diff mbox

[PATCHv2] bond_alb: don't disable softirq under bond_alb_xmit

Message ID 1326146497-27711-1-git-send-email-maxim.uvarov@oracle.com
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Maxim Uvarov Jan. 9, 2012, 10:01 p.m. UTC 
No need to lock soft irqs under bond_alb_xmit()
which already has softirq disabled.

Changes:
1. add non-bh/bh version to tlb_clear_slave()

2. represent BH and non BH hash table locks
_lock_rx_hashtbl_bh/_unlock_rx_hashtbl_bh
_lock_rx_hashtbl/_unlock_rx_hashtbl
_lock_tx_hashtbl_bh/_unlock_tx_hashtbl_bh
_lock_tx_hashtbl/_unlock_tx_hashtbl

Signed-off-by: Maxim Uvarov <maxim.uvarov@oracle.com>
Signed-off-by: Cong Wang <amwang@redhat.com>
---
 drivers/net/bonding/bond_alb.c |  112 +++++++++++++++++++++++++++-------------
 1 files changed, 76 insertions(+), 36 deletions(-)

Comments

Andy Gospodarek Jan. 9, 2012, 10:20 p.m. UTC | #1 
On Mon, Jan 09, 2012 at 02:01:37PM -0800, Maxim Uvarov wrote:
> No need to lock soft irqs under bond_alb_xmit()
> which already has softirq disabled.
> 
> Changes:
> 1. add non-bh/bh version to tlb_clear_slave()
> 
> 2. represent BH and non BH hash table locks
> _lock_rx_hashtbl_bh/_unlock_rx_hashtbl_bh
> _lock_rx_hashtbl/_unlock_rx_hashtbl
> _lock_tx_hashtbl_bh/_unlock_tx_hashtbl_bh
> _lock_tx_hashtbl/_unlock_tx_hashtbl
> 
> Signed-off-by: Maxim Uvarov <maxim.uvarov@oracle.com>
> Signed-off-by: Cong Wang <amwang@redhat.com>
> ---
>  drivers/net/bonding/bond_alb.c |  112 +++++++++++++++++++++++++++-------------
>  1 files changed, 76 insertions(+), 36 deletions(-)
> 

Did you test this with CONFIG_DEBUG_LOCKDEP=y at all?

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Maxim Uvarov Jan. 10, 2012, 3:09 a.m. UTC | #2 
On 01/09/2012 02:20 PM, Andy Gospodarek wrote:
> On Mon, Jan 09, 2012 at 02:01:37PM -0800, Maxim Uvarov wrote:
>> No need to lock soft irqs under bond_alb_xmit()
>> which already has softirq disabled.
>>
>> Changes:
>> 1. add non-bh/bh version to tlb_clear_slave()
>>
>> 2. represent BH and non BH hash table locks
>> _lock_rx_hashtbl_bh/_unlock_rx_hashtbl_bh
>> _lock_rx_hashtbl/_unlock_rx_hashtbl
>> _lock_tx_hashtbl_bh/_unlock_tx_hashtbl_bh
>> _lock_tx_hashtbl/_unlock_tx_hashtbl
>>
>> Signed-off-by: Maxim Uvarov<maxim.uvarov@oracle.com>
>> Signed-off-by: Cong Wang<amwang@redhat.com>
>> ---
>>   drivers/net/bonding/bond_alb.c |  112 +++++++++++++++++++++++++++-------------
>>   1 files changed, 76 insertions(+), 36 deletions(-)
>>
>
> Did you test this with CONFIG_DEBUG_LOCKDEP=y at all?
>

Just did some simple tests on bond0 with LOCKDEP. Do not see any warnings.

