kernel-packages team mailing list archive

[Tim Gardner <tim.gardner@xxxxxxxxxxxxx>]

** Also affects: linux (Ubuntu Xenial)
   Importance: Medium
     Assignee: Canonical Kernel Team (canonical-kernel-team)
       Status: Triaged

** Changed in: linux (Ubuntu Xenial)
       Status: Triaged => Fix Committed

** Changed in: linux (Ubuntu Xenial)
     Assignee: Canonical Kernel Team (canonical-kernel-team) => Tim Gardner (timg-tpi)

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1556096

Title:
  [Ubuntu 16.04.1] RELEASE and ACQUIRE atomics on Power

Status in linux package in Ubuntu:
  Fix Committed
Status in linux source package in Xenial:
  Fix Committed

Bug description:
  Will Deacon proposed the RELEASE/ACQUIRE/RELAXED semantics for atomic
  operations(https://lkml.org/lkml/2015/8/6/498), which has been already
  merged in 4.3-rc1. More people tend to use it for better
  performance(https://lkml.org/lkml/2015/9/20/47). These semantics are
  proposed for better fitting the need of high scalability.

   However, the efficient implementation hasn't been merged yet. As we
  have a weaker memory order model on POWER, if we didn't implement
  these atomics efficently, there would be no benefit we can gain from
  these new operations.

  A proposal of an efficient implemenation has been sent out to
  upstream(https://lkml.org/lkml/2015/9/16/527), hopefully will get
  merged. Still there may be more atomics having RELEASE and ACQUIRE
  semantics, and there need more effort to implement them correctly.
  Also, we need to spread these new atomic variants to LTC kernel team
  to understand what they can provide and how to use them.

  > Was this feature already accepted upstream?

  Not yet, still waiting to get into the maintainer's tree, but there is
  no obvious objection to this feature.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1556096/+subscriptions