[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5fbd5ff1-8cb8-425f-be5f-7ed9fe4edf1c@paulmck-laptop>
Date: Mon, 21 Oct 2024 17:21:27 -0700
From: "Paul E. McKenney" <paulmck@...nel.org>
To: Andrii Nakryiko <andrii.nakryiko@...il.com>
Cc: rcu@...r.kernel.org, linux-kernel@...r.kernel.org, kernel-team@...a.com,
rostedt@...dmis.org, peterz@...radead.org, andrii@...nel.org
Subject: Re: [PATCH rcu] srcu: Guarantee non-negative return value from
srcu_read_lock()
On Mon, Oct 21, 2024 at 04:50:44PM -0700, Andrii Nakryiko wrote:
> On Mon, Oct 21, 2024 at 3:13 PM Paul E. McKenney <paulmck@...nel.org> wrote:
> >
> > For almost 20 years, the int return value from srcu_read_lock() has
> > been always either zero or one. This commit therefore documents the
> > fact that it will be non-negative.
> >
> > Signed-off-by: Paul E. McKenney <paulmck@...nel.org>
> > Cc: Peter Zijlstra <peterz@...radead.org>
> > Cc: Andrii Nakryiko <andrii@...nel.org
> >
> > diff --git a/include/linux/srcu.h b/include/linux/srcu.h
> > index bab1dae3f69e6..512a8c54ba5ba 100644
> > --- a/include/linux/srcu.h
> > +++ b/include/linux/srcu.h
> > @@ -238,13 +238,14 @@ void srcu_check_read_flavor(struct srcu_struct *ssp, int read_flavor);
> > * a mutex that is held elsewhere while calling synchronize_srcu() or
> > * synchronize_srcu_expedited().
> > *
> > - * The return value from srcu_read_lock() must be passed unaltered
> > - * to the matching srcu_read_unlock(). Note that srcu_read_lock() and
> > - * the matching srcu_read_unlock() must occur in the same context, for
> > - * example, it is illegal to invoke srcu_read_unlock() in an irq handler
> > - * if the matching srcu_read_lock() was invoked in process context. Or,
> > - * for that matter to invoke srcu_read_unlock() from one task and the
> > - * matching srcu_read_lock() from another.
> > + * The return value from srcu_read_lock() is guaranteed to be
> > + * non-negative. This value must be passed unaltered to the matching
> > + * srcu_read_unlock(). Note that srcu_read_lock() and the matching
> > + * srcu_read_unlock() must occur in the same context, for example, it is
> > + * illegal to invoke srcu_read_unlock() in an irq handler if the matching
> > + * srcu_read_lock() was invoked in process context. Or, for that matter to
> > + * invoke srcu_read_unlock() from one task and the matching srcu_read_lock()
> > + * from another.
>
> For uprobe work I'm using __srcu_read_lock() and __srcu_read_unlock().
> Presumably the same non-negative index will be returned/consumed there
> as well, right? Can we add a blurb to that effect for them as well?
Does the change shown below cover it?
> Otherwise LGTM, thanks!
>
> Acked-by: Andrii Nakryiko <andrii@...nel.org>
Thank you -- I will apply on my next rebase.
Thanx, Paul
> > */
> > static inline int srcu_read_lock(struct srcu_struct *ssp) __acquires(ssp)
> > {
------------------------------------------------------------------------
diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
index 07147efcb64d3..3d587bf2b2c12 100644
--- a/kernel/rcu/srcutree.c
+++ b/kernel/rcu/srcutree.c
@@ -738,7 +738,8 @@ EXPORT_SYMBOL_GPL(srcu_check_read_flavor);
/*
* Counts the new reader in the appropriate per-CPU element of the
* srcu_struct.
- * Returns an index that must be passed to the matching srcu_read_unlock().
+ * Returns a guaranteed non-negative index that must be passed to the
+ * matching srcu_read_unlock().
*/
int __srcu_read_lock(struct srcu_struct *ssp)
{
Powered by blists - more mailing lists