Maxim.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andy Gospodarek Jan. 10, 2012, 5:52 p.m. UTC | #3 
On Mon, Jan 09, 2012 at 07:09:32PM -0800, Maxim Uvarov wrote:
> On 01/09/2012 02:20 PM, Andy Gospodarek wrote:
> >On Mon, Jan 09, 2012 at 02:01:37PM -0800, Maxim Uvarov wrote:
> >>No need to lock soft irqs under bond_alb_xmit()
> >>which already has softirq disabled.
> >>
> >>Changes:
> >>1. add non-bh/bh version to tlb_clear_slave()
> >>
> >>2. represent BH and non BH hash table locks
> >>_lock_rx_hashtbl_bh/_unlock_rx_hashtbl_bh
> >>_lock_rx_hashtbl/_unlock_rx_hashtbl
> >>_lock_tx_hashtbl_bh/_unlock_tx_hashtbl_bh
> >>_lock_tx_hashtbl/_unlock_tx_hashtbl
> >>
> >>Signed-off-by: Maxim Uvarov<maxim.uvarov@oracle.com>
> >>Signed-off-by: Cong Wang<amwang@redhat.com>
> >>---
> >>  drivers/net/bonding/bond_alb.c |  112 +++++++++++++++++++++++++++-------------
> >>  1 files changed, 76 insertions(+), 36 deletions(-)
> >>
> >
> >Did you test this with CONFIG_DEBUG_LOCKDEP=y at all?
> >
> 
> Just did some simple tests on bond0 with LOCKDEP. Do not see any warnings.
> 

OK, I will test it today and let you know how it looks.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller Jan. 11, 2012, 8:52 p.m. UTC | #4 
From: Maxim Uvarov <maxim.uvarov@oracle.com>
Date: Mon,  9 Jan 2012 14:01:37 -0800

> No need to lock soft irqs under bond_alb_xmit()
> which already has softirq disabled.
> 
> Changes:
> 1. add non-bh/bh version to tlb_clear_slave()
> 
> 2. represent BH and non BH hash table locks
> _lock_rx_hashtbl_bh/_unlock_rx_hashtbl_bh
> _lock_rx_hashtbl/_unlock_rx_hashtbl
> _lock_tx_hashtbl_bh/_unlock_tx_hashtbl_bh
> _lock_tx_hashtbl/_unlock_tx_hashtbl
> 
> Signed-off-by: Maxim Uvarov <maxim.uvarov@oracle.com>
> Signed-off-by: Cong Wang <amwang@redhat.com>

Applied, thanks.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andy Gospodarek Jan. 11, 2012, 9:26 p.m. UTC | #5 
On Wed, Jan 11, 2012 at 12:52:46PM -0800, David Miller wrote:
> From: Maxim Uvarov <maxim.uvarov@oracle.com>
> Date: Mon,  9 Jan 2012 14:01:37 -0800
> 
> > No need to lock soft irqs under bond_alb_xmit()
> > which already has softirq disabled.
> > 
> > Changes:
> > 1. add non-bh/bh version to tlb_clear_slave()
> > 
> > 2. represent BH and non BH hash table locks
> > _lock_rx_hashtbl_bh/_unlock_rx_hashtbl_bh
> > _lock_rx_hashtbl/_unlock_rx_hashtbl
> > _lock_tx_hashtbl_bh/_unlock_tx_hashtbl_bh
> > _lock_tx_hashtbl/_unlock_tx_hashtbl
> > 
> > Signed-off-by: Maxim Uvarov <maxim.uvarov@oracle.com>
> > Signed-off-by: Cong Wang <amwang@redhat.com>
> 
> Applied, thanks.

Thanks, Dave.  I'm still testing, but everything looks good so far.  I
ran into an issue with my test system and it took longer to get going
than I expected.  I should be done later tonight and will post if I find
anything wrong.

It's probably too late, but you can feel free to add a

Signed-off-by: Andy Gospodarek <andy@greyhouse.net>

if you want.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
index 106b88a..342626f 100644
--- a/drivers/net/bonding/bond_alb.c
+++ b/drivers/net/bonding/bond_alb.c
@@ -99,16 +99,26 @@  static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
 
 /*********************** tlb specific functions ***************************/
 
-static inline void _lock_tx_hashtbl(struct bonding *bond)
+static inline void _lock_tx_hashtbl_bh(struct bonding *bond)
 {
 	spin_lock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
 }
 
-static inline void _unlock_tx_hashtbl(struct bonding *bond)
+static inline void _unlock_tx_hashtbl_bh(struct bonding *bond)
 {
 	spin_unlock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
 }
 
+static inline void _lock_tx_hashtbl(struct bonding *bond)
+{
+	spin_lock(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
+}
+
+static inline void _unlock_tx_hashtbl(struct bonding *bond)
+{
+	spin_unlock(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
+}
+
 /* Caller must hold tx_hashtbl lock */
 static inline void tlb_init_table_entry(struct tlb_client_info *entry, int save_load)
 {
@@ -129,14 +139,13 @@  static inline void tlb_init_slave(struct slave *slave)
 	SLAVE_TLB_INFO(slave).head = TLB_NULL_INDEX;
 }
 
