Home » Mailing list archives » linux-netdev » 2009 » March » 4

Re: [PATCH] net: more timeouts that reach -1

Previous thread: [PATCH] ns83820: remove some pointless conditionals before kfree_skb() by Wei Yongjun on Wednesday, February 25, 2009 - 3:47 am. (1 message)

Next thread: Kernel problem by Denis Romanenko on Wednesday, February 25, 2009 - 4:43 am. (24 messages)
[view original message]
From: Roel Kluin
Subject: [PATCH] net: more timeouts that reach -1
Date: Wednesday, February 25, 2009 - 4:15 am

These were not previously reported by me.
------------------------------>8-------------8<---------------------------------
with while (timeout-- > 0); timeout reaches -1 after the loop, so the tests
below are off by one.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
---
 drivers/net/arm/ks8695net.c |    2 +-
 drivers/net/jme.c           |    3 ++-
 drivers/net/ucc_geth_mii.c  |    2 +-
 3 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/net/arm/ks8695net.c b/drivers/net/arm/ks8695net.c
index 1cf2f94..f3a1274 100644
--- a/drivers/net/arm/ks8695net.c
+++ b/drivers/net/arm/ks8695net.c
@@ -560,7 +560,7 @@ ks8695_reset(struct ks8695_priv *ksp)
 		msleep(1);
 	}
 
-	if (reset_timeout == 0) {
+	if (reset_timeout < 0) {
 		dev_crit(ksp->dev,
 			 "Timeout waiting for DMA engines to reset\n");
 		/* And blithely carry on */
diff --git a/drivers/net/jme.c b/drivers/net/jme.c
index 08b3405..0173ed0 100644
--- a/drivers/net/jme.c
+++ b/drivers/net/jme.c
@@ -957,7 +957,8 @@ jme_process_receive(struct jme_adapter *jme, int limit)
 		goto out_inc;
 
 	i = atomic_read(&rxring->next_to_clean);
-	while (limit-- > 0) {
+	while (limit > 0) {
+		limit--;
 		rxdesc = rxring->desc;
 		rxdesc += i;
 
diff --git a/drivers/net/ucc_geth_mii.c b/drivers/net/ucc_geth_mii.c
index 5463591..7b1b46c 100644
--- a/drivers/net/ucc_geth_mii.c
+++ b/drivers/net/ucc_geth_mii.c
@@ -123,7 +123,7 @@ static int uec_mdio_reset(struct mii_bus *bus)
 
 	mutex_unlock(&bus->mdio_lock);
 
-	if (timeout <= 0) {
+	if (timeout < 0) {
 		printk(KERN_ERR "%s: The MII Bus is stuck!\n", bus->name);
 		return -EBUSY;
 	}
--


[ message continues ]
[view original message]
From: Guo-Fu Tseng
Subject: Re: [PATCH] net: more timeouts that reach -1
Date: Friday, February 27, 2009 - 4:43 am

There should be no difference after this modification.


Guo-Fu Tseng

--


[ message continues ]
[view original message]
From: roel kluin
Subject: Re: [PATCH] net: more timeouts that reach -1
Date: Friday, February 27, 2009 - 8:37 am

There is:
In the last iteration limit is 1 during the test before it is decremented to 0.

rxdesc = rxring->desc;
rxdesc += i;

If then we break out of the loop by the 'goto out;', we continue with:

out:
        atomic_set(&rxring->next_to_clean, i);

out_inc:
        atomic_inc(&jme->rx_cleaning);


Roel
--


[ message continues ]
[view original message]
From: Guo-Fu Tseng
Subject: Re: [PATCH] net: more timeouts that reach -1
Date: Friday, February 27, 2009 - 8:51 am

I see.
But the correct patch should be following one, right?

===================================================================
--- jme.c	(revision 580)
+++ jme.c	(working copy)
@@ -958,13 +958,14 @@
 		goto out_inc;
 
 	i = atomic_read(&rxring->next_to_clean);
-	while (limit-- > 0) {
+	while (limit > 0) {
 		rxdesc = rxring->desc;
 		rxdesc += i;
 
 		if ((rxdesc->descwb.flags & RXWBFLAG_OWN) ||
 		!(rxdesc->descwb.desccnt & RXWBDCNT_WBCPL))
 			goto out;
+		--limit;
 
 		desccnt = rxdesc->descwb.desccnt & RXWBDCNT_DCNT;
 



Guo-Fu Tseng

--


[ message continues ]
[view original message]
From: Roel Kluin
Subject: [PATCH v3] net: more timeouts that reach -1
Date: Friday, February 27, 2009 - 9:43 am

Correct, thanks.
Here are all three patches again with another issue I spotted in 
drivers/net/ucc_geth_mii.c:

After my patch it still wouldn't err, because timeout was unsigned.
------------------------------>8-------------8<---------------------------------
with while (timeout-- > 0); timeout reaches -1 after the loop, so the tests
below are off by one. also don't do an '< 0' test on an unsigned.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
---
diff --git a/drivers/net/arm/ks8695net.c b/drivers/net/arm/ks8695net.c
index 1cf2f94..f3a1274 100644
--- a/drivers/net/arm/ks8695net.c
+++ b/drivers/net/arm/ks8695net.c
@@ -560,7 +560,7 @@ ks8695_reset(struct ks8695_priv *ksp)
 		msleep(1);
 	}
 
