Re: [PATCH 2/2] Documentation: RCU: Refer to ptr_eq()

[Alan Stern]

On Sat, Sep 28, 2024 at 09:51:28AM -0400, Mathieu Desnoyers wrote:
> Refer to ptr_eq() in the rcu_dereference() documentation.
> 
> ptr_eq() is a mechanism that preserves address dependencies when
> comparing pointers, and should be favored when comparing a pointer
> obtained from rcu_dereference() against another pointer.
> 
> Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@xxxxxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxx>
> Cc: Will Deacon <will@xxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Boqun Feng <boqun.feng@xxxxxxxxx>
> Cc: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
> Cc: John Stultz <jstultz@xxxxxxxxxx>
> Cc: Neeraj Upadhyay <Neeraj.Upadhyay@xxxxxxx>
> Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
> Cc: Boqun Feng <boqun.feng@xxxxxxxxx>
> Cc: Frederic Weisbecker <frederic@xxxxxxxxxx>
> Cc: Joel Fernandes <joel@xxxxxxxxxxxxxxxxx>
> Cc: Josh Triplett <josh@xxxxxxxxxxxxxxxx>
> Cc: Uladzislau Rezki <urezki@xxxxxxxxx>
> Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
> Cc: Lai Jiangshan <jiangshanlai@xxxxxxxxx>
> Cc: Zqiang <qiang.zhang1211@xxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Waiman Long <longman@xxxxxxxxxx>
> Cc: Mark Rutland <mark.rutland@xxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Vlastimil Babka <vbabka@xxxxxxx>
> Cc: maged.michael@xxxxxxxxx
> Cc: Mateusz Guzik <mjguzik@xxxxxxxxx>
> Cc: Gary Guo <gary@xxxxxxxxxxx>
> Cc: Jonas Oberhauser <jonas.oberhauser@xxxxxxxxxxxxxxx>
> Cc: rcu@xxxxxxxxxxxxxxx
> Cc: linux-mm@xxxxxxxxx
> Cc: lkmm@xxxxxxxxxxxxxxx
> ---
>  Documentation/RCU/rcu_dereference.rst | 34 +++++++++++++++++++++++----
>  1 file changed, 29 insertions(+), 5 deletions(-)
> 
> diff --git a/Documentation/RCU/rcu_dereference.rst b/Documentation/RCU/rcu_dereference.rst
> index 2524dcdadde2..c36b8d1721f6 100644
> --- a/Documentation/RCU/rcu_dereference.rst
> +++ b/Documentation/RCU/rcu_dereference.rst
> @@ -104,11 +104,13 @@ readers working properly:
>  	after such branches, but can speculate loads, which can again
>  	result in misordering bugs.
>  
> --	Be very careful about comparing pointers obtained from
> -	rcu_dereference() against non-NULL values.  As Linus Torvalds
> -	explained, if the two pointers are equal, the compiler could
> -	substitute the pointer you are comparing against for the pointer
> -	obtained from rcu_dereference().  For example::
> +-	Use relational operators which preserve address dependencies
> +	(such as "ptr_eq()") to compare pointers obtained from

Nit: ptr_eq() is an inline function, not a relational operator.  Say 
"operations that" instead of "relational operators which".

> +	rcu_dereference() against non-NULL values or against pointers
> +	obtained from prior loads. As Linus Torvalds explained, if the
> +	two pointers are equal, the compiler could substitute the
> +	pointer you are comparing against for the pointer obtained from
> +	rcu_dereference().  For example::
>  
>  		p = rcu_dereference(gp);
>  		if (p == &default_struct)
> @@ -125,6 +127,23 @@ readers working properly:
>  	On ARM and Power hardware, the load from "default_struct.a"
>  	can now be speculated, such that it might happen before the
>  	rcu_dereference().  This could result in bugs due to misordering.
> +	Performing the comparison with "ptr_eq()" ensures the compiler
> +	does not perform such transformation.
> +
> +	If the comparison is against a pointer obtained from prior
> +	loads, the compiler is allowed to use either register for the

This is true even when the comparison is against a pointer obtained from 
a later load.  Just say "another pointer" instead of "a pointer obtained 
from prior loads".  (And why would someone need multiple loads to 
obtain a single pointer?)

Also, say "pointer" instead of "register".

> +	following accesses, which loses the address dependency and
> +	allows weakly-ordered architectures such as ARM and PowerPC
> +	to speculate the address-dependent load before rcu_dereference().
> +	For example::
> +
> +		p1 = READ_ONCE(gp);
> +		p2 = rcu_dereference(gp);
> +		if (p1 == p2)
> +			do_default(p2->a);

Here you should say that the compiler could use p1->a rather than p2->a, 
destroying the address dependency.  That's the whole point of this; you 
shouldn't skip over it.

> +
> +	Performing the comparison with "ptr_eq()" ensures the compiler
> +	preserves the address dependencies.
>  
>  	However, comparisons are OK in the following cases:
>  
> @@ -204,6 +223,11 @@ readers working properly:
>  		comparison will provide exactly the information that the
>  		compiler needs to deduce the value of the pointer.
>  
> +	When in doubt, use relational operators that preserve address

Again, "operations" instead of "relational operators".

Alan Stern

> +	dependencies (such as "ptr_eq()") to compare pointers obtained
> +	from rcu_dereference() against non-NULL values or against
> +	pointers obtained from prior loads.
> +
>  -	Disable any value-speculation optimizations that your compiler
>  	might provide, especially if you are making use of feedback-based
>  	optimizations that take data collected from prior runs.  Such
> -- 
> 2.39.2
>