-/* Caller must hold bond lock for read */
-static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_load)
+/* Caller must hold bond lock for read, BH disabled */
+static void __tlb_clear_slave(struct bonding *bond, struct slave *slave,
+			 int save_load)
 {
 	struct tlb_client_info *tx_hash_table;
 	u32 index;
 
-	_lock_tx_hashtbl(bond);
-
 	/* clear slave from tx_hashtbl */
 	tx_hash_table = BOND_ALB_INFO(bond).tx_hashtbl;
 
@@ -151,8 +160,15 @@  static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
 	}
 
 	tlb_init_slave(slave);
+}
 
-	_unlock_tx_hashtbl(bond);
+/* Caller must hold bond lock for read */
+static void tlb_clear_slave(struct bonding *bond, struct slave *slave,
+			 int save_load)
+{
+	_lock_tx_hashtbl_bh(bond);
+	__tlb_clear_slave(bond, slave, save_load);
+	_unlock_tx_hashtbl_bh(bond);
 }
 
 /* Must be called before starting the monitor timer */
@@ -169,7 +185,7 @@  static int tlb_initialize(struct bonding *bond)
 		       bond->dev->name);
 		return -1;
 	}
-	_lock_tx_hashtbl(bond);
+	_lock_tx_hashtbl_bh(bond);
 
 	bond_info->tx_hashtbl = new_hashtbl;
 
@@ -177,7 +193,7 @@  static int tlb_initialize(struct bonding *bond)
 		tlb_init_table_entry(&bond_info->tx_hashtbl[i], 0);
 	}
 
-	_unlock_tx_hashtbl(bond);
+	_unlock_tx_hashtbl_bh(bond);
 
 	return 0;
 }
@@ -187,12 +203,12 @@  static void tlb_deinitialize(struct bonding *bond)
 {
 	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
 
-	_lock_tx_hashtbl(bond);
+	_lock_tx_hashtbl_bh(bond);
 
 	kfree(bond_info->tx_hashtbl);
 	bond_info->tx_hashtbl = NULL;
 
-	_unlock_tx_hashtbl(bond);
+	_unlock_tx_hashtbl_bh(bond);
 }
 
 static long long compute_gap(struct slave *slave)
@@ -226,15 +242,13 @@  static struct slave *tlb_get_least_loaded_slave(struct bonding *bond)
 	return least_loaded;
 }
 
-/* Caller must hold bond lock for read */
-static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u32 skb_len)
+static struct slave *__tlb_choose_channel(struct bonding *bond, u32 hash_index,
+						u32 skb_len)
 {
 	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
 	struct tlb_client_info *hash_table;
 	struct slave *assigned_slave;
 
-	_lock_tx_hashtbl(bond);
-
 	hash_table = bond_info->tx_hashtbl;
 	assigned_slave = hash_table[hash_index].tx_slave;
 	if (!assigned_slave) {
@@ -263,22 +277,46 @@  static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
 		hash_table[hash_index].tx_bytes += skb_len;
 	}
 
-	_unlock_tx_hashtbl(bond);
-
 	return assigned_slave;
 }
 
+/* Caller must hold bond lock for read */
+static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index,
+					u32 skb_len)
+{
+	struct slave *tx_slave;
+	/*
+	 * We don't need to disable softirq here, becase
+	 * tlb_choose_channel() is only called by bond_alb_xmit()
+	 * which already has softirq disabled.
+	 */
+	_lock_tx_hashtbl(bond);
+	tx_slave = __tlb_choose_channel(bond, hash_index, skb_len);
+	_unlock_tx_hashtbl(bond);
+	return tx_slave;
+}
+
 /*********************** rlb specific functions ***************************/
-static inline void _lock_rx_hashtbl(struct bonding *bond)
+static inline void _lock_rx_hashtbl_bh(struct bonding *bond)
 {
 	spin_lock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
 }
 
-static inline void _unlock_rx_hashtbl(struct bonding *bond)
+static inline void _unlock_rx_hashtbl_bh(struct bonding *bond)
 {
 	spin_unlock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
 }
 
