Skip to contentSkip to conversation

Regression fixes

GIT PULL Paul E. McKenney 2 participants 2 messages Oct 29, 2014 – Oct 30, 2014

lore
Share & tools
lore
#0 lore raw

Hello, Ingo,

This pull request contains a couple of fixes for RCU regressions:

1. Complete the work of commit dd56af42bd82 (rcu: Eliminate deadlock
between CPU hotplug and expedited grace periods), which was intended to allow synchronize_sched_expedited() to be safely used when holding locks acquired by CPU-hotplug notifiers. This commit makes the put_online_cpus() avoid the deadlock instead of just handling the get_online_cpus().

2. Complete the work of commit 35ce7f29a44a (rcu: Create rcuo
kthreads only for onlined CPUs), which was intended to allow RCU to avoid allocating unneeded kthreads on systems where the firmware says that there are more CPUs than are really present. This commit makes rcu_barrier() aware of the mismatch, so that it doesn't hang waiting for non-existent CPUs. (And yes, I do need to force this non-existent-CPU condition into my testing.)

Both of these changes have been tested by the submitters and have been exposed to -next testing. These changes are needed to avoid system hangs, and these system hangs have actually occurred during mainline testing.

These changes are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git urgent-for-mingo

for you to fetch changes up to d7e29933969e5ca7c112ce1368a07911f4485dc2:

rcu: Make rcu_barrier() understand about missing rcuo kthreads (2014-10-28 13:24:13 -0700)

----------------------------------------------------------------
Paul E. McKenney (2):
rcu: More on deadlock between CPU hotplug and expedited grace periods
rcu: Make rcu_barrier() understand about missing rcuo kthreads

 include/trace/events/rcu.h | 18 +++++++++---------
 kernel/cpu.c               | 14 +++++++++++++-
 kernel/rcu/tree.c          | 15 ++++++++++-----
 kernel/rcu/tree.h          |  1 +
 kernel/rcu/tree_plugin.h   | 33 +++++++++++++++++++++++++++++++++
 5 files changed, 66 insertions(+), 15 deletions(-)
#1 lore raw
Quoted 36 lines · Paul E. McKenney

Hello, Ingo,

This pull request contains a couple of fixes for RCU regressions:

1. Complete the work of commit dd56af42bd82 (rcu: Eliminate deadlock
between CPU hotplug and expedited grace periods), which was intended to allow synchronize_sched_expedited() to be safely used when holding locks acquired by CPU-hotplug notifiers. This commit makes the put_online_cpus() avoid the deadlock instead of just handling the get_online_cpus().

2. Complete the work of commit 35ce7f29a44a (rcu: Create rcuo
kthreads only for onlined CPUs), which was intended to allow RCU to avoid allocating unneeded kthreads on systems where the firmware says that there are more CPUs than are really present. This commit makes rcu_barrier() aware of the mismatch, so that it doesn't hang waiting for non-existent CPUs. (And yes, I do need to force this non-existent-CPU condition into my testing.)

Both of these changes have been tested by the submitters and have been exposed to -next testing. These changes are needed to avoid system hangs, and these system hangs have actually occurred during mainline testing.

These changes are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git urgent-for-mingo

for you to fetch changes up to d7e29933969e5ca7c112ce1368a07911f4485dc2:

rcu: Make rcu_barrier() understand about missing rcuo kthreads (2014-10-28 13:24:13 -0700)

----------------------------------------------------------------
Paul E. McKenney (2):
rcu: More on deadlock between CPU hotplug and expedited grace periods
rcu: Make rcu_barrier() understand about missing rcuo kthreads

Quoted 7 lines
 include/trace/events/rcu.h | 18 +++++++++---------
 kernel/cpu.c               | 14 +++++++++++++-
 kernel/rcu/tree.c          | 15 ++++++++++-----
 kernel/rcu/tree.h          |  1 +
 kernel/rcu/tree_plugin.h   | 33 +++++++++++++++++++++++++++++++++
 5 files changed, 66 insertions(+), 15 deletions(-)

Pulled, thanks a lot Paul!

Ingo