On 8/19/25 6:00 AM, Sebastian Andrzej Siewior wrote:
On 2025-08-18 14:06:39 [-0400], Waiman Long wrote:
index 80c914f6eae7a..37b6a5670c2fa 100644
--- a/Documentation/locking/locktypes.rst
+++ b/Documentation/locking/locktypes.rst
@@ -204,6 +204,27 @@ per-CPU data structures on a non PREEMPT_RT kernel.
local_lock is not suitable to protect against preemption or interrupts on a
PREEMPT_RT kernel due to the PREEMPT_RT specific spinlock_t semantics.
+CPU local scope and bottom-half
+-------------------------------
+
+Per-CPU variables that are accessed only in softirq context should not rely on
+the assumption that this context is implicitly protected due to being
+non-preemptible. In a PREEMPT_RT kernel, softirq context is preemptible, and
+synchronizing every bottom-half-disabled section via implicit context results
+in an implicit per-CPU "big kernel lock."
+
+A local_lock_t together with local_lock_nested_bh() and
+local_unlock_nested_bh() for locking operations help to identify the locking
+scope.
+
+When lockdep is enabled, these functions verify that data structure access
+occurs within softirq context.
+Unlike local_lock(), local_unlock_nested_bh() does not disable preemption and
+does not add overhead when used without lockdep.
Should it be local_lock_nested_bh()? It doesn't make sense to compare
local_unlock_nested_bh() against local_lock(). In a PREEMPT_RT kernel,
local_lock() disables migration but not preemption.
Yes, it should have been the lock and not the unlock part. I mention
just preemption part here because it focuses on the !RT part compared to
local_lock() and that it adds no overhead.
The PREEMPT_RT part below mentions that it behaves as a real lock so
that should be enough (not to mention the migration part (technically
migration must be already disabled so we could omit disabling migration
here but it is just a counter increment/ decrement at this point so we
don't win much by doing so)).
I made the following:
@@ -219,11 +219,11 @@ scope.
When lockdep is enabled, these functions verify that data structure access
occurs within softirq context.
-Unlike local_lock(), local_unlock_nested_bh() does not disable preemption and
+Unlike local_lock(), local_lock_nested_bh() does not disable preemption and
does not add overhead when used without lockdep.
On a PREEMPT_RT kernel, local_lock_t behaves as a real lock and
-local_unlock_nested_bh() serializes access to the data structure, which allows
+local_lock_nested_bh() serializes access to the data structure, which allows
removal of serialization via local_bh_disable().
raw_spinlock_t and spinlock_t
Good?
LGTM, thanks!
Cheers,
Longman