+static inline void _lock_rx_hashtbl(struct bonding *bond)
+{
+	spin_lock(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
+}
+
+static inline void _unlock_rx_hashtbl(struct bonding *bond)
+{
+	spin_unlock(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
+}
+
 /* when an ARP REPLY is received from a client update its info
  * in the rx_hashtbl
  */
@@ -288,7 +326,7 @@  static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
 	struct rlb_client_info *client_info;
 	u32 hash_index;
 
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	hash_index = _simple_hash((u8*)&(arp->ip_src), sizeof(arp->ip_src));
 	client_info = &(bond_info->rx_hashtbl[hash_index]);
@@ -303,7 +341,7 @@  static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
 		bond_info->rx_ntt = 1;
 	}
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 }
 
 static void rlb_arp_recv(struct sk_buff *skb, struct bonding *bond,
@@ -401,7 +439,7 @@  static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
 	u32 index, next_index;
 
 	/* clear slave from rx_hashtbl */
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	rx_hash_table = bond_info->rx_hashtbl;
 	index = bond_info->rx_hashtbl_head;
@@ -432,7 +470,7 @@  static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
 		}
 	}
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 
 	write_lock_bh(&bond->curr_slave_lock);
 
@@ -489,7 +527,7 @@  static void rlb_update_rx_clients(struct bonding *bond)
 	struct rlb_client_info *client_info;
 	u32 hash_index;
 
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	hash_index = bond_info->rx_hashtbl_head;
 	for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
@@ -507,7 +545,7 @@  static void rlb_update_rx_clients(struct bonding *bond)
 	 */
 	bond_info->rlb_update_delay_counter = RLB_UPDATE_DELAY;
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 }
 
 /* The slave was assigned a new mac address - update the clients */
@@ -518,7 +556,7 @@  static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
 	int ntt = 0;
 	u32 hash_index;
 
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	hash_index = bond_info->rx_hashtbl_head;
 	for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
@@ -538,7 +576,7 @@  static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
 		bond_info->rlb_update_retry_counter = RLB_UPDATE_RETRY;
 	}
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 }
 
 /* mark all clients using src_ip to be updated */
@@ -709,7 +747,7 @@  static void rlb_rebalance(struct bonding *bond)
 	int ntt;
 	u32 hash_index;
 
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	ntt = 0;
 	hash_index = bond_info->rx_hashtbl_head;
@@ -727,7 +765,7 @@  static void rlb_rebalance(struct bonding *bond)
 	if (ntt) {
 		bond_info->rx_ntt = 1;
 	}
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 }
 
 /* Caller must hold rx_hashtbl lock */
@@ -751,7 +789,7 @@  static int rlb_initialize(struct bonding *bond)
 		       bond->dev->name);
 		return -1;
 	}
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	bond_info->rx_hashtbl = new_hashtbl;
 
@@ -761,7 +799,7 @@  static int rlb_initialize(struct bonding *bond)
 		rlb_init_table_entry(bond_info->rx_hashtbl + i);
 	}
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 
 	/* register to receive ARPs */
 	bond->recv_probe = rlb_arp_recv;
@@ -773,13 +811,13 @@  static void rlb_deinitialize(struct bonding *bond)
 {
 	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
 
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	kfree(bond_info->rx_hashtbl);
 	bond_info->rx_hashtbl = NULL;
 	bond_info->rx_hashtbl_head = RLB_NULL_INDEX;
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 }
 
 static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
@@ -787,7 +825,7 @@  static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
 	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
 	u32 curr_index;
 
-	_lock_rx_hashtbl(bond);
+	_lock_rx_hashtbl_bh(bond);
 
 	curr_index = bond_info->rx_hashtbl_head;
 	while (curr_index != RLB_NULL_INDEX) {
@@ -812,7 +850,7 @@  static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
 		curr_index = next_index;
 	}
 
-	_unlock_rx_hashtbl(bond);
+	_unlock_rx_hashtbl_bh(bond);
 }
 
 /*********************** tlb/rlb shared functions *********************/
@@ -1320,7 +1358,9 @@  int bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
 		res = bond_dev_queue_xmit(bond, skb, tx_slave->dev);
 	} else {
 		if (tx_slave) {
-			tlb_clear_slave(bond, tx_slave, 0);
+			_lock_tx_hashtbl(bond);
+			__tlb_clear_slave(bond, tx_slave, 0);
+			_unlock_tx_hashtbl(bond);
 		}
 	}