-	if (reset_timeout == 0) {
+	if (reset_timeout < 0) {
 		dev_crit(ksp->dev,
 			 "Timeout waiting for DMA engines to reset\n");
 		/* And blithely carry on */
diff --git a/drivers/net/jme.c b/drivers/net/jme.c
index 08b3405..a6e1a35 100644
--- a/drivers/net/jme.c
+++ b/drivers/net/jme.c
@@ -957,13 +957,14 @@ jme_process_receive(struct jme_adapter *jme, int limit)
 		goto out_inc;
 
 	i = atomic_read(&rxring->next_to_clean);
-	while (limit-- > 0) {
+	while (limit > 0) {
 		rxdesc = rxring->desc;
 		rxdesc += i;
 
 		if ((rxdesc->descwb.flags & cpu_to_le16(RXWBFLAG_OWN)) ||
 		!(rxdesc->descwb.desccnt & RXWBDCNT_WBCPL))
 			goto out;
+		--limit;
 
 		desccnt = rxdesc->descwb.desccnt & RXWBDCNT_DCNT;
 
diff --git a/drivers/net/ucc_geth_mii.c b/drivers/net/ucc_geth_mii.c
index 5463591..0ada4ed 100644
--- a/drivers/net/ucc_geth_mii.c
+++ b/drivers/net/ucc_geth_mii.c
@@ -107,7 +107,7 @@ int uec_mdio_read(struct mii_bus *bus, int mii_id, int regnum)
 static int uec_mdio_reset(struct mii_bus *bus)
 {
 	struct ucc_mii_mng __iomem *regs = (void __iomem *)bus->priv;
-	unsigned int timeout = PHY_INIT_TIMEOUT;
+	int timeout = PHY_INIT_TIMEOUT;
 
 	mutex_lock(&bus->mdio_lock);
 
@@ -123,7 +123,7 @@ static int uec_mdio_reset(struct mii_bus *bus)
 
 ...
[ message continues ]
[view original message]
From: David Miller
Subject: Re: [PATCH v3] net: more timeouts that reach -1
Date: Wednesday, March 4, 2009 - 1:10 am

From: Roel Kluin <roel.kluin@gmail.com>

I'll make sure to integrate this version of your patch instead
of the original which I just said I applied :-)
--


[ message continues ]
[view original message]
From: Roel Kluin
Subject: [PATCH v2] net: more timeouts that reach -1
Date: Friday, February 27, 2009 - 8:16 am

I just noticed that in drivers/net/ucc_geth_mii.c after my patch this still
won't give an err, because timeout is unsigned
------------------------------>8-------------8<---------------------------------
with while (timeout-- > 0); timeout reaches -1 after the loop, so the tests
below are off by one. also don't do an '< 0' test on an unsigned.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
---
diff --git a/drivers/net/arm/ks8695net.c b/drivers/net/arm/ks8695net.c
index 1cf2f94..f3a1274 100644
--- a/drivers/net/arm/ks8695net.c
+++ b/drivers/net/arm/ks8695net.c
@@ -560,7 +560,7 @@ ks8695_reset(struct ks8695_priv *ksp)
 		msleep(1);
 	}
 
-	if (reset_timeout == 0) {
+	if (reset_timeout < 0) {
 		dev_crit(ksp->dev,
 			 "Timeout waiting for DMA engines to reset\n");
 		/* And blithely carry on */
diff --git a/drivers/net/jme.c b/drivers/net/jme.c
index 08b3405..0173ed0 100644
--- a/drivers/net/jme.c
+++ b/drivers/net/jme.c
@@ -957,7 +957,8 @@ jme_process_receive(struct jme_adapter *jme, int limit)
 		goto out_inc;
 
 	i = atomic_read(&rxring->next_to_clean);
-	while (limit-- > 0) {
+	while (limit > 0) {
+		limit--;
 		rxdesc = rxring->desc;
 		rxdesc += i;
 
diff --git a/drivers/net/ucc_geth_mii.c b/drivers/net/ucc_geth_mii.c
index 5463591..0ada4ed 100644
--- a/drivers/net/ucc_geth_mii.c
+++ b/drivers/net/ucc_geth_mii.c
@@ -107,7 +107,7 @@ int uec_mdio_read(struct mii_bus *bus, int mii_id, int regnum)
 static int uec_mdio_reset(struct mii_bus *bus)
 {
 	struct ucc_mii_mng __iomem *regs = (void __iomem *)bus->priv;
-	unsigned int timeout = PHY_INIT_TIMEOUT;
+	int timeout = PHY_INIT_TIMEOUT;
 
 	mutex_lock(&bus->mdio_lock);
 
@@ -123,7 +123,7 @@ static int uec_mdio_reset(struct mii_bus *bus)
 
 	mutex_unlock(&bus->mdio_lock);
 
-	if (timeout <= 0) {
+	if (timeout < 0) {
 		printk(KERN_ERR "%s: The MII Bus is stuck!\n", bus->name);
 		return -EBUSY;
 	}

--


[ message continues ]
[view original message]
From: David Miller
Subject: Re: [PATCH] net: more timeouts that reach -1
Date: Wednesday, March 4, 2009 - 1:05 am

From: Roel Kluin <roel.kluin@gmail.com>

Applied, thanks.
--


[ message continues ]
Previous thread: [PATCH] ns83820: remove some pointless conditionals before kfree_skb() by Wei Yongjun on Wednesday, February 25, 2009 - 3:47 am. (1 message)

Next thread: Kernel problem by Denis Romanenko on Wednesday, February 25, 2009 - 4:43 am. (24 messages)