ipv6: sit: Replace rcu_assign_pointer() with RCU_INIT_POINTER()
Subsystems (2)
PATCH 1/1 ipv6: sit: Replace rcu_assign_pointer() with RCU_INIT_POINTER()
According to RCU_INIT_POINTER()'s block comment 3.a, it can be used if
"3. The referenced data structure has already been exposed to readers either
at compile time or via rcu_assign_pointer() -and-
a. You have not made -any- reader-visible changes to this structure since
then".
This case fulfills the conditions above because between the rtnl_dereference() call and the rcu_assign_pointer() call there is no update of that value. Therefore, this patch makes the replacement.
The following Coccinelle semantic patch was used:
@@
@@
- rcu_assign_pointer
+ RCU_INIT_POINTER
(...,
(
rtnl_dereference(...)
|
rcu_dereference_protected(...)
) )
--- net/ipv6/sit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
net/ipv6/sit.c +1 -1
@@ -167,7 +167,7 @@ static void ipip6_tunnel_link(struct sit_net *sitn, struct ip_tunnel *t) {struct ip_tunnel __rcu **tp =ipip6_bucket (sitn, t); -rcu_assign_pointer (t->next,rtnl_dereference (*tp)); +RCU_INIT_POINTER (t->next,rtnl_dereference (*tp));rcu_assign_pointer (*tp, t); }
Signature · 1 line
1.9.1
David Miller
maintainerFrom: Andreea-Cristina Bernat <[email protected]>
Date: Tue, 26 Aug 2014 16:46:32 +0300
Quoted 23 lines
According to RCU_INIT_POINTER()'s block comment 3.a, it can be used if
"3. The referenced data structure has already been exposed to readers either
at compile time or viarcu_assign_pointer()-and-
a. You have not made -any- reader-visible changes to this structure since
then".This case fulfills the conditions above because between the
rtnl_dereference()call and thercu_assign_pointer()call there is no update of that value. Therefore, this patch makes the replacement.The following Coccinelle semantic patch was used:
@@
@@-
rcu_assign_pointer
+RCU_INIT_POINTER
(...,
(
rtnl_dereference(...)
|
rcu_dereference_protected(...)
) )
Quoted 1 line
Signed-off-by Andreea-Cristina Bernat <[email protected]>
These changes are not correct.
We need the memory barrier added by rcu_assign_pointer(), which occurs before the assignment, otherwise the RCU pointer update might be visible before the stores that initialize the rest of the structure in question.
I'm not applying any of these patches, sorry.