only cpu-hot-add, cpu-ids are
always
contiguous even after cpu hot add.
In enterprise,
On 2015/04/01 12:02, Tejun Heo wrote:
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
Now, hot-added cpus will have the lowest free cpu id.
Because of this, in most of systems which has only cpu-hot-add, cpu-ids are
always
contiguous even after cpu hot add.
In enterprise,
On 2015/04/02 10:36, Gu Zheng wrote:
Hi Kame, TJ,
On 04/01/2015 04:30 PM, Kamezawa Hiroyuki wrote:
On 2015/04/01 12:02, Tejun Heo wrote:
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
Now, hot-added cpus will have the lowest free cpu id.
Because of this, in most of
Hi Kame, TJ,
On 04/01/2015 04:30 PM, Kamezawa Hiroyuki wrote:
On 2015/04/01 12:02, Tejun Heo wrote:
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
Now, hot-added cpus will have the lowest free cpu id.
Because of this, in most of systems which has only cpu-hot-add,
On Tue, Mar 31, 2015 at 11:02:42PM -0400, Tejun Heo wrote:
> Ugh... so, cpu number allocation on hot-add is part of userland
> interface that we're locked into? Tying hotplug and id allocation
> order together usually isn't a good idea. What if the cpu up fails
> while running the notifiers?
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
> Now, hot-added cpus will have the lowest free cpu id.
>
> Because of this, in most of systems which has only cpu-hot-add, cpu-ids are
> always
> contiguous even after cpu hot add.
> In enterprise, this would be considered as
On 2015/04/01 0:28, Tejun Heo wrote:
Hello, Kamezawa.
On Tue, Mar 31, 2015 at 03:09:05PM +0900, Kamezawa Hiroyuki wrote:
But this may be considered as API change for most hot-add users.
Hmm... Why would it be? What can that possibly break?
Now, hot-added cpus will have the lowest free
Hello, Kamezawa.
On Tue, Mar 31, 2015 at 03:09:05PM +0900, Kamezawa Hiroyuki wrote:
> But this may be considered as API change for most hot-add users.
Hmm... Why would it be? What can that possibly break?
> So, for now, I vote for detemining ids at online but record it is a good way.
If we
On 2015/03/30 18:49, Gu Zheng wrote:
Hi Kame-san,
On 03/27/2015 12:42 AM, Kamezawa Hiroyuki wrote:
On 2015/03/27 0:18, Tejun Heo wrote:
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool->node) for all the possible cpu's
per cpu workqueue
On 2015/03/30 18:49, Gu Zheng wrote:
Hi Kame-san,
On 03/27/2015 12:42 AM, Kamezawa Hiroyuki wrote:
On 2015/03/27 0:18, Tejun Heo wrote:
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool-node) for all the possible cpu's
per cpu workqueue
On Tue, Mar 31, 2015 at 11:02:42PM -0400, Tejun Heo wrote:
Ugh... so, cpu number allocation on hot-add is part of userland
interface that we're locked into? Tying hotplug and id allocation
order together usually isn't a good idea. What if the cpu up fails
while running the notifiers? The ID
On 2015/04/01 0:28, Tejun Heo wrote:
Hello, Kamezawa.
On Tue, Mar 31, 2015 at 03:09:05PM +0900, Kamezawa Hiroyuki wrote:
But this may be considered as API change for most hot-add users.
Hmm... Why would it be? What can that possibly break?
Now, hot-added cpus will have the lowest free
On Wed, Apr 01, 2015 at 11:55:11AM +0900, Kamezawa Hiroyuki wrote:
Now, hot-added cpus will have the lowest free cpu id.
Because of this, in most of systems which has only cpu-hot-add, cpu-ids are
always
contiguous even after cpu hot add.
In enterprise, this would be considered as
Hello, Kamezawa.
On Tue, Mar 31, 2015 at 03:09:05PM +0900, Kamezawa Hiroyuki wrote:
But this may be considered as API change for most hot-add users.
Hmm... Why would it be? What can that possibly break?
So, for now, I vote for detemining ids at online but record it is a good way.
