diff options
| author | Frederic Weisbecker <[email protected]> | 2021-10-11 16:51:30 +0200 |
|---|---|---|
| committer | Paul E. McKenney <[email protected]> | 2021-12-07 16:24:44 -0800 |
| commit | 118e0d4a1bc85d4ecea0427e440a72d21ffbfa6a (patch) | |
| tree | 201c06531e250335ce5120a0331ed0c301f13a05 /scripts/gcc-plugins/randomize_layout_plugin.c | |
| parent | 614ddad17f22a22e035e2ea37a04815f50362017 (diff) | |
rcu/nocb: Make local rcu_nocb_lock_irqsave() safe against concurrent deoffloading
rcu_nocb_lock_irqsave() can be preempted between the call to
rcu_segcblist_is_offloaded() and the actual locking. This matters now
that rcu_core() is preemptible on PREEMPT_RT and the (de-)offloading
process can interrupt the softirq or the rcuc kthread.
As a result we may locklessly call into code that requires nocb locking.
In practice this is a problem while we accelerate callbacks on rcu_core().
Simply disabling interrupts before (instead of after) checking the NOCB
offload state fixes the issue.
Reported-and-tested-by: Valentin Schneider <[email protected]>
Tested-by: Sebastian Andrzej Siewior <[email protected]>
Signed-off-by: Frederic Weisbecker <[email protected]>
Cc: Valentin Schneider <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Sebastian Andrzej Siewior <[email protected]>
Cc: Josh Triplett <[email protected]>
Cc: Joel Fernandes <[email protected]>
Cc: Boqun Feng <[email protected]>
Cc: Neeraj Upadhyay <[email protected]>
Cc: Uladzislau Rezki <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Signed-off-by: Paul E. McKenney <[email protected]>
Diffstat (limited to 'scripts/gcc-plugins/randomize_layout_plugin.c')
0 files changed, 0 insertions, 0 deletions