diff mbox

eql: dont rely on HZ=100

Message ID 1332838320.3248.6.camel@edumazet-laptop
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Eric Dumazet March 27, 2012, 8:52 a.m. UTC 
Le mardi 27 mars 2012 à 09:17 +0100, David Woodhouse a écrit :

> Thanks. I blame the baby for that one too. Once she's been taken
> swimming and dropped off at nursery and I've had at least two more cups
> of tea, I'll have another go :)
> 

Looking at this driver, it seems it depends on HZ=100 ?

[PATCH] eql: dont rely on HZ=100

HZ is more likely to be 1000 these days.

timer handlers are run from softirq, no need to disable bh

skb priority 1 is TC_PRIO_FILLER

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
 drivers/net/eql.c      |    7 ++++---
 include/linux/if_eql.h |    2 +-
 2 files changed, 5 insertions(+), 4 deletions(-)



--
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

Comments

David Miller March 28, 2012, 2:47 a.m. UTC | #1 
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 27 Mar 2012 01:52:00 -0700

> [PATCH] eql: dont rely on HZ=100
> 
> HZ is more likely to be 1000 these days.
> 
> timer handlers are run from softirq, no need to disable bh
> 
> skb priority 1 is TC_PRIO_FILLER
> 
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>

Applied.

Strange we export this EQL_DEFAULT_RESCHED_IVAL to userspace, where
HZ will take on a different value I think, something akin to USER_HZ.

Anyways, that's a different unrelated problem to the bug you're
fixing.
--
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/eql.c b/drivers/net/eql.c
index a59cf96..f219d38 100644
--- a/drivers/net/eql.c
+++ b/drivers/net/eql.c
@@ -125,6 +125,7 @@ 
 #include <linux/if.h>
 #include <linux/if_arp.h>
 #include <linux/if_eql.h>
+#include <linux/pkt_sched.h>
 
 #include <asm/uaccess.h>
 
@@ -143,7 +144,7 @@  static void eql_timer(unsigned long param)
 	equalizer_t *eql = (equalizer_t *) param;
 	struct list_head *this, *tmp, *head;
 
-	spin_lock_bh(&eql->queue.lock);
+	spin_lock(&eql->queue.lock);
 	head = &eql->queue.all_slaves;
 	list_for_each_safe(this, tmp, head) {
 		slave_t *slave = list_entry(this, slave_t, list);
@@ -157,7 +158,7 @@  static void eql_timer(unsigned long param)
 		}
 
 	}
-	spin_unlock_bh(&eql->queue.lock);
+	spin_unlock(&eql->queue.lock);
 
 	eql->timer.expires = jiffies + EQL_DEFAULT_RESCHED_IVAL;
 	add_timer(&eql->timer);
@@ -341,7 +342,7 @@  static netdev_tx_t eql_slave_xmit(struct sk_buff *skb, struct net_device *dev)
 		struct net_device *slave_dev = slave->dev;
 
 		skb->dev = slave_dev;
-		skb->priority = 1;
+		skb->priority = TC_PRIO_FILLER;
 		slave->bytes_queued += skb->len;
 		dev_queue_xmit(skb);
 		dev->stats.tx_packets++;
diff --git a/include/linux/if_eql.h b/include/linux/if_eql.h
index 79c4f26..18a5d02 100644
--- a/include/linux/if_eql.h
+++ b/include/linux/if_eql.h
@@ -22,7 +22,7 @@ 
 #define EQL_DEFAULT_SLAVE_PRIORITY 28800
 #define EQL_DEFAULT_MAX_SLAVES     4
 #define EQL_DEFAULT_MTU            576
-#define EQL_DEFAULT_RESCHED_IVAL   100
+#define EQL_DEFAULT_RESCHED_IVAL   HZ
 
 #define EQL_ENSLAVE     (SIOCDEVPRIVATE)
 #define EQL_EMANCIPATE  (SIOCDEVPRIVATE + 1)