RFC Real-time elimination of RCU_SOFTIRQ
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v2 RFC Latest4 patches 0 repliesNew
- v1 RFC2 patches 8 repliesSuperseded
Patches in v2 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/4 | rcu: Enable elimination of Tree-RCU softirq processingDocumentation/ kernel/ kernel/ +1 more | +146 | 0 replies | |
| 2/4 | rcu: Check for wakeup-safe conditions in rcu_read_unlock_special()include/ kernel/ | +15 | 0 replies | |
| 3/4 | rcu: Only do rcu_read_unlock_special() wakeups if expeditedkernel/ | +10 | 0 replies | |
| 4/4 | rcu: Allow rcu_read_unlock_special() to raise_softirq() if in_irq()kernel/ | +1 | 0 replies |
Reviews trailers per patch
None of the 4 patches has a Reviewed-by or Acked-by.
| Patch | Mike Galbraith |
|---|---|
| 1/4 | T-b |
| 2/4 | |
| 3/4 | |
| 4/4 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH RFC v2 tip/core/rcu 0/4] Real-time elimination of RCU_SOFTIRQ
Hello!
This series is v2 of an early posting of code to add a boot-time option to move RCU's softirq processing to per-CPU kthreads. This is not done by default for performance reasons, nor are these reasons theoretical. In fact, earlier attempts to do just this were not met with silence.
The patches are as follows:
1. Enable elimination of Tree-RCU softirq processing via a new
rcutree.use_softirq kernel boot parameter. This defaults to 1, so boot with "rcutree.use_softirq=0" to move RCU_SOFTIRQ work to the rcuc kthreads. Courtesy of Sebastian Andrzej Siewior.
2. Improve wakeup-safety checks in rcu_read_unlock_special(),
thus allowing both rcuc kthreads and reasonably snappy expedited RCU grace periods.
3. Only do rcu_read_unlock_special() wakeups if expedited.
4. Allow rcu_read_unlock_special() to raise_softirq() if in_irq().
Changes since v1:
o Inspired by feedback from Peter Zijlstra, avoid expensive wakeups
unless there is a high probability that the current task or CPU is blocking an expedited grace period.
o Account for the possibility of long-term scheduling-clock-free
kernel sojourns in NO_HZ_FULL kernels.
o As an exception to the above, allow use of raise_softirq_irqoff()
when in an interrupt handler, in which case this is an extremely low-overhead option.
Thanx, Paul
------------------------------------------------------------------------
Documentation/admin-guide/kernel-parameters.txt | 6 include/linux/sched.h | 2 kernel/rcu/tree.c | 138 +++++++++++++++++-- kernel/rcu/tree.h | 2 kernel/rcu/tree_plugin.h | 167 +++++------------------- 5 files changed, 172 insertions(+), 143 deletions(-)