Re: [PATCH tip/core/rcu 0/55] Torture-test changes for 3.15

From: Paul E. McKenney
Date: Mon Feb 17 2014 - 20:19:13 EST


On Mon, Feb 17, 2014 at 04:48:00PM -0800, Josh Triplett wrote:
> On Mon, Feb 17, 2014 at 04:41:45PM -0800, Josh Triplett wrote:
> > On Mon, Feb 17, 2014 at 02:12:31PM -0800, Paul E. McKenney wrote:
> > > Hello!
> > >
> > > This series contains rcutorture changes, including adding a simple
> > > locktorture. Creating this locktorture while sharing the rcutorture
> > > infrastructure was the main point of this patch, but this effort
> > > uncovered a number of shortcomings in rcutorture, which this series
> > > also fixes.
> > >
> > > 1-6. Usability improvements in rcutorture scripting.
> > >
> > > 7-13. Enable concurrent rcutorture runs on systems with sufficient
> > > numbers of CPUs.
> > >
> > > 14. Print the results directory at the end of the test.
> > >
> > > 15,17-25,27-28,30,32,37-41,46-48.
> > > Abstract facilities from rcutorture module and scripting for later
> > > use by locktorture.
> > >
> > > 16. Don't create a results directory for dryruns.
> > >
> > > 26. Print date and time of each phase of torturing.
> > >
> > > 29. Issue a diagnostic if something does a system shutdown while
> > > rcutorture is running.
> > >
> > > 31. Apply ACCESS_ONCE() to racy fullstop accesses.
> > >
> > > 33. Clean up rcu_torture_init() error handling.
> > >
> > > 34. Announce kthread creation.
> > >
> > > 35. Clean up a number of rcutorture shutdown races, unifying the
> > > required shutdown actions into a new torture_kthread_stopping()
> > > function.
> > >
> > > 36. Add a missing return statement in rcu_torture_barrier_init().
> > >
> > > 42. Create a minimal locktorture module.
> > >
> > > 43-44. Add an on-purpose buggy RCU implementation to rcutorture to help
> > > test the tests.
> > >
> > > 45. Create a file for Kconfig parameters that are commmon across all
> > > rcutorture tests.
> > >
> > > 49. Add beginning set of config files for locktorture.
> > >
> > > 50. Avoid SEGV when cleanup-hooks function pointer is NULL.
> > >
> > > 51. Add locktorture plugin for kvm_recheck.sh.
> > >
> > > 52. Rename TREE_RCU-Kconfig.txt to avoid confusing scripts that
> > > look for Kconfig files, courtesy of Paul Bolle.
> > >
> > > 53. Retain output from kvm-test-1-run.sh script.
> > >
> > > 54. Add an on-purpose buggy lock implementation to locktorture to
> > > help test the tests.
> > >
> > > 55. Save kvm.sh progress messages to log so that they can be used
> > > for later timing analysis.
> >
> > I replied with comments on 15, 17, 19, 21, 35, and 37.
> >
> > This series also appears to be missing patches numbered 12, 16, 18, 26,
> > 38, 39, 43, and 53. (In some cases those patch numbers appear above but
> > the corresponding patches don't appear in subsequent mails; in other
> > cases those numbers appear missing above as well.)
> >
> > For the remaining patches:
> > Reviewed-by: Josh Triplett <josh@xxxxxxxxxxxxxxxx>
>
> It appears that the patches that appeared "missing" showed up after all,
> but for some reason those mails were delayed.
> Reviewed-by: Josh Triplett <josh@xxxxxxxxxxxxxxxx>
> for those as well.

Whew! And thank you again!

Thanx, Paul

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/