Hi all,
This is the long awaited report of the percpu-rwsem rework (sorry Juri).
IIRC (I really have trouble keeping up momentum on this series) I've addressed
all previous comments by Oleg and Davidlohr and Waiman and hope we can stick
this in tip/locking/core for inclusion in the next merge.
It has been cooked (thoroughly) in PREEMPT_RT, and not found wanting.
Any objections to me stuffing it in so we can all forget about it properly?
On 1/31/20 10:07 AM, Peter Zijlstra wrote:
> Hi all,
>
> This is the long awaited report of the percpu-rwsem rework (sorry Juri).
>
> IIRC (I really have trouble keeping up momentum on this series) I've addressed
> all previous comments by Oleg and Davidlohr and Waiman and hope we can stick
> this in tip/locking/core for inclusion in the next merge.
>
> It has been cooked (thoroughly) in PREEMPT_RT, and not found wanting.
>
> Any objections to me stuffing it in so we can all forget about it properly?
>
>
The patchset looks good to me.
Acked-by: Waiman Long <[email protected]>
Cheers,
Longman
On Fri, 31 Jan 2020, Peter Zijlstra wrote:
>Hi all,
>
>This is the long awaited report of the percpu-rwsem rework (sorry Juri).
>
>IIRC (I really have trouble keeping up momentum on this series) I've addressed
>all previous comments by Oleg and Davidlohr and Waiman and hope we can stick
>this in tip/locking/core for inclusion in the next merge.
>
>It has been cooked (thoroughly) in PREEMPT_RT, and not found wanting.
>
>Any objections to me stuffing it in so we can all forget about it properly?
Feel free to add my:
Reviewed-by: Davidlohr Bueso <[email protected]>
Thanks,
Davidlohr
On 2020-01-31 16:07:03 [+0100], Peter Zijlstra wrote:
> Hi all,
>
> This is the long awaited report of the percpu-rwsem rework (sorry Juri).
>
> IIRC (I really have trouble keeping up momentum on this series) I've addressed
> all previous comments by Oleg and Davidlohr and Waiman and hope we can stick
> this in tip/locking/core for inclusion in the next merge.
>
> It has been cooked (thoroughly) in PREEMPT_RT, and not found wanting.
I did not suck it into -RT earlier since it look like work-in-progress
(based on the review). Now if you feel confident, I will suck it in.
Thank you.
> Any objections to me stuffing it in so we can all forget about it properly?
Sebastian
Hi,
On 31/01/20 16:07, Peter Zijlstra wrote:
> Hi all,
>
> This is the long awaited report of the percpu-rwsem rework (sorry Juri).
>
> IIRC (I really have trouble keeping up momentum on this series) I've addressed
> all previous comments by Oleg and Davidlohr and Waiman and hope we can stick
> this in tip/locking/core for inclusion in the next merge.
>
> It has been cooked (thoroughly) in PREEMPT_RT, and not found wanting.
>
> Any objections to me stuffing it in so we can all forget about it properly?
FWIW, backported and tested again on downstream PREEMPT_RT kernel.
locktorture didn't find any problem and latencies look good.
Tested-by: Juri Lelli <[email protected]>
Thanks!
Juri
On Fri, Jan 31, 2020 at 04:07:03PM +0100, Peter Zijlstra wrote:
> This is the long awaited report of the percpu-rwsem rework (sorry Juri).
>
> IIRC (I really have trouble keeping up momentum on this series) I've addressed
> all previous comments by Oleg and Davidlohr and Waiman and hope we can stick
> this in tip/locking/core for inclusion in the next merge.
>
> It has been cooked (thoroughly) in PREEMPT_RT, and not found wanting.
>
> Any objections to me stuffing it in so we can all forget about it properly?
Whole series looks fine to me and it also passes my arm64 build tests, so:
Acked-by: Will Deacon <[email protected]>
Will