Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1605140 > unrolled thread
| Started by | Michalis Kokologiannakis <mixaskok@gmail.com> |
|---|---|
| First post | 2017-03-20 22:40 +0100 |
| Last post | 2017-03-21 19:20 +0100 |
| Articles | 2 — 2 participants |
Back to article view | Back to linux.kernel
[PATCH] doc: Update the comparisons rule in rcu_dereference.txt Michalis Kokologiannakis <mixaskok@gmail.com> - 2017-03-20 22:40 +0100
Re: [PATCH] doc: Update the comparisons rule in rcu_dereference.txt "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-03-21 19:20 +0100
| From | Michalis Kokologiannakis <mixaskok@gmail.com> |
|---|---|
| Date | 2017-03-20 22:40 +0100 |
| Subject | [PATCH] doc: Update the comparisons rule in rcu_dereference.txt |
| Message-ID | <tndbY-2bk-27@gated-at.bofh.it> |
When an RCU-protected pointer is fetched but never dereferenced rcu_access_pointer() should be used in place of rcu_dereference(). This commit explicitly records this very fact in Documentation/ RCU/rcu_dereference.txt, in order to prevent the usage of rcu_dereference() in comparisons. Signed-off-by: Michalis Kokologiannakis <mixaskok@gmail.com> --- Documentation/RCU/rcu_dereference.txt | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/Documentation/RCU/rcu_dereference.txt b/Documentation/RCU/rcu_dereference.txt index c0bf244..b2a613f 100644 --- a/Documentation/RCU/rcu_dereference.txt +++ b/Documentation/RCU/rcu_dereference.txt @@ -138,6 +138,15 @@ o Be very careful about comparing pointers obtained from This sort of comparison occurs frequently when scanning RCU-protected circular linked lists. + Note that if checks for being within an RCU read-side + critical section are not required and the pointer is never + dereferenced, rcu_access_pointer() should be used in place + of rcu_dereference(). The rcu_access_pointer() primitive + does not require an enclosing read-side critical section, + and also omits the smp_read_barrier_depends() included in + rcu_dereference(), which in turn should provide a small + performance gain in some CPUs (e.g., the DEC Alpha). + o The comparison is against a pointer that references memory that was initialized "a long time ago." The reason this is safe is that even if misordering occurs, the -- 2.1.4
[toc] | [next] | [standalone]
| From | "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
|---|---|
| Date | 2017-03-21 19:20 +0100 |
| Message-ID | <tnwxX-6Zw-7@gated-at.bofh.it> |
| In reply to | #1605140 |
On Mon, Mar 20, 2017 at 10:38:35PM +0100, Michalis Kokologiannakis wrote: > When an RCU-protected pointer is fetched but never dereferenced > rcu_access_pointer() should be used in place of rcu_dereference(). > This commit explicitly records this very fact in Documentation/ > RCU/rcu_dereference.txt, in order to prevent the usage of > rcu_dereference() in comparisons. > > Signed-off-by: Michalis Kokologiannakis <mixaskok@gmail.com> Queued for review, thank you! Thanx, Paul > --- > Documentation/RCU/rcu_dereference.txt | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/Documentation/RCU/rcu_dereference.txt b/Documentation/RCU/rcu_dereference.txt > index c0bf244..b2a613f 100644 > --- a/Documentation/RCU/rcu_dereference.txt > +++ b/Documentation/RCU/rcu_dereference.txt > @@ -138,6 +138,15 @@ o Be very careful about comparing pointers obtained from > This sort of comparison occurs frequently when scanning > RCU-protected circular linked lists. > > + Note that if checks for being within an RCU read-side > + critical section are not required and the pointer is never > + dereferenced, rcu_access_pointer() should be used in place > + of rcu_dereference(). The rcu_access_pointer() primitive > + does not require an enclosing read-side critical section, > + and also omits the smp_read_barrier_depends() included in > + rcu_dereference(), which in turn should provide a small > + performance gain in some CPUs (e.g., the DEC Alpha). > + > o The comparison is against a pointer that references memory > that was initialized "a long time ago." The reason > this is safe is that even if misordering occurs, the > -- > 2.1.4 >
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web