2014-08-22 14:08:08

by Andreea Bernat

Subject: [PATCH] nfs: delegation: Replace rcu_assign_pointer() with RCU_INIT_POINTER()

The use of "rcu_assign_pointer()" is NULLing out the pointer.
According to RCU_INIT_POINTER()'s block comment:
"1. This use of RCU_INIT_POINTER() is NULLing out the pointer"
it is better to use it instead of rcu_assign_pointer() because it has a
smaller overhead.

The following Coccinelle semantic patch was used:
@@
@@

- rcu_assign_pointer
+ RCU_INIT_POINTER
(..., NULL)

Signed-off-by: Andreea-Cristina Bernat <[email protected]>
---
fs/nfs/delegation.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
index 5d8ccec..349e63b 100644
--- a/fs/nfs/delegation.c
+++ b/fs/nfs/delegation.c
@@ -249,7 +249,7 @@ nfs_detach_delegation_locked(struct nfs_inode *nfsi,
list_del_rcu(&delegation->super_list);
delegation->inode = NULL;
nfsi->delegation_state = 0;
- rcu_assign_pointer(nfsi->delegation, NULL);
+ RCU_INIT_POINTER(nfsi->delegation, NULL);
spin_unlock(&delegation->lock);
return delegation;
}
--
1.9.1

2014-08-22 14:49:10

by Trond Myklebust

Subject: Re: [PATCH] nfs: delegation: Replace rcu_assign_pointer() with RCU_INIT_POINTER()

On Fri, Aug 22, 2014 at 10:07 AM, Andreea-Cristina Bernat
<[email protected]> wrote:
> The use of "rcu_assign_pointer()" is NULLing out the pointer.
> According to RCU_INIT_POINTER()'s block comment:
> "1. This use of RCU_INIT_POINTER() is NULLing out the pointer"
> it is better to use it instead of rcu_assign_pointer() because it has a
> smaller overhead.
>
> The following Coccinelle semantic patch was used:
> @@
> @@
>
> - rcu_assign_pointer
> + RCU_INIT_POINTER
> (..., NULL)
>
> Signed-off-by: Andreea-Cristina Bernat <[email protected]>
> ---
> fs/nfs/delegation.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
> index 5d8ccec..349e63b 100644
> --- a/fs/nfs/delegation.c
> +++ b/fs/nfs/delegation.c
> @@ -249,7 +249,7 @@ nfs_detach_delegation_locked(struct nfs_inode *nfsi,
> list_del_rcu(&delegation->super_list);
> delegation->inode = NULL;
> nfsi->delegation_state = 0;
> - rcu_assign_pointer(nfsi->delegation, NULL);
> + RCU_INIT_POINTER(nfsi->delegation, NULL);
> spin_unlock(&delegation->lock);
> return delegation;
> }
> --
> 1.9.1

Sorry, but NACK.

We do need the memory barriers here, as the nfs_inode is globally
visible, and other threads may be checking nfsi->delegation_state
and/or delegation->inode in order to conclude whether or not there is
still a delegation associated with this inode.

Cheers
Trond
--
Trond Myklebust

Linux NFS client maintainer, PrimaryData

[email protected]