On 03/07/2017 07:29 AM, Launchpad Bug Tracker wrote:
> bugproxy (bugproxy) has assigned this bug to you for Ubuntu:
>
> There is a performance problem issue with h-prod. I can see that h-prod
> calls are only waking hardware threads from a h-cede call when the
> caller is on the same core as the thread being awoken (sibling theads).
> For cross core prods the target thread remains in h-cede until its next
> decrementer interrupt. That is causing guest OS some significant
> dispatching delays.
>
> The fix is now upstream as commit 8464c8842de2 ("KVM: PPC: Book3S HV:
> Fix H_PROD to actually wake the target vcpu", 2016-12-06) in v4.11-rc1.
>
> ** Affects: ubuntu
> Importance: Undecided
> Assignee: Taco Screen team (taco-screen-team)
> Status: New
>
>
> ** Tags: architecture-ppc64le bugnameltc-152309 severity-high targetmilestone-inin16042
--
Michael Hohnbaum
OIL Program Manager
Power (ppc64el) Development Project Manager
Canonical, Ltd.
Leann,
A kernel patch request from IBM.
On 03/07/2017 07:29 AM, Launchpad Bug Tracker wrote: ppc64le bugnameltc-152309 severity-high targetmilestone -inin16042
> bugproxy (bugproxy) has assigned this bug to you for Ubuntu:
>
> There is a performance problem issue with h-prod. I can see that h-prod
> calls are only waking hardware threads from a h-cede call when the
> caller is on the same core as the thread being awoken (sibling theads).
> For cross core prods the target thread remains in h-cede until its next
> decrementer interrupt. That is causing guest OS some significant
> dispatching delays.
>
> The fix is now upstream as commit 8464c8842de2 ("KVM: PPC: Book3S HV:
> Fix H_PROD to actually wake the target vcpu", 2016-12-06) in v4.11-rc1.
>
> ** Affects: ubuntu
> Importance: Undecided
> Assignee: Taco Screen team (taco-screen-team)
> Status: New
>
>
> ** Tags: architecture-
--
Michael Hohnbaum
OIL Program Manager
Power (ppc64el) Development Project Manager
Canonical, Ltd.