On 08/05/2011 02:44 PM, Peter Zijlstra wrote:

Hi,

Since yesterday is old, and yesterday's kernels aren't interesting,
here's a fresh one: 3.0.1-rt8.

Hmmm, what are reasonable values for CONFIG_RCU_BOOST_PRIO and 
CONFIG_RCU_BOOST_DELAY? (or at least what are you using?). I think 
CONFIG_RCU_BOOST should be working now so I should reenable the option 
for my next build.

-- Fernando

Aside from the obvious rebase to 3.0.1 changes include:

  - a fix from Rostedt for his cpupri rework involving memory ordering
  - a fix from me regarding this highmem muck, seeing as I mostly
    wrecked things last time around :-)
  - some lockdep annotations from upstream for things people were
    seeing here as well
  - a patch from Clark adding a sysfs file to recognize a PREEMP_RT
    kernel for udev -- is there really no saner solution !?
  - and a fresh raw_spinlock_t conversion of cpufreq_driver_lock, done
    just minutes before releasing this, to keep life interesting.

Still working on the known issues from last time..

The patch is at:

   http://www.kernel.org/pub/linux/kernel/projects/rt/patch-3.0.1-rt8.patch.bz2

The quilt series can be found here:

   http://www.kernel.org/pub/linux/kernel/projects/rt/patches-3.0.1-rt8.tar.bz2

Once kernel.org mirrors catch up that is..