[PATCH] e1000: ring buffers resources cleanup

[PATCH] e1000: ring buffers resources cleanup

[Vasily Averin]

[-- Attachment #1: Type: text/plain, Size: 267 bytes --]

Memory leak was found in 2.6.18-rc4 and e1000 7.2.7 from sourceforge:
We should free resources allocated for previous rings if following allocation fails.

Signed-off-by: Vasily Averin <vvs@sw.ru>

Thank you,
	Vasily Averin
SWsoft Virtuozzo/OpenVZ Linux kernel team


[-- Attachment #2: diff-e1000-ringresources-20060818 --]
[-- Type: text/plain, Size: 714 bytes --]

--- linux-2.6.18-rc4/drivers/net/e1000/e1000_main.c.irsrs	2006-08-18 16:58:51.000000000 +0400
+++ linux-2.6.18-rc4/drivers/net/e1000/e1000_main.c	2006-08-18 18:53:05.000000000 +0400
@@ -1398,6 +1398,9 @@ e1000_setup_all_tx_resources(struct e100
 		if (err) {
 			DPRINTK(PROBE, ERR,
 				"Allocation for Tx Queue %u failed\n", i);
+			for (i-- ; i >= 0; i--)
+				e1000_free_tx_resources(adapter,
+							&adapter->tx_ring[i]);
 			break;
 		}
 	}
@@ -1656,6 +1659,9 @@ e1000_setup_all_rx_resources(struct e100
 		if (err) {
 			DPRINTK(PROBE, ERR,
 				"Allocation for Rx Queue %u failed\n", i);
+			for (i-- ; i >= 0; i--)
+				e1000_free_rx_resources(adapter,
+							&adapter->rx_ring[i]);
 			break;
 		}
 	}

[-- Attachment #3: Type: text/plain, Size: 373 bytes --]

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

[-- Attachment #4: Type: text/plain, Size: 164 bytes --]

_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel

[PATCH] e1000: ring buffers resources cleanup

[Vasily Averin]

[-- Attachment #1: Type: text/plain, Size: 241 bytes --]

irq leak was found in 2.6.18-rc4 and e1000 7.2.7 from sourceforge:
if e1000_up fails in e1000_open() we do not free allocated irq

Signed-off-by: Vasily Averin <vvs@sw.ru>

Thank you,
	Vasily Averin
SWsoft Virtuozzo/OpenVZ Linux kernel team

[-- Attachment #2: diff-e1000-irqleak-20060818 --]
[-- Type: text/plain, Size: 593 bytes --]

--- linux-2.6.18-rc4/drivers/net/e1000/e1000_main.c.oirq	2006-08-18 18:53:05.000000000 +0400
+++ linux-2.6.18-rc4/drivers/net/e1000/e1000_main.c	2006-08-18 19:18:22.000000000 +0400
@@ -1208,7 +1208,7 @@ e1000_open(struct net_device *netdev)
 
 	err = e1000_request_irq(adapter);
 	if (err)
-		goto err_up;
+		goto err_req_irq;
 
 	e1000_power_up_phy(adapter);
 
@@ -1229,6 +1229,8 @@ e1000_open(struct net_device *netdev)
 	return E1000_SUCCESS;
 
 err_up:
+	e1000_free_irq(adapter);
+err_req_irq:
 	e1000_free_all_rx_resources(adapter);
 err_setup_rx:
 	e1000_free_all_tx_resources(adapter);

Re: [PATCH] e1000: ring buffers resources cleanup

[Vasily Averin]

[-- Attachment #1: Type: text/plain, Size: 1404 bytes --]

Hello Joe,

Joe Perches wrote:
> On Fri, 2006-08-18 at 19:02 +0400, Vasily Averin wrote:
>>Memory leak was found in 2.6.18-rc4 and e1000 7.2.7 from sourceforge:
>>We should free resources allocated for previous rings if following allocation fails.
> 
> Did you read the comment headers in the function?
> 
>  * If this function returns with an error, then it's possible one or
>  * more of the rings is populated (while the rest are not).  It is the
>  * callers duty to clean those orphaned rings.

Thank you for your notice.
I believe this comment is incorrect: if some function returns an error it should
restore original state on exit, otherwise can lead to resource leaks. Also I
would note that this requirements is not accomplished in current driver version:
e1000_setup_all_Xx_resources functions are called in two places: in
e1000_set_ringparam() and in e1000_open() and in both cases nobody cleans those
orphaned rings.

Therefore I think it make sense to remove these comments too.

Andrew, could you please use attached patch instead previous version?
---
Memory leak was found in 2.6.18-rc4 and e1000 7.2.7 from sourceforge:
We should free resources allocated for previous rings if following allocation
fails. Also incorrect comments in e1000_setup_all_Xx_resources() are removed

Signed-off-by: Vasily Averin <vvs@sw.ru>

Thank you,
	Vasily Averin
SWsoft Virtuozzo/OpenVZ Linux kernel team

[-- Attachment #2: diff-e1000-ringresources-20060819 --]
[-- Type: text/plain, Size: 1454 bytes --]

--- linux-2.6.18-rc4/drivers/net/e1000/e1000_main.c.irsrs	2006-08-18 16:58:51.000000000 +0400
+++ linux-2.6.18-rc4/drivers/net/e1000/e1000_main.c	2006-08-19 22:54:00.000000000 +0400
@@ -1381,10 +1381,6 @@ setup_tx_desc_die:
  * 				  (Descriptors) for all queues
  * @adapter: board private structure
  *
- * If this function returns with an error, then it's possible one or
- * more of the rings is populated (while the rest are not).  It is the
- * callers duty to clean those orphaned rings.
- *
  * Return 0 on success, negative on failure
  **/
 
@@ -1398,6 +1394,9 @@ e1000_setup_all_tx_resources(struct e100
 		if (err) {
 			DPRINTK(PROBE, ERR,
 				"Allocation for Tx Queue %u failed\n", i);
+			for (i-- ; i >= 0; i--)
+				e1000_free_tx_resources(adapter,
+							&adapter->tx_ring[i]);
 			break;
 		}
 	}
@@ -1639,10 +1638,6 @@ setup_rx_desc_die:
  * 				  (Descriptors) for all queues
  * @adapter: board private structure
  *
- * If this function returns with an error, then it's possible one or
- * more of the rings is populated (while the rest are not).  It is the
- * callers duty to clean those orphaned rings.
- *
  * Return 0 on success, negative on failure
  **/
 
@@ -1656,6 +1651,9 @@ e1000_setup_all_rx_resources(struct e100
 		if (err) {
 			DPRINTK(PROBE, ERR,
 				"Allocation for Rx Queue %u failed\n", i);
+			for (i-- ; i >= 0; i--)
+				e1000_free_rx_resources(adapter,
+							&adapter->rx_ring[i]);
 			break;
 		}
 	}