Hi Marcelo,
a few hints on rcuification, sorry I reviewed the code so late:
On Fri, Jun 5, 2015, at 19:08, mleitner@xxxxxxxxxx wrote:
> From: Marcelo Ricardo Leitner <marcelo.leitner@xxxxxxxxx>
>
> That's needed for the next patch, so we break the lock inversion between
> netns_sctp->addr_wq_lock and socket lock on
> sctp_addr_wq_timeout_handler(). With this, we can traverse addr_waitq
> without taking addr_wq_lock, taking it just for the write operations.
>
> Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@xxxxxxxxx>
> ---
>
> Notes:
> v2->v3:
> placed break statement on sctp_free_addr_wq_entry()
> removed unnecessary spin_lock noticed by Neil
>
> include/net/netns/sctp.h | 2 +-
> net/sctp/protocol.c | 80
> +++++++++++++++++++++++++++++-------------------
> 2 files changed, 49 insertions(+), 33 deletions(-)
>
> diff --git a/include/net/netns/sctp.h b/include/net/netns/sctp.h
> index
> 3573a81815ad9e0efb6ceb721eb066d3726419f0..9e53412c4ed829e8e45777a6d95406d490dbaa75
> 100644
> --- a/include/net/netns/sctp.h
> +++ b/include/net/netns/sctp.h
> @@ -28,7 +28,7 @@ struct netns_sctp {
> * It is a list of sctp_sockaddr_entry.
> */
> struct list_head local_addr_list;
> - struct list_head addr_waitq;
> + struct list_head __rcu addr_waitq;
> struct timer_list addr_wq_timer;
> struct list_head auto_asconf_splist;
> spinlock_t addr_wq_lock;
> diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
> index
> 53b7acde9aa37bf3d4029c459421564d5270f4c0..9954fb8c9a9455d5ad7a627e2d7f9a1fef861fc2
> 100644
> --- a/net/sctp/protocol.c
> +++ b/net/sctp/protocol.c
> @@ -593,15 +593,47 @@ static void sctp_v4_ecn_capable(struct sock *sk)
> INET_ECN_xmit(sk);
> }
>
> +static void sctp_free_addr_wq(struct net *net)
> +{
> + struct sctp_sockaddr_entry *addrw;
> +
> + spin_lock_bh(&net->sctp.addr_wq_lock);
Instead of holding spin_lock_bh you need to hold rcu_read_lock_bh, so
kfree_rcu does not call free function at once (in theory ;) ).
> + del_timer(&net->sctp.addr_wq_timer);
> + list_for_each_entry_rcu(addrw, &net->sctp.addr_waitq, list) {
> + list_del_rcu(&addrw->list);
> + kfree_rcu(addrw, rcu);
> + }
> + spin_unlock_bh(&net->sctp.addr_wq_lock);
> +}
> +
> +/* As there is no refcnt on sctp_sockaddr_entry, we must check inside
> + * the lock if it wasn't removed from addr_waitq already, otherwise we
> + * could double-free it.
> + */
> +static void sctp_free_addr_wq_entry(struct net *net,
> + struct sctp_sockaddr_entry *addrw)
> +{
> + struct sctp_sockaddr_entry *temp;
> +
> + spin_lock_bh(&net->sctp.addr_wq_lock);
I don't think this spin_lock operation is needed. The del_timer
functions do synchronize themselves.
> + list_for_each_entry_rcu(temp, &net->sctp.addr_waitq, list) {
> + if (temp == addrw) {
> + list_del_rcu(&addrw->list);
> + kfree_rcu(addrw, rcu);
> + break;
> + }
> + }
> + spin_unlock_bh(&net->sctp.addr_wq_lock);
> +}
> +
> static void sctp_addr_wq_timeout_handler(unsigned long arg)
> {
> struct net *net = (struct net *)arg;
> - struct sctp_sockaddr_entry *addrw, *temp;
> + struct sctp_sockaddr_entry *addrw;
> struct sctp_sock *sp;
>
> - spin_lock_bh(&net->sctp.addr_wq_lock);
> -
> - list_for_each_entry_safe(addrw, temp, &net->sctp.addr_waitq,
> list) {
> + rcu_read_lock_bh();
> + list_for_each_entry_rcu(addrw, &net->sctp.addr_waitq, list) {
> pr_debug("%s: the first ent in wq:%p is addr:%pISc for cmd:%d at "
> "entry:%p\n", __func__, &net->sctp.addr_waitq, &addrw->a.sa,
> addrw->state, addrw);
> @@ -647,35 +679,20 @@ static void sctp_addr_wq_timeout_handler(unsigned
> long arg)
> #if IS_ENABLED(CONFIG_IPV6)
> free_next:
> #endif
> - list_del(&addrw->list);
> - kfree(addrw);
> - }
> - spin_unlock_bh(&net->sctp.addr_wq_lock);
> -}
> -
> -static void sctp_free_addr_wq(struct net *net)
> -{
> - struct sctp_sockaddr_entry *addrw;
> - struct sctp_sockaddr_entry *temp;
> -
> - spin_lock_bh(&net->sctp.addr_wq_lock);
> - del_timer(&net->sctp.addr_wq_timer);
> - list_for_each_entry_safe(addrw, temp, &net->sctp.addr_waitq,
> list) {
> - list_del(&addrw->list);
> - kfree(addrw);
> + sctp_free_addr_wq_entry(net, addrw);
> }
> - spin_unlock_bh(&net->sctp.addr_wq_lock);
> + rcu_read_unlock_bh();
> }
>
This code looks strange to me: You rcu_read_lock_bh and walk addr_waitq
list just to pass in pointers to the sctp_free_addr_wq_entry free
function, which then walks the list again just to compare the pointer?
> /* lookup the entry for the same address in the addr_waitq
> - * sctp_addr_wq MUST be locked
> + * rcu read MUST be locked
> */
> static struct sctp_sockaddr_entry *sctp_addr_wq_lookup(struct net *net,
> struct sctp_sockaddr_entry *addr)
> {
> struct sctp_sockaddr_entry *addrw;
>
> - list_for_each_entry(addrw, &net->sctp.addr_waitq, list) {
> + list_for_each_entry_rcu(addrw, &net->sctp.addr_waitq, list) {
> if (addrw->a.sa.sa_family != addr->a.sa.sa_family)
> continue;
> if (addrw->a.sa.sa_family == AF_INET) {
> @@ -702,7 +719,7 @@ void sctp_addr_wq_mgmt(struct net *net, struct
> sctp_sockaddr_entry *addr, int cm
> * new address after a couple of addition and deletion of that address
> */
>
> - spin_lock_bh(&net->sctp.addr_wq_lock);
> + rcu_read_lock_bh();
> /* Offsets existing events in addr_wq */
> addrw = sctp_addr_wq_lookup(net, addr);
> if (addrw) {
> @@ -710,22 +727,21 @@ void sctp_addr_wq_mgmt(struct net *net, struct
> sctp_sockaddr_entry *addr, int cm
> pr_debug("%s: offsets existing entry for %d, addr:%pISc "
> "in wq:%p\n", __func__, addrw->state, &addrw->a.sa,
> &net->sctp.addr_waitq);
> -
> - list_del(&addrw->list);
> - kfree(addrw);
> + sctp_free_addr_wq_entry(net, addrw);
> }
> - spin_unlock_bh(&net->sctp.addr_wq_lock);
> + rcu_read_unlock_bh();
> return;
> }
> + rcu_read_unlock_bh();
>
> /* OK, we have to add the new address to the wait queue */
> addrw = kmemdup(addr, sizeof(struct sctp_sockaddr_entry), GFP_ATOMIC);
> - if (addrw == NULL) {
> - spin_unlock_bh(&net->sctp.addr_wq_lock);
> + if (!addrw)
> return;
> - }
> addrw->state = cmd;
> - list_add_tail(&addrw->list, &net->sctp.addr_waitq);
> +
> + spin_lock_bh(&net->sctp.addr_wq_lock);
> + list_add_tail_rcu(&addrw->list, &net->sctp.addr_waitq);
list_rcu functions can in general run concurrently without spin_locks
taken, is this one necessary?
Bye,
Hannes
--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