If we know
Hi Kame-san,
On 03/27/2015 12:42 AM, Kamezawa Hiroyuki wrote:
> On 2015/03/27 0:18, Tejun Heo wrote:
>> Hello,
>>
>> On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
>>> wq generates the numa affinity (pool->node) for all the possible cpu's
>>> per cpu workqueue at init stage, that
Hi Kame-san,
On 03/27/2015 12:42 AM, Kamezawa Hiroyuki wrote:
> On 2015/03/27 0:18, Tejun Heo wrote:
>> Hello,
>>
>> On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
>>> wq generates the numa affinity (pool->node) for all the possible cpu's
>>> per cpu workqueue at init stage, that
Hi Kame-san,
On 03/27/2015 12:42 AM, Kamezawa Hiroyuki wrote:
On 2015/03/27 0:18, Tejun Heo wrote:
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool-node) for all the possible cpu's
per cpu workqueue at init stage, that means the affinity
Hi Kame-san,
On 03/27/2015 12:42 AM, Kamezawa Hiroyuki wrote:
On 2015/03/27 0:18, Tejun Heo wrote:
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool-node) for all the possible cpu's
per cpu workqueue at init stage, that means the affinity
On 2015/03/27 0:18, Tejun Heo wrote:
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool->node) for all the possible cpu's
per cpu workqueue at init stage, that means the affinity of currently un-present
ones' may be incorrect, so we need to
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
> wq generates the numa affinity (pool->node) for all the possible cpu's
> per cpu workqueue at init stage, that means the affinity of currently
> un-present
> ones' may be incorrect, so we need to update the pool->node for the new
Hi Kame-san,
On 03/26/2015 11:12 AM, Kamezawa Hiroyuki wrote:
> On 2015/03/26 11:17, Gu Zheng wrote:
>> Yasuaki Ishimatsu found that with node online/offline, cpu<->node
>> relationship is established. Because workqueue uses a info which was
>> established at boot time, but it may be changed by
On 2015/03/27 0:18, Tejun Heo wrote:
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool-node) for all the possible cpu's
per cpu workqueue at init stage, that means the affinity of currently un-present
ones' may be incorrect, so we need to
Hi Kame-san,
On 03/26/2015 11:12 AM, Kamezawa Hiroyuki wrote:
On 2015/03/26 11:17, Gu Zheng wrote:
Yasuaki Ishimatsu found that with node online/offline, cpu-node
relationship is established. Because workqueue uses a info which was
established at boot time, but it may be changed by node
Hello,
On Thu, Mar 26, 2015 at 01:04:00PM +0800, Gu Zheng wrote:
wq generates the numa affinity (pool-node) for all the possible cpu's
per cpu workqueue at init stage, that means the affinity of currently
un-present
ones' may be incorrect, so we need to update the pool-node for the new added
On 2015/03/26 11:17, Gu Zheng wrote:
> Yasuaki Ishimatsu found that with node online/offline, cpu<->node
> relationship is established. Because workqueue uses a info which was
> established at boot time, but it may be changed by node hotpluging.
>
> Once pool->node points to a stale node,
Yasuaki Ishimatsu found that with node online/offline, cpu<->node
relationship is established. Because workqueue uses a info which was
established at boot time, but it may be changed by node hotpluging.
Once pool->node points to a stale node, following allocation failure
happens.
==
SLUB:
Yasuaki Ishimatsu found that with node online/offline, cpu-node
relationship is established. Because workqueue uses a info which was
established at boot time, but it may be changed by node hotpluging.
Once pool-node points to a stale node, following allocation failure
happens.
==
SLUB:
On 2015/03/26 11:17, Gu Zheng wrote:
Yasuaki Ishimatsu found that with node online/offline, cpu-node
relationship is established. Because workqueue uses a info which was
established at boot time, but it may be changed by node hotpluging.
Once pool-node points to a stale node, following
This is v4. Thank you for hints/commentes to previous versions.
I think this versions only contains necessary things and not invasive.
Tested several patterns of node hotplug and seems to work well.
Changes since v3
- removed changes against get_unbound_pool()
- remvoed codes in cpu offline
This is v4. Thank you for hints/commentes to previous versions.
I think this versions only contains necessary things and not invasive.
Tested several patterns of node hotplug and seems to work well.
Changes since v3
- removed changes against get_unbound_pool()
- remvoed codes in cpu offline
32 matches
Mail